Multi-chain multi-target bispecific antigen-binding molecules with increased selectivity

Abstract

The present invention provides a multi-chain multi-target bispecific antigen-binding molecule comprising a first bispecific entity and a second bispecific entity, each comprising a domain that binds to a target and a second domain that binds to an extracellular epitope of the human and Macaca CD3ε chains, wherein both bispecific entities are linked to each other by a spacer that separates the first bispecific entity from the second bispecific entity. Furthermore, the present invention provides a polynucleotide encoding the multi-target bispecific antigen-binding molecule, a vector comprising this polynucleotide, a host cell expressing this construct, and a pharmaceutical composition comprising the same.

Description

Technical Field

[0001] The present invention relates to biotechnological products and methods, in particular, multi-chain multi-target antigen-binding molecules, their preparation and their use.

Background Art

[0002] The redirection of T cell activity against tumor cells by bispecific molecules that do not depend on the specificity of the T cell receptor is an evolving approach in immuno-oncology (Frankel SR, Baeuerle PA. Targeting T cells to tumor cells using bispecific antibodies. Curr Opin Chem Biol 2013;17:385-92). Such novel protein-based pharmaceuticals can typically bind simultaneously to two different types of antigens. They are known in several structural formats and are currently being explored for applications in cancer immunotherapy and drug delivery (Fan, Gaowei; Wang, Zujian; Hao, Mingju; Li, Jinming (2015). “Bispecific antibodies and their applications”. Journal of Hematology & Oncology. 8:130).

[0003] Bispecific molecules useful in immuno-oncology can be antigen-binding polypeptides such as antibodies, for example, IgG-like bispecific antibodies (i.e., full-length bispecific antibodies), or non-IgG-like bispecific antibodies that are not full-length antigen-binding molecules. Full-length bispecific antibodies typically retain the conventional monoclonal antibody (mAb) structure of two Fab arms and one Fc region, except that the two Fab sites bind to different antigens. Non-full-length bispecific antibodies may completely lack the Fc region. These include chemically linked Fabs consisting only of the Fab region, as well as various types of bivalent and trivalent single-chain variable fragments (scFvs). There are also fusion proteins that mimic the variable domains of two antibodies. An example of such a format is the bispecific T cell engager (BiTE®) (Yang, Fa; Wen, Weihong; Qin, Weijun (2016). “Bispecific Antibodies as a Development Platform for New Concepts and Treatment Strategies”. International Journal of Molecular Sciences. 18(1):48).

[0004] Molecules derived from exemplary bispecific antibodies, such as BiTE® molecules, are recombinant protein constructs composed of two covalently linked antibody-derived binding domains. One binding domain of a BiTE® antigen-binding molecule is specific for a selected tumor-associated surface antigen on a target cell, and the second binding domain is specific for CD3, a subunit of the T cell receptor complex on T cells. Due to these special designs, BiTE® antigen-binding molecules are uniquely suited to transiently bind T cells to target cells while potently activating the cytotoxic ability of T cells against their native target cells. Further important development of first-generation BiTE® antigen-binding molecules, developed clinically as AMG 103 and AMG 110 (see International Publication No. WO 99 / 54440 pamphlet and International Publication No. WO 2005 / 040220 pamphlet), provided bispecific antigen-binding molecules that bind to a context-independent epitope at the N-terminus of the CD3ε chain (International Publication No. WO 2008 / 119567 pamphlet). BiTE® antigen-binding molecules that bind to this selected epitope not only show no interspecies specificity for the CD3ε chain of humans and Macaca, or Callithrix jacchus, Saguinus oedipus or Saimiri sciureus, but also do not show non-specific activation of T cells to the same extent as observed with previous-generation T cell engaging antibodies, as they recognize this specific epitope (instead of the epitopes of the CD3 binders previously described in bispecific T cell engaging molecules). This reduction in T cell activation is associated with a decrease or reduction in the redistribution of T cells in patients, which has been identified as a risk of side effects, for example, in pasuximab.

[0005] The antibody-based molecules described in International Publication No. WO 2008 / 119567 are characterized by rapid clearance from the body. Thus, they can rapidly reach most parts of the body, but their in vivo applicability may be limited due to their short in vivo persistence. On the other hand, their concentrations in the body can be immediately adapted and finely adjusted. Since this single-chain molecule of low molecular weight has a short in vivo half-life, long-term administration by continuous intravenous infusion has been used to achieve a therapeutic effect. However, bispecific antigen-binding molecules with more favorable pharmacokinetic properties, such as a longer half-life as described in International Publication No. WO 2017 / 134140, are available. A long half-life is typically useful for in vivo applications of immunoglobulins, particularly small-sized antibody fragments or constructs, for example, for patient compliance.

[0006] One of the difficult problems currently occurring in antibody-based immuno-oncology is tumor escape. Such tumor escape occurs when genetic and epigenetic modifications accumulate (even if the immune system is induced or guided by some antibody-based immunotherapies), and the immune editing process cannot completely eradicate tumors using several mechanisms to gain the upper hand. (Keshavarz-Fathi, Mahsa; Rezaei, Nima (2019) “Vaccines for Cancer Immunotherapy”). Generally, the following four mechanisms that interfere with an effective anti-tumor immune response are known: (1) defective processing or presentation of tumor antigens, (2) lack of activation mechanisms, (3) inhibitory mechanisms and immunosuppressive states, and (4) resistant tumor cells. In particular, regarding the first mechanism, due to genetic instability, tumor mutations, and immune system avoidance, tumor antigens may exist in new forms. Epitope-negative tumor cells remain latent and, as a result, show resistance to immune rejection. These have developed following the elimination of epitope-positive tumor cells, similar to Darwin's theory of natural selection. As a result, antibody-based immunotherapy against antigens on tumor cells becomes ineffective when such tumor cells no longer express the corresponding antigens due to tumor escape. This antigen deficiency is understood in this specification as the driving force of tumor escape and is therefore used interchangeably. Therefore, in order to effectively prevent tumor escape, it is necessary to provide an improved antibody-based cancer immunotherapy that addresses the problem of antigen deficiency.

[0007] A more pressing challenge for the broad use of immuno-oncology with T cell engaging bispecific molecules would be the availability of suitable targets (Bacac et al., Clin Cancer Res;22(13)July 1,2016). For example, targets for solid tumors can be overexpressed on tumor cells but are lower and can be expressed at significant levels on non-malignant primary cells of important tissues. In nature, according to Bacac et al., T cells can distinguish between cells with high antigen expression and cells with low antigen expression by means of a T cell receptor (TCR) with relatively low affinity (which can still achieve high avidity binding to target cells expressing a sufficiently high level of the target antigen). There is a great need for T cell engaging bispecific molecules that can facilitate this discrimination and thereby maximize the window between the killing of high target-expressing cells and low target-expressing cells. One approach being considered in the art is to use dual targeting of two antigens, which may lead to improved target selectivity compared to normal tissues expressing only one of both target antigens or both target antigens at low levels. This effect is thought to depend on the avidity component mediated by the simultaneous binding of the bsAb to both antigens on the same cell. With respect to such dual targeting, several multispecific monoclonal antibodies (mAbs) or other immunoconstructs are known in the art. WO 2014 / 116846 pamphlet teaches a multispecific binding protein comprising a first binding site that specifically binds to a target cell antigen, a second binding site that specifically binds to a cell surface receptor on an immune cell, and a third binding site that specifically binds to a cell surface regulator on an immune cell. US 2017 / 0022274 A discloses a trivalent T cell redirecting complex comprising a bispecific antibody having two binding sites for a tumor-associated antigen (TAA) and one binding site for a T cell.

[0008] However, dual targeting alone, as in the case of the above-mentioned molecules, may be insufficient for effective target selectivity (Mazor et al, mAbs 7:3, 461-469; May / June 2015). In particular, the composition of the bsAb binding domain, i.e., being monovalent vs. bivalent, is an important factor. Furthermore, since it is also necessary to consider the potential risk profile for serious immunological side effects such as cytokine release syndrome (CRS), simply providing bispecific molecules with multiple valences may not lead to clinically suitable therapeutics. Therefore, despite the preclinical and clinical success of antibody-based immunotherapeutics achieved so far, notable limitations such as differences in response between individuals and cancer types still exist. Dose-limiting toxicity can be a factor restricting the effectiveness of antibody-based immunotherapeutics, so not all patients respond to therapy at a safe dose that is available. Therefore, it is also necessary to reduce dose-limiting toxicity in antibody-based immunotherapeutics and make such therapies available to more patients suffering from various proliferative diseases.

[0009] Various multispecific antibodies or antibody fragments are known in the art, some of which address T cells, but a multi-target bispecific molecule that addresses the need to overcome dose-limiting toxicity in T cell redirected immunotherapeutics by expanding the therapeutic area and is a stable and immediately available therapeutic system has not been proposed so far.

Prior Art Documents

Patent Documents

[0010]

Patent Document 1

Patent Document 2

Patent Document 3

Patent Document 4

[0011] [Non-Patent Document 1] Frankel SR, Baeuerle PA. Targeting T cells to tumor cells using bispecific antibodies. Curr Opin Chem Biol 2013;17:385 - 92 [Non-Patent Document 2] Fan, Gaowei; Wang, Zujian; Hao, Mingju; Li, Jinming(2015). “Bispecific antibodies and their applications”. Journal of Hematology & Oncology. 8:130 [Non-Patent Document 3] Yang, Fa; Wen, Weihong; Qin, Weijun(2016). “Bispecific Antibodies as a Development Platform for New Concepts and Treatment Strategies”. International Journal of Molecular Sciences. 18(1):48 [Non-Patent Document 4] Keshavarz-Fathi, Mahsa; Rezaei, Nima(2019) “Vaccines for Cancer Immunotherapy” [Non-Patent Document 5] Bacac et al., Clin Cancer Res;22(13) July 1, 2016 [Non-Patent Document 6] Mazor et al, mAbs 7:3, 461 - 469; May / June 2015

Summary of the Invention

Means for Solving the Problems

[0012] In view of the various unmet needs described above, it is an object of the present invention to provide a molecule comprising at least two polypeptide chains, i.e., a multi-chain molecule, preferably an antigen-binding molecule. The molecule of the present invention is further preferably a bispecific molecule, such as a T cell engaging molecule. Furthermore, the molecule of the present invention is preferably multi-targeted and can, for example, typically immunospecifically bind to at least two antigens on target cells typically associated with one or more diseases. The molecule of the present invention can more preferably typically immunospecifically bind simultaneously to two antigens on effector cells, preferably for use in the treatment of said one or more diseases. Accordingly, the present invention preferably provides a multi-targeted bispecific antigen-binding molecule comprising at least one polypeptide, the molecule comprising at least five characteristic structural entities, namely, (i.) a first domain that binds to a target cell surface antigen (e.g., a first tumor-associated antigen, TAA), (ii.) a second domain that binds to an extracellular epitope of a human (and preferably non-human primate, such as Macaca) CD3 chain (the first binding domain and the second binding domain together form a first bispecific entity), (iii.) a spacer that links and is sufficiently spaced apart from the second bispecific entity and the first bispecific entity, (iv.) a third domain that binds to the same or preferably a different target cell surface antigen (e.g., a second TAA), and (v.) a fourth domain that binds to an extracellular epitope of a human (and preferably non-human primate, such as Macaca) CD3 chain. Preferably, the domains are (i.) scFV domains each composed of VH and VL domains in the amino to carboxyl direction, with short peptide linkers that are flexible but link the VL of the first domain to the VH of the second domain and the VL of the third domain to the VH of the fourth domain, respectively, and / or (ii.) Fab domains comprising a first polypeptide monomer comprising VL and CL domains and a second polypeptide monomer comprising VH and CH domains.Surprisingly, multi-chain multi-target bispecific antigen-binding molecules as described herein typically enable T cells to respectively identify the killing of cells expressing only one or both targets typically associated with a particular disease, thereby expanding the therapeutic area and reducing the risk of off-target toxicity and side effects. Furthermore, the present invention provides a polynucleotide encoding a multi-target bispecific antigen-binding molecule, a vector containing this polynucleotide, a host cell expressing this construct, and a pharmaceutical composition containing the same.

[0013] In a first aspect, in the context of the present invention, it is contemplated to provide a multi-chain multi-target bispecific molecule comprising at least two polypeptide chains, the molecule being (i.) a first binding domain that binds to a first target cell surface antigen (TAA1), (ii.) a second binding domain that binds to an extracellular epitope of a human and / or Macaca CD3 chain, (iii.) a third binding domain that binds to a second target cell surface antigen (TAA2), and (iv.) a fourth binding domain that binds to an extracellular epitope of a human and / or Macaca CD3 chain comprising, the first binding domain and the second binding domain form a first bispecific entity, the third binding domain and the fourth binding domain form a second bispecific entity, the molecule further comprises a spacer entity, and the spacer entity is (1.) (a.) an Fc domain comprising a first polypeptide monomer and a second polypeptide monomer each containing a hinge, CH2 domain, and CH3 domain, wherein the first polypeptide monomer and the second polypeptide monomer form a heterodimer; the heterodimer is - (i.) D399K, K409D, K392D, and E356K, (ii.) D399K, K409D, K392D, E357K, K370D, and E356K, (iii.) D399K, K409D, K392D, E356K and K439D, (iv.) D399K, K409D, and K392D, (v.) D399K, K409D, K392D, E357K and E370K, (vi.) D399K, K409D, K392D, E357K, K370E and K360E, (vii.) D399K, K409D, K392D, E357K, K370E, E356K, and K439E, and (viii.) D399K, K409D, K392D, E357K, K370E, K360E, E356K, and K439D, which are charge pair mutations, preferably including K392D, K409D and / or K439D mutations in the CH3 domain of the first polypeptide monomer, and E356K and / or D399K mutations in the CH3 domain of the second polypeptide monomer, where the positions are according to EU numbering, charge pair mutations; or - Preferably including T366S, L368A and Y407V mutations in the first polypeptide monomer, and T366W mutation in the second monomer, where the positions are according to EU numbering, knob-into-hole mutations formed by, Fc domain; (b.) A human serum albumin (HSA) domain comprising a first polypeptide monomer and a second polypeptide monomer, where the first polypeptide monomer and the second polypeptide monomer each correspond to an HSA subdomain and form a native HSA-like heterodimer, human serum albumin (HSA) domain; and (c.) A Fab comprising a first polypeptide monomer and a second polypeptide monomer, where the first polypeptide monomer comprises VL and CL domains, the second polypeptide monomer comprises VH and CH1 domains, and the CL domain and the CH1 domain are linked by a disulfide bridge, Fab selected from, dimerization domain; Comprising two N - termini and two C - termini respectively, with at least one N - terminus and one C - terminus each being linked to a bispecific entity, and any one of the first, second, third, and fourth domains can be selected from any form of binding domain, preferably selected from Fab and single - chain domains, and the single - chain domain is preferably selected from single - chain Fv (scFv) and scFab, a dimerization domain; (2.) Selected from ubiquitin, β2 - microglobulin, VH - only domain, PSI domain derived from the Met receptor, fibronectin type III domain derived from tenascin, granulocyte - macrophage colony - stimulating factor (GM - CSF), interleukin - 4, CD137L extracellular domain, interleukin - 2, PD - 1 binding domain derived from human programmed cell death 1 ligand 1 (PDL1), Tim - 3 (AS 24 - 130), MiniSOG, programmed cell death protein 1 (PD1) domain, human serum albumin (HSA), or a single - chain domain comprising two polypeptide monomers each containing a hinge, CH2, and CH3 domains, a hinge, and further CH2 and CH3 domains, wherein the two polypeptide monomers are fused to each other via a peptide linker, Comprising one N - terminus and one C - terminus each linked to a bispecific entity, with at least one of the first, second, third, and fourth binding domains being a double - chain Fab, and any one of the remaining at least three binding domains can be selected from any form of binding domain, preferably selected from Fab and single - chain domains, preferably selected from scFv and scFab, a single - chain domain selected from; The distance between the Cα atom of the first amino acid located at the N-terminus of the spacer entity and the Cα atom of the last amino acid at the C-terminus is separated by at least 30 Å. The spacer entity separates the first bispecific entity and the second bispecific entity by a distance of at least about 50 Å. The distances indicated are preferably understood as (i.) the distance between the first binding domain and the third binding domain, or (ii.) the distance between the centers of mass of the first bispecific entity and the second bispecific entity. The spacer entity is disposed between the first bispecific entity and the second bispecific entity.

[0014] In said aspect, it is also contemplated in the context of the present invention to provide a multichain multi-target bispecific antigen-binding molecule. When the spacer is a single-chain domain, the arrangement of the binding domains in amino-to-carboxyl order is (i.) a first domain and a second domain, a spacer, a third domain and a fourth domain (ii.) a first domain and a second domain, a spacer, a fourth domain and a third domain (iii.) a second domain and a first domain, a spacer, a third domain and a fourth domain, and (iv.) a second domain and a first domain, a spacer, a fourth domain and a third domain selected from the group consisting of.

[0015] In said aspect, it is also contemplated in the context of the present invention to provide a multichain multi-target bispecific antigen-binding molecule. When the spacer is a single-chain domain, the arrangement of the binding domains in amino-to-carboxyl order is (i.) a first domain in the form of a Fab, a second domain in the form of a scFv, a spacer, a third domain in the form of a Fab, and a fourth domain in the form of a scFv (e.g., FIG. 3B); (ii.) a first domain in the form of a Fab, a second domain in the form of a Fab, a spacer, a third domain in the form of a Fab, and a fourth domain in the form of a Fab (e.g., FIG. 3D); (iii) a first domain in the form of a scFv, a second domain in the form of a Fab, a spacer, a third domain in the form of a scFv, and a fourth domain in the form of a Fab (e.g., FIG. 3H); (iv) a first domain in the form of a scFv, a second domain in the form of a scFv, a spacer, a third domain in the form of a scFv, and a fourth domain in the form of a Fab; (v) a first domain in the form of a scFv, a second domain in the form of a scFv, a spacer, a third domain in the form of a Fab, and a fourth domain in the form of a scFv; (vi) a first domain in the form of a Fab, a second domain in the form of a scFv, a spacer, a third domain in the form of a scFv, and a fourth domain in the form of a scFv; and (vii) a first domain in the form of a scFv, a second domain in the form of a Fab, a spacer, a third domain in the form of a scFv, and a fourth domain in the form of a scFv selected from the group consisting of, each scFv comprises, in amino to carboxyl order, VH, a linker and VL, or VL, a linker and VH, preferably VH, a linker and VL.

[0016] In the above aspect, it is also contemplated in the context of the present invention to provide a multichain multi-target bispecific antigen-binding molecule. When the spacer is a dimerization domain, the arrangement of the binding domains in amino to carboxyl order is (i) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that forms a Fab together with the first chain, a second domain in the form of a scFv, a first polypeptide monomer of the spacer dimerization domain, a third chain comprising VH and CH1 of the third domain that forms a Fab together with the VL and CL of the third domain of the fourth chain and the second polypeptide monomer of the spacer dimerization domain, and a fourth chain comprising VL and CL of the third domain and a fourth domain in the form of a scFv (e.g., FIG. 3A); (ii.) a first domain in the form of a Fab, a second domain in the form of a Fab, a spacer, a third domain in the form of a Fab, and a fourth domain in the form of a Fab (e.g., FIG. 3C); (iii.) a second domain in the form of a scFv, VH and CH1 of the first domain that forms a Fab with the second chain, a first chain comprising a first polypeptide monomer of the spacer dimerization domain, a second chain comprising VL and CL of the first domain, a second polypeptide monomer of the spacer dimerization domain, a third chain comprising VH and CH1 of the third domain that forms a Fab with VL and CL of the third domain of the fourth chain, and a fourth chain comprising VL and CL of the third domain and a fourth domain in the form of a scFv (e.g., FIG. 3E); (iv.) a second domain in the form of a scFv, VH and CH1 of the first domain that forms a Fab with the second chain, a first chain comprising a first polypeptide monomer of the spacer dimerization domain, a second chain comprising VL and CL of the first domain, a second polypeptide monomer of the spacer dimerization domain, a fourth domain in the form of a scFv, a third chain comprising VH and CH1 of the third domain that forms a Fab with VL and CL of the third domain of the fourth chain, and a fourth chain comprising VL and CL of the third domain (e.g., FIG. 3F); (v.) a second domain in the form of a scFv, VH and CH1 of the first domain that forms a Fab with the second chain, a first chain comprising a first polypeptide monomer of the spacer dimerization domain, a second chain comprising VL and CL of the first domain, a second polypeptide monomer of the spacer dimerization domain, a fourth domain in the form of a scFv, a third chain comprising VH and CH1 of the third domain that forms a Fab with VL and CL of the third domain of the fourth chain, and a fourth chain comprising VL and CL of the third domain (e.g., FIG. 3G) selected from the group consisting of; Each scFv comprises, in order from amino to carboxyl, VH, a linker and VL, or VL, a linker and VH, preferably VH, a linker and VL.

[0017] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule. When the spacer is a dimerization domain, the arrangement of the binding domains in the order from amino to carboxyl is (i.) a first chain comprising a first domain in the form of a scFv, a first polypeptide monomer of the spacer dimerization domain, and a third domain in the form of a scFv, and a second chain comprising a second domain in the form of a scFv, a second polypeptide monomer of the spacer dimerization domain, and a fourth domain in the form of a scFv (e.g., Figure 2A); (ii.) a first chain comprising a first domain in the form of a scFv, a first polypeptide monomer of the spacer dimerization domain, and a second domain in the form of a scFv, and a second chain comprising a third domain in the form of a scFv, a second polypeptide monomer of the spacer dimerization domain, and a fourth domain in the form of a scFv (e.g., Figure 2B); (iii.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that forms a Fab together with the first chain, a first polypeptide monomer of the spacer dimerization domain, and a third domain in the form of a scFv, and a third chain comprising a second domain in the form of a scFv, a second polypeptide monomer of the spacer dimerization domain, and a fourth domain in the form of a scFv (e.g., Figure 2C); (iv.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that forms a Fab together with the first chain, a first polypeptide monomer of the spacer dimerization domain, and a second domain in the form of a scFv, and a third chain comprising a fourth domain in the form of a scFv, a second polypeptide monomer of the spacer dimerization domain, and a third domain in the form of a scFv (e.g., Figure 2D); (v.) A first chain comprising VL and CL of a first domain, VH and CH1 of a first domain that forms a Fab together with the first chain, a second chain comprising a first polypeptide monomer of a spacer dimerization domain, a second domain in the form of a scFv, a second polypeptide monomer of the spacer dimerization domain, a third domain in the form of a scFv, and a third chain comprising a fourth domain in the form of a scFv (e.g., FIG. 2E); (vi.) A first chain comprising VL and CL of a first domain, VH and CH1 of a first domain that forms a Fab together with the first chain, a first polypeptide monomer of a spacer dimerization domain, a third domain in the form of a scFv, and a fourth domain in the form of a scFv, a second chain, a second domain in the form of a scFv, and a third chain comprising a second polypeptide monomer of the spacer dimerization domain (e.g., FIG. 2F); (vii) A first chain comprising VL and CL of a first domain, VH and CH1 of a first domain that forms a Fab together with the first chain, a first polypeptide monomer of a spacer dimerization domain, a second chain comprising VH and CH1 of a third domain that forms a Fab together with the third chain, a third chain comprising VL and CL of the third domain, VH and CH1 of a second domain that forms a Fab together with the fifth chain, a second polypeptide monomer of the spacer dimerization domain, a fourth (forth) chain comprising VH and CH1 of a third domain that forms a Fab together with the sixth chain, a fifth chain comprising VL and CL of the second domain, and a sixth chain comprising VL and CL of the fourth domain (e.g., FIG. 2G); (viii.) A first chain comprising VL and CL of a first domain, VH and CH1 of a first domain that forms a Fab together with the first chain, a first polypeptide monomer of a spacer dimerization domain, a second chain comprising a second domain in the form of a scFv, a fourth domain in the form of a scFv, a second polypeptide monomer of the spacer dimerization domain, a third chain comprising VH and CH1 of a third domain that forms a Fab together with the fourth chain, and a fourth chain comprising VL and CL of the third domain (e.g., FIG. 2H); (ix.) A first chain comprising a first domain in scFv format, a first polypeptide monomer of a spacer dimerization domain, a third domain in scFv format; a VH and CH1 of a second domain that forms a Fab with a third chain, a second polypeptide monomer of a spacer dimerization domain, a second chain comprising a VH and CH1 of a fourth domain that forms a Fab with a fourth chain; a third chain comprising a VL and CL of the second domain; and a fourth chain comprising a VL and CL of the fourth domain (e.g., FIG. 2I); (x.) A first chain comprising a VL and CL of a first domain; a VH and CH1 of a first domain that forms a Fab with the first chain, a first polypeptide monomer of a spacer dimerization domain, a second chain comprising a third domain in scFv format; a VH and CH1 of a second domain that forms a Fab with a fourth chain, a second polypeptide monomer of a spacer dimerization domain, a third chain comprising a fourth domain in scFv format; and a fourth chain comprising a VL and CL of the second domain (e.g., FIG. 2J); (xi.) A first chain comprising a VL and CL of a first domain; a VH and CH1 of a first domain that forms a Fab with the first chain, a first polypeptide monomer of a spacer dimerization domain, a second chain comprising a second domain in scFv format; a VH and CH1 of a fourth domain that forms a Fab with a fourth chain, a second polypeptide monomer of a spacer dimerization domain, a third chain comprising a third domain in scFv format; and a fourth chain comprising a VL and CL of the fourth domain (e.g., FIG. 2K); (xii.) A first chain comprising a VL and CL of a first domain; a VH and CH1 of a first domain that forms a Fab with the first chain, a first polypeptide monomer of a spacer dimerization domain, a second chain comprising a second domain in scFv format; a VH and CH1 of a second domain that forms a Fab with a fourth chain, a second polypeptide monomer of a spacer dimerization domain, a third chain comprising a third domain and a fourth domain in scFv format; and a fourth chain comprising a VL and CL of the second domain (e.g., FIG. 2L); (xiii.) a first chain comprising VL and CL of the first domain, VH and CH1 of the first domain that forms a Fab together with the first chain, a first polypeptide monomer of the spacer dimerization domain, a third domain in the form of a scFv, a second chain comprising a fourth domain in the form of a scFv, VH and CH1 of the second domain that forms a Fab together with the fourth chain, a third chain comprising a second polypeptide monomer of the spacer dimerization domain, and a fourth chain comprising VL and CL of the second domain (e.g., FIG. 2M) selected from the group consisting of; Each scFv comprises, in the N- to C-direction, VH, a linker, and VL, or VL, a linker, and VH, preferably VH, a linker, and VL.

[0018] In said aspect, it is also contemplated to provide an antigen-binding molecule in the context of the present invention, the spacer entity being a globular protein, the distance between the Cα atom of the first amino acid located at the N-terminus and the Cα atom of the last amino acid at the C-terminus being at least 20 Å, preferably at least 30 Å, more preferably at least 50 Å, so as to effectively separate the first bispecific entity and the second bispecific entity by preferably at least 50 Å.

[0019] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, and the spacer entity that effectively separates the first bispecific entity and the second bispecific entity, when the spacer is single-stranded, is ubiquitin, β2-microglobulin, SAND domain, green fluorescent protein (GFP), VHH antibody llama domain, PSI domain derived from Met receptor, fibronectin type III domain derived from tenascin, granulocyte macrophage colony-stimulating factor (GM-CSF), interleukin-4, CD137L extracellular domain, interleukin-2, PD-1 binding domain derived from human programmed cell death 1 ligand 1 (PDL1), Tim-3 (AS 24-130), MiniSOG, programmed cell death protein 1 (PD1) domain, human serum albumin (HSA) or a derivative of any of the aforementioned spacer entities, a multimer of a rigid linker, and an Fc domain or its dimer or trimer, and each Fc domain includes two polypeptide monomers each containing a hinge, CH2 and CH3 domains, a hinge and further CH2 and CH3 domains, respectively, and the two polypeptide monomers are fused to each other via a peptide linker or are linked together by non-covalent CH3-CH3 interactions and / or covalent disulfide bonds to form a heterodimer.

[0020] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, and the spacer entity when the single strand is at least one Fc domain is preferably one domain or two or three covalently linked domains, and they or each of them, in the order from amino to carboxyl: hinge-CH2-CH3-linker-hinge-CH2-CH3 include.

[0021] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, and each of the polypeptide monomers in the spacer entity has an amino acid sequence that is at least 90% identical to a sequence selected from the group consisting of SEQ ID NOs: 17-24, and preferably, each of the polypeptide monomers has an amino acid sequence selected from SEQ ID NOs: 17-24.

[0022] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, and the CH2 domain in the spacer contains an intradomain cysteine disulfide bridge.

[0023] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, and the single-chain spacer entity contains an amino acid sequence selected from the group consisting of SEQ ID NOs: 13 and 15-16 and 25-34, ubiquitin (SEQ ID NO: 1081), β2-microglobulin (SEQ ID NO: 1083), SAND domain (SEQ ID NO: 1084), green fluorescent protein (GFP) (SEQ ID NO: 1085), VHH antibody llama domain (SEQ ID NO: 1086), PSI domain derived from the Met receptor (SEQ ID NO: 1087), fibronectin type III domain derived from tenascin (SEQ ID NO: 1088), granulocyte macrophage colony-stimulating factor (GM-CSF) (SEQ ID NO: 1089), interleukin-4 (SEQ ID NO: 1090), CD137L extracellular domain (SEQ ID NO: 1091), interleukin-2 (SEQ ID NO: 1092), PD-1 binding domain derived from human programmed cell death 1 ligand 1 (PDL1) (SEQ ID NO: 1093), Tim-3 (AS 24-130) (SEQ ID NO: 1094), MiniSOG (SEQ ID NO: 1095), programmed cell death protein 1 (PD1) domain (SEQ ID NO: 16), human serum albumin (has, SEQ ID NO: 15) or an amino acid having at least 90%, preferably 95% or even 98% sequence identity thereto, preferably scFc (SEQ ID NO: 25).

[0024] In the above aspect, it is also assumed that the first peptide monomer of the first peptide chain is SEQ ID NO: 35 and the second peptide monomer of the second peptide chain is SEQ ID NO: 36, and the two peptide monomers preferably form a heterodimer.

[0025] In the above aspect, the antigen-binding molecule (i) the first domain and the third domain contain variable domains from two antibodies, and the second domain and the fourth domain contain variable domains from two antibodies; (ii) the first domain and the third domain contain variable domains from one antibody, and the second domain and the fourth domain contain variable domains from two antibodies; (iii) the first domain and the third domain contain variable domains from two antibodies, and the second domain and the fourth domain contain variable domains from one antibody; or (iv) the first domain contains a variable domain from one antibody and the third domain contains a variable domain from one antibody is also assumed to be characterized thereby.

[0026] In the above aspect, in the context of the present invention, it is also assumed to provide an antigen-binding molecule comprising two polypeptide chains, the first polypeptide chain comprising VH of the first domain, VH of the second domain, a first polypeptide monomer preferably comprising hinge, CH2 and CH3 domains, VH of the third domain, and VH of the fourth domain; the second polypeptide chain comprising VL of the first domain, VL of the second domain, a first polypeptide monomer preferably comprising hinge, CH2 and CH3 domains, VL of the third domain, and VL of the fourth domain, preferably, the first polypeptide monomer and the second polypeptide monomer form a heterodimer, thereby linking the first polypeptide chain and the second polypeptide chain.

[0027] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, and the first, second, third, and fourth binding domains each contain a VH domain and a VL domain in the order from amino to carboxyl, and VH and VL within each domain are connected by a peptide linker, preferably a flexible linker containing serine, glutamine, and / or glycine as amino acid components, preferably only serine (Ser, S) or glutamine (Gln, Q) and glycine (Gly, G), more preferably (G4S)n or (G4Q)n, and even more preferably containing SEQ ID NO: 1 or 3.

[0028] In the above aspect, in the context of the present invention, it is also contemplated to provide a peptide linker, and the peptide linker contains or consists of S(G4X)n and (G4X)n (X is selected from the group consisting of Q, T, N, C, G, A, V, I, L, and M, n is an integer selected from integers 1 to 20, preferably n is 1, 2, 3, 4, 5, or 6, and preferably X is Q), and preferably the peptide linker is (G4X)n (n is 3 and X is Q).

[0029] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, and the peptide linker between the first binding domain and the second binding domain and between the third binding domain and the fourth binding domain preferably contains serine, glutamine, and / or glycine or glutamic acid, alanine, and lysine as amino acid components, and preferably is a flexible linker selected from the group consisting of SEQ ID NOs: 1 to 4, 6 to 12, and 1125.

[0030] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, and the peptide linker between the first binding domain or the second binding domain and the spacer, and / or between the third binding domain and the fourth binding domain and the spacer is preferably a short linker rich in low molecular weight and / or hydrophilic amino acids, preferably glycine, and preferably SEQ ID NO: 5.

[0031] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, and either the first target cell surface antigen or the second target cell surface antigen is selected from the group consisting of CS1, BCMA, CDH3, FLT3, CD123, CD20, CD22, EpCAM, MSLN, and CLL1.

[0032] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, and the first target cell surface antigen and the second target cell surface antigen are not the same.

[0033] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, and the first target cell surface antigen and the second target cell surface antigen are the same.

[0034] In the above aspect, in the context of the present invention, it is also contemplated to provide the antigen-binding molecule of claim 1, wherein the first binding domain can bind to the first target cell surface antigen, and the third binding domain can simultaneously bind to the second target cell surface antigen, preferably, the first target cell surface antigen and the second target cell surface antigen are on the same target cell.

[0035] In the above aspect, in the context of the present invention, it is also contemplated to provide the antigen-binding molecule of claim 1, and the first target cell surface antigen and the second target cell surface antigen are each selected from the group consisting of CS1 and BCMA, BCMA and CS1, FLT3 and CD123, CD123 and FLT3, CD20 and CD22, CD22 and CD20, EpCAM and MSLN, MSLN and EpCAM, MSLN and CDH3, CDH3 and MSLN, FLT3 and CLL1, and CLL1 and FLT3.

[0036] In the above aspect, in the context of the present invention, it is also contemplated to provide the antigen-binding molecule of claim 1, wherein the first target cell surface antigen and / or the second target cell surface antigen is human MSLN (selected from SEQ ID NOs: 1181, 1182, and 1183), and the first binding domain and / or the third binding domain of the antigen-binding molecule of the present invention binds to the human MSLN epitope cluster E1 (SEQ ID NO: 1175, aa positions 296-346 according to Kabat) as determined by the mouse chimeric sequence analysis described herein, preferably not binding to the human MSLN epitope clusters E2 (SEQ ID NO: 1176, aa positions 247-384 according to Kabat), E3 (SEQ ID NO: 1177, aa positions 385-453 according to Kabat), E4 (SEQ ID NO: 1178, aa positions 454-501 according to Kabat), and / or E5 (SEQ ID NO: 1179, aa positions 502-545 according to Kabat).

[0037] In the above aspect, in the context of the present invention, it is also contemplated to provide the antigen-binding molecule of claim 1, wherein the first target cell surface antigen and / or the second target cell surface antigen is human CDH3 (SEQ ID NO: 1170), and the first binding domain and / or the third binding domain of the antigen-binding molecule of claim 1 binds to the human CDH3 epitope clusters D2B (SEQ ID NO: 1171, aa positions 253-290 according to Kabat), D2C (SEQ ID NO: 1172, aa positions 291-327 according to Kabat), D3A (SEQ ID NO: 1173, aa positions 328-363 according to Kabat), and D4B (SEQ ID NO: 1174, aa positions 476-511 according to Kabat), preferably D4B (SEQ ID NO: 1174, aa positions 476-511 according to Kabat) as determined by the mouse chimeric sequence analysis described herein.

[0038] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, wherein both the second binding domain and the fourth binding domain (CD3 binding domain) are (i.) about 1.2×10 measured by surface plasmon resonance (SPR). -8An affinity lower than that characterized by the KD value of M, or (ii.) having an affinity characterized by the KD value of M as measured by SPR. -8 having an affinity characterized by the KD value of M.

[0039] In said aspect, it is also contemplated to provide an antigen-binding molecule in the context of the present invention, wherein the second binding domain and the fourth binding domain (CD3 binding domain) have an affinity of about 1.0×10 -7 ~5.0×10 -6 M, preferably about 1.0 to 3.0×10 -6 M, more preferably about 2.5×10 -6 having an affinity characterized by the KD value of M.

[0040] In said aspect, it is also contemplated to provide an antigen-binding molecule in the context of the present invention, wherein the second binding domain and the fourth binding domain (CD3 binding domain) have an affinity of about 1.0×10 -7 ~5.0×10 -6 M, preferably about 1.0 to 3.0×10 -6 M, more preferably about 2.5×10 -6 having an affinity characterized by the KD value of M.

[0041] In said aspect, it is also contemplated to provide an antigen-binding molecule in the context of the present invention, and each of the second binding domain and the fourth binding domain (CD3 binding domain) individually has an activity at least 10-fold, preferably at least about 50-fold, or more preferably at least about 100-fold lower than that of a single CD3 binding domain comprising a VH according to SEQ ID NO: 43 and a VL according to SEQ ID NO: 44 (i.e., in the context of single targeting as opposed to the context of dual targeting).

[0042] In said aspect, it is also contemplated to provide an antigen-binding molecule in the context of the present invention, and the second domain and the fourth domain are effector binding domains that bind to a CD3ε chain comprising or consisting of a VH region linked to a VL region. i) The VH region is the CDR-H1 sequence of X1YAX2N (X1 is K, V, S, G, R, T, or I; X2 is M or I); the CDR-H2 sequence of RIRSKYNNYATYYADX1VKX2 (X1 is S or Q; X2 is D, G, K, S, or E); and the CDR-H3 sequence of HX1NFGNSYX2SX3X4AY (X1 is G, R, or A; X2 is I, L, V, or T; X3 is Y, W or F; X4 is W, F or Y) and includes; ii) The VL region is the CDR-L1 sequence of X1SSTGAVTX2X3X4YX5N (X1 is G, R, or A; X2 is S or T; X3 is G or S; X4 is N or Y; X5 is P or A); the CDR-L2 sequence of X1TX2X3X4X5X6 (X1 is G or A; X2 is K, D, or N; X3 is F, M or K; X4 is L or R; X5 is A, P, or V; X6 is P or S); and the CDR-L3 sequence of X1LWYSNX2WV (X1 is V, A, or T; X2 is R or L) and includes; iii) One or more of the CDR sequences of the VH region of i) and / or the VL region of ii) include one amino acid substitution selected from X24V and X24F in CDR-H1; D15 and X116A in CDR-H2; H1, X12E, F4, and N6 in CDR-H3; and X11L and W3 in CDR-L3 or a combination thereof.

[0043] In the above aspect, it is also contemplated to provide an antigen-binding molecule in the context of the present invention. The second binding domain and the fourth binding domain include a VH region comprising CDR-H1, CDR-H2, and CDR-H3 selected from SEQ ID NOs: 37-39, 45-47, 53-55, 61-63, 69-71, 436-438, 1126-1128, 1136-1138, 1142-1144, 1148-1150, and 1217-1219, and a VL region comprising CDR-L1, CDR-L2, and CDR-L3 selected from SEQ ID NOs: 40-42, 48-50, 56-58, 64-66, 72-74, 439-441, 1129-1131, 1139-1141, 1145-1147, 1151-1153, and 1220-1222, preferably 61-63 and 64-66 or 1217-1219 and 1220-1222.

[0044] In the above aspect, it is also contemplated to provide an antigen-binding molecule in the context of the present invention. The second binding domain and the fourth binding domain include a VH region selected from SEQ ID NOs: 43, 51, 59, 67, 75, 442, 1132, and 1223, preferably 67 or 1223.

[0045] In the above aspect, it is also contemplated to provide an antigen-binding molecule in the context of the present invention. The second binding domain and the fourth binding domain include a VL region selected from SEQ ID NOs: 44, 52, 60, 68, 76, 443, 1133, and 1224, preferably 68 or 1224.

[0046] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, wherein the second binding domain and the fourth binding domain comprise a VH region selected from SEQ ID NOs: 43, 51, 59, 67, 75, 442, 1132 and 1223, preferably 67, and a VL region selected from SEQ ID NOs: 44, 52, 60, 68, 76, 443, 1133 and 1224, preferably 68. When the VH region is 1132 and the VL region is 1133, the second binding domain and / or the fourth binding domain further comprise, as an scFab domain, a CH1 domain of SEQ ID NO: 1134 and a CLK domain of SEQ ID NO: 1135. The VH and VL regions are preferably linked to each other by a linker selected from SEQ ID NOs: 1, 3 and 1125. Alternatively, the VH of VH-CH1 of the second domain and the fourth domain is SEQ ID NO: 1223, the CH1 of VH-CH1 of the second domain and the fourth domain is SEQ ID NO: 1224, the VL of VL-CL of the second domain and the fourth domain is SEQ ID NO: 1225, and the CL of VL-CL of the second domain and the fourth domain is SEQ ID NO: 1226. In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, wherein the first (target) binding domain and / or the third (target) binding domain binds to CDH3, with SEQ ID NO: 1154 as CDR-H1 (the number after X in X1 (the number after "X" indicates the order of "X" in the corresponding amino acid sequence in the N-to-C direction of the sequence listing) is S or N; X2 is Y or S; X3 is P or W; X4 is I or M; X5 is Y, N or H); SEQ ID NO: 1155 as CDR-H2 (X1 is K, V, N or R; X2 is A, D, R, Y, S, W or H; X3 is Y, S, P, G or T; X4 is S, G or K; X5 is A, V, D, K, G, or T; X6 is A, V, D, K, S, G or H; X7 is Y, G, or E; X8 is K, I, or N; X9 is A, S, or N; X10 is S, Q or G; X11 is S or K; X12 is F or V; X13 is K or Q); and SEQ ID NO: 1156 as CDR-H3 (X1 is F or Q; X2 is R, K, S or W; X3 is G or D;X4 is Y, P or R; X5 is R, S, G, N or T; X6 is Y, A or H; X7 is F, L or M; X8 is A or V; X9 is Y or V), and includes a VH region; the first (target) binding domain and / or the third (target) binding domain binds to CDH3, and has SEQ ID NO: 1158 as CDR-L1 (X1 is K or R; X2 is A or S; X3 is Q, D, S, G or E; X4 is S, D or N; X5 is V, L or I; X6 is K, Y, S or H; X7 is S or N; X8 is F, L or M; X9 is A, N or H); SEQ ID NO: 1159 as CDR-L2 (X1 is Y, G, W, N; X2 is T or A; X3 is S or K; X4 is T, N or R; X5 is L or R; X6 is E, A, V or H; X7 is S or E); and SEQ ID NO: 1160 as CDR-L3 (X1 is Q or V; X2 is Q, N or H; X3 is F, L, Y, W, N or H; X4 is A, D, Y, S or N; X5 is Q, R, S, G, W or M; X6 is T, Y or F; X7 is F, Y or L), and includes a VL region.;

[0047] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, wherein the first (target) binding domain and / or the third (target) binding domain binds to MSLN, and the sequence number 1162 as CDR-H1 (where X1 (the number after "X" indicates the order of "X" in the corresponding amino acid sequence from N to C in the sequence listing) is S, G or D; X2 is Y, A, G or F; X3 is I, W, or M; X4 is V, S, G, T, or H); the sequence number 1163 as CDR-H2 (where X1 is A, S, N, W, Y, or V; X2 is Y, S or N; X3 is Y, G, P, or S; X4 is D, H, S, or N; X5 is G or S; X6 is E, G or S; X7 is G, S, N, F, T or Q; X8 is S, W, K, D, I or T; X9 is Y or N; X10 is A or N; X11 is A, P, N, D, E, I or Q; X12 is D, A, S or K; X13 is V, L, or F; X14 is K or Q; X15 is G or S); and the sequence number 1164 as CDR-H3 (where X1 is D, E or V; X2 is R, G, or E; X3 is Y, A, or N; X4 is S, Y, V, or H; X5 is A, P, F, Y, or H; X6 is R or S; X7 is E or G; X8 is Y or L; X9 is R, Y or L; X10 is Y or G; X11 is D or Y; X12 is R, Y, or F; X13 is M, S, F, D or Y; X14 is A, G, S, or T; X15 is L, M, or F; X16 is Y, I or V) are included; the first (target) binding domain and / or the third (target) binding domain binds to MSLN, and the sequence number 1166 as CDR-L1 (where X1 is A or S; X2 is G or S; X3 is E or Q; X4 is G, S or K; X5 is I, L, V or F; X6 is R, G or S; X7 is D, S, N or T; X8 is A, S, K or T; X9 is Y or W; X10 is V or L; X11 is Y or A); the sequence number 1167 as CDR-L2 (where X1 is A, G or Q; X2 is A or S;X3 is S or T; X4 is G, S, K, I or T; X5 is R or L; X6 is A, P or Q; X7 is S or T); and a VL region comprising the sequence of SEQ ID NO: 1168 as CDR-L3 (X1 is A or Q; X2 is Y, S, A, or T; X3 is G, E, Y, H or Q; X4 is A or S; X5 is S, T or F; X6 is -, P or T; X7 is R, A, L or F; X8 is V or T).;

[0048] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, wherein the first (target) binding domain and / or the third (target) binding domain binds to CDH3, and SEQ ID NO: 1157 (the numbers after "X" indicate the order of "X" in the corresponding amino acid sequence in the N- to C-direction of the sequence listing) (X1 is Q or E; X2 is V, L; X3 is Q, E; X4 is A or G; X5 is G or E; X6 is V or L;X7 is K or V, X8 is K or Q, X9 is A or G, X10 is V or L, X11 is K or R, X12 is V or L, X13 is A or K, X14 is Y or F, X15 is T or S, X16 is T or S, X17 is S or N, X18 is Y or S, X19 is P or W, X20 is I or M, X21 is Y, N or H, X22 is T or A, X23 is Q or K, X24 is V or M, X25 is S or G, X26 is K, V, N or R, X27 is A, D, R, Y, S, W or H, X28 is Y, S, P, Gr or T, X29 is S, K, or G, X30 is A, V, D, K, or T, X31 is A, -, D, K, S, G, or H, X32 is Y, G, or E, X33 is K, I, or N, X34 is A, S, or N, X35 is S, Q, or G, X36 is S or K, X37 is F or V, X38 is Q or K, X39 is F or V, X40 is I or M, X41 is T or S, X42 is V, I or R, X43 is T, K or N, X44 is T, A, S or K, X45 is S or N, X46 is A, V or L, X47 is L or M, X48 is Q or E, X49 is L or M, X50 is S or N, X51 is S or R, X52 is T or R, X53 is A or S, X54 is G, D, or E, X55 is T or S, X56 is T, K, or R, X57 is S, Q, W, or R, X58 is -, D, or G, X59 is Y, P, or R, X60 is F, S, G, N or T, X61 is Y, A, or H, X62 is A, -, or V, X63 is F or M, X64 is Y or V; X65 is T, L or M) VH region and; SEQ ID NO: 1161 (X1 is D or E; X2 is Q or V; X3 is L, M; X4 is A, S or D; X5 is F, S or T; X6 is A, S; X7 is A, V; X8 is P, V, L; X9 is D, E; X10 is A, V;X11 is I, L; X12 is T, S, N; X13 is K, R; X14 is A, S; X15 is Q, D, S, G or E; X16 is S, D, N; X17 is V, I or L; X18 is -, K, Y, S or H; X19 is S, N; X20 is F, L, M; X21 is A, N, H; X22 is K, Q; X23 is A, P, V; X24 is K, R; X25 is I, V; X26 is Y, G, W, N; X27 is T, A; X28 is S, K; X29 is T, N, R; X30 is L, R; X31 is E, A, V, H; X32 is S, E; X33 is A, S, V, D; X34 is D, E; X35 is T, K; X36 is S, R; X37 is A, S, P; X38 is F, V; X39 is A, G; X40 is T, V; X41 is Q, V; X42 is Q, N, H; X43 is F, L, Y, W, N, H; X44 is A, D, Y, S, N; X45 is Q, R, S, G, W, M; X46 is F, Y, T; X47 is F, Y, L; X48 is V, L; X49 is D or E) and includes the VL region (all aas at each position, even if not explicitly stated, alternatively mean "or").;

[0049] In the foregoing aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, wherein the first (target) binding domain and / or the third (target) binding domain binds to MSLN, and SEQ ID NO: 1165 (the numbers after "X" indicate the order of "X" in the corresponding amino acid sequence in the N-to-C direction of the sequence listing) (X1 is E, Q; X2 is V, L, Q; X3 is E, Q; X4 is A, G, P; X5 is E, G; X6 is V, L; X7 is V, K; X8 is K, Q; X9 is G, S; X10 is E, A, G, R; X11 is S, T; X12 is V, L; X13 is R, S, K; X14 is V, L; X15 is S, T; X16 is A, K, T; X17 is A, V; X18 is Y, I, F; X19 is S, T; X20 is S, F; X21 is S, T; X22 is D, G, S; X23 is Y, G, A, F; X24 is I, W, M; X25 is G, S, V, T, H; X26 is I, V; X27 is A, P; X28 is M, K, Q; X29 is G, C; X30 is I, M, V, L; X31 is A, G, S; X32 is A, S, N, W, Y, V; X33 is Y, S, N; X34 is Y, G, P, S; X35 is D, H, S, N; X36 is G, S; X37 is E, G, S; X38 is G, S, N, F, T, Q; X39 is S, K, W, D, I, -, T; X40 is Y, N; X41 is A, N; X42 is A, P, N, E, D, I, Q; X43 is D, A, S, K; X44 is V, L, F; X45 is K, Q; X46 is G, S; X47 is V, F; X48 is I, M; X49 is S, T; X50 is R, V; X51 is N, T; X52 is A, S; X53 is I, K; X54 is S, N; X55 is S, T, Q; X56 is A, L, F; X57 is Y, S, F; X58 is L, M; X59 is E, K, Q; X60 is M, L; X61 is S, N; X62 is R, S; X63 is V, L; X64 is R, T;X65 is A, S; X66 is D, A, E; X67 is R, K; X68 is D, E, V, L; X69 is E, R, G, P; X70 is R, A, N, Y; X71 is G, S, Y, V, H; X72 is A, P, F, D, Y; X73 is R, G; X74 is M, R, S, D; X75 is E, G; X76 is Y, L; X77 is Y, F; X78 is Y, S, F; X79 is A, G, S, T, H; X80 is L, M, F; X81 is Y, I, V; X82 is L, M, T) and the VH region of SEQ ID NO: 1169 (the number after "X" indicates the order of "X" in the corresponding amino acid sequence from N to C in the sequence listing) (X1 is E, S, D; X2 is Y, I, L; X3 is E, -, V, T; X4 is V, L, M; X5 is P, S; X6 is G, S; X7 is S, T; X8 is V, L; X9 is A, V, L; X10 is P, V; X11 is E, Q, D; X12 is R, T; X13 is A, V; X14 is S, T; X15 is I, L; X16 is S, T; X17 is A, S; X18 is G, S; X19 is E, Q; X20 is G, S, K; X21 is I, V, L, F; X22 is R, G, S; X23 is D, S, -; X24 is A, S, N, K, T; X25 is Y, W, M; X26 is V, L; X27 is Y, A; X28 is K, Q; X29 is A, S, V; X30 is R, V, K; X31 is V, L; X32 is A, G, Q; X33 is A, S; X34 is S, T; X35 is G, S, K, I, T; X36 is R, L; X37 is A, P, Q; X38 is S, T; X39 is I, V; X40 is E, S, D; X41 is G, N; X42 is N, T; X43 is D, T; X44 is A, F; X45 is R, G, S; X46 is L, T; X47 is E, Q; X48 is A, P; X49 is E, M; X50 is E, F; X51 is D, V, T; X52 is A, Q;X53 consists of Y, S, A, T; X54 consists of G, E, Y, H, Q; X55 consists of A, S; X56 consists of S, T, F; X57 consists of P, T; X58 consists of R, A, L, F; X59 consists of V, T; X60 consists of P, C; X61 consists of V, L; X62 consists of E, T; X63 consists of I, V; X64 consists of L, K) and includes the VL region (all aas for each position, even if not explicitly described, alternatively mean "or").;

[0050] In said aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, and the first (target) binding domain and / or the third (target) binding domain are CDR-H1, CDR-H2 and CDR-H3 selected from SEQ ID NOs: 77-79, 86-88, 95-97, 103-105, 111-113, 119-121, 127-129, 135-137, 143-145, 151-153, 159-161, 168-170, 177-179, 185-187, 194-196, 203-205, 212-214, 221-223, 230-232, 238-240, 334-336, 356-358, 365-367, 376-378, 385-387, and 194, 432 and 196, or any combination of CDR-H1, CDR-H2 and CDR-H3 as disclosed together in Sequence Listing 6, preferably 86-88, 194, ~196 or 1227-1229 and 1237-1239, and includes a VH region containing the same.

[0051] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, and the first (target) binding domain and / or the third (target) binding domain comprises CDR-L1, CDR-L2 and CDR-L3 selected from SEQ ID NOs: 80-82, 89-91, 98-100, 106-108, 114-116, 122-124, 130-132, 138-140, 146-148, 154-156, 162-164, 171-173, 180-182, 188-190, 197-199, 206-208, 215-217, 224-226, 233-235, 241-243, 337-339, 359-361, 368-370, 379-381, 388-390, or any combination of CDR-H1, CDR-H2 and CDR-H3 as disclosed together in Sequence Listing 6, preferably 89-91 and 197-199 or 1230-1232 and 1240-1242, and comprises a VL region.

[0052] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, and the first (target) binding domain and / or the third (target) binding domain comprises a VH region selected from SEQ ID NOs: 83, 92, 101, 109, 117, 125, 133, 141, 149, 157, 165, 174, 183, 191, 200, 209, 218, 227, 236, 244, 340, 362, 371, 382, 391 and 433, preferably 433 and 92 or 1233+1235 and 1243+1245 (VH and CH1 in Fab) for the first binding domain and the third binding domain, respectively.

[0053] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, and the first (target) binding domain and / or the third (target) binding domain comprises a VL region selected from SEQ ID NO: 84, 93, 102, 110, 118, 126, 134, 142, 150, 158, 166, 175, 184, 192, 201, 210, 219, 228, 237, 245, 341, 363, 372, 383, 392, preferably, with respect to the first binding domain and the third binding domain respectively, 200 and 93 or 1234 + 1236 and 1244 + 1246 (VL and CL in Fab).

[0054] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule, and the first (target) binding domain and / or the third (target) binding domain comprises a VL region with increased stability by a single amino acid substitution (E to I) selected from SEQ ID NO: 85, 94, 193, 202, 211, 220, 229, 364, 384, 393, preferably 94 and 202.

[0055] In the above aspect, in the context of the present invention, it is also contemplated to provide an antigen-binding molecule comprising a combination of amino acid sequences selected from the group consisting of SEQ ID NO: 1259 and 1251, 1247 and 1248, 1249 and 1250, 1254, 1255 and 1253, 1252, 1257, 1253 and 1256, and 1254, 1258, 1253 and 1256.

[0056] In a second aspect, in the context of the present invention, it is further contemplated to provide a polynucleotide encoding an antigen-binding molecule of the present invention, preferably selected from SEQ ID NO: 1070 - 1072 and 1074.

[0057] In a third aspect, in the context of the present invention, it is also contemplated to provide a vector comprising the polynucleotide of the present invention.

[0058] In a fourth aspect, in the context of the present invention, it is further contemplated to provide a host cell transformed or transfected with a polynucleotide or vector of the present invention.

[0059] In a fifth aspect, in the context of the present invention, it is also contemplated to provide a process for generating an antigen-binding molecule of the present invention, the process comprising culturing a host cell of the present invention under conditions that allow expression of the antigen-binding molecule, and recovering the generated antigen-binding molecule from the culture.

[0060] In a sixth aspect, in the context of the present invention, it is further contemplated to provide a pharmaceutical composition comprising an antigen-binding molecule of the present invention or an antigen-binding molecule generated according to the process of the present invention.

[0061] In the said aspect, in the context of the present invention, it is also contemplated that the pharmaceutical composition is stable at about -20°C for at least 4 weeks.

[0062] In the context of the present invention, it is further contemplated to provide an antigen-binding molecule of the present invention or an antigen-binding molecule generated according to the process of the present invention for use in the prevention, treatment, or amelioration of a disease selected from a proliferative disease, a neoplastic disease, cancer, or an immune disorder.

[0063] In the said aspect, in the context of the present invention, it is also contemplated that the disease is preferably acute myeloid leukemia (AML), non-Hodgkin lymphoma (NHL), non-small cell lung cancer (NSCLC), pancreatic cancer, and colorectal cancer (CRC). In a seventh aspect, in the context of the present invention, a method for the treatment or amelioration of a proliferative disease, the method comprising administering to a subject in need thereof a molecule comprising at least one polypeptide chain, the molecule being (i.) preferably, a first binding domain comprising a paratope that specifically binds to a first target cell surface antigen (e.g., TAA1), (ii.) preferably, a second binding domain comprising a paratope that specifically binds to an extracellular epitope of the human (and preferably Macaca) CD3ε chain. (iii.) Preferably, a third binding domain comprising a paratope that specifically binds to a second target cell surface antigen (e.g., TAA2), and (iv.) Preferably, a fourth binding domain comprising a paratope that specifically binds to an extracellular epitope of the human (and preferably Macaca) CD3ε chain comprising the first binding domain and the second binding domain form a first bispecific entity, and the third binding domain and the fourth binding domain form a second bispecific entity, it is further envisioned that the molecule comprises a spacer entity having a molecular weight greater than at least about 5 kDa and / or an amino acid length greater than 50, the spacer entity separating the first bispecific entity and the second bispecific entity by at least about 50 Å (the distance between the centers of mass of the first bispecific entity and the second bispecific entity), and this spacer entity is disposed between the first bispecific entity and the second bispecific entity.

[0064] In said aspect, in the context of the present invention, a method of addressing pathologically relevant tissues and disease-related targets significantly co-expressed on one or more physiological tissues by providing a multichain multi-target bispecific antigen-binding molecule of the type described herein, wherein the molecule addresses (i.) a target expressed on both the disease-related tissue and the physiological tissue, and (ii.) a further target related to the disease but not expressed on the physiological tissue of (i.), and preferably such a target is MSLN, to avoid the formation of intraperitoneal adhesions and / or fibrosis is also envisioned.

[0065] In said aspect, in the context of the present invention, the disease is preferably a neoplastic disease, cancer, or immune disorder, and the method comprises administering to a subject in need thereof an antigen-binding molecule of the present invention, or an antigen-binding molecule produced according to the process of the present invention, and it is also envisioned that the disease is preferably acute myeloid leukemia, non-Hodgkin lymphoma, non-small cell lung cancer, pancreatic cancer, and / or colorectal cancer.

[0066] In the above aspect, in the context of the present invention, it is also contemplated that TAA1 and TAA2 are preferably selected from EpCAM and MSLN, MSLN and EpCAM, MSLN and CDH3, CDH3 and MSLN, FLT3 and CLL1, and CLL1 and FLT3.

[0067] In an eighth aspect, in the context of the present invention, it is also contemplated to provide a kit comprising an antigen-binding molecule of the present invention, or an antigen-binding molecule produced according to the process of the present invention, a polynucleotide of the present invention, a vector of the present invention, and / or a host cell of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0068]

Figure 1

Figure 2-1

Figure 2-2

Figure 3-1

Figure 3-2

Figure 4-1

Figure 4-2

Figure 4-3

Figure 4-4

Figure 4-5

Figure 5-1

Figure 5-2

Figure 5-3

Figure 5-4

Figure 5-5

Figure 6-1

Figure 6-2

Modes for Carrying Out the Invention

[0069] In the context of the present invention, there is provided a multichain multi-target bispecific molecule comprising at least five characteristic structural entities, namely, (i.) a first domain that binds to a target cell surface antigen (e.g., a first tumor-associated antigen, TAA), (ii.) a second domain that binds to an extracellular epitope of the human (and preferably non-human, e.g., Macaca) CD3ε chain (the first binding domain and the second binding domain together form a first bispecific entity), (iii.) a spacer that links but separates the first bispecific entity and a second bispecific entity, (iv.) a third domain that binds to the same or preferably a different target cell surface antigen (e.g., a second TAA), and (v.) a fourth domain that binds to an extracellular epitope of the human (and preferably non-human, e.g., Macaca) CD3ε chain. Molecules of the form of the present invention typically exhibit the advantage of being characterized by avidity-derived potency and specificity from two targets co-expressed on the target cells, thereby typically resulting in a reduction of cytokine release (and related clinically significant side effects such as CRS) that is not desirable, while at the same time ensuring effective anti-tumor activity, preferably against solid tumors such as colorectal cancer, non-small cell lung cancer, and pancreatic cancer.

[0070] The fact that the bispecific (T cell engaging) multichain multitarget (antigen binding molecule) molecules according to the invention bring about a dual avidity effect on both the target cell conjugate side and the effector cell conjugate side by virtue of those particular formats that result in efficient and complementary target cell killing is a surprising finding in the context of the present invention. This effect is promoted by molecular formats that specifically target two (different) antigens on one target cell, such as cancer cells, and, in contrast, non-target cells are not targeted as significantly, but at the same time mediate a strong T cell response against said target cells. By being able to address two target antigens simultaneously, the likelihood of targeting target cells associated with a disease rather than physiological cells is very high when two TAAs typically associated with the target cells associated with the disease are selected. Thus, the T cell engaging multichain multitarget molecules according to the invention provide both improved efficacy and safety with respect to existing bispecific antibodies or antibody-derived constructs that engage T cells. The above-mentioned advantageous properties are preferably achieved by the fact that the multichain multitarget bispecific molecules of the invention contain two bispecific entities each containing a target binding domain and an effector (CD3) binding domain that are able to act in a pathophysiological environment while simultaneously complementing each other and without (e.g., sterically) interfering with each other. The above-mentioned interaction of the two bispecific entities within one multichain multitarget bispecific molecule of the invention means that the target binding domain (e.g., the first domain) and the effector CD3 binding domain (e.g., the second domain) of the first bispecific entity interact with their corresponding binding partners to form a cytolytic synapse between the target cell and the T cell without interfering with the interaction with the target binding domain (e.g., the third domain) and the effector domain (e.g., the fourth domain) of the second bispecific entity. However, in order to provide the desired action, and as a result, the therapeutic function, preferably both target binding domains of both the first bispecific entity and the second bispecific entity need to bind to their corresponding targets in order to fully involve the effector CD3 binding domains of the first bispecific entity and the second bispecific entity.Furthermore, it was a surprising discovery that two respective bispecific entities must be functionally preserved by being structurally separated in a molecular form in a specific manner in order to benefit from the dual avidity effect required to achieve the extraordinary effectiveness and implied safety described herein. In particular, it was surprising that two bispecific entities, including a target binding domain and a CD3 binding domain, did not need to be present in one chain at the N-terminus and C-terminus of a (central) spacer in order to be structurally arranged to act as described herein. Both the target conjugate and / or the CD3 conjugate of one or both bispecific entities can be Fabs, i.e., each can include a chain. Even more surprisingly, the spacer can be double-stranded and preferably in the form of a hetero Fc. In such a case, the bispecific entity simultaneously holds the two domains of each bispecific entity in place so that they act together, and isolates the two bispecific entities from each other so that they act without interfering with each other, provided that the two domains are not present on the same chain but are held in proximity together by a four-part spacer (e.g., hetero Fc). By using this spacer and domain arrangement, and preferably two different TAA binding domains that are two low-affinity CD3 binding domains (preferably binding to CD3ε), a surprising technical effect is achieved of improving the selectivity of target cells and reducing the risk of serious side effects, namely unwanted cytokine release. At the same time, the multichain molecules of the present invention can be produced well with respect to both yield and purity.

[0071] In addition to, or instead of, the improved specificity and thus improved safety described herein, as a secondary effect, the possibility of targeting target cells such as cancer cells with a multichain multi-target antigen-binding molecule compared to a single target molecule is greatly increased because antigen deficiency occurs in such target cells, and thus tumor escape from an effective anti-tumor therapy is facilitated because one antigen effective for targeting remains on the cells where antigen avoidance has occurred. The effect is achieved, for example, when both CD3 binders have low affinity, in that the activity is increased compared to a molecule that contains only one CD3 binder and / or target binder and does not contain two bispecific entities that are linked but spaced apart, such as a CD3 binding domain containing VH and VL of SEQ ID NO: 67 and 68 linked by the linker of SEQ ID NO: 1 or 3, respectively.

[0072] The above - described findings underlying the present invention are surprising when viewed in light of the teachings of the prior art. For example, antigen - binding formats each containing two or more target - binding domains and effector - binding domains are known in the art (e.g., the Adaptir™ format). However, in such formats, two bispecific entities that can interact individually with their corresponding targets and effectors and function together simultaneously are not provided. As a result, a significant selectivity gap in the advantages of multi - chain multi - target molecules is effectively provided, such that a dual - avidity effect cannot be achieved on both the target - conjugate side and the effector - conjugate side to the extent of effectively providing a large selectivity gap in the advantages of multi - chain multi - target molecules. According to the present invention, the two bispecific entities need to be separated from each other by a certain distance, preferably at least 50 Å, more preferably at least 60, 70, 80, 90 or at least 100 Å. The distance [Å] indicated between the two bispecific entities is typically understood, in the context of the present invention, as the distance between the centers of mass of the two bispecific entities respectively. Generally, the center of mass (COM) of a mass distribution in space (here, a bispecific entity comprising a binding domain that binds to a target cell - surface antigen and a binding domain that binds to an extracellular epitope of the human (and preferably Macaca) CD3ε chain, where both binding domains are preferably selected from Fab or single - domain formats, preferably scFv and scFab formats, and are linked by a peptide linker) is understood as a specific point at which the sum of the weighted relative positions of the distributed mass is zero. This distance is typically determined by modeling the molecule using generally recognized modeling programs (MD / visualization software), such as PyMOL (PyMOL Molecular Graphics System, version 2.3.3, Schroedinger, LLC.) based on the input structure of the COM and typically an ensemble of snapshot structures from MD simulations. The mass of each atom is typically part of a basic "force field" as generally known in the art. Alternatively and / or in addition, the distance can be determined by crystallography, cryo - electron microscopy, or nuclear magnetic resonance analysis techniques.

[0073] Typical methods for obtaining distances by molecular modeling as shown in the present invention are as follows: 1) Obtaining the atomic-level structure of a complete bispecific antigen-binding molecule. The source of the structure is a. Protein X-ray crystallography that enables visualization of the amino acid backbone and side chains, preferably with a resolution of less than 5 Å; b. Cryo-electron microscopy (cryo-EM) that enables visualization of the amino acid backbone and side chains, preferably with a resolution of less than 5 Å; c. In silico homology modeling of the entire molecule based on a single highly homogeneous crystal and / or cryo-EM structure (preferably with a sequence identity of more than 60%); d. In silico homology modeling including associating two or more experimental structures and can be selected from the group consisting of. The structure is preferably identical or highly homologous (preferably with a sequence identity of more than 60%) to the domain found in the complete bispecific antigen-binding molecule. When there is no experimental linker conformation, it is preferable to refine the model by explicit solvent molecular dynamics (MD) simulation (preferably with a simulation length of at least 100 ns, except when energy convergence is required more quickly). The simulation is performed by state-of-the-art software (e.g., Schrodinger, Amber, Gromacs, NAMD, or equivalents) with parameters corresponding to room temperature and pressure. No artificial forces are applied during the simulation (i.e., preferably methods such as metadynamics or steered molecular dynamics are excluded). Similarly, it is preferable not to impose artificial geometric constraints on the molecule. 2) Identifying the center of mass (COM) of the relevant molecular domain. This identification is typically performed with the MD software used or visualization tools such as PyMOL or VMD. The center of mass can be defined as the pseudo atom or non-hydrogen atom closest to the true COM. The domain linker is typically not considered part of the domain. 3) Use the same software to report the distance between two COMs (in angstroms, Å). When refining the homology model using MD simulations (as described in 1d), the median distance for snapshots of multiple simulations is reported. To further reduce potential inaccuracies in the initial model, at least the first 10% of the simulation, preferably up to 50% if the signal changes significantly, is skipped when calculating the median distance between COMs and when extracting snapshots for visualizing the MD simulation.

[0074] Unless otherwise indicated, the distance [Å] in the context of the present invention is the median distance as determined by MD simulations.

[0075] The preferred distance between a first bispecific entity and a second bispecific entity as disclosed herein is facilitated by a spacer entity (i.e., a spacer) between the two bispecific entities that separates the two bispecific entities and holds them in separate positions. The spacer is of a certain size, preferably at least greater than 5 kDa, more preferably at least about 10, 15, 20, 25, 30, 35, 40, 45 or even at least 50 kDa, thereby preventing undesirable interactions between the two separate bispecific entities. The preferred range of the molecular size of the spacer is from about 15 to 200 kDa, preferably from about 15 to 150 kDa, to facilitate the isolation of the two bispecific entities according to the invention and to maintain a high overall activity of the molecule. Typically, a spacer that is too large, e.g., a spacer larger than about 200 kDa, may affect the ability of the two bispecific entities to bind to two target surface structures on the same target cell, and as a result, the overall activity of the molecule against the target cell may be reduced. Thus, the typical maximum preferred size with respect to the molecular weight of the spacer is about 200 kDa, preferably about 150 or 120 kDa, even more preferably about 100 kDa. A typical spacer of the maximum preferred size is a bispecific scFc domain as disclosed herein of about 105.7 kDa (two scFcs linked to each other forming one large single-chain spacer).Exemplary-sized spacers that typically sufficiently isolate two bispecific entities are the PSI region of the Met receptor of about 5.3 kDa, ubiquitin of about 8.6 kDa, fibronectin type III domain from tenascin of about 10.1 kDa, SAND domain of about 11 kDa, β2-microglobulin of about 11.9 kDa, Tim-3 (aa 24-130) of about 12.2 kDa, MiniSOG of about 13.3 kDa, a 12.1 kDa SpyCatcher associated with a 1.7 kDa SpyTag preferably linked together by isopeptide bond formation to form a 13.8 kDa two-chain spacer, VHH antibody llama domain of about 14 kDa, PD-1 binding domain from human programmed cell death 1 ligand (PDL1) of about 14.4 kDa, granulocyte macrophage colony-stimulating factor (GM-CSF) of about 14.5 kDa, interleukin-4 of about 15 kDa, interleukin-2 of about 15.4 kDa, CD137L (4-1BBL; TNFSF9) extracellular domain of about 17.7 kDa, programmed cell death protein 1 (PD-1) of about 16.6 kDa, green fluorescent protein (GFP) of about 26.3 kDa, a single-chain Fc region (scFc) as described herein of about 52.8 kDa (about 54.6 kDa including N-terminal and C-terminal linkers (G4S)3 respectively), human serum albumin (HSA) of about 66.5 kDa (about 68.3 kDa including N-terminal and C-terminal linkers (G4S)3 respectively), and a double scFc of about 105.7 kDa (two scFcs linked to each other to form one larger single-chain spacer, about 107.5 kDa including N-terminal and C-terminal linkers (G4S)3 respectively). Generally, the more rigid the spacer, the shorter the median distance required, and if not, a safety margin for flexible spacers needs to be included.

[0076] Also, spacers that are preferred in the context of the present invention, such as globular domains, typically have N- and C-termini that are not too close spatially to each other in order to effectively separate two bispecific entities according to the present invention. In this regard, the spacer typically exhibits a distance between the N- and C-termini that is significantly greater than 10 Å. A distance between the N- and C-termini of a spacer of about 10 Å or less is considered "close". Thus, a spacer in the context of the present invention preferably has a distance between the alpha-carbon atom of the first amino acid located at the N-terminus and the alpha-carbon atom of the last amino acid at the C-terminus of at least 20 Å, more preferably at least 30 Å, even more preferably at least 50 Å, and that distance typically ensures that the first bispecific entity and the second bispecific entity are separated by at least 50 Å as described herein. Alpha-carbon (α-carbon) is understood herein as a term applied to proteins and amino acids. The alpha-carbon is the backbone carbon preceding the carbonyl carbon atom within the molecule. Thus, reading along the backbone of a typical protein, a sequence such as -[N-Cα-carbonyl C]n- (when reading in the N-to-C direction) would be obtained. The alpha-carbon is where different substituents are attached to different amino acids. That is, the groups hanging off the chain at the alpha-carbon confer diversity to the amino acids. Thus, in the context of the present invention, a spacer is less preferred if the distance between the alpha-carbon atom of the first amino acid located at the N-terminus and the alpha-carbon atom of the last amino acid at the C-terminus is too close, i.e., if the distance is, for example, only about 10 Å, even if it has a size of at least 5 kDa and a length exceeding 50 aa. For example, preferred spacers exhibit typical distances between the alpha-carbon atom of the first amino acid located at the N-terminus and the alpha-carbon atom of the last amino acid at the C-terminus such as: scFc (based on the 5G4S crystal structure): 89 Å, HSA (based on the 5VNW crystal structure): 77 Å, ubiquitin (based on the 1UBQ crystal structure): 37 Å, and SAND (based on the 1OQJ crystal structure): 32 Å.In contrast, HSP70-1 (based on the 3JXU crystal structure) shows a distance of only 9 Å between the alpha-carbon atom of the first amino acid located at the N-terminus and the alpha-carbon atom of the last amino acid at the C-terminus. At the same time, HSP70-1, in the context of the present invention, has a median distance of only about 48 Å between the COMs of the first bispecific entity and the second bispecific entity, which is less than the threshold median distance of 50 Å and is significantly lower than the typical median distance of about 60 - 100 Å between the COMs of two bispecific entities facilitated by preferred spacers such as scFc, HSA, ubiquitin, and SAND. Among them, scFc (SEQ ID NO: 25) is preferred.

[0077] Alternatively, a non-globular but rigid linker can function as a spacer in the context of the present invention that separates two bispecific entities. Such linkers include (PA)25P (SEQ ID NO: 1097) and A(EAAAK)4ALEA(EAAAK)4A (SEQ ID NO: 1096), even when the Mw is less than 5 kDa (4.3 kDa herein) and the amino acid length is only about 50 or less (51 and 46 aa respectively). However, such spacers are typically less preferred than globular domains that preferably further increase the half-life.

[0078] Also, as contemplated in connection with the present invention, the spacer between two bispecific entities is typically a polypeptide comprising more than 50 amino acids, preferably at least about 75, 100, 150, 200, 250, 300, 350, 400, 450 or at least 500 amino acids. The preferred range of the amino acid length of the spacer is from about 100 to 1500 amino acids, preferably from about 100 to 1000 amino acids, more preferably from about 250 to 650 amino acids, to facilitate the isolation of the two bispecific entities according to the present invention. This distance is preferably to maintain a high overall activity of the whole molecule according to the present invention (not necessarily of the individual separated bispecific entities, which may individually have low affinity (and low activity) to enhance the specificity for double positive target cells), typically less than 20 pM, preferably less than 5 pM, more preferably less than 1 pM. Typically, a spacer that is too large, for example, a spacer longer than about 1500 amino acids, may affect the ability of the two bispecific entities to bind to the surface structures of the two targets on the same target cell, and as a result, the overall binding activity of the molecule to the target cell may be reduced. Thus, the typical maximum preferred length of the spacer is about 1500 amino acids, more preferably about 1000 amino acids. Exemplary amino acid lengths of the spacer that sufficiently isolate the two bispecific entities are about (ECD25 - 167) 143aa of PD-1, about 484aa (about 514aa having (G4S)3 at the N-terminus and C-terminus, respectively) of scFc as described herein, about 585aa (about 615aa having (G4S)3 at the N-terminus and C-terminus, respectively) of HSA, and about 968aa (about 998aa having (G4S)3 at the N-terminus and C-terminus, respectively) of bispecific scFc.Additional spacers include ubiquitin of approximately 76 aa, fibronectin type III domain derived from tenascin of approximately 90 aa, SAND domain of approximately 90 or 97 aa, β2-microglobulin of approximately 100 aa, Tim-3 (aa 24-130) of approximately 108 aa, MiniSOG of approximately 115 aa, SpyCatcher of approximately 113 aa associated with SpyTag of approximately 14 aa preferably linked together by isopeptide bond formation to form a 127-dimer spacer, VHH antibody llama domain of approximately 129 aa, PD-1 binding domain derived from human programmed cell death 1 ligand (PDL1) of approximately 126 aa, granulocyte macrophage colony-stimulating factor (GM-CSF) of approximately 127 aa, interleukin-4 of approximately 129 aa, interleukin-2 of approximately 133 aa, CD137L (4-1BBL; TNFSF9) extracellular domain of approximately 167 aa, and green fluorescent protein (GFP) of approximately 238 aa.

[0079] The composition and arrangement of the amino acid sequence of the preferred spacer preferably impart a certain rigidity and are not characterized by high mobility. Rigidity in the context of the present invention typically exists when a spacer with a molecular weight of more than 50 aa and / or exceeding 5 kDa allows the two bispecific entities of the molecule according to the present invention to have a maximum distance between the centers of mass of less than 200% (i.e., twice) the median distance. Thus, a preferred rigid spacer in the context of the present invention does not extend further than about 100% of its median distance, more preferably about 80% or less (each calculated as the distance between the centers of mass of the two bispecific entities). Thus, a preferred rigid spacer in the context of the present invention that separates two bispecific entities by about 100 Å (median distance) does not extend further than 200 Å (maximum distance). For example, the typical median distance between the centers of mass of the bispecific entities of a molecule having the format of the present invention containing scFc (such as SEQ ID NO: 25) as a spacer is about 101 Å. However, the maximum distance in such a case is typically about 182 Å, i.e., about 100% or even slightly about 80% compared to the median distance. Such a spacer is considered rigid in the context of the present invention. In contrast, for example, a molecule containing (G4S) 10 (SEQ ID NO: 8) exhibits a typical median distance of about 48 Å and a maximum distance of about 179 Å. Thus, (G4S) 10Spacers such as do not exhibit high mobility and do not exhibit the preferred spacer rigidity as an advantageous feature according to the present invention. In this regard, the amino acid sequence of the spacer is typically rich in proline and may not be rich in serine and glycine. For example, a spacer that is a folded polypeptide having a secondary structure (e.g., a helical structure) or a tertiary structure forming, for example, a three-dimensional protein domain structure is particularly envisaged, such that their configuration ensures a certain rigidity and preferably confers further advantageous effects such as an extended in vivo half-life of a multichain multi-target bispecific molecule as a therapeutic agent. A typical domain structure includes a hydrophobic core having a hydrophilic surface. In the context of the present invention, a protein having the structure of a globular protein is preferred as the spacer. A globular protein is understood in the context of the present invention to be a spherical ("globular-like") protein and is one of the general types of proteins. A globular protein in the context of the present invention may be characterized by a globin fold. Particular envisaged are PD-1 or HSA domains that are spacers comprising an Fc domain or a part or parts thereof. Also envisaged are spacers comprising a combination of different globular proteins or parts thereof, and it is even more preferred to include an Fc receptor binding function to increase the half-life of the molecules according to the present invention. The format described herein for separating two bispecific entities has characteristic advantages. When there is only one target addressed by the first binding domain, the first domain "uses" only the second domain to engage T cells and does not use the fourth domain, or the third domain uses the fourth domain but does not use the second domain (or only to a much smaller extent by the spacer). When there is only one target, efficient T cell engagement is blocked by the Kd of the preferably low affinity CD3 conjugate as disclosed herein. Thus, selectivity is increased compared to other (bis)targeting molecules.

[0080] In the case where both targets are present, the multichain multi-target bispecific T cell engager of the present invention binds more tightly to the target cell (due to increased avidity) and can engage T cells using both low-affinity CD3-binding domains (such as I2L) of the present invention (also due to increased avidity). For example, a second domain that binds to the CD3 domain on effector T cells and a third domain that binds to the target antigen are less likely to form a cytolytic synapse and thus do not act together as a bispecific entity, and otherwise would result in a less beneficial cytotoxic activity profile. This has the advantage that the first and fourth domains do not remain "unused" (meaning that the effect of double avidity due to the dual binding of the target-binding domain and the effector-binding domain respectively cannot fully utilize the effect). Similarly, the theoretical interaction between the first domain that binds to the target antigen and the fourth domain that binds to the CD3 domain on effector T cells is hindered, thereby ultimately preventing the second and third domains from being used to form cytolytic synapses with the "partner domains" intended for their respective bispecific entities.

[0081] Typically, the advantageous avidity effect conferred by the multichain multi-target bispecific molecule according to the present invention is based on an activity factor or "selectivity gap" indicated by the difference in the activity of the molecule in double-positive cells, i.e., (i.) two different targets of a combination that are overexpressed on the cell type to be targeted and associated with a particular disease, and / or (ii.) target cells having one target at an overexpression level. In either case, the molecule according to the present invention that targets two (preferably different) targets simultaneously is preferably bound to such target cells and results in a more prominent T cell response compared to either non-target cells expressing only one of the two targets or one target at a low expression level. Preferred for the multichain multi-target bispecific molecule of the present invention, for example, a lower EC 50The activity from the perspective of increased cytotoxicity as determined by the value is at least 5-fold, preferably 10-fold, more preferably 30-fold, 50-fold, 80-fold or even 100-fold greater in target cells (e.g., characterized by expressing either two different targets or one target at a high level) than in non-target cells (e.g., characterized by expressing only one of two targets or only one target at a low level). The selectivity gap in the context of the present invention is preferably greater than 100-fold. The selectivity gap (which can also be defined as an activity gap) is contemplated in the context of the present invention to be at least 250-fold, 500-fold, 750-fold or even 1000-fold, whereby the effectiveness and safety of the present multi-chain multi-target bispecific molecule are significantly improved as compared to various forms of single-target bispecific molecules.

[0082] In a further aspect contemplated in the context of the present invention, the double avidity effect conferred by the form of the multi-chain multi-target antigen-binding molecule using low-affinity, preferably both a target antigen conjugate and a CD3 effector conjugate, is further supported. In the context of the present invention, a CD3 conjugate having an affinity with a KD of less than 1.2×10 -8 M is preferred. A CD3 conjugate having an activity 10-fold lower, more preferably 50-fold lower, or even more preferably 100-fold lower than the activity of a CD3 conjugate having a KD of 1.2×10 -8 is particularly preferred. Without wishing to be bound by theory, for example, in order to avoid off-target toxicity and related side effects, two conjugates with relatively balanced, i.e., typically low to moderate affinity, preferably low affinity, are such that (even when only one target on a cell that may be a physiological non-target cell that should not be targeted is bound, it will cause cytolytic activity), the avidity effect is more pronounced when binding to two targets on the same target cell as compared to a conjugate with a mixed affinity, or typically a high affinity.

[0083] Accordingly, the multichain multi-target bispecific antigen-binding molecules according to the invention, which bind to two (preferably different) targets on the target cell and exhibit significant cytotoxic activity, preferably exhibit fewer side effects than single-target bispecific antigen-binding molecules that bring effector T cells and target cells together. This is demonstrated, for example, by a significant reduction in the release of important cytokines, IL-2, IL-6, IL-10, TNFα and IFNγ, which are indicators of side effects at the clinical stage. For example, the release of IL-6 typically decreases when using the multichain multi-target bispecific antigen-binding molecules according to the invention as compared to the corresponding single-target bispecific molecules. As is known in the art, extremely elevated interleukin 6 (IL-6) levels are seen in patients with CRS, and thus IL-6 is thought to play an important role in the pathophysiology of CRS (Shimabukuro-Vornhagen et al. Journal for ImmunoTherapy of Cancer (2018) 6:56). Since CRS is a serious side effect in immunotherapy, such a reduction indicates less CRS at the clinical stage.

[0084] Furthermore, in order to exhibit significant cytotoxic activity, the multichain multi-target bispecific antigen-binding molecules according to the invention that bind to two (preferably different) targets on the target cell preferably show fewer side effects than single-target bispecific antigen-binding molecules from the perspective of toxic tissue damage. It was a surprising discovery that multivalent molecules in the form as described herein show higher tolerance, i.e., higher doses can be administered than the corresponding single-target bispecific molecules without clinical findings such as tissue damage tested by histopathological examination. For example, a dose of 1.5 μg / kg of the MSLN single-target bispecific antigen-binding molecule (SEQ ID NO: 1183) did not show tolerance and resulted in death, while on the other hand, a dose of 0.1 μg / kg showed tolerance. Conversely, the multichain multi-target CDH3-MSLN bispecific molecule according to the invention (SEQ ID NO: 251) showed tolerance at a maximum dose of 1000 μg / kg. The histopathological changes seen with the single-target molecule were generally more severe at a dose of 1.5 μg / kg than those caused by the multichain multi-target molecule at 1000 μg / kg each. Adhesive or irreversible fibrotic changes such as those induced by the single-target molecule were not present after treatment with the multichain multi-target molecule. Thus, the tolerance of the multichain multi-target molecules according to the invention is, for example, 600-fold (histopathology) to, for example, 10,000-fold (tolerance) higher than that of the corresponding single-target molecules, despite comparable in vitro potency against tumor cells. Thus, the multichain multi-target molecules of the invention are particularly suitable in the therapeutic setting for addressing targets that are not only significantly present on disease-related (pathophysiological) tissues but also significantly or predominantly present on physiological tissues that should not be targeted by cytotoxic immunotherapy. This is the case, for example, with MSLN, which is typically expressed in mesothelial cells that form the lining of several body cavities: the pleura (the pleural cavity around the lungs), the peritoneum (the abdominopelvic cavity including the mesentery, omentum, falciform ligament, and serosa of the uterus), and the pericardium (around the heart). Addressing targets such as MSLN by cytotoxic immunotherapy entails a risk of serious side effects such as intraperitoneal adhesions and / or fibrosis. Intraperitoneal adhesions are understood herein as pathological scars formed between intraperitoneal organs.Subsequently, adhesions may occur in the presence of intraperitoneal inflammation, causing the peritoneal surfaces to adhere to each other. Adhesions can cause problems if they restrict the free movement of organs by scarring (Mutsaers S.E., Prele C.M, Pengelly, S., Herrick, S.E. Mesothelial cells and peritoneal homeostasis. Fertil Steril 2016, 106(5)1018 - 1024). Fibrosis is understood herein as a common pathological outcome of several etiological states that cause chronic tissue damage and is typically defined as the excessive deposition of extracellular matrix (ECM) components, leading to the formation of scar tissue over time and ultimately causing organ damage and organ failure (Maurizio Parola, Massimo Pinzani, Pathophysiology of Organ and Tissue Fibrosis, Molecular Aspects of Medicine 2019, (65)1). Accordingly, the present invention also provides a method of addressing patho - physiological tissues and disease - related targets significantly co - expressed on one or more physiological tissues by providing a multichain multi - target bispecific antigen - binding molecule in the form described herein, wherein the molecule addresses (i.) targets expressed on both disease - related and physiological tissues, and (ii.) additional targets related to the disease but not expressed on the physiological tissues of (i.), and preferably provides a method of avoiding the formation of intraperitoneal adhesions and / or fibrosis, such that such targets are MSLN.

[0085] The bispecific antigen - binding molecules according to the present invention are expected to have cross - reactivity against cynomolgus tumor - related antigens such as, for example, CDH3, MSLN, CD20, CD22, FLT3, CLL1, and EpCAM. It is particularly contemplated in the context of the present invention that it is possible to address two targets simultaneously with one multichain multi - target bispecific antigen - binding molecule.

[0086] Alternatively, in addition to the main advantage of increased selectivity as described herein, dual targeting can mitigate the lack of access to one target when it can cause sufficient residual effects by targeting the remaining target. Examples include (i) the presence of a soluble target that "masks" the target on the target cell by binding to the antigen-binding molecule without enabling any therapeutic effect on the remaining molecule, and (ii) antigen deficiency as a driver for tumor escape (reducing target expression on the target cell).

[0087] For example, the multichain multi-target antigen-binding molecules according to the present invention, such as constructs induced against MSLN as TAA1 and CDH3 as TAA2, are useful for the treatment, amelioration or prevention of cancers selected from the group consisting of cancer, particularly lung cancer, head and neck cancer, primary or secondary CNS tumors, primary or secondary brain tumors, primary CNS lymphoma, spinal cord axis tumors, brainstem gliomas, pituitary adenomas, adrenocortical carcinomas, esophageal cancer, colon cancer, breast cancer, ovarian cancer, NSCLC (non-small cell lung cancer), SCLC (small cell lung cancer), endometrial cancer, cervical cancer, uterine cancer, transitional cell carcinoma, bone cancer, pancreatic cancer, skin cancer, cutaneous or intraocular melanoma, hepatocellular carcinoma, cholangiocarcinoma, gallbladder cancer, kidney cancer, rectal cancer, cancer of the anal region, gastric cancer, gastrointestinal tract (stomach, colorectal, and duodenum) cancer, small intestine cancer, cholangiocarcinoma, urethral cancer, renal cell carcinoma, endometrial carcinoma, thyroid cancer, testicular cancer, cutaneous squamous cell carcinoma, melanoma, gastric cancer, prostate cancer, bladder cancer, osteosarcoma, mesothelioma, Hodgkin's disease, non-Hodgkin's lymphoma, chronic or acute leukemia, chronic myelogenous leukemia, lymphocytic lymphoma, multiple myeloma, fibrosarcoma, neuroblastoma, retinoblastoma, and soft tissue sarcoma.

[0088] Thus, it is particularly contemplated in the context of the present invention that multichain multi-target antigen-binding molecules that preferably address two different target cell surface antigens are highly specific for their target cells and are therefore preferably safe in their therapeutic use. Efficacy from the perspective of inhibiting tumor growth has been demonstrated in vivo in a mouse model.

[0089] Target cell surface antigens that are preferred in the context of the present invention are MSLN, CDH3, FLT3, CLL1, EpCAM, CD20, and CD22. Typically, target cell surface antigens in the context of the present invention are tumor-associated antigens (TAAs). CD20, a B-lymphocyte antigen, or CD20, is expressed on the surface of all B cells from the pro-B stage (CD45R+, CD117+) and its concentration gradually increases until maturation. CD22, or surface antigen classification 22, is a molecule belonging to the SIGLEC family of lectins. This molecule is found on the surface of mature B cells and to a relatively low extent on some immature B cells. Fms-like tyrosine kinase 3 (FLT3) is also known as surface antigen classification 135 (CD135), receptor-type tyrosine-protein kinase FLT3, or fetal liver kinase-2 (Flk2). FLT3 is a cytokine receptor belonging to class III of the receptor tyrosine kinase. CD135 is a receptor for the cytokine Flt3 ligand (FLT3L). The FLT3 gene is frequently mutated in acute myeloid leukemia (AML). C-type lectin-like receptor (CLL1) is also known as CLEC12A or MICL. CLL1 contains an ITIM motif in the cytoplasmic tail that can be associated with the signaling phosphatases SHP-1 and SHP-2. Human MICL is expressed as a monomer mainly on myeloid cells, including granulocytes, monocytes, macrophages, and dendritic cells, and is associated with AML. Mesothelin (MSLN) is a 40 kDa protein expressed in mesothelial cells and overexpressed in some human tumors. Cadherin-3 (CDH3), also known as P-cadherin, is a calcium-dependent intercellular adhesion glycoprotein composed of five extracellular cadherin repeat sequences, a transmembrane region, and a highly conserved cytoplasmic tail. CDH3 is associated with some types of tumors. Epithelial cell adhesion molecule (EpCAM) is a transmembrane glycoprotein that mediates Ca2+-independent homotypic cell-cell adhesion in epithelia. EpCAM is oncogenic and is thought to contribute to tumor formation and metastasis of carcinomas.

[0090] Furthermore, in the context of the present invention, it is optionally but preferably envisaged that the multi-chain multi-target antigen-binding molecule is provided in a globular protein structure such as a spacer, preferably a dimerizing Fc domain such as an scFc domain or a hetero Fc, which also increases the half-life of the molecule and enables intravenous administration at a frequency of once a week, once every two weeks, once every three weeks, or even once every four weeks, or less frequently.

[0091] For example, in order to determine the epitopes of preferred multi-chain multi-target antigen-binding molecules according to the present invention derived from CDH3, MSLN or CD20 epitopes, mapping as described herein was performed. Preferred bispecific antigen-binding molecules having a target conjugate for CD20 are derived from all of the epitope clusters E1A, E2B and E2C. An epitope cluster is understood herein as a continuous sequence of amino acids within a target (disclosed herein and defined by their positions according to Kabat) where the entire target conjugate of a multi-chain multi-target bispecific antigen-binding molecule as described herein is essentially no longer bound when a continuous sequence of amino acids of the human target is replaced by the corresponding continuous sequence of amino acids of the mouse target. Thus, the method of epitope clusters is understood herein as mouse chimeric sequence analysis. The method is described, for example, by Muenz et al. Cancer Cell International 2010, 10:44 and was applied as described in detail in the examples for CDH3 and MSLN.

[0092] Preferred epitope clusters are D4B in the case of CDH3 as described herein and E1 in the case of MSLN as described herein. As illustrated in the Examples, the selectivity gap of the multichain multi-target bispecific antigen-binding molecules of the invention is typically even greater (compared to comparable single-target bispecific antigen-binding molecules), and thus it is more preferred when the MSLN target conjugate addresses the E1 epitope cluster and the CDH3 target conjugate addresses the D4B epitope cluster. Addressing other epitope clusters also results in a surprisingly high selectivity gap and associated advantages in terms of efficacy and tolerability / safety, particularly for molecules with a high selectivity gap and thus including target conjugates that address E1 and D4B. Such molecules include, for example, molecules having MSLN target conjugates comprising CDR H1-H3 of SEQ ID NOs: 774-776 and CDR L1-L3 of 777-779 (and corresponding VH and VL of 780 and 781), CDR H1-H3 of SEQ ID NOs: 782-784 and CDR L1-L3 of 785-787 (and corresponding VH and VL of 788 and 789), CDR H1-H3 of SEQ ID NOs: 806-808 and CDR L1-L3 of 809-811 (and corresponding VH and VL of 812 and 813), CDR H1-H3 of SEQ ID NOs: 838-840 and CDR L1-L3 of 841-843 (and corresponding VH and VL of 844 and 845), CDR H1-H3 of SEQ ID NOs: 862-864 and CDR L1-L3 of 865-867 (and corresponding VH and VL of 868 and 869), CDR H1-H3 of SEQ ID NOs: 894-896 and CDR L1-L3 of 897-899 (and corresponding VH and VL of 900 and 901), CDR H1-H3 of SEQ ID NOs: 950-952 and CDR L1-L3 of 953-955 (and corresponding VH and VL of 956 and 957), CDR H1-H3 of SEQ ID NOs: 1030-1032 and CDR L1-L3 of 1033-1035 (and corresponding VH and VL of 1036 and 1037), or CDR H1-H3 of SEQ ID NOs: 86-88 and CDR L1-L3 of 89-91 (and corresponding VH of 92 and VL of 93 or 94).Preferred examples of CDH3 conjugates that bind to preferred DB4 epitope clusters include CDR H1-H3 of SEQ ID NOs: 194, 432, and 196 and CDR L1-L3 of 197-199 (and corresponding VH and VL of 433 and 200). Further target conjugates that preferably bind to the preferred epitope cluster of D4B are specified herein, for example, as CH3 15-E11 CC and CH3 24-D7 CC.

[0093] It is particularly surprising that the multichain multi-target antigen-binding molecules according to the invention can preferably bind to two different targets simultaneously. Simultaneous binding has been demonstrated herein for several targets. However, this is surprising considering the typical distances between targets. For example, CD20 contains only two small extracellular domains of 6aa and 47aa. In contrast, CD22 contains an extracellular domain of 7Ig domain length containing 676aa. However, despite the significant differences in extracellular size and composition, the multichain multi-target antigen-binding molecules according to the invention can successfully cope simultaneously with both CD20 and CD22, which are TAAs, due to the advantages of high efficacy and low toxicity.

[0094] Preferred multi-chain multi-target antigen-binding molecules are expected in the context of the present invention to not only exhibit a preferred ratio of cytotoxicity to affinity, but also to exhibit sufficient stability characteristics to facilitate practical handling during formulation, storage, and administration of said constructs. Sufficient stability is characterized, for example, by a high monomer content after standard preparation (i.e., non-aggregated and / or non-associated native molecules) of at least 65%, more preferably at least 70%, even more preferably at least 75% when measured by preparative size exclusion chromatography (SEC). Also, for example, the turbidity measured at 340 nm as light absorption at a concentration of 2.5 mg / ml should preferably be 0.025 or less, more preferably 0.020, in order to conclude that essentially no unwanted aggregates are present. Advantageously, a high monomer content is maintained after incubation under stress conditions such as freeze / thaw or incubation at 37 °C or 40 °C. Furthermore, multi-chain multi-target antigen-binding molecules according to the present invention typically have a heat stability that is at least equivalent to or even higher than that of bispecific antigen-binding molecules having only one target-binding domain but otherwise containing a CD3-binding domain and a half-life extending scFc domain (i.e., not overly structurally complex). One skilled in the art would expect that the more structurally complex a protein-based molecule is, the more susceptible it is to thermal and other degradation, i.e., the lower its heat stability. However, surprisingly, in contrast to this case, the multi-chain multi-target bispecific antigen-binding molecules according to the present invention exhibit a heat stability that is at least equivalent to or even better than that of single-chain molecules. Even when tested for long-term storage stability and freeze-thaw stability, the molecules of the present invention advantageously exhibit characteristics that are at least equivalent to or even better than those of single-chain molecules having the same binding domains. Preferably, the molecules of the present invention also show less reduction of monomers after storage and a higher protein homogeneity than each single-chain bispecific antigen-binding molecule, i.e., a molecule containing, for example, the same target-binding and CD3-binding moieties as disclosed herein.

[0095] In one embodiment, the present invention provides a multi-chain multi-target bispecific antigen-binding molecule that includes all four such domains. In a preferred embodiment, the domains of (i), (ii), (iii), and (iv) are arranged as described in FIGS. 1, 2, and 3.

[0096] The term "polypeptide" is understood herein to be an organic polymer that includes at least one continuous, unbranched chain of amino acids. In the context of the present invention, polypeptides that include two or more amino acid chains are likewise contemplated. The amino acid chains of a polypeptide typically include at least 50 amino acids, preferably at least 100, 200, 300, 400, or 500 amino acids. In the context of the present invention, it is also contemplated that the amino acid chains of a polymer are linked to entities that are not composed of amino acids.

[0097] The term "antigen-binding polypeptide" according to the present invention is preferably a polypeptide that binds immunospecifically to its target or antigen. An antigen-binding polypeptide typically includes or comprises domains that include or are derived from the heavy-chain variable region (VH) and / or the light-chain variable region (VL) of an antibody. The polypeptides according to the present invention include the minimal structural requirements of an antibody that enable immunospecific target binding. This minimal requirement can be defined, for example, by the presence of at least three light-chain CDRs (i.e., CDR1, CDR2, and CDR3 of the VL region) and / or three heavy-chain CDRs (i.e., CDR1, CDR2, and CDR3 of the VH region), and preferably can be defined by the presence of all six CDRs. The antigen-binding molecules of the present invention are thus preferably T-cell engaging polypeptides that can be characterized by the presence of three or six CDRs in one or both binding domains, and those skilled in the art know where (in what order) those CDRs are located within the binding domains. Preferably, an "antigen-binding molecule" is understood as an "antigen-binding polypeptide" in the context of the present invention. In an alternative embodiment, the antigen-binding polypeptide of the present invention may be an aptamer.

[0098] Alternatively, a molecule in the context of the present invention is an antigen-binding polypeptide corresponding to an "antibody construct", typically referring to a molecule whose structure and / or function is based on that of an antibody (e.g., a full-length immunoglobulin molecule or a whole immunoglobulin molecule). Thus, an antigen-binding molecule can bind to its specific target or antigen and / or is derived from the variable heavy chain (VH) and / or variable light chain (VL) domains of an antibody or a fragment thereof. Further, a domain that binds to a binding partner according to the present invention is understood herein as the binding domain of an antigen-binding molecule according to the present invention. Typically, the binding domain according to the present invention comprises the minimal structural requirements of an antibody that enable target binding. This minimal requirement can be defined, for example, by the presence of at least three light chain CDRs (i.e., CDR1, CDR2, and CDR3 of the VL region), and / or three heavy chain CDRs (i.e., CDR1, CDR2, and CDR3 of the VH region), and preferably can be defined by the presence of all six CDRs. An alternative approach to defining the minimal structural requirements of an antibody is to define the epitope of the antibody within the structure of a particular target, the protein domain (epitope cluster) of the target protein containing the epitope region, respectively, or by reference to a specific antibody that competes with the defined epitope of the antibody. Antibodies that can serve as the basis for the constructs according to the present invention include, for example, monoclonal antibodies, recombinant antibodies, chimeric antibodies, deimmunized antibodies, humanized antibodies, and human antibodies.

[0099] In the context of the present invention, the polypeptides of the present invention bind to their respective target structures in a specific manner. Preferably, the polypeptides according to the present invention each contain one paratope for a binding domain that "specifically or immunospecifically binds to", "specifically or immunospecifically recognizes", or "(specifically or immunospecifically) reacts with" their respective target structures. This means that, according to the present invention, the polypeptide or its binding domain interacts or (immunologically) specifically interacts with a given epitope on the target molecule (antigen) and CD3, respectively. This interaction or association occurs more frequently, more rapidly, more persistently, more affinity, or in some combination of these parameters, with respect to the epitope on a particular target, compared to alternative substances (non-target molecules). However, due to sequence similarity between homologous proteins in different species, the binding domain that binds (immunologically) specifically to its target (such as a human target) may cross-react with homologous target molecules from different species (e.g., from non-human primates). Thus, the term "specific / immunologically specific binding" may include the binding of the binding domain to epitopes of two or more species and / or epitopes that are structurally related. The term "(immunologically) selectively binds" excludes binding to epitopes that are structurally related.

[0100] The binding domain of the antigen-binding molecule according to the present invention may include, for example, the group of CDRs referred to above. Preferably, these CDRs are included within the frameworks of the antibody light chain variable region (VL) and the antibody heavy chain variable region (VH), however, it is not necessary to include both. The Fd fragment has, for example, two VH regions and often retains the antigen-binding function of a part of the intact antigen-binding domain. Further examples of antibody fragments, antibody variants, or forms of the binding domain include: (1) the Fab fragment, which is a monovalent fragment having VL, VH, CL, and CH1 domains; (2) the F(ab’)2 fragment, which is a divalent fragment having two Fab fragments linked by a disulfide bridge in the hinge region; (3) the Fd fragment having two VH and CH1 domains; (4) the Fv fragment having the VL domain and the VH domain of a single arm of the antibody; (5) the dAb fragment having a VH domain (Ward et al., (1989) Nature 341:544-546); (6) the isolated complementarity-determining region (CDR), and (7) the single-chain Fv (scFv), with the latter being preferred (for example, those derived from scFV libraries). Examples of embodiments of the antigen-binding molecule according to the present invention are described, for example, in WO 00 / 006605 pamphlet, WO 2005 / 040220 pamphlet, WO 2008 / 119567 pamphlet, WO 2010 / 037838 pamphlet, WO 2013 / 026837 pamphlet, WO 2013 / 026833 pamphlet, US Patent Application Publication No. 2014 / 0308285, US Patent Application Publication No. 2014 / 0302037, WO 2014 / 144722 pamphlet, WO 2014 / 151910 pamphlet, and WO 2015 / 048272 pamphlet.

[0101] In addition, the definition of "binding domain" or "domain that binds to" also includes fragments of full-length antibodies, such as VH, VHH, VL, (s)dAb, Fv, Fd, Fab, Fab’, F(ab’)2, or "r IgG" ("half-antibody"), etc. The antigen-binding molecule according to the present invention can also be a modified fragment of an antibody, also called an antibody variant, such as scFv, di-scFv or bi(s)-scFv, scFv-Fc, scFv-zip, scFab, Fab2, Fab3, diabody, single-chain diabody, tandem diabody (Tandab’s), tandem di-scFv, tandem tri-scFv, "multibody", such as tribody or tetrabody, and single-domain antibodies, such as nanobody, or a single variable domain antibody containing only one variable domain that can be a VHH, VH or VL that specifically binds to an antigen or epitope independent of other V regions or domains. Typically, the binding domain of the present invention includes a paratope that facilitates binding to its binding partner.

[0102] As used herein, the terms "single-chain Fv", "single-chain antibody", or "scFv" refer to an antibody fragment that is a single polypeptide chain containing variable regions derived from both the heavy and light chains but lacking constant regions. Typically, a single-chain antibody further includes a polypeptide linker between the VH domain and the VL domain, thereby enabling the formation of a desired structure that can bind to an antigen. Single-chain antibodies are discussed in detail by Pluckthun in The Pharmacology of Monoclonal Antibodies, vol. 113, Rosenberg and Moore eds., Springer-Verlag, New York, pp. 269-315 (1994). Various methods for making single-chain antibodies are known, for example, as described in U.S. Patent No. 4,694,778 and U.S. Patent No. 5,260,203; International Publication No. 88 / 01649 pamphlet; Bird (1988) Science 242:423-442; Huston et al. (1988) Proc. Natl. Acad. Sci. USA 85:5879-5883; Ward et al. (1989) Nature 334:54454; Skerra et al. (1988) Science 242:1038-1041. In certain embodiments, single-chain antibodies may also be bispecific, multispecific, human, and / or humanized, and / or synthetic.

[0103] In the context of the present invention, a paratope is understood to be a part of a polypeptide as described herein and an antigen-binding site that recognizes and binds an antigen. A paratope is typically a small region of at least about 5 amino acids. A paratope as understood herein typically includes a portion of the sequences of the heavy (VH) and light (VL) chains derived from an antibody. Each binding domain of the molecules according to the present invention comprises a paratope that includes a set of six complementarity-determining regions (CDR loops), three of each being contained within the VH and VL sequences derived from an antibody.

[0104] Furthermore, the definition of the term "antigen-binding molecule" preferably includes polyvalent / multivalent constructs and thus bispecific molecules, where bispecificity means specifically binding to two cell types (i.e., target cells and effector cells) containing different antigen structures. Since the antigen-binding molecules of the present invention are preferably multi-chain and multi-target, they are typically also polyvalent / multivalent molecules, i.e., they specifically bind to four different binding domains in the context of the present invention, which are three or more antigen structures, preferably two target-binding domains and two CD3-binding domains. The term "multi-chain multi-target bispecific antigen-binding molecule" includes the terms "multi-chain multi-target bispecific T cell engaging molecule" and "multi-chain multi-target bispecific T cell engaging polypeptide (MMBiTEP)". Preferred "multi-chain multi-target bispecific antigen-binding molecules" are "multi-chain multi-target bispecific T cell engaging molecules" or "multi-chain multi-target bispecific T cell engaging polypeptides (MMBiTEP)". The term "multi-chain multi-target bispecific T cell engaging molecule" is understood to include the term "multi-chain multi-target bispecific T cell engaging polypeptide". Furthermore, the definition of the term "antigen-binding molecule" includes not only molecules containing only one peptide chain, but also molecules consisting of two or more polypeptide chains, and these chains may be the same (homo-dimer, homo-trimer or homo-oligomer) or different (hetero-dimer, hetero-trimer or hetero-oligomer). Such molecules containing two or more polypeptide chains (i.e., typically two chains) have these chains typically bound to each other as a hetero-dimer by charge-pair binding, for example, within a hetero-Fc entity that functions as a spacer and half-life extension moiety between two bispecific entities as described herein.Examples of antigen-binding molecules identified above, such as antibody-based molecules and variants or derivatives thereof, are described, inter alia, in Harlow and Lane, Antibodies a laboratory manual, CSHL Press (1988) and Using Antibodies: a laboratory manual, CSHL Press (1999), Kontermann and Duebel, Antibody Engineering, Springer, 2nd ed. 2010 and Little, Recombinant Antibodies for Immunotherapy, Cambridge University Press 2009.

[0105] The term "bispecific", as used herein, refers to an antigen-binding molecule that is "at least bispecific", i.e., one that addresses two different cell types, namely a target cell and an effector cell, and includes at least a first binding domain and a third binding domain, and a second binding domain and a fourth binding domain, wherein at least two of the binding domains preferably bind to two antigens or targets selected from CD20, CD22, FLT3, MSLN, CDH3, CLL1 and EpCAM, and the other two binding domains of the same molecule bind to another antigen (here CD3) on an effector cell, typically a T cell. Thus, the antigen-binding molecules according to the invention comprise specificity for at least two different antigens or targets. For example, it is preferred that the two domains do not bind to one or more extracellular epitopes of CD3ε of the species as described herein.

[0106] The term "target cell surface antigen" refers to an antigen structure that is expressed by a cell and is present on the cell surface such that an antigen-binding molecule as described herein can reach it. Preferred target cell surface antigens in the context of the present invention are tumor-associated antigens (TAAs). TAAs can be proteins (preferably the extracellular part of a protein), or carbohydrate structures (preferably the carbohydrate structure of a protein such as a glycoprotein). TAAs are preferably tumor antigens. The term "bispecific antigen-binding molecule" of the present invention also encompasses bispecific multichain multi-target antigen-binding molecules, for example, trispecific antigen-binding molecules (the latter containing three binding domains), or constructs having four or more (e.g., four, five...) specificities.

[0107] In the context of the present invention, molecules that are "multi-target" are preferred and are understood herein to "typically target at least two targets (e.g., TAAs) per molecule of the invention per target cell". In this regard, multi-target molecules such as antigen-binding molecules are specific for two (typically identical) effector structures on effector cells such as CD3, more preferably CD3 epsilon (CD3ε is always included when referring to "CD3" in the present invention), and at least two target cell surface antigens. The specificities are conferred by each binding domain as defined herein. Typically, "multi-target" refers to a molecule that is specific for at least two (preferably different) target cell surface antigens (e.g., TAAs), thereby conferring the preferred properties of the multi-target antigen-binding molecules according to the present invention, in other words, reducing antigen deficiency and increasing selectivity, i.e., the molecules of the present invention have binding domains and, since the target cells are associated with a disease, selectivity is conferred to kill target cells co-expressing the targets. Thus, the therapeutic scope of the molecules of the present invention is increased compared to single-target bispecific molecules, and typically higher drug tolerance as demonstrated herein is brought about.

[0108] The T cell-engaging antigen-binding molecules according to the invention, such as multi-chain polypeptides, are preferably bispecific and are herein understood to typically comprise one domain that binds to at least one target antigen and another domain that binds to CD3. Thus, it does not occur naturally and its function is significantly different from naturally occurring products. Thus, the polypeptides according to the invention are artificial "hybrid" polypeptides comprising at least two individual binding domains having different specificities and are thus bispecific. Bispecific antigen-binding molecules can be generated by various methods including fusion of hybridomas or ligation of Fab' fragments. See, for example, Songsivilai & Lachmann, Clin. Exp. Immunol. 79:315-321 (1990).

[0109] At least four binding domains and variable domains (VH / VL) of the antigen-binding molecule of the present invention typically include a peptide linker (spacer peptide). The term "peptide linker" in the context of the present invention includes an amino acid sequence that links the amino acid sequences of one (variable and / or binding) domain of the antigen-binding molecule of the present invention to another (variable and / or binding) domain. The peptide linker between the first domain and the second domain and between the third domain and the fourth domain is (the first domain and the third domain can preferably bind simultaneously to preferably two different targets (e.g., TAA1 and TAA2) on the same cell), preferably mobile and of limited length (e.g., 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17 or 18 amino acids). The peptide linker can also be used to fuse a spacer to other domains of the antigen-binding molecule of the present invention. The essential technical feature of such a peptide linker is that it does not contain polymerization activity. Suitable peptide linkers are those described in U.S. Patent Nos. 4,751,180 and 4,935,233, or International Publication No. 88 / 09344 pamphlet. The peptide linker can also be used to attach other domains or modules or regions (such as a half-life extension domain) to the antigen-binding molecule of the present invention. However, typically, the linker between the first target-binding domain and the second target-binding domain is different from the intrabody linker that links VH and VL within the target-binding domain. This difference is that the linker between the first binding domain and the second binding domain has one more amino acid compared to the intrabody linker (e.g., 6 and 5 amino acids for GGGGS and SGGGGS respectively). Thus, surprisingly, mobility and stability are simultaneously imparted to a specific antigen-binding molecule format as described herein.Since the spacer contributes to constructing at least one continuous polypeptide chain that links two bispecific entities and preferably contains four binding domains or portions thereof, and thus also functions as a linker, the spacer (or spacer entity as a synonym) between two bispecific entities as described herein is a particular embodiment of a linker. However, in addition, the spacer functions as an entity that spatially separates two bispecific entities. Thus, the spacer in the context of the present invention is a particular embodiment of a linker, which thereby contributes to linking two binding domains (of two different bispecific entities) together with two shorter and more flexible linkers at both ends, but most importantly, so that the two bispecific entities can act advantageously as described herein, for example, separating the bispecific entities so that they can exhibit a surprisingly high selectivity gap.

[0110] The antigen-binding molecule of the present invention is preferably an "antigen-binding molecule generated in vitro". This term refers to an antigen-binding molecule as defined above, wherein all or part of the variable region (e.g., at least one CDR) is generated in non-immune cells by selection, e.g., in vitro phage display, protein chips, or any other method capable of testing candidate sequences for antigen-binding ability. Thus, this term preferably excludes sequences generated only by genomic rearrangement in the immune cells of an animal. A "recombinant antibody" is an antibody produced by using recombinant DNA technology or genetic engineering.

[0111] The term "monoclonal antibody" (mAb) or monoclonal antibodies derived from antigen-binding molecules as used herein refers to antibodies obtained from a substantially homogeneous population of antibodies, i.e., individual antibodies that comprise a population that is identical except for possible naturally occurring mutations and / or post-translational modifications (e.g., isomerization, amidation) that may be present in minor amounts. Monoclonal antibodies are highly specific and are induced against a single antigenic site or determinant on an antigen, as opposed to conventional (polyclonal) antibody preparations that typically include different antibodies induced against different determinants (or epitopes). In addition to their specificity, monoclonal antibodies are advantageous in that they are synthesized by hybridoma culture and thus are not subject to contamination by other immunoglobulins. The modifier "monoclonal" indicates the property of an antibody obtained from a substantially homogeneous population of antibodies and should not be construed as requiring that the antibody be produced by any particular method.

[0112] To prepare monoclonal antibodies, any technique that provides antibodies produced by culturing of continuous cell lines can be used. For example, the monoclonal antibodies used can be made by the hybridoma method first described by Koehler et al., Nature, 256:495 (1975), or by recombinant DNA methods (see, e.g., U.S. Patent No. 4,816,567). Examples of additional techniques for producing human monoclonal antibodies include the trioma technique, the human B-cell hybridoma technique (Kozbor, Immunology Today 4 (1983), 72), and the EBV-hybridoma technique (Cole et al., Monoclonal Antibodies and Cancer Therapy, Alan R. Liss, Inc. (1985), 77-96).

[0113] Next, hybridomas can be screened using standard methods such as enzyme-linked immunosorbent assay (ELISA) and surface plasmon resonance analysis, such as Biacore™, to identify one or more hybridomas that produce antibodies that specifically bind to a particular antigen. For example, any form of related antigen, such as a recombinant antigen, a naturally occurring form, any variant or fragment thereof, and its antigenic peptide, may be used as an immunogen. Surface plasmon resonance, such as that employed in the Biacore system, may be used to enhance the efficiency of phage antibodies that bind to epitopes of the target cell surface antigen (Schier, Human Antibodies Hybridomas 7(1996),97-105; Malmborg, J. Immunol. Methods 183(1995),7-13).

[0114] Another exemplary method for generating monoclonal antibodies involves screening protein expression libraries, such as phage display or ribosome display libraries. Phage display is described, for example, in Ladner et al., U.S. Patent No. 5,223,409, Smith (1985) Science 228:1315-1317, Clackson et al., Nature, 352:624-628 (1991), and Marks et al., J. Mol. Biol., 222:581-597 (1991).

[0115] In addition to the use of display libraries, related antigens can be used to immunize non-human animals, such as rodents (e.g., mice, hamsters, rabbits, or rats). In one embodiment, the non-human animal comprises at least a portion of a human immunoglobulin gene. For example, a mouse strain defective in the production of mouse antibodies can be modified using a large fragment of the human Ig (immunoglobulin) locus. Hybridoma technology can be used to generate and select antigen-specific monoclonal antibodies derived from genes having the desired specificity. See, e.g., XENOMOUSE™, Green et al. (1994) Nature Genetics 7:13-21, U.S. Patent Application Publication No. 2003-0070185, International Publication No. 96 / 34096 Pamphlet, and International Publication No. 96 / 33735 Pamphlet.

[0116] Alternatively, monoclonal antibodies can be obtained from non-human animals and then modified, such as humanization, deimmunization, chimerization, etc., using recombinant DNA techniques known in the art. Examples of modified antigen-binding molecules include humanized variants of non-human antibodies, "affinity matured" antibodies (see, e.g., Hawkins et al. J. Mol. Biol. 254, 889-896 (1992) and Lowman et al., Biochemistry 30, 10832-10837 (1991)), and antibody variants with modified effector functions (see, e.g., U.S. Patent No. 5,648,260, Kontermann and Duebel (2010) supra, and Little (2009) supra).

[0117] In immunology, affinity maturation is the process by which B cells produce antibodies with increased affinity for an antigen during the course of an immune response. Upon repeated exposure to the same antigen, the host continuously produces antibodies with higher affinity. Similar to the natural prototype, in vitro affinity maturation is based on the principles of mutation and selection. In vitro affinity maturation has been successfully used to optimize antibodies, antigen-binding molecules, and antibody fragments. Random mutations within the CDRs are introduced using radiation, chemical mutagens, or error-prone PCR. In addition, genetic diversity can be increased by chain shuffling. Two or three rounds of mutation and selection using display methods such as phage display typically yield antibody fragments with affinities in the low nanomolar range.

[0118] Preferred types of amino acid substitution variants of antigen-binding molecules include substitutions of one or more hypervariable region residues of a parental antibody (e.g., a humanized or human antibody). Generally, the resulting variants selected for further development have improved biological properties compared to the parental antibodies from which they were generated. Convenient methods for generating such substitution variants include affinity maturation using phage display. Briefly, several hypervariable region sites (e.g., 6-7 sites) are mutated to generate all possible amino acid substitutions at each site. The antibody variants so generated are displayed in monovalent form from filamentous phage particles as fusions with the gene III product of M13 packaged within each particle. The phage-displayed variants are then screened for their biological activities (e.g., binding affinity) as disclosed herein. To identify candidate hypervariable region sites to modify, the alanine scanning mutagenesis method can be performed to identify hypervariable region residues that contribute significantly to antigen binding. Alternatively or in addition, it may be beneficial to analyze the crystal structure of the antigen-antibody complex to identify contact points between the binding domain and, for example, human CS1, BCMA, CD20, CD22, FLT3, CD123, CDH3, MSLN, CLL1 or EpCAM. Such contact residues and adjacent residues are candidates for substitution by the techniques detailed herein. Once such variants have been generated, screening as described herein can be performed on a panel of variants, and antibodies having excellent properties in one or more relevant assays can be selected for further development.

[0119] The monoclonal antibodies and antigen-binding molecules of the present invention include, in particular, "chimeric" antibodies (immunoglobulins) in which a portion of the heavy chain and / or light chain is identical or homologous to the corresponding sequence in an antibody derived from a particular species or belonging to a particular antibody class or subclass, while the remainder of the chain is derived from a different species or belongs to a different antibody class or subclass as long as the desired biological activity is exhibited, and is identical or homologous to the corresponding sequence in fragments of such antibodies (U.S. Patent No. 4,816,567; Morrison et al., Proc. Natl. Acad. Sci. USA, 81:6851-6855 (1984)). Chimeric antibodies for the purposes herein include "primatized" antibodies that include variable domain antigen-binding sequences derived from non-human primates (e.g., Old World monkeys, apes, etc.) and human constant region sequences. Various methods for making chimeric antibodies have been described. See, for example, Morrison et al., Proc. Natl. Acad. Sci. U.S.A. 81:6851, 1985; Takeda et al., Nature 314:452, 1985, Cabilly et al., U.S. Patent No. 4,816,567; Boss et al., No. 4,816,397; Tanaguchi et al., European Patent No. 0171496; No. 0173494; and British Patent No. 2177096.

[0120] An antibody, antigen-binding molecule, antibody fragment, or antibody variant may also be modified by, for example, the method of specific deletion of human T cell epitopes (a method called "deimmunization") as disclosed in WO 98 / 52976 pamphlet or WO 00 / 34317 pamphlet. Briefly described, it is possible to analyze whether the heavy and light chain variable domains of an antibody are peptides that bind to MHC class II, and these peptides represent potential T cell epitopes (defined in WO 98 / 52976 pamphlet and WO 00 / 34317 pamphlet). To detect potential T cell epitopes, a computer modeling technique called "peptide threading" can be applied as described in WO 98 / 52976 pamphlet and WO 00 / 34317 pamphlet. In addition, databases of human MHC class II-binding peptides can be searched for motifs present in the VH and VL sequences. These motifs can bind to any of the 18 major MHC class II DR allotypes and thus become potential T cell epitopes. The detected potential T cell epitopes can be eliminated by substituting a small number of amino acid residues in the variable domain or, preferably, by substituting a single amino acid. Typically, conservative substitutions are made. Although not all, in many cases, common amino acids can be used at positions in the human germline antibody sequences. Human germline sequences are disclosed, for example, in Tomlinson, et al. (1992) J. Mol. Biol. 227:776-798; Cook, G.P. et al. (1995) Immunol. Today Vol. 16(5):237-242; and Tomlinson et al. (1995) EMBO J. 14:14:4628-4638. The V BASE directory provides a comprehensive directory of human immunoglobulin variable region sequences (edited by Tomlinson, LA. et al., MRC Centre for Protein Engineering, Cambridge, UK). These sequences can be used, for example, as a source of human sequences for framework regions and CDRs.For example, a consensus human framework region as described in U.S. Patent No. 6,300,064 can also be used.

[0121] A "humanized" antibody, antigen-binding molecule, variant, or fragment thereof (e.g., Fv, Fab, Fab', F(ab')2, or other antigen-binding subsequences of an antibody) is an antibody or immunoglobulin that contains minimal sequences from non-human immunoglobulins and is mostly human sequences. In most cases, a humanized antibody is a human immunoglobulin (recipient antibody) in which residues from the hypervariable regions (also called CDRs) of the recipient are replaced with residues from the hypervariable regions of a non-human (e.g., rodent) species such as mouse, rat, hamster, or rabbit that have the desired specificity, affinity, and capacity (donor antibody). In some cases, residues in the Fv framework region (FR) of the human immunoglobulin are replaced with the corresponding non-human residues. Further, a "humanized antibody" may also, as used herein, contain residues not found in either the recipient antibody or the donor antibody. These modifications are made to further refine and optimize the performance of the antibody. A humanized antibody may also contain at least a portion of an immunoglobulin constant region (Fc), typically that of a human immunoglobulin. For further details, see Jones et al., Nature, 321:522-525 (1986); Reichmann et al., Nature, 332:323-329 (1988); and Presta, Curr. Op. Struct. Biol., 2:593-596 (1992).

[0122] A humanized antibody or fragment thereof can be generated by replacing the sequence of the Fv variable domain that is not directly involved in antigen binding with an equivalent sequence derived from a human Fv variable domain. Exemplary methods for generating a humanized antibody or fragment thereof are provided by Morrison (1985) Science 229:1202-1207; Oi et al. (1986) Bio Techniques 4:214; and U.S. Patent Nos. 5,585,089, 5,693,761, 5,693,762, 5,859,205, and 6,407,213. These methods involve isolating, manipulating, and expressing a nucleic acid sequence encoding all or part of an immunoglobulin Fv variable domain derived from at least one of the heavy or light chains. Such nucleic acids may be obtained from sources other than just hybridomas that produce antibodies against a given target as described above. Recombinant DNA encoding the humanized antibody molecule can then be cloned into an appropriate expression vector.

[0123] Humanized antibodies may also be prepared using transgenic animals such as mice that express human heavy and light chain genes but are unable to express endogenous mouse immunoglobulin heavy and light chain genes. An exemplary CDR grafting method that can be used for the preparation of humanized antibodies described herein has been described by Winter (U.S. Patent No. 5,225,539). All of the CDRs of a particular human antibody may be replaced with at least a portion of non-human CDRs, or only a portion of the CDRs may be replaced with non-human CDRs. It is only necessary to replace the number of CDRs necessary for the humanized antibody to bind to a given antigen.

[0124] Humanized antibodies can be optimized by introducing conservative substitutions, consensus sequence substitutions, germline substitutions, and / or revertant mutations. Such modified immunoglobulin molecules can be made by any of several techniques known in the art (e.g., Teng et al., Proc. Natl. Acad. Sci. U.S.A., 80:7308-7312, 1983; Kozbor et al., Immunology Today, 4:7279, 1983; Olsson et al., Meth. Enzymol., 92:3-16, 1982, and European Patent No. 239400).

[0125] The terms “human antibody,” “human antigen-binding molecule,” and “human binding domain” include antibodies, antibody-binding molecules, and binding domains having antibody regions such as variable and constant regions or domains that substantially correspond to immunoglobulin sequences of the human germline known in the art, e.g., those described in Kabat et al. (1991) (supra). The human antibodies, antigen-binding molecules, or binding domains of the present invention may contain amino acid residues not encoded by human germline immunoglobulin sequences (e.g., mutations introduced by in vitro random mutagenesis or site-directed mutagenesis, or by somatic mutations in vivo), e.g., in the CDRs, particularly CDR3. The human antibodies, antigen-binding molecules, or binding domains may have at least one, two, three, four, five, or more positions replaced with amino acid residues not encoded by human germline immunoglobulin sequences. The definitions of human antibodies, antigen-binding molecules, and binding domains, as used herein, also contemplate fully human antibodies that contain only the human sequences of non-artificial and / or genetically modified antibodies, such as those that can be obtained by using techniques or systems such as Xenomouse. Preferably, a “fully human antibody” does not contain amino acid residues not encoded by human germline immunoglobulin sequences.

[0126] In some embodiments, the antigen-binding molecules of the invention are "isolated" or "substantially pure" antigen-binding molecules. "Isolated" or "substantially pure", when used in the context of the description of the antigen-binding molecules disclosed herein, means an antigen-binding molecule that has been identified, separated and / or recovered from the components of its production environment. Preferably, the antigen-binding molecule does not or substantially does not associate with all other components derived from its production environment. Contaminating components of its production environment, such as components resulting from recombinant transfected cells, are typically materials that interfere with the diagnostic or therapeutic use of the polypeptide and may include enzymes, hormones, and other proteinaceous or non-proteinaceous solutes. The antigen-binding molecule can constitute, for example, at least about 5% by weight, or at least about 50% by weight, of the total proteins in a given sample. It is understood that an isolated protein can constitute from 5% to 99.9% by weight of the total protein content, depending on the circumstances. The polypeptide can be produced at significantly high concentrations by using an inducible or highly expressing promoter such that it is produced at an increased concentration level. This definition includes the production of antigen-binding molecules in a wide variety of organisms and / or host cells known in the art. In preferred embodiments, the antigen-binding molecule is purified to a sufficient extent to obtain at least 15 residues of the N-terminal or internal amino acid sequence by using a spinning cup sequenator, or to homogeneity by SDS-PAGE under non-reducing or reducing conditions using Coomassie blue or preferably silver staining. However, typically, an isolated antigen-binding molecule is prepared by at least one purification step.

[0127] The term "binding domain" is considered, in the context of the present invention, to be a domain that binds to (specifically), interacts with, or recognizes a given target epitope or a given target site on a target molecule (antigen), such as CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, MSLN, or EpCAM, respectively, and CD3. Typically, the structure and function of the first binding domain and the third binding domain or the second binding domain and the fourth binding domain (e.g., recognizing CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, MSLN, or EpCAM), and also preferably the structure and / or function of the effector binding domain (typically the second binding domain and the fourth binding domain or the first binding domain and the third binding domain that recognize CD3), are based on the structure and / or function of an antibody, such as a full-length or whole immunoglobulin molecule, and / or are derived from the variable heavy chain (VH) and / or variable light chain (VL) domains of an antibody or a fragment thereof. Preferably, the target cell surface antigen binding domain is characterized by the presence of three light chain CDRs (i.e., CDR1, CDR2, and CDR3 of the VL region), and / or three heavy chain CDRs (i.e., CDR1, CDR2, and CDR3 of the VH region). The effector (typically CD3) binding domain also preferably includes the minimum structural requirements of an antibody that enable target binding. More preferably, the second binding domain includes at least three light chain CDRs (i.e., CDR1, CDR2, and CDR3 of the VL region), and / or three heavy chain CDRs (i.e., CDR1, CDR2, and CDR3 of the VH region). The first binding domain and / or the second binding domain are envisioned to be made or obtained by phage display or library screening methods, in addition to grafting CDR sequences from existing (monoclonal) antibodies onto a scaffold.

[0128] According to the present invention, the binding domain is in the form of one or more polypeptides. Such polypeptides may include a protein portion and a non-protein portion (e.g., a chemical linker such as a chemical crosslinking agent like glutaraldehyde). Proteins (including fragments thereof, preferably biologically active fragments, and peptides having usually less than 30 amino acids) include two or more amino acids linked to each other via covalent peptide bonds (resulting in a chain of amino acids).

[0129] The term "polypeptide" as used herein refers to a group of molecules consisting of usually more than 30 amino acids. Polypeptides may further form multimers such as dimers, trimers, and higher-order oligomers, i.e., they may consist of two or more polypeptide molecules. The polypeptide molecules forming such dimers, trimers, etc. may or may not be identical. The corresponding higher-order structures of such multimers are thus referred to as homodimers or heterodimers, homotrimers or heterotrimers, etc. An example of a heteromultimer is an antibody molecule consisting in its native form of two identical polypeptide light chains and two identical polypeptide heavy chains. The terms "peptide", "polypeptide", and "protein" also refer to native modified peptides / polypeptides / proteins modified by post-translational modifications such as, for example, glycosylation, acetylation, phosphorylation, etc. "Peptide", "polypeptide", or "protein" may also be chemically modified such as pegylated as referred to herein. Such modifications are known in the art and are described below herein.

[0130] Preferably, the binding domain that binds to any of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, and EpCAM, and / or the binding domain that binds to CD3ε, is a human binding domain. Antibodies and antigen-binding molecules comprising at least one human binding domain avoid some of the problems associated with antibodies or antigen-binding molecules having non-human variable and / or constant regions, such as rodents (e.g., mouse, rat, hamster, or rabbit). The presence of such rodent-derived proteins can lead to rapid clearance of the antibody or antigen-binding molecule or can cause an immune response in the patient against the antibody or antigen-binding molecule. To avoid using rodent-derived antibodies or antigen-binding molecules, human or fully human antibodies / antigen-binding molecules can be generated by introducing human antibody functions into rodents such that the rodents produce fully human antibodies.

[0131] The ability to clone and reconstruct megabase-sized human loci in yeast artificial chromosomes (YACs) and introduce them into the mouse germ cell lineage not only reveals functional elements of very large or coarsely mapped loci, but also provides a powerful approach for generating useful models of human disease. Furthermore, using such techniques to replace mouse loci with their human equivalents allows for unique insights into the expression and regulation of human gene products during development, their communication with other systems, and their involvement in disease induction and progression.

[0132] An important practical application of such a strategy is to "humanize" the murine humoral immune system. By introducing the human Ig locus into mice with inactivated endogenous immunoglobulin (Ig) genes, an opportunity is provided to study the mechanisms underlying their role in B cell development, as well as the programmed expression and construction of antibodies. Furthermore, such a strategy makes it possible to provide an ideal source for the generation of fully human monoclonal antibodies (mAbs), which are important milestones in realizing the potential of antibody therapies in human diseases. Fully human antibodies or antigen-binding molecules are expected to minimize the immunogenicity and allergic reactions inherent to murine mAbs or mAbs derived from mice, thereby increasing the efficacy and safety of the administered antibody / antigen-binding molecule. The use of fully human antibodies or antigen-binding molecules is expected to provide significant advantages in the treatment of chronic and recurrent human diseases that require repeated administration of compounds, such as inflammation, autoimmunity, and cancer.

[0133] One approach to this goal is to modify mouse strains that are defective in producing mouse antibodies with large fragments of the human Ig locus, which was predicated on the prediction that such mice would produce a broad repertoire of human antibodies in the absence of mouse antibodies. The large human Ig fragments retain not only the extensive diversity of the variable genes but also the appropriate regulation of antibody production and expression. By taking advantage of the mouse mechanisms for antibody diversification and selection and the lack of immune tolerance to human proteins, the human antibody repertoires replicated in these mouse strains should produce antibodies with high affinity for any antigen of interest, including human antigens. Using hybridoma technology, antigen-specific human mAbs with the desired specificity can be readily generated and selected. This general strategy has been demonstrated in connection with the generation of the first XenoMouse mouse strain (see Green et al. Nature Genetics 7:13-21 (1994)). This XenoMouse strain has been modified by YACs containing germline configuration fragments of 245 kb and 190 kb in size of the human heavy chain locus and the kappa light chain locus, respectively, containing the core sequences of the variable and constant regions. This YAC containing human Ig has been shown to be compatible with the mouse system with respect to both antibody rearrangement and expression and was able to replace the inactivated mouse Ig genes. This was demonstrated by their ability to induce B cell development to produce an adult-like human repertoire of fully human antibodies and to generate antigen-specific human mAbs. These results also suggested that by introducing large portions of the human Ig locus containing multiple V genes, additional regulatory elements, and human Ig constant regions, a substantially complete repertoire characteristic of the human humoral response to infection and immunization could be reproduced. In recent years, building on the work of Green et al., over approximately 80% of the human antibody repertoire has been introduced by introducing megabase-sized germline configuration YAC fragments of the human heavy chain locus and the kappa light chain locus, respectively.See Mendez et al., Nature Genetics 15:146-156 (1997) and U.S. Patent Application No. 08 / 759,620.

[0134] The production of XenoMouse animals is further discussed and detailed in U.S. Patent Application Nos. 07 / 466,008, 07 / 610,515, 07 / 919,297, 07 / 922,649, 08 / 031,801, 08 / 112,848, 08 / 234,145, 08 / 376,279, 08 / 430,938, 08 / 464,584, 08 / 464,582, 08 / 463,191, 08 / 462,837, 08 / 486,853, 08 / 486,857, 08 / 486,859, 08 / 462,513, 08 / 724,752, and 08 / 759,620; as well as U.S. Patents Nos. 6,162,963, 6,150,584, 6,114,598, 6,075,181, and 5,939,598; and Japanese Patent Publications Nos. 3068180B2, 3068506B2, and 3068507B2. See also Mendez et al., Nature Genetics 15:146-156 (1997) and Green and Jakobovits, J. Exp. Med. 188:483-495 (1998), European Patent No. 0463151B1, International Publications Nos. 94 / 02602, 96 / 34096, 98 / 24893, 00 / 76310, and 03 / 47336.

[0135] In an alternative approach, a method using "miniloci" has been utilized by other companies including GenPharm International, Inc. In the minilocus method, an exogenous Ig locus is mimicked by including small pieces (individual genes) derived from the Ig locus. Therefore, one or more VH genes, one or more DH genes, one or more JH genes, a μ constant region, and a second constant region (preferably a γ constant region) are formed in a construct for insertion into an animal. This method is described in U.S. Patent No. 5,545,807 to Surani et al., and U.S. Patent Nos. 5,545,806, 5,625,825, 5,625,126, 5,633,425, 5,661,016, 5,770,429, 5,789,650, 5,814,318, 5,877,397, 5,874,299, and 6,255,458 to Lonberg and Kay respectively, U.S. Patent Nos. 5,591,669 and 6,023,010 to Krimpenfort and Berns, U.S. Patent Nos. 5,612,205, 5,721,367, and 5,789,215 to Berns et al., U.S. Patent No. 5,643,763 to Choi and Dunn, and GenPharm International's U.S. Patent Applications Nos. 07 / 574,748, 07 / 575,962, 07 / 810,279, 07 / 853,408, 07 / 904,068, 07 / 990,860, 08 / 053,131, 08 / 096,762, 08 / 155,301, 08 / 161,739, 08 / 165,699, 08 / 209,741.Also, refer to European Patent No. 0546073B1, International Publication Nos. WO 92 / 03918, WO 92 / 22645, WO 92 / 22647, WO 92 / 22670, WO 93 / 12227, WO 94 / 00569, WO 94 / 25585, WO 96 / 14436, WO 97 / 13852, and WO 98 / 24884, as well as U.S. Patent No. 5,981,175. Further, refer to Taylor et al. (1992), Chen et al. (1993), Tuaillon et al. (1993), Choi et al. (1993), Lonberg et al. (1994), Taylor et al. (1994), and Tuaillon et al. (1995), Fishwild et al. (1996).

[0136] Kirin has also demonstrated the production of human antibodies derived from mice into which large chromosomal fragments or entire chromosomes have been introduced by microcell fusion. Refer to European Patent Application Publication Nos. 0773288 and 0843961. Xenerex Biosciences is developing a technology for generating promising human antibodies. In this technology, SCID mice are reconstituted with human lymphocytes, such as B cells and / or T cells. Then, when the mice are immunized with an antigen, an immune response against the antigen can be elicited. Refer to U.S. Patent Nos. 5,476,996, 5,698,767, and 5,958,765.

[0137] The human anti-mouse antibody (HAMA) reaction has been a factor for the industry to prepare chimeric antibodies or humanized antibodies by other methods. However, particularly when antibodies are used chronically or in multiple doses, a specific human anti-chimeric antibody (HACA) reaction is expected to be observed. Therefore, in order to nullify the concern and / or effect of the HAMA or HACA reaction, it is desirable to provide an antigen-binding molecule comprising a human binding domain for CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM and a human binding domain for CD3ε.

[0138] According to the present invention, the terms “(specifically) bind”, “(specifically) recognize”, “(specifically) induced” and “(specifically) react” mean that the binding domain interacts or specifically interacts with a given epitope or a given target site on a target molecule (antigen), herein CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM respectively, and CD3ε as an effector.

[0139] The term “epitope” refers to a site on an antigen to which a binding domain, such as an antibody or immunoglobulin, or a derivative, fragment, or variant of an antibody or immunoglobulin, specifically binds. An “epitope” is antigenic, and thus the term “epitope” may also be referred to herein as an “antigenic structure” or “antigenic determinant”. Therefore, the binding domain is an “antigen interaction site”. The binding / interaction is also understood to define “specific recognition”.

[0140] An “epitope” can be formed by both contiguous amino acids or discontinuous amino acids juxtaposed by the tertiary folding of a protein. A “linear epitope” is an epitope that includes an epitope recognized by the primary sequence of amino acids. A linear epitope typically includes at least 3 or at least 4, more commonly at least 5, or at least 6, or at least 7, for example about 8 to about 10 amino acids within a unique sequence.

[0141] A "conformational epitope" is, in contrast to a linear epitope, an epitope in which the primary sequence of the amino acids containing the epitope is not the only element defining the recognized epitope (e.g., an epitope in which the primary sequence of amino acids is not necessarily recognized by a binding domain). Typically, conformational epitopes contain a larger number of amino acids compared to linear epitopes. With respect to the recognition of conformational epitopes, the binding domain recognizes the three-dimensional structure of the antigen (preferably a peptide or protein or a fragment thereof) (in the context of the present invention, the antigen structure for one of the binding domains is included within the target cell surface antigen protein). For example, when a protein molecule folds to form a three-dimensional structure, the specific amino acids and / or polypeptide backbone forming the conformational epitope come into juxtaposition, thereby enabling an antibody to recognize the epitope. Methods for determining the conformation of an epitope include, but are not limited to, X-ray crystallography, two-dimensional nuclear magnetic resonance (2D-NMR) spectroscopy, site-directed spin labeling, and electron paramagnetic resonance (EPR) spectroscopy.

[0142] Methods of epitope mapping are described below: When regions (adjacent amino acid stretches) in human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM proteins are exchanged or replaced with corresponding regions in non-human and non-primate CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM (e.g., mouse CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM, but other ones such as chicken, rat, hamster, rabbit, etc. can also be considered), a decrease in the binding of the binding domain is expected unless the binding domain is cross-reactive with the non-human, non-primate CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM used. This decrease is preferably at least 10%, 20%, 30%, 40%, or 50% compared to the binding to each region in the human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM protein, with the binding to each region in the human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM protein being set as 100%; more preferably at least 60%, 70%, or 80%; and most preferably 90%, 95%, or even 100%. Expression of the aforementioned chimeras of human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM / non-human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM in CHO cells is envisioned. Also envisioned is fusing the chimeras of human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM / non-human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM with transmembrane domains and / or cytoplasmic domains of different membrane-bound proteins such as EpCAM.

[0143] In alternative or additional methods of epitope mapping, several truncated forms of the extracellular domains of human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM can be generated to determine the specific regions recognized by the binding domain. In these truncated forms, various extracellular CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM domains / subdomains or regions are deleted stepwise from the N-terminus. The truncated CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM are expected to be expressed in CHO cells. Also, the truncated CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM are also expected to be fused with the transmembrane domain and / or cytoplasmic domain of a different membrane-bound protein such as EpCAM. Also, the truncated CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM are also expected to include a signal peptide domain at their N-terminus, for example, a signal peptide derived from the mouse IgG heavy chain signal peptide. Furthermore, the truncated CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM are expected to include a v5 domain at the N-terminus (following the signal peptide), by which it can be confirmed that they are correctly expressed on the cell surface. Those truncated forms of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM that no longer include the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM regions recognized by the binding domain are expected to exhibit a decrease or loss of binding.The reduction in binding is preferably at least 10%, 20%, 30%, 40% or 50%, more preferably at least 60%, 70%, 80%, and most preferably 90%, 95% or even 100% when the binding to the entire human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM protein (or its extracellular region or domain) is taken as 100.

[0144] A further method for determining the contribution of specific residues of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM to recognition by an antigen-binding molecule or binding domain is alanine scanning, which involves replacing each residue to be analyzed with alanine, for example, by site-directed mutagenesis (see, for example, Morrison KL & Weiss GA. Curr Opin Chem Biol. 2001 Jun;5(3):302-7). The reason alanine is used is that it is not bulky, is chemically inert, and yet mimics the methyl functional group of many other amino acids' secondary structure criteria. In many cases, if it is desirable to retain the size of the residue to be mutated, bulky amino acids such as valine or leucine can be used. Alanine scanning is a well-established technique that has been used for a long time.

[0145] The interaction between the binding domain and the epitope or region containing the epitope means that the binding domain shows a distinct affinity for the epitope / region containing the epitope on a specific protein or antigen (herein, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM and CD3, respectively), and generally does not show significant reactivity with proteins or antigens other than CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM or CD3. "Distinct affinity" includes bindings with a strong affinity of about 10 -6 M (KD) or more. Preferably, the binding affinity is about 10 -12 ~10 -8 M, 10-12 ~10 -9 M, 10 -12 ~10 -10 M, 10 -11 ~10 -8 M, preferably about 10 -11 ~10 -9 When it is M, the binding is considered specific. Whether the binding domain reacts specifically with or specifically binds to the target can be easily verified, inter alia, by comparing the reaction of the binding domain with the target protein or antigen with the reaction of the binding domain with a protein or antigen other than CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM or CD3. Preferably, the binding domain of the present invention does not bind essentially or substantially to a protein or antigen other than CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM or CD3 (i.e., the first binding domain cannot bind to a protein other than CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, or EpCAM, and the second binding domain cannot bind to a protein other than CD3). Having superior affinity characteristics compared to other HLE forms is an expected feature of the antigen-binding molecule according to the present invention. Such superior affinity suggests, as a result, an extended half-life in vivo. The longer the half-life of the antigen-binding molecule according to the present invention, the more the administration period and frequency can be reduced, which typically contributes to an improvement in patient compliance. This is particularly important because the antigen-binding molecule of the present invention is particularly beneficial for cancer patients who are extremely debilitated or have multiple diseases.

[0146] The terms "essentially / substantially non - binding" or "unable to bind" mean that the binding domain of the present invention does not bind to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM or a protein or antigen other than CD3 as an effector, that is, when the binding to each of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM or CD3 as an effector is set to 100%, it shows a reactivity of more than 30%, preferably more than 20%, more preferably more than 10%, particularly preferably more than 9%, 8%, 7%, 6% or 5% to a protein or antigen other than CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM or CD3 as an effector.

[0147] Specific binding is thought to be brought about by specific motifs within the amino acid sequences of the binding domain and the antigen. Thus, binding is achieved not only as a result of their primary, secondary and / or tertiary structures, but also as a result of secondary modifications of said structures. Specific interaction between the antigen - interacting site and its specific antigen can result in simple binding of said site to the antigen. Furthermore, specific interaction between the antigen - interacting site and its specific antigen can alternatively or additionally result in initiation of a signal, for example, by induction of conformational change of the antigen, oligomerization of the antigen, etc.

[0148] The term "variable" indicates variability within a sequence and refers to a part of an antibody domain or immunoglobulin domain (i.e., the "variable domain") that is involved in determining the specificity and binding affinity of a particular antibody. A single antigen - binding site is formed by the pairing of a variable heavy chain (VH) and a variable light chain (VL).

[0149] Variability is not uniformly distributed throughout the variable domains of antibodies but is concentrated in the subdomains of each of the heavy and light chain variable regions. These subdomains are called "hypervariable regions" or "complementary determining regions" (CDRs). The more conserved (i.e., non-hypervariable) portions of the variable domains are called "framework" regions (FRM or FR) and provide a scaffold for the six CDRs within the three-dimensional space forming the antigen-binding surface. The naturally occurring variable domains of the heavy and light chains each contain four FRM regions (FR1, FR2, FR3, and FR4), mostly in a β-sheet configuration, connected by three hypervariable regions that form loop connections and in some cases form part of the β-sheet structure. The hypervariable regions of each chain are held together in close proximity to those of the other chain by the FRM and contribute to the formation of the antigen-binding site (see Kabat et al., supra).

[0150] The terms "CDR" and its plural form "CDRs" refer to the complementary determining regions, three of which constitute the binding characteristics of the light chain variable region (CDR-L1, CDR-L2, and CDR-L3) and three of which constitute the binding characteristics of the heavy chain variable region (CDR-H1, CDR-H2, and CDR-H3). The CDRs contain most of the residues responsible for the specific interaction of the antibody with the antigen and thus contribute to the functional activity of the antibody molecule, i.e., the CDRs are the major determinants of antigen specificity.

[0151] The precise definition of the CDR boundaries and lengths follows various classification and numbering schemes. Thus, the CDRs can be referred to by Kabat, Chothia, contact, or any other arbitrary boundary definition, including the numbering scheme described herein. Despite differences in boundaries, each of these schemes has some overlap in the portions that make up the so-called "hypervariable regions" within the variable array. Thus, the definitions of CDRs by these schemes can differ in length and in the boundary regions relative to the adjacent framework regions. See, for example, Kabat (a method based on interspecies sequence variability), Chothia (a method based on crystallographic studies of antigen-antibody complexes), and / or MacCallum (supra Kabat et al.,; Chothia et al., J. Mol. Biol, 1987, 196:901-917; and MacCallum et al., J. Mol. Biol, 1996, 262:732). Yet another criterion for characterizing the antigen-binding site is the AbM definition used in Oxford Molecular's AbM antibody modeling software. See, for example, Protein Sequence and Structure Analysis of Antibody Variable Domains. In: Antibody Engineering Lab Manual (Ed.: Duebel, S. and Kontermann, R., Springer-Verlag, Heidelberg). Two residue identification techniques can be combined to define hybrid CDRs as long as they define regions that overlap but are not identical. However, numbering according to the so-called Kabat system is preferred.

[0152] Typically, a CDR forms a loop structure that can be classified as a canonical structure. The term "canonical structure" refers to the backbone conformation adopted by the antigen-binding (CDR) loop. From comparative structure studies, it has been found that 5 out of 6 antigen-binding loops have only a limited repertoire of useful conformations. Each canonical structure can be characterized by the torsional angles of the polypeptide backbone. Thus, corresponding loops between antibodies can have very similar three-dimensional structures, despite the high variability of the amino acid sequence in most of the loop (Chothia and Lesk, J. Mol. Biol., 1987, 196:901; Chothia et al., Nature, 1989, 342:877; Martin and Thornton, J. Mol. Biol, 1996, 263:800). Furthermore, there is a relationship between the adopted loop structure and the amino acid sequence surrounding it. The conformation of a particular canonical class is determined by the length of the loop and the amino acid residues present at important positions within the loop and within the conserved framework (i.e., outside the loop). Thus, based on the presence of these important amino acid residues, an assignment to a particular canonical class can be made.

[0153] The term "canonical structure" can also include considerations regarding the linear sequence of an antibody, as classified, for example, by Kabat (Kabat et al., supra). The Kabat numbering scheme is a widely adopted standard for numbering the amino acid residues of antibody variable domains in a consistent manner and is a preferred scheme for application to the present invention, as also mentioned elsewhere in this specification. Further structural considerations can also be used to determine the canonical structure of an antibody. For example, differences that are not fully reflected by the Kabat numbering method can be accounted for by the numbering scheme of Chothia et al., and / or elucidated by other techniques, such as crystallography and two-dimensional or three-dimensional computer modeling). Thus, a given antibody sequence can be classified into canonical classes that can, inter alia, (e.g., based on a desire to include various canonical structures in a library) identify appropriate chassis sequences. The Kabat numbering of the amino acid sequence of an antibody, and structural considerations such as those described by Chothia et al., supra, and their meaning with respect to the interpretation of the canonical aspects of antibody structure, are described in the literature. The subunit structures and three-dimensional arrangements of various classes of immunoglobulins are well known in the art. For an overview of antibody structure, see Antibodies: A Laboratory Manual, Cold Spring Harbor Laboratory, eds. Harlow et al., 1988.

[0154] The CDR3 of the light chain and especially the CDR3 of the heavy chain can be the most important determinants in antigen binding within the light and heavy chain variable regions. In some antigen-binding molecules, the heavy chain CDR3 is thought to constitute the major contact area between the antigen and the antibody. In vitro selection schemes that vary only the CDR3 can be used to change the binding properties of an antibody or to determine which residues contribute to antigen binding. Thus, CDR3 is typically the greatest source of molecular diversity within the antibody binding site. For example, H3 can be as short as about two amino acid residues or more than 26 amino acids.

[0155] In classical full-length antibodies or immunoglobulins, each light (L) chain is linked to the heavy (H) chain by one covalent disulfide bond, while the two H chains are linked to each other by one or more disulfide bonds depending on the H chain isotype. The CH domain closest to VH is usually designated CH1. The constant ("C") domains do not directly participate in antigen binding but exhibit various effector functions such as antibody-dependent, cell-mediated cytotoxicity, and complement activation. The Fc region of an antibody is contained within the heavy chain constant domains and can interact, for example, with Fc receptors located on the cell surface.

[0156] The sequences of antibody genes after construction and somatic mutation are extremely diverse, and these diversified genes are estimated to encode 10 10 different antibody molecules (Immunoglobulin Genes, 2 nd(ed., eds. Jonio et al., Academic Press, San Diego, CA, 1995). Thus, the immune system provides a repertoire of immunoglobulins. The term "repertoire" refers to at least one nucleotide sequence that is derived, in whole or in part, from at least one sequence encoding at least one immunoglobulin. This sequence can be generated by rearranging in vivo the V, D, and J segments of the heavy chain and the V and J segments of the light chain. Alternatively, this sequence can be generated from cells that respond to in vitro stimulation, for example, that cause rearrangement. Alternatively, some or all of this sequence may be obtained by DNA splicing, nucleotide synthesis, mutagenesis, and other methods (see, for example, U.S. Patent No. 5,565,332). The repertoire can contain only one sequence or can contain multiple sequences including sequences in a genetically diverse collection.

[0157] As used herein, the terms "Fc portion" or "Fc monomer" refer to a polypeptide comprising at least one domain having the function of the CH2 domain of an immunoglobulin molecule and at least one domain having the function of the CH3 domain. As is clear from the term "Fc monomer", the polypeptide containing these CH domains is a "polypeptide monomer". The Fc monomer includes a fragment of the constant region of an immunoglobulin excluding at least the first constant region immunoglobulin domain (CH1) of the heavy chain, but maintains at least a functional portion of one CH2 domain and a functional portion of one CH3 domain, and can be a polypeptide in which the CH2 domain is on the amino-terminal side of the CH3 domain. In a preferred embodiment of this definition, the Fc monomer can be a polypeptide constant region comprising a part of the Ig-Fc hinge region, the CH2 region, and the CH3 region, with the hinge region on the amino-terminal side of the CH2 domain. The hinge region of the present invention is envisioned to facilitate dimerization. Such Fc polypeptide molecules can be obtained, for example, by papain digestion of the immunoglobulin region (although, of course, a dimer of two Fc polypeptides results), but are not limited thereto. In another embodiment of this definition, the Fc monomer can be a polypeptide region comprising a part of the CH2 region and the CH3 region. Such Fc polypeptide molecules can be obtained, for example, by pepsin digestion of an immunoglobulin molecule, but are not limited thereto. In one embodiment, the polypeptide sequence of the Fc monomer is substantially similar to the Fc polypeptide sequences of the Fc regions of IgG1, IgG2, IgG3, IgG4, IgM, IgA, IgD, and IgE (see, for example, Padlan, Molecular Immunology, 31(3), 169-217 (1993)). Because there are certain variations among immunoglobulins, and for clarity only, the Fc monomer refers to the last two heavy chain constant region immunoglobulin domains of IgA, IgD, and IgG, and the last three heavy chain constant region immunoglobulin domains of IgE and IgM. As mentioned, the Fc monomer can also include a flexible hinge on the N-terminal side of these domains.In the case of IgA and IgM, the Fc monomer may contain a J chain. In the case of IgG, the Fc portion includes the CH2 and CH3 of the immunoglobulin domain, and the hinge between the first two domains and CH2. Although the boundaries of the Fc portion may vary, an example of the human IgG heavy chain Fc portion containing the functional hinge, CH2, and CH3 domains can be defined, for example, as residues D231 to P476 (corresponding to D234 in Table 1 below of the hinge domain), up to L476 at the carboxyl terminus of the CH3 domain (in the case of IgG4) (the numbering is according to Kabat). Two Fc portions or Fc monomers fused to each other via a peptide linker are preferred examples of spacers between the two bispecific entities of the antigen-binding molecule of the present invention, and this can also be defined as an scFc domain.

[0158] In one embodiment of the present invention, it is contemplated that the scFc domains as disclosed herein, Fc monomers fused to each other, are included only in the spacer of the antigen-binding molecule.

[0159] In accordance with the present invention, the hinge region of IgG can be identified by similarity using the Kabat numbering as set forth in Table 1. In accordance with the foregoing, in the case of the hinge domain / region of the present invention, the minimum requirement is assumed to include amino acid residues corresponding to the IgG1 sequence stretch of D231 D234~P243 according to Kabat numbering. Similarly, the hinge domain / region of the present invention is assumed to include or consist of the hinge sequence DKTHTCPPCP (SEQ ID NO: 330) of IgG1 (corresponding to D234~P243 as a stretch as shown in Table 1 below, but variants of the said sequence are also assumed if the hinge region still promotes dimerization). In a preferred embodiment of the present invention, the glycosylation site at Kabat position 314 of the CH2 domain in the spacer of the antigen-binding molecule is removed by N314X substitution (X is any amino acid other than Q). The said substitution is preferably N314G substitution. In a more preferred embodiment, the CH2 domain further includes the following substitutions (positions according to Kabat), V321C and R309C (by these substitutions, intradomain cysteine disulfide bridges are introduced at Kabat positions 309 and 321).

[0160] The spacer of the antigen-binding molecule of the present invention may also be an scFc domain that includes or consists of, in order from amino to carboxyl, DKTHTCPPCP (SEQ ID NO: 330) (i.e., hinge)-CH2-CH3-linker-DKTHTCPPCP (SEQ ID NO: 330) (i.e., hinge)-CH2-CH3. The peptide linker of the antigen-binding molecule described above is, in a preferred embodiment, the amino acid sequence Gly-Gly-Gly-Gly-Ser, i.e., Gly4Ser (SEQ ID NO: 7), or a polymer thereof, i.e., (Gly4Ser)x (x is an integer of 5 or more (e.g., 5, 6, 7, 8, etc., or more), and 6 is preferred ((Gly4Ser)6)). According to the present invention, Ser can advantageously be replaced with Gln as disclosed herein. The construct may further include the aforementioned substitution: N314X, preferably N314G, and / or additional substitutions V321C and R309C. In a preferred embodiment of the antigen-binding molecule defined above in this specification, the second domain is expected to bind to the extracellular epitope of the human and / or Macaca CD3ε chain.

[0161]

Table 1

[0162] In a further embodiment of the present invention, the hinge domain / region includes or consists of the hinge sequence ERKCCVECPPCP (SEQ ID NO: 331) of the IgG2 subtype, the hinge sequence ELKTPLDTTHTCPRCP (SEQ ID NO: 332) or ELKTPLGDTTHTCPRCP (SEQ ID NO: 333) of the IgG3 subtype, and / or the hinge sequence ESKYGPPCPSCP (SEQ ID NO: 444) of the IgG4 subtype. The hinge sequence of the IgG1 subtype may be the following sequence, EPKSCDKTHTCPPCP (shown in Table 1 and SEQ ID NO: 445). Thus, these core hinge regions are also envisioned in the context of the present invention.

[0163] The positions and sequences of the IgG CH2 and IgG CD3 domains can be identified by similarity using the Kabat numbering described in Table 2.

[0164]

Table 2

[0165] In one embodiment of the present invention, in the CH3 domain of the first or both Fc monomers, the amino acid residues highlighted in bold are deleted.

[0166] The peptide linker that fuses the polypeptide monomers of the spacer (the "Fc portion" or "Fc monomer") to each other preferably contains at least 25 amino acid residues (such as 25, 26, 27, 28, 29, 30, etc.). This peptide linker more preferably contains at least 30 amino acid residues (such as 30, 31, 32, 33, 34, 35, etc.). Also, it is preferred that the linker contains a maximum of 40 amino acid residues, more preferably a maximum of 35 amino acid residues, and most preferably exactly 30 amino acid residues. A preferred embodiment of such a peptide linker is characterized by the amino acid sequence Gly-Gly-Gly-Gly-Ser, i.e., Gly4Ser (SEQ ID NO: 7), or a polymer thereof, i.e., (Gly4Ser)x where x is an integer of 5 or more (for example, 6, 7, or 8). Preferably, the integer is 6 or 7, and more preferably the integer is 6.

[0167] When using a linker to fuse a first domain to a second domain and / or a third domain to a fourth domain and / or the second and third domains to a spacer, the linker preferably has a length and sequence sufficient for each of the first and second domains to reliably retain their binding specificities based on their differences independently of each other. Regarding the peptide linker that links at least two binding domains (or two variable domains) of the antigen-binding molecule of the present invention, those peptide linkers preferably contain only a few amino acid residues (for example, those containing 12 or fewer amino acid residues). Therefore, peptide linkers of 12, 11, 10, 9, 8, 7, 6, or 5 amino acid residues are preferred. A putative peptide linker having less than 5 amino acids contains 4, 3, 2, or 1 amino acid, and a Gly-rich linker is preferred. A preferred embodiment of the peptide linker for fusing the first domain and the second domain is shown in SEQ ID NO: 1. A preferred linker embodiment of the peptide linker for fusing the second domain and the third domain to a spacer is the (Gly)4-linker, also called the G4-linker.

[0168] In the context of being particularly preferred in relation to one of the "peptide linkers" described above, the "single" amino acid is Gly. Thus, the peptide linker can consist of the single amino acid Gly. In a preferred embodiment of the present invention, the peptide linker is characterized by the amino acid sequence Gly-Gly-Gly-Gly-Ser, i.e., Gly4Ser (SEQ ID NO: 1), or a polymer thereof, i.e., (Gly4Ser)x (x is an integer of 1 or more (e.g., 2 or 3)). Preferred linkers are shown in SEQ ID NOs: 1 to 12. The characteristics of the peptide linker that do not promote secondary structure are known in the art and are described, for example, in Dall’Acqua et al. (Biochem. (1998) 37, 9266-9273), Cheadle et al. (Mol Immunol (1992) 29, 21-30), and Raag and Whitlow (FASEB (1995) 9(1), 73-80). Furthermore, a peptide linker that does not promote any secondary structure is preferred. The mutual linkage of the domains can be provided, for example, by the genetic manipulation described in the examples. Methods for preparing fusion and operably linked bispecific single-chain constructs and expressing them in mammalian cells or bacteria are well known in the art (e.g., WO 99 / 54440 pamphlet or Sambrook et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, 2001).

[0169] In a preferred embodiment of the antigen-binding molecule of the present invention, the first domain and the second domain form an antigen-binding molecule in a form selected from the group consisting of (scFv)2, scFv-single domain mAb, diabody, and oligomers of any of these forms.

[0170] According to a particularly preferred embodiment, and as described in the appended examples, the first and second domains of the antigen-binding molecule of the present invention are "bispecific single-chain antigen-binding molecules", more preferably bispecific "single-chain Fv" (scFv). The two domains of the Fv fragment, VL and VH, are encoded by separate genes, but they can be joined by a synthetic linker as described hereinbefore which makes it possible to produce them as a single protein chain using recombinant methods; see, for example, Huston et al. (1988) Proc. Natl. Acad. Sci USA 85:5879-5883. These antibody fragments are obtained using conventional techniques known to those skilled in the art, and the fragments are evaluated for function in the same manner as full or whole-length antibodies. Thus, a single-chain variable fragment (scFv) is a fusion protein of the variable region of the heavy chain (VH) and the variable region of the light chain (VL) of an immunoglobulin, and is usually linked by a short linker peptide of about 10 to about 25 amino acids, preferably about 15 to 20 amino acids. The linker usually consists of glycine for flexibility and is rich in serine or threonine for solubility, and can be either connecting the N-terminus of VH to the C-terminus of VL or vice versa. This protein retains the specificity of the original immunoglobulin despite removal of the constant regions and introduction of the linker.

[0171] Bispecific single-chain antigen-binding molecules are known in the art and are described in WO 99 / 54440, Mack, J. Immunol. (1997), 158, 3965-3970, Mack, PNAS (1995), 92, 7021-7025, Kufer, Cancer Immunol. Immunother., (1997), 45, 193-197, Loeffler, Blood, (2000), 95, 6, 2098-2103, Bruehl, Immunol., (2001), 166, 2420-2426, Kipriyanov, J. Mol. Biol., (1999), 293, 41-56. Techniques described for generating single-chain antibodies (see, inter alia, U.S. Pat. No. 4,946,778, Kontermann and Duebel (2010) supra, and Little (2009) supra) can be adapted to generate single-chain antigen-binding molecules that specifically recognize a selected target.

[0172] A bivalent (also called divalent) or bispecific single-chain variable fragment (bi-scFv or di-scFv having the format (scFv)2) can be designed by linking two scFv molecules (e.g., by a linker as previously described herein). When these two scFv molecules have the same binding specificity, the resulting (scFv)2 molecule is preferably called bivalent (i.e., having a valence of 2 for the same target epitope). When these two scFv molecules have different binding specificities, the resulting (scFv)2 molecule is preferably called bispecific. This linking can be done by generating a single peptide chain having two VH regions and two VL regions to obtain a tandem scFv (see, for example, Kufer P. et al., (2004) Trends in Biotechnology 22(5):238-244). Another possibility is to generate an scFv molecule using a linker peptide that is too short (e.g., about 5 amino acids) to fold the two variable regions together and then dimerize this scFv. This type is known as a diabody (see, for example, Hollinger, Philipp et al., (July 1993) Proceedings of the National Academy of Sciences of the United States of America 90(14):6444-8).

[0173] In accordance with the present invention, any of the first, second, third, and / or fourth domains may each comprise a single domain antibody, a variable domain of a single domain antibody, or at least a CDR of a single domain antibody. A single domain antibody comprises only one (monomeric) antibody variable domain that can independently selectively bind a specific antigen to which other V regions or domains can bind. The first single domain antibodies were designed from heavy chain antibodies found in camels, which are called V H H fragments. Cartilaginous fish also have heavy chain antibodies (IgNAR), from which V NARSingle-domain antibodies, also known as fragments, can be obtained. An alternative approach is, for example, to split the dimeric variable domains from common immunoglobulins derived from humans or rodents into monomers, thereby obtaining VH or VL as single-domain Abs. Currently, most of the research on single-domain antibodies is based on the heavy-chain variable region, but nanobodies derived from the light chain have also been shown to specifically bind to target epitopes. Examples of single-domain antibodies are referred to as sdAbs, nanobodies, or single variable domain antibodies.

[0174] Thus, (single-domain mAb)2 is a monoclonal antigen-binding molecule composed of (at least) two single-domain monoclonal antibodies individually selected from the group comprising V H V L V H H and V NAR The linker is preferably in the form of a peptide linker. Similarly, "scFv-single-domain mAb" is a monoclonal antigen-binding molecule composed of at least one single-domain antibody as described above and one scFv molecule as described above. Again, in this case, the linker is preferably in the form of a peptide linker.

[0175] Whether an antigen-binding molecule competes with another given antigen-binding molecule for binding can be measured in a competition assay such as a competitive ELISA or a cell-based competition assay. Avidin-binding microparticles (beads) can also be used. Similar to an ELISA plate coated with avidin, each of these beads can be used as a substrate when reacting with a biotinylated protein, and the assay can be performed thereon. The antigen is coated on the beads and then pre-coated with the first antibody. A secondary antibody is added to confirm any additional binding. Possible means for reading include flow cytometry.

[0176] T cells, or T lymphocytes, are a type of lymphocyte (which itself is a type of white blood cell) that plays a central role in cellular immunity. There are several subsets of T cells, each with a different function. T cells can be distinguished from other lymphocytes, such as B cells and NK cells, by the presence of the T cell receptor (TCR) on their cell surface. The TCR is involved in the recognition of antigens bound to major histocompatibility complex (MHC) molecules and is composed of two different protein chains. In 95% of T cells, the TCR is composed of an alpha (α) chain and a beta (β) chain. When the TCR binds to an antigen peptide and MHC (peptide / MHC complex), the T lymphocyte is activated through a series of biochemical events mediated by associated enzymes, co-receptors, specialized adapter molecules, and activated or released transcription factors.

[0177] The CD3 receptor complex is a protein complex composed of four chains. In mammals, this complex contains the CD3γ (gamma) chain, the CD3δ (delta) chain, and two CD3ε (epsilon) chains. These chains associate with the T cell receptor (TCR) and the so-called ζ (zeta) chain to form the T cell receptor-CD3 complex and generate activation signals in T lymphocytes. The CD3γ (gamma), CD3δ (delta), and CD3ε (epsilon) chains are cell surface proteins of the highly related immunoglobulin superfamily that contain a single extracellular immunoglobulin domain. The intracellular tail of the CD3 molecule contains a single conserved motif known as the immunoreceptor tyrosine-based activation motif, or ITAM for short, which is essential for the signaling ability of the TCR. The CD3 epsilon molecule is a polypeptide encoded by the CD3E gene located on human chromosome 11. The most preferred epitope of CD3 epsilon is contained within the range of amino acid residues 1 to 27 of the extracellular domain of human CD3 epsilon. The antigen-binding molecule according to the present invention is typically and preferably not expected to show much unwanted non-specific T cell activation in a particular immunotherapy. In other words, this means reducing the risk of side effects.

[0178] Lysis of redirected target cells via recruitment of T cells by a multi-lock multi-target minimal bispecific antigen-binding molecule involves the formation of a cytolytic synapse and the delivery of perforin and granzyme. The bound T cells are capable of continuously lysing target cells and are not affected by immune escape mechanisms that interfere with the processing and presentation of peptide antigens or clonal T cell differentiation (see, for example, WO 2007 / 042261 pamphlet).

[0179] The cytotoxicity mediated by the antigen-binding molecules of the present invention can be measured in various ways. Effector cells can be, for example, stimulated enriched (human) CD8-positive T cells or unstimulated (human) peripheral blood mononuclear cells (PBMCs). If the target cells are of macaque origin or express or are transfected with macaque CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM and are bound by the first domain, the effector cells must also be of macaque origin such as a macaque T cell line (e.g., 4119LnPx). The target cells must express CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM, e.g., at least the extracellular domain of human or macaque CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM. The target cells can be cell lines (such as CHO) stably or transiently transfected with CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM, e.g., human or macaque CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM. Usually, EC 50The value is expected to be lower in target cell lines that express CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM at higher levels on the cell surface. The effector to target cell (E:T) ratio is usually about 10:1, but can vary. The cytotoxic activity of the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM bispecific antigen-binding molecule is 51 measurable by a Cr release assay (incubation time of about 18 hours) or a FACS-based cytotoxicity assay (incubation time of about 48 hours). The incubation time of the assay (cytotoxic reaction) can also be changed. Other methods of measuring cytotoxicity are known to those skilled in the art and include MTT or MTS assays, ATP-based assays including bioluminescence assays, sulforhodamine B (SRB) assays, WST assays, clonogenic assays, and ECIS technology.

[0180] The cytotoxic activity mediated by the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM×CD3 bispecific antigen-binding molecule of the present invention is preferably measured by a cell-based cytotoxicity assay. The cytotoxic activity is 51 also measurable by a Cr release assay. The cytotoxic activity is EC 50 represented by a value that corresponds to the half maximal effective concentration (the concentration of the antigen-binding molecule that induces a cytotoxic reaction midway between the baseline and the maximum value). Preferably, the EC 50 value of the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM×CD3 bispecific antigen-binding molecule is ≦5000 pM or ≦4000 pM, more preferably ≦3000 pM or ≦2000 pM, even more preferably ≦1000 pM or ≦500 pM, even more preferably ≦400 pM or ≦300 pM, even more preferably ≦200 pM, even more preferably ≦100 pM, even more preferably ≦50 pM, even more preferably ≦20 pM or ≦10 pM, and most preferably ≦5 pM.

[0181] the above-mentioned given EC 50 values can be measured in various assays. When using stimulated / concentrated CD8 + T cells as effector cells, those skilled in the art recognize that the EC 50 value can be expected to be lower compared to unstimulated PBMC. Furthermore, the EC 50 value can be expected to be lower when the target cells express a large number of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM, compared to rats with low target expression. For example, when stimulated / concentrated human CD8 + T cells are used as effector cells (and cells transfected with CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM (e.g., CHO cells) or any of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM-positive human cells are used as target cells), the EC of the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM×CD3 bispecific antigen-binding molecule 50 value is preferably ≦1000 pM, more preferably ≦500 pM, even more preferably ≦250 pM, even more preferably ≦100 pM, even more preferably ≦50 pM, even more preferably ≦10 pM, and most preferably ≦5 pM. When human PBMC are used as effector cells, the EC of the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM×CD3 bispecific antigen-binding molecule 50The value is preferably ≤5000 pM or ≤4000 pM (particularly when the target cells are CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM positive cell lines), more preferably ≤2000 pM (particularly when the target cells are cells transfected with CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM (e.g., CHO cells)), even more preferably ≤1000 pM or ≤500 pM, still even more preferably ≤200 pM, still even more preferably ≤150 pM, still even more preferably ≤100 pM, and most preferably ≤50 pM or less. When a cynomolgus T cell line such as LnPx4119 is used as the effector cell and a cell line transfected with cynomolgus CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM (e.g., CHO cells) is used as the target cell line, the EC of the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM×CD3 bispecific antigen-binding molecule 50 The value is preferably ≤2000 pM or ≤1500 pM, more preferably ≤1000 pM or ≤500 pM, even more preferably ≤300 pM or ≤250 pM, still even more preferably ≤100 pM, and most preferably ≤50 pM.

[0182] Preferably, the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM×CD3 bispecific antigen-binding molecule of the present invention does not induce / mediate, or essentially does not induce / mediate, the lysis of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM-negative cells such as CHO cells. The terms "does not induce lysis", "essentially does not induce lysis", "does not mediate lysis", or "essentially does not mediate lysis" mean that when the lysis of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM-positive human cell lines is taken as 100%, the antigen-binding molecule of the present invention does not induce or mediate lysis of more than 30% of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM-negative cells, preferably not more than 20%, more preferably not more than 10%, particularly preferably not more than 9%, 8%, 7%, 6%, or 5%. This usually applies to a concentration of the antigen-binding molecule of up to 500 nM. Those skilled in the art know methods for measuring cell lysis without further effort. Further, specific instructions on methods for measuring cell lysis are taught herein.

[0183] The difference in cytotoxic activity between the monomeric and dimeric isoforms of an individual CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM×CD3 bispecific antigen-binding molecule is referred to as the "efficacy gap". This efficacy gap can be calculated, for example, as the ratio between the EC 50 value of the monomeric form and the EC 50 value of the dimeric form of the molecule. The efficacy gap of the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM×CD3 bispecific antigen-binding molecule of the present invention is preferably ≦5, more preferably ≦4, even more preferably ≦3, even more preferably ≦2, and most preferably ≦1.

[0184] The first, second, third and / or fourth binding domains of the antigen-binding molecule of the present invention are preferably heterospecific within members of the mammalian order of primates. Heterospecific CD3 binding domains are, for example, those described herein and in WO 2008 / 119567 pamphlet. According to one embodiment, the first binding domain and the third binding domain, in addition to binding to human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM and human CD3, also bind to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM / CD3 of primates including (but not limited to) New World primates (e.g., marmoset (Callithrix jacchus), cotton-top tamarin (Saguinus Oedipus) or squirrel monkey (Saimiri sciureus)), Old World primates (e.g., baboon and macaque), langur, and non-human homininae.

[0185] In one embodiment of the antigen-binding molecule of the present invention, the first domain binds to human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM, and further binds to cynomolgus CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM (e.g., Macaca fascicularis CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM), and more preferably further binds to cynomolgus CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM expressed on the surface of cells such as, for example, CHO cells or 293 cells. The affinity of the first domain for CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM (preferably human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM) is preferably ≤100 nM or ≤50 nM, more preferably ≤25 nM or ≤20 nM, more preferably ≤15 nM or ≤10 nM, even more preferably ≤5 nM, even more preferably ≤2.5 nM or ≤2 nM, even more preferably ≤1 nM, even more preferably ≤0.6 nM, even more preferably ≤0.5 nM, and most preferably ≤0.4 nM. The affinity can be measured, for example, by a BIAcore assay or a Scatchard assay. Other methods for determining affinity are also well known to those skilled in the art. The affinity of the first domain for cynomolgus CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM is preferably ≤15 nM, more preferably ≤10 nM, even more preferably ≤5 nM, even more preferably ≤1 nM, even more preferably ≤0.5 nM, even more preferably ≤0.1 nM, and most preferably ≤0.05 nM or even ≤0.01 nM.

[0186] Preferably, the affinity gap of the antigen-binding molecule according to the present invention for the binding of cynomolgus CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM to human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM [cynomolgus CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM:human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM] (determined by surface plasmon resonance analysis such as BiaCore™ or Scatchard analysis) is <100, preferably <20, more preferably <15, even more preferably <10, even more preferably <8, more preferably <6, and most preferably <2. The preferred range of the affinity gap of the antigen-binding molecule according to the present invention for the binding of cynomolgus CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM to human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM is 0.1 to 20, more preferably 0.2 to 10, even more preferably 0.3 to 6, even more preferably 0.5 to 3 or 0.5 to 2.5, and most preferably 0.5 to 2 or 0.6 to 2.

[0187] The second and fourth binding domains of the antigen-binding molecule of the present invention typically bind to human CD3 epsilon and / or Macaca CD3 epsilon. In a preferred embodiment, when a selectivity gap is achieved, the second and fourth binding domains, or the first and third binding domains, further bind to Callithrix jacchus, Saguinus oedipus or Saimiri sciureus CD3 epsilon. Callithrix jacchus and Saguinus oedipus are both New World primates belonging to the family Callitrichidae, while Saimiri sciureus is a New World primate belonging to the family Cebidae. The binding domain may preferably be selected from the sequences specified herein as "I2L" (or the synonymous "I2L0"), "I2M" and "I2M2", more preferably "I2L" or "I2L0".

[0188] In the antigen-binding molecule of the present invention, the preferred second and fourth binding domains that bind to the extracellular epitopes of the human and / or Macaca CD3 epsilon chain preferably comprise a VL region comprising CDR-L1, CDR-L2 and CDR-L3 selected from: (a) a VL region comprising CDR-L1, CDR-L2 and CDR-L3 selected from SEQ ID NOs: 40-42, 48-50, 56-58, 64-66, 72-74, 439-441, preferably 64-66; (b) CDR-L1 shown in SEQ ID NO: 27 of WO 2008 / 119567 pamphlet, CDR-L2 shown in SEQ ID NO: 28 of WO 2008 / 119567 pamphlet, and CDR-L3 shown in SEQ ID NO: 29 of WO 2008 / 119567 pamphlet; (c) CDR-L1 shown in SEQ ID NO: 117 of WO 2008 / 119567 pamphlet, CDR-L2 shown in SEQ ID NO: 118 of WO 2008 / 119567 pamphlet, and CDR-L3 shown in SEQ ID NO: 119 of WO 2008 / 119567 pamphlet; (d) CDR-L1 shown in SEQ ID NO: 153 of WO 2008 / 119567 pamphlet, CDR-L2 shown in SEQ ID NO: 154 of WO 2008 / 119567 pamphlet, and CDR-L3 shown in SEQ ID NO: 155 of WO 2008 / 119567 pamphlet; and (e) A VL region containing CDR-L1, CDR-L2, and CDR-L3 of SEQ ID NOs: 420 to 422.

[0189] In a more preferred embodiment of the antigen-binding molecule of the present invention, the second domain and the fourth domain that bind to the extracellular epitope of the human and / or Macaca CD3 epsilon chain preferably comprise a VH region containing CDR-H1, CDR-H2, and CDR-H3 selected from the following: (a) A VH region containing CDR-H1, CDR-H2, and CDR-H3 selected from SEQ ID NOs: 37 to 39, 45 to 47, 53 to 55, 61 to 63, 69 to 71, and 436 to 438, preferably 61 to 63; (b) CDR-H1 shown in SEQ ID NO: 12 of WO 2008 / 119567 pamphlet, CDR-H2 shown in SEQ ID NO: 13 of WO 2008 / 119567 pamphlet, and CDR-H3 shown in SEQ ID NO: 14 of WO 2008 / 119567 pamphlet; (c) CDR-H1 shown in SEQ ID NO: 30 of WO 2008 / 119567 pamphlet, CDR-H2 shown in SEQ ID NO: 31 of WO 2008 / 119567 pamphlet, and CDR-H3 shown in SEQ ID NO: 32 of WO 2008 / 119567 pamphlet; (d) CDR-H1 shown in SEQ ID NO: 48 of WO 2008 / 119567 pamphlet, CDR-H2 shown in SEQ ID NO: 49 of WO 2008 / 119567 pamphlet, and CDR-H3 shown in SEQ ID NO: 50 of WO 2008 / 119567 pamphlet; (e) CDR-H1 shown in SEQ ID NO: 66 of WO 2008 / 119567 pamphlet, CDR-H2 shown in SEQ ID NO: 67 of WO 2008 / 119567 pamphlet, and CDR-H3 shown in SEQ ID NO: 68 of WO 2008 / 119567 pamphlet; (f) CDR-H1 shown in SEQ ID NO: 84 of WO 2008 / 119567 pamphlet, CDR-H2 shown in SEQ ID NO: 85 of WO 2008 / 119567 pamphlet, and CDR-H3 shown in SEQ ID NO: 86 of WO 2008 / 119567 pamphlet; (g) CDR-H1 shown in SEQ ID NO: 102 of WO 2008 / 119567 pamphlet, CDR-H2 shown in SEQ ID NO: 103 of WO 2008 / 119567 pamphlet, and CDR-H3 shown in SEQ ID NO: 104 of WO 2008 / 119567 pamphlet; (h) CDR-H1 shown in SEQ ID NO: 120 of WO 2008 / 119567 pamphlet, CDR-H2 shown in SEQ ID NO: 121 of WO 2008 / 119567 pamphlet, and CDR-H3 shown in SEQ ID NO: 122 of WO 2008 / 119567 pamphlet; (i) CDR-H1 shown in SEQ ID NO: 138 of WO 2008 / 119567 pamphlet, CDR-H2 shown in SEQ ID NO: 139 of WO 2008 / 119567 pamphlet, and CDR-H3 shown in SEQ ID NO: 140 of WO 2008 / 119567 pamphlet; (j) CDR-H1 shown in SEQ ID NO: 156 of WO 2008 / 119567 pamphlet, CDR-H2 shown in SEQ ID NO: 157 of WO 2008 / 119567 pamphlet, and CDR-H3 shown in SEQ ID NO: 158 of WO 2008 / 119567 pamphlet; (k) CDR-H1 shown in SEQ ID NO: 174 of WO 2008 / 119567 pamphlet, CDR-H2 shown in SEQ ID NO: 175 of WO 2008 / 119567 pamphlet, and CDR-H3 shown in SEQ ID NO: 176 of WO 2008 / 119567 pamphlet; and (l) A VH region comprising CDR-H1, CDR-H2 and CDR-H3 of SEQ ID NOs: 423 to 425.

[0190] In a preferred embodiment of the antigen-binding molecule of the present invention, the above three groups of VL CDRs are combined with the above ten groups of VH CDRs in the third binding domain to form (30) groups each containing CDR-L1 to 3 and CDR-H1 to 3.

[0191] In the antigen-binding molecule of the present invention, the third domain that binds to CD3 is shown in SEQ ID NOs: 17, 21, 35, 39, 53, 57, 71, 75, 89, 93, 107, 111, 125, 129, 143, 147, 161, 165, 179 or 183 of WO 2008 / 119567 pamphlet, or preferably, it preferably comprises a VL region selected from the group consisting of SEQ ID NOs: 44, 52, 60, 68 and 76, preferably 68, according to the present invention.

[0192] The third domain that binds to CD3 preferably also comprises a VH region selected from the group consisting of SEQ ID NOs: 15, 19, 33, 37, 51, 55, 69, 73, 87, 91, 105, 109, 123, 127, 141, 145, 159, 163, 177 or 181 of WO 2008 / 119567 pamphlet, or preferably, SEQ ID NOs: 43, 51, 59, 67 and 75, preferably 67, according to the present invention.

[0193] More preferably, the antigen-binding molecule of the present invention is characterized by a second domain and a fourth domain that binds to CD3 and preferably comprises a VL region and a VH region selected from the group consisting of: (a) A VL region selected from SEQ ID NOs: 44, 52, 60, 68, 76, and 443, and a VH region selected from SEQ ID NOs: 43, 51, 59, 67, 75, and 442; (b) A VL region shown in SEQ ID NO: 17 or 21 of WO 2008 / 119567 pamphlet and a VH region shown in SEQ ID NO: 15 or 19 of WO 2008 / 119567 pamphlet; (c) A VL region shown in SEQ ID NO: 35 or 39 of WO 2008 / 119567 pamphlet and a VH region shown in SEQ ID NO: 33 or 37 of WO 2008 / 119567 pamphlet; (d) A VL region shown in SEQ ID NO: 53 or 57 of WO 2008 / 119567 pamphlet and a VH region shown in SEQ ID NO: 51 or 55 of WO 2008 / 119567 pamphlet; (e) A VL region shown in SEQ ID NO: 71 or 75 of WO 2008 / 119567 pamphlet and a VH region shown in SEQ ID NO: 69 or 73 of WO 2008 / 119567 pamphlet; (f) A VL region shown in SEQ ID NO: 89 or 93 of WO 2008 / 119567 pamphlet and a VH region shown in SEQ ID NO: 87 or 91 of WO 2008 / 119567 pamphlet; (g) A VL region shown in SEQ ID NO: 107 or 111 of WO 2008 / 119567 pamphlet and a VH region shown in SEQ ID NO: 105 or 109 of WO 2008 / 119567 pamphlet; (h) A VL region shown in SEQ ID NO: 125 or 129 of WO 2008 / 119567 pamphlet and a VH region shown in SEQ ID NO: 123 or 127 of WO 2008 / 119567 pamphlet; (i) A VL region shown in SEQ ID NO: 143 or 147 of WO 2008 / 119567 pamphlet and a VH region shown in SEQ ID NO: 141 or 145 of WO 2008 / 119567 pamphlet; (j) The VL region shown in SEQ ID NO: 161 or 165 of WO 2008 / 119567 pamphlet and the VH region shown in SEQ ID NO: 159 or 163 of WO 2008 / 119567 pamphlet; and (k) The VL region shown in SEQ ID NO: 179 or 183 of WO 2008 / 119567 pamphlet and the VH region shown in SEQ ID NO: 177 or 181 of WO 2008 / 119567 pamphlet.

[0194] Also, in relation to the antigen-binding molecule of the present invention, the second and fourth domains that bind to CD3 and contain the VL region shown in SEQ ID NO: 68 and the VH region shown in SEQ ID NO: 67 are also preferred.

[0195] According to a preferred embodiment of the antigen-binding molecule of the present invention, the first domain and / or the third domain has the following form: The pair of the VH region and the VL region is in the form of a single-chain antibody (scFv). The VH and VL regions are arranged in the order of VH-VL or VL-VH. It is preferred that the VH region is arranged at the N-terminus of the linker sequence and the VL region is arranged at the C-terminus of the linker sequence.

[0196] The present invention further provides an antigen-binding molecule comprising or having an amino acid sequence (fully bispecific antigen-binding molecule) selected from the group consisting of 673, 676, 679, 682, 685, 688, 691, 694, 697, 700, 703, 706, 709, 712, 715, 718, 721, 724, 727, 730, 733, 736, 739, 742, 745, 748, 751, 754, 757, 760, 763, 766, 769, 772, 775, 778, 781, 784, 787, 790, 793, 796, 799, 802, 805, 808, 811, 814, 817, 820, 823, 826, 829, 832, 835, 838, 841, 844, 847, 850, 853, 856, 859, 862, 865, 868, 871, 1437, 1440, 1443, 1446, 1449, 1452, 1455, 1458, 1461, 1464, 1467, 1470, 1473, 1476, 1479, 1482, 1485, 1488, 1499, 1667, 1670, 1673, 1676, 1679, 1682, 1685, 1688, 1691, 1694, 1697, 1700, 1703, 1706, 1709, 1712, 1715, 1718, 1721, 1724, 1727, 1730, 1733, 1736, 1739, 1742, 1745, 1748, 1751, 1754, 1757, 1760, 1763, 1766, 1769, 1772, 1775, 1778, 1781, 1784, 1787, 1790, 1793, 1796, 1799, 1802, 1805, 1808, 1811, 1814, 1817, 1820, 1823, 1826, and 1829, preferably any of 1437, or having an amino acid sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identity with said sequence.

[0197] Covalent modifications of the antigen-binding molecule are also included within the scope of the present invention and are generally carried out, although not necessarily, after translation. For example, some types of covalent modifications of the antigen-binding molecule are introduced into the molecule by reacting specific amino acid residues of the antigen-binding molecule with an organic derivatizing agent capable of reacting with the selected side chain or N-terminal or C-terminal residue.

[0198] Cysteinyl residues are most commonly derivatized by reaction with α-haloacetates (and corresponding amines), such as chloroacetic acid or chloroacetamide, to yield carboxymethyl or carboxamidomethyl derivatives. Cysteinyl residues may also be derivatized by reaction with bromotrifluoroacetone, α-bromo-β-(5-imidazolyl)propionic acid, chloroacetyl phosphate, N-alkylmaleimide, 3-nitro-2-pyridyl disulfide, methyl 2-pyridyl disulfide, p-chloromercuribenzoic acid, 2-chloromercuri-4-nitrophenol, or chloro-7-nitrobenzo-2-oxa-1,3-diazole.

[0199] Histidyl residues are derivatized by reaction with diethyl pyrocarbonate at pH 5.5 - 7.0, because this reagent is relatively specific for the histidyl side chain. Para-bromophenacyl bromide is also useful, and the reaction is preferably carried out in 0.1 M sodium cacodylate at pH 6.0. Lysinyl and amino-terminal residues react with succinic or other carboxylic acid anhydrides. Derivatization with these agents has the effect of inverting the charge of the lysinyl residues. Other suitable reagents for derivatizing alpha-amino-containing residues include imidoesters such as methyl picolinimidate; pyridoxal phosphate; pyridoxal; chloroborohydride; trinitrobenzenesulfonic acid; O-methylisourea; 2,4-pentanedione; and transaminase-catalyzed reaction with glyoxylate.

[0200] Arginyl residues are modified by reaction with one or several conventional reagents, in particular phenylglyoxal, 2,3-butanedione, 1,2-cyclohexanedione, and ninhydrin. Because of the high pKa of the guanidine functional group, derivatization of arginine residues requires that the reaction be carried out under alkaline conditions. Furthermore, these reagents may react with the lysine groups and arginine epsilon-amino groups.

[0201] Specific modification of tyrosine residues can be carried out for the purpose of introducing a spectral label into the tyrosine residue, especially by reacting with an aromatic diazonium compound or tetranitromethane. Most commonly, N-acetylimidazole and tetranitromethane are used to form O-acetyltyrosyl species and 3-nitro derivatives, respectively. To prepare labeled proteins used in radioimmunoassays 125 I or 131 the above-mentioned chloramine T method of iodinating tyrosine residues using I is preferred.

[0202] Carboxyl side groups (aspartyl or glutamyl) are selectively modified by reacting with a carbodiimide (R'-N=C=N-R') (R and R' are optionally different alkyl groups, for example, 1-cyclohexyl-3-(2-morpholinyl-4-ethyl)carbodiimide or 1-ethyl-3-(4-azonia-4,4-dimethylpentyl)carbodiimide). Furthermore, aspartyl residues and glutamyl residues are converted to asparaginyl residues and glutaminyl residues by reacting with ammonium ions.

[0203] Derivatization with bifunctional substances is useful for crosslinking the antigen-binding molecules of the present invention to water-insoluble support matrices or support surfaces used in various methods. Commonly used crosslinking agents include, for example, the following: 1,1-bis(diazoacetyl)-2-phenylethane, glutaraldehyde, N-hydroxysuccinimide esters, such as esters with 4-azidosalicylic acid, homobifunctional imide esters containing disuccinimidyl esters, such as 3,3-dithiobis(succinimidyl propionate), and bifunctional maleimides, such as bis-N-maleimide-1,8-octane. Derivatizing agents such as methyl-3-[(p-azidophenyl)dithio]propionimidate can provide photoactivatable intermediates capable of forming crosslinks in the presence of light. Alternatively, reactive water-insoluble matrices such as cyanogen bromide-activated carbohydrates, as well as reactive substrates as described in U.S. Patent Nos. 3,969,287; 3,691,016; 4,195,128; 4,247,642; 4,229,537; and 4,330,440 are utilized for protein immobilization.

[0204] Glutaminyl and asparaginyl residues are frequently deamidated to the corresponding glutamyl and aspartyl residues, respectively. Alternatively, these residues are deamidated under weakly acidic conditions. Any form of these residues is included within the scope of the present invention.

[0205] Other modifications include hydroxylation of proline and lysine, phosphorylation of the hydroxyl groups of serine or threonyl residues, methylation of the α-amino groups of lysine, arginine, and histidine side chains (T.E. Creighton, Proteins: Structure and Molecular Properties, W.H. Freeman & Co., San Francisco, 1983, pp. 79-86), acetylation of the N-terminal amine, and amidation of any C-terminal carboxyl group.

[0206] Another type of covalent modification of antigen-binding molecules that is included within the scope of the present invention involves changing the glycosylation pattern of the protein. As is known in the art, the glycosylation pattern may depend on both the protein sequence (e.g., the presence or absence of specific glycosylated amino acid residues discussed below) or the host cell or organism producing the protein. Specific expression systems are discussed below.

[0207] Glycosylation of polypeptides is typically either N-linked or O-linked. N-linked refers to a sugar moiety being attached to the side chain of an asparagine residue. The tripeptide sequences asparagine-X-serine and asparagine-X-threonine (where X is any amino acid except proline) are recognition sequences for the enzymatic attachment of a sugar moiety to the asparagine side chain. Thus, the presence of any of these tripeptide sequences in a polypeptide creates potential glycosylation sites. O-linked glycosylation refers to one of the sugars N-acetylgalactosamine, galactose, or xylose being attached to a hydroxy amino acid, most commonly serine or threonine, although 5-hydroxyproline or 5-hydroxylysine may also be used.

[0208] The addition of glycosylation sites to antigen-binding molecules is conveniently achieved by modifying the amino acid sequence to contain one or more of the tripeptide sequences described above (in the case of N-linked glycosylation sites). This modification can be done by adding or substituting one or more serine or threonine residues to the initiation sequence (in the case of O-linked glycosylation sites). Briefly, it is preferred to change the amino acid sequence of the antigen-binding molecule by changing the DNA encoding the polypeptide at the DNA level, in particular by mutating the DNA with preselected bases so as to generate codons that translate into the desired amino acids.

[0209] Another means of increasing the number of sugar chains on an antigen-binding molecule is by chemically or enzymatically attaching glycosides to the protein. These procedures are advantageous in that they do not require production of the protein in a glycosylation-capable host cell for N-linked and O-linked glycosylation. Depending on the binding mode used, the sugar can be added to (a) arginine and histidine, (b) free carboxyl groups, (c) free sulfhydryl groups such as cysteine, (d) free hydroxyl groups such as serine, threonine or hydroxyproline, (e) aromatic residues such as phenylalanine, tyrosine or tryptophan, or (f) the amide group of glutamine. These methods are described in WO 87 / 05330 pamphlet and Aplin and Wriston, 1981, CRC Crit. Rev. Biochem., pp. 259-306.

[0210] Removal of the sugar chains present on the starting antigen-binding molecule can be carried out chemically or enzymatically. In chemical deglycosylation, it is necessary to expose the protein to the compound trifluoromethanesulfonic acid or an equivalent compound. By this treatment, most or all of the sugars except the bound sugar (N-acetylglucosamine or N-acetylgalactosamine) are cleaved while the polypeptide remains intact. Chemical deglycosylation is described in Hakimuddin et al., 1987, Arch. Biochem. Biophys. 259:52 and Edge et al., 1981, Anal. Biochem. 118:131. Enzymatic cleavage of sugar chains in polypeptides can be achieved by using various endoglycosidases and exoglycosidases as described in Thotakura et al., 1987, Meth. Enzymol. 138:350. Glycosylation at potential glycosylation sites can be avoided by using tunicamycin, a compound described in Duskin et al., 1982, J. Biol. Chem. 257:3105. Tunicamycin blocks the formation of protein-N-glycoside bonds.

[0211] Other modifications of the antigen-binding molecule are also contemplated herein. For example, another type of covalent modification of the antigen-binding molecule involves linking the antigen-binding molecule to various non-proteinaceous polymers, including polymers such as polyethylene glycol, polypropylene glycol, polyoxyalkylene, or a copolymer of polyethylene glycol and polypropylene glycol, as described in U.S. Patent Nos. 4,640,835; 4,496,689; 4,301,144; 4,670,417; 4,791,192, or 4,179,337, but not limited thereto. In addition, as is known in the art, amino acid substitutions can be made at various positions within the antigen-binding molecule to facilitate the addition of polymers such as PEG.

[0212] In some embodiments, the covalent modification of the antigen-binding molecule of the invention includes the addition of one or more labels. The label group may be attached to the antigen-binding molecule via spacer arms of various lengths so as to reduce potential steric hindrance. Various methods for labeling proteins are known in the art and can be used in practicing the present invention. The terms "label" or "label group" refer to any detectable label. Generally, labels are classified into various classes depending on the assay for detecting them, and examples include, but are not limited to, the following. a) Radioisotopes or radionuclides (e.g., 3 H, 14 C, 15 N, 35 S, 89 Zr, 90 Y, 99 Tc, 111 In, 125 I, 131 I) and other isotope labels that can be radioisotopes or heavy isotopes b) Magnetic labels (e.g., magnetic particles) c) Redox-active moieties d) Optical dyes (including, but not limited to, chromophores, phosphors, and fluorophores), such as fluorescent groups (e.g., FITC, rhodamine, lanthanide phosphors), chemiluminescent groups, and fluorophores that can be either "small molecule" fluorophores or proteinaceous fluorophores e) Enzyme groups (e.g., horseradish peroxidase, β - galactosidase, luciferase, alkaline phosphatase) f) Biotinylated groups g) A predetermined polypeptide epitope recognized by a secondary reporter (e.g., leucine zipper pair sequences, binding sites of secondary antibodies, metal - binding domains, epitope tags, etc.).

[0213] "Fluorescent label" means any molecule that can be detected by its intrinsic fluorescence properties. Suitable fluorescent labels include, but are not limited to, fluorescein, rhodamine, tetramethylrhodamine, eosin, erythrosin, coumarin, methylcoumarins, pyrene, malachite green, stilbene, lucifer yellow, Cascade BlueJ, Texas Red, IAEDANS, EDANS, BODIPY FL, LC Red 640, Cy 5, Cy 5.5, LC Red 705, Oregon Green, Alexa - Fluor dyes (Alexa Fluor 350, Alexa Fluor 430, Alexa Fluor 488, Alexa Fluor 546, Alexa Fluor 568, Alexa Fluor 594, Alexa Fluor 633, Alexa Fluor 660, Alexa Fluor 680), Cascade Blue, Cascade Yellow, and R - phycoerythrin (PE) (Molecular Probes, Eugene, OR), FITC, rhodamine, and Texas Red (Pierce, Rockford, IL), Cy5, Cy5.5, Cy7 (Amersham Life Science, Pittsburgh, PA). Suitable optical dyes containing fluorophores are described in the Molecular Probes Handbook by Richard P. Haugland.

[0214] Suitable proteinaceous fluorescent labels include, but are not limited to, GFP from the genus Renilla, Ptilosarcus, or Aequorea species (Chalfie et al., 1994, Science 263:802-805), green fluorescent protein including EGFP (Clontech Laboratories, Inc., Genbank accession number U55762), blue fluorescent protein (BFP, Quantum Biotechnologies, Inc. 1801 de Maisonneuve Blvd. West, 8th Floor, Montreal, Quebec, Canada H3H 1J9; Stauber, 1998, Biotechniques 24:462-471; Heim et al., 1996, Curr. Biol. 6:178-182), enhanced yellow fluorescent protein (EYFP, Clontech Laboratories, Inc.), luciferase (Ichiki et al., 1993, J. Immunol. 150:5408-5417), β-galactosidase (Nolan et al., 1988, Proc. Natl. Acad. Sci. U.S.A. 85:2603-2607), and Renilla (WO 92 / 15673 pamphlet, WO 95 / 07463 pamphlet, WO 98 / 14605 pamphlet, WO 98 / 26277 pamphlet, WO 99 / 49019 pamphlet, U.S. Patent No. 5,292,658; No. 5,418,155; No. 5,683,888; No. 5,741,668; No. 5,777,079; No. 5,804,387; No. 5,874,304; No. 5,876,995; No. 5,925,558).

[0215] The antigen-binding molecules of the present invention may also contain, for example, additional domains that are useful for isolating the molecule or that are related to the molecule's tailored pharmacokinetic profile. Domains useful for isolating the antigen-binding molecule can be selected from isolation methods, such as peptide motifs that can be captured on an isolation column or a moiety introduced adjunctively. Non-limiting embodiments of such additional domains include Myc tag, HAT tag, HA tag, TAP tag, GST tag, chitin-binding domain (CBD tag), maltose-binding protein (MBP tag), Flag tag, Strep tag and its variants (e.g., StrepII tag), and peptide motifs known as His tag. All of the antigen-binding molecules disclosed herein may contain a His tag domain, which is generally known as a repeat of consecutive His residues (preferably 5, more preferably 6 His residues (hexahistidine)) within the amino acid sequence of the molecule. The His tag may be located, for example, at the N-terminus or C-terminus of the antigen-binding molecule, preferably at the C-terminus. Most preferably, the hexahistidine tag (HHHHHH) (SEQ ID NO: 16) is linked to the C-terminus of the antigen-binding molecule according to the present invention via a peptide bond. In addition, for sustained release and improvement of the pharmacokinetic profile, a conjugate system of PLGA-PEG-PLGA may be combined with the polyhistidine tag.

[0216] Amino acid sequence modifications of the antigen-binding molecules described herein are also contemplated. For example, it may be desirable to improve the binding affinity and / or other biological properties of the antigen-binding molecule. Variants of the amino acid sequence of the antigen-binding molecule are prepared by introducing appropriate nucleotide changes into the nucleic acid of the antigen-binding molecule or by synthesizing the peptide. All of the following amino acid sequence modifications should result in antigen-binding molecules that still retain the desired biological activity (binding to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM and CD3) of the unmodified parent molecule.

[0217] The term "amino acid" or "amino acid residue" typically refers to an amino acid having a definition recognized in the art, such as an amino acid selected from the group consisting of alanine (Ala or A); arginine (Arg or R); asparagine (Asn or N); aspartic acid (Asp or D); cysteine (Cys or C); glutamine (Gln or Q); glutamic acid (Glu or E); glycine (Gly or G); histidine (His or H); isoleucine (Ile or I); leucine (Leu or L); lysine (Lys or K); methionine (Met or M); phenylalanine (Phe or F); proline (Pro or P); serine (Ser or S); threonine (Thr or T); tryptophan (Trp or W); tyrosine (Tyr or Y); and valine (Val or V), although modified amino acids, synthetic amino acids, or rare amino acids may be used if desired. Generally, amino acids can be classified by the presence of a nonpolar side chain (e.g., Ala, Cys, Ile, Leu, Met, Phe, Pro, Val); a negatively charged side chain (e.g., Asp, Glu); a positively charged side chain (e.g., Arg, His, Lys); or an uncharged polar side chain (e.g., Asn, Cys, Gln, Gly, His, Met, Phe, Ser, Thr, Trp, and Tyr).

[0218] Amino acid modifications include, for example, deletions from residues within the amino acid sequence of an antigen-binding molecule, and / or insertions into residues, and / or substitutions of residues. Any combination of deletions, insertions, and substitutions may be made to arrive at the final construct, provided that the final construct has the desired properties. Also, by changing the amino acids, it is possible to change post-translational processes of the antigen-binding molecule, such as changing the number or position of glycosylation sites.

[0219] For example, (naturally, depending on their lengths), 1, 2, 3, 4, 5, or 6 amino acids may be inserted, substituted, or deleted in each of the CDRs, while in each of the FRs, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or 25 amino acids may be inserted, substituted, or deleted. Preferably, the insertion of amino acid sequences into the antigen-binding molecule includes not only amino-terminal fusions and / or carboxyl-terminal fusions within the length range of polypeptides containing 100 or more residues from 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 residues, but also in-sequence insertions of single or multiple amino acid residues. The insertion variants of the antigen-binding molecules of the present invention include fusions of enzymes to the N-terminus or C-terminus of the antigen-binding molecule, or fusions to polypeptides.

[0220] The most important sites for substitution mutagenesis include the CDRs of the heavy and / or light chains, particularly the hypervariable regions (however, not limited thereto), and modifications of the FRs in the heavy and / or light chains are also contemplated. The substitutions are preferably conservative substitutions as described herein. Preferably, depending on the lengths of the CDR or FR, 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 amino acids may be substituted in the CDR, while in the framework region (FR), 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or 25 amino acids may be substituted. For example, if the CDR sequence contains 6 amino acids, it is envisioned that 1, 2, or 3 of these amino acids will be substituted. Similarly, if the CDR sequence contains 15 amino acids, it is envisioned that 1, 2, 3, 4, 5, or 6 of these amino acids will be substituted.

[0221] A method useful for identifying specific residues or regions of an antigen-binding molecule that are preferred positions for mutagenesis is what is called the "alanine scanning mutagenesis method" as described by Cunningham and Wells in Science, 244:1081-1085 (1989). In this method, residues or a group of target residues within the antigen-binding molecule are identified (e.g., charged residues such as arg, asp, his, lys, and glu), and these are replaced with neutral or negatively charged amino acids (most preferably, alanine or polyalanine) that affect the interaction between the amino acid and the epitope.

[0222] Subsequently, by introducing additional or other variants into the substitution site, i.e., instead of the substitution site, the range of those amino acid positions that show functional sensitivity to the substitution is stringently selected. Thus, the site or region into which the amino acid sequence variation is introduced is predetermined, but the nature of the variation itself does not need to be predetermined. For example, to analyze or optimize the performance of a mutation at a given site, alanine scanning or random mutagenesis can be performed at the target codon or target region to screen whether the expressed antigen-binding molecule variant has the optimal combination of desired activities. Techniques for performing substitution mutations at a predetermined site within DNA having a known sequence are well known, such as M13 primer mutagenesis and PCR mutagenesis. Screening of mutants is performed using assays of antigen-binding activity such as binding to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM or CD3.

[0223] Generally, when one or more or all of the CDRs of the heavy chain and / or light chain have amino acid substitutions, the resulting "substituted" sequence is preferably at least 60% or 65%, more preferably 70% or 75%, even more preferably 80% or 85%, particularly preferably 90% or 95% identical to the "original" CDR sequence. This means that the degree of identity with the "substituted" sequence depends on the length of the CDR. For example, for a CDR having 5 amino acids, it is preferable that the substituted sequence is 80% identical to it. Thus, the CDRs of the antigen-binding molecule may have various degrees of identity to their substituted sequences. For example, CDR-L1 may have 80% while CDR-L3 may have 90%.

[0224] Preferred substitutions (or replacements) are conservative substitutions. However, any substitution (including non-conservative substitutions or one or more of the "exemplary substitutions" listed in Table 3 below) is contemplated as long as the antigen-binding molecule retains the ability to bind to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM via a first domain and to bind to CD3 epsilon via a second domain, and / or as long as its CDRs have identity to the substituted sequence (at least 60% or 65%, more preferably 70% or 75%, even more preferably 80% or 85%, particularly preferably 90% or 95% identical to the "original" CDR sequence).

[0225] Conservative substitutions are shown under the heading "Preferred Substitutions" in Table 3. If such substitutions result in a change in biological activity, one may introduce substantial changes such as those referred to as "exemplary substitutions" in Table 3 or those further described below with reference to classes of amino acids, and screen the product for the desired characteristics.

[0226]

Table 3

[0227] Substantial modifications in the biological properties of the antigen-binding molecules of the present invention are achieved by selecting substitutions that differ significantly in their effect on (a) the structure of the polypeptide backbone of the substitution region as a sheet-like or helical three-dimensional structure, (b) the charge or hydrophobicity of the molecule at the target site, or (c) the maintenance of the bulkiness of the side chains. Naturally occurring residues are classified into the following groups based on common side-chain properties: (1) hydrophobic: norleucine, met, ala, val, leu, ile; (2) neutral hydrophobic: cys, ser, thr; asn, gln (3) acidic: asp, glu; (4) basic: his, lys, arg; (5) residues affecting the orientation of the chain: gly, pro; and (6) aromatic: trp, tyr, phe.

[0228] Non-conservative substitutions involve exchanging a member of one of these classes for another. To avoid abnormal cross-linking, any cysteine residues not involved in maintaining the proper three-dimensional structure of the antigen-binding molecule can generally be replaced with serine to improve the oxidative stability of the molecule. Conversely, the stability of an antibody can be improved by adding cysteine bonds to it (especially when the antibody is an antibody fragment such as an Fv fragment).

[0229] Regarding amino acid sequences, sequence identity and / or similarity are determined by standard techniques known in the art, such as, but not limited to, the local sequence identity algorithm of Smith and Waterman, 1981, Adv. Appl. Math. 2:482, the sequence identity alignment algorithm of Needleman and Wunsch, 1970, J. Mol. Biol. 48:443, the similarity search method of Pearson and Lipman, 1988, Proc. Nat. Acad. Sci. U.S.A. 85:2444, the computer implementations of these algorithms (GAP, BESTFIT, FASTA, and TFASTA in the Wisconsin Genetics Software Package, Genetics Computer Group, 575 Science Drive, Madison, Wis.), by using the Best Fit sequence program described in Devereux et al., 1984, Nucl. Acid Res. 12:387-395, preferably with default settings, or by visual inspection. Preferably, the percent identity is calculated by FastDB based on the following parameters: a mismatch penalty of 1; a gap penalty of 1; a gap size penalty of 0.33; and a joining penalty of 30, “Current Methods in Sequence Comparison and Analysis,” Macromolecule Sequencing and Synthesis, Selected Methods and Applications, pp 127-149 (1988), Alan R. Liss, Inc.

[0230] An example of a useful algorithm is PILEUP. PILEUP generates a multiple sequence alignment from a group of related sequences using progressive pairwise alignment. This also allows plotting a tree showing the clustering relationships used to generate the alignment. PILEUP uses a simplified version of the progressive alignment method of Feng & Doolittle, 1987, J. Mol. Evol. 35:351-360. This method is similar to that described by Higgins and Sharp, 1989, CABIOS 5:151-153. Useful PILEUP parameters include a default gap weight of 3.00, a default gap length weight of 0.10, and weighted end gaps.

[0231] Another example of a useful algorithm is the BLAST algorithm described in Altschul et al., 1990, J. Mol. Biol. 215:403-410; Altschul et al., 1997, Nucleic Acids Res. 25:3389-3402; and Karin et al., 1993, Proc. Natl. Acad. Sci. U.S.A. 90:5873-5787. A particularly useful BLAST program is the WU-BLAST-2 program obtained from Altschul et al., 1996, Methods in Enzymology 266:460-480. WU-BLAST-2 uses several search parameters, most of which are set to default values. Tunable parameters are set to the following values: overlap span = 1, overlap fraction = 0.125, word threshold (T) = II. The HSP S parameter and the HSP S2 parameter are dynamic values that are constructed by the program itself according to the composition of a particular sequence and the composition of the particular database against which the target sequence is being searched, but the values can be adjusted to increase sensitivity.

[0232] An additional useful algorithm is gapped BLAST as reported in Altschul et al., 1993, Nucl. Acids Res. 25:3389-3402. Gapped BLAST uses the BLOSUM-62 substitution score; the threshold T parameter is set to 9, the cost of gap length k is 10 + k in the two-hit method that results in no-gap extensions, Xu is set to 16, and Xg is set to 40 in the database search stage and 67 in the output stage of the algorithm. Gapped alignments are brought about by scores corresponding to approximately 22 bits.

[0233] Generally, the amino acid homology, similarity, or identity between individual variant CDRs or VH / VL sequences is at least 60% relative to the sequences shown herein, more typically the homology or identity is at least 65% or 70%, more preferably at least 75% or 80%, even more preferably at least 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, and nearly 100%. Similarly, "percent nucleic acid sequence identity (%)" with respect to the nucleic acid sequences of the binding proteins identified herein is defined as the percentage of nucleotide residues in a candidate sequence that are identical to the nucleotide residues in the coding sequence of the antigen-binding molecule. As a specific method, the BLASTN module of WU-BLAST-2 with default parameters set such that the overlap span and overlap fraction are 1 and 0.125, respectively, is utilized.

[0234] Generally, the nucleic acid sequence homology, similarity or identity between the nucleotide sequence encoding an individual variant CDR or VH / VL sequence and the nucleotide sequences shown herein is at least 60%, more typically the homology or identity is at least 65%, 70%, 75%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% and preferably increasing to almost 100%. Thus, a “variant CDR” or “variant VH / VL region” has a specific homology, similarity, or identity to the parental CDR / VH / VL of the invention and shares a biological function including, but not limited to, at least 60%, 65%, 70%, 75%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% of the specificity and / or activity of the parental CDR or VH / VL.

[0235] In one embodiment, the percentage identity of the antigen-binding molecule according to the present invention to the human germline is ≧70% or ≧75%, more preferably ≧80% or ≧85%, even more preferably ≧90%, and most preferably ≧91%, ≧92%, ≧93%, ≧94%, ≧95% or even ≧96%. Identity to the human antibody germline gene product is considered an important feature for reducing the risk that a therapeutic protein will induce an immune response to the drug in the patient being treated. Hwang & Foote (“Immunogenicity of engineered antibodies”; Methods 36 (2005) 3-10) demonstrated that reducing the non-human portion of a drug antigen-binding molecule results in a reduced risk of inducing anti-drug antibodies in the patient being treated. By comparing a vast number of clinically evaluated antibody drugs and their respective immunogenicity data, it has been shown that humanization of the V region of an antibody results in less immunogenicity of the protein (average 5.1% of patients) compared to antibodies bearing unmodified non-human V regions (average 23.59% of patients). Therefore, for V region-based protein therapeutics in the form of antigen-binding molecules, it is desirable to have a high degree of identity to the human sequence. To determine this germline identity, the V region of VL can be aligned with the amino acid sequences of human germline V and J segments (http: / / vbase.mrc-cpe.cam.ac.uk / ) using Vector NTI software, and the percentage of the amino acid sequence can be calculated by dividing the number of identical amino acid residues by the total number of amino acid residues in VL. The same method is possible for the VH segment (http: / / vbase.mrc-cpe.cam.ac.uk / ), with the exception that VH CDR3 is highly diverse and may be excluded because there is no existing alignment partner for human germline VH CDR3. Subsequently, recombinant techniques can be used to increase the sequence identity to the human antibody germline gene.

[0236] In a further embodiment, the bispecific antigen-binding molecule of the present invention exhibits a high monomer yield under standard research scale conditions (e.g., a standard two-step purification process). Preferably, the monomer yield of the antigen-binding molecule according to the present invention is ≧0.25 mg per liter of supernatant, more preferably ≧0.5 mg per liter, even more preferably ≧1 mg per liter, and most preferably ≧3 mg per liter of supernatant.

[0237] Similarly, the yield of the dimer antigen-binding molecule isoform, and thus the percentage of monomers of the antigen-binding molecule (i.e., monomer / (monomer + dimer)) can be determined. The productivity of the monomer and dimer antigen-binding molecules, as well as the calculated percentage of monomers, can be obtained, for example, in the SEC purification step of the culture supernatant obtained from production on a standardized research scale in roller bottles. In one embodiment, the percentage of monomers of the antigen-binding molecule is ≧80%, more preferably ≧85%, even more preferably ≧90%, and most preferably ≧95%.

[0238] In one embodiment, the antigen-binding molecule preferably has a plasma stability (ratio of EC50 in the presence of plasma to EC50 in the absence of plasma) of ≦5 or ≦4, more preferably ≦3.5 or ≦3, even more preferably ≦2.5 or ≦2, and most preferably ≦1.5 or ≦1. The plasma stability of the antigen-binding molecule is determined by incubating the construct in human plasma at 37° C. for 24 hours, followed by 51It can be verified by determining the EC50 in a chromium release cytotoxicity assay. The effector cells in the cytotoxicity assay can be stimulated enriched human CD8+ T cells. The target cells can be, for example, CHO cells transfected with human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM. The ratio of effector cells to target cells (E:T) can be selected as 10:1 or 5:1. The human plasma pool used for this purpose is derived from the blood of healthy donors collected with syringes coated with EDTA. The cellular components are removed by centrifugation, and the upper plasma phase is recovered and subsequently pooled. As a control, the antigen-binding molecule is diluted with RPMI-1640 medium immediately before the cytotoxicity assay. Plasma stability is calculated as the ratio of EC50 (after plasma incubation) to EC50 (control).

[0239] It is even more preferable that the conversion rate of the monomer of the antigen-binding molecule of the present invention to the dimer is low. The conversion rate can be measured under different conditions and analyzed by high performance size exclusion chromatography. For example, the incubation of the monomer isoform of the antigen-binding molecule can be carried out at 37°C for 7 days in an incubator at a concentration of, for example, 100 μg / ml or 250 μg / ml. Under these conditions, the antigen-binding molecule of the present invention preferably shows a dimer percentage of ≦5%, more preferably ≦4%, even more preferably ≦3%, even more preferably ≦2.5%, even more preferably ≦2%, even more preferably ≦1.5%, most preferably ≦1% or ≦0.5% or even 0%.

[0240] Furthermore, the bispecific antigen-binding molecule of the present invention preferably exhibits a very low dimer conversion rate after several freeze / thaw cycles. For example, the monomer of the antigen-binding molecule is adjusted to a concentration of 250 μg / ml in a common formulation buffer, subjected to 3 freeze / thaw cycles (freezing at -80°C for 30 minutes followed by thawing at room temperature for 30 minutes), and then high-speed SEC is performed to determine the percentage of the initial monomeric antigen-binding molecule that has been converted to the dimeric antigen-binding molecule. Preferably, the percentage of the dimer of the bispecific antigen-binding molecule is, for example, ≦5% after 3 freeze / thaw cycles, more preferably ≦4%, even more preferably ≦3%, even more preferably ≦2.5%, even more preferably ≦2%, even more preferably ≦1.5%, and most preferably ≦1% or even ≦0.5%.

[0241] The bispecific antigen-binding molecule of the present invention preferably exhibits good thermal stability with an aggregation temperature of ≧45°C or ≧50°C, more preferably ≧52°C or ≧54°C, even more preferably ≧56°C or ≧57°C, and most preferably ≧58°C or ≧59°C. From the perspective of the aggregation temperature of the antibody, the thermal stability parameter can be determined as follows: Transfer an antibody solution with a concentration of 250 μg / ml to a single-use cuvette and place it in a dynamic light scattering (DLS) apparatus. While constantly acquiring the measured radius, heat the sample from 40°C to 70°C at a heating rate of 0.5°C / min. An increase in the radius indicates protein melting and aggregation, and this is used to calculate the aggregation temperature of the antibody.

[0242] Alternatively, the melting temperature curve can be measured by differential scanning calorimetry (DSC) to determine the inherent biophysical protein stability of the antigen-binding molecule. These experiments are performed using a MicroCal LLC (Northampton, MA, U.S.A) VP-DSC instrument. The energy uptake of a sample containing the antigen-binding molecule is recorded from 20°C to 90°C and compared to a sample containing only the formulation buffer. The antigen-binding molecule is adjusted to a final concentration of 250 μg / ml, for example, in SEC running buffer. The temperature of the entire sample is increased stepwise to record each melting curve. At each temperature T, the energy uptake of the sample and the formulation buffer standard are recorded. The difference in energy uptake Cp (kcal / mole / °C) of the sample minus the standard is plotted for each temperature. The melting temperature is defined as the temperature at the first maximum in energy uptake.

[0243] The CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM×CD3 bispecific antigen-binding molecules of the present invention are also expected to have a turbidity (measured by OD340 after concentrating the purified monomeric antigen-binding molecule to 2.5 mg / ml and incubating overnight) of ≤0.2, preferably ≤0.15, more preferably ≤0.12, even more preferably ≤0.1, and most preferably ≤0.08.

[0244] In a further embodiment, the antigen-binding molecule according to the invention is stable at physiological pH or a pH slightly lower than that (i.e., about pH 7.4 to 6.0). The higher the resistance shown by the antigen-binding molecule at a non-physiological pH (e.g., about pH 6.0), the higher the recovery rate of the antigen-binding molecule eluted from the ion-exchange column with respect to the total amount of the loaded protein. The recovery rate of the antigen-binding molecule from an ion (e.g., cation) exchange column at about pH 6.0 is preferably ≧30%, more preferably ≧40%, more preferably ≧50%, even more preferably ≧60%, even more preferably ≧70%, even more preferably ≧80%, even more preferably ≧90%, even more preferably ≧95%, and most preferably ≧99%.

[0245] The bispecific antigen-binding molecule of the present invention is further contemplated to exhibit therapeutic efficacy or antitumor activity. This can be evaluated, for example, in tests disclosed in the following generalized examples of advanced human tumor xenograft models.

[0246] On day 1 of the test, 5×10 6 cells of a human target cell antigen (CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM as used herein) -positive cancer cell line are subcutaneously injected into the right dorsal flank of female NOD / SCID mice. When the average tumor volume reaches about 100 mm 3 about 2×10 7 human CD3-positive T cells proliferated in vitro are transplanted into the mice by injecting the cells into the peritoneal cavity of the animals. Mice in vehicle control group 1 are not given effector cells and are used as a non-transplanted control for comparison with (effector cell-given) vehicle control group 2 to monitor the effect of T cells alone on tumor growth. Treatment with the bispecific antigen-binding molecule is carried out when the average tumor volume reaches about 200 mm 3It starts when it reaches. The average tumor size of each treatment group on the treatment start date must not be statistically different from any other group (analysis of variance). Mice are treated with 0.5 mg / kg / day of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM and a CD3 bispecific antigen-binding molecule by intravenous bolus injection for about 15 - 20 days. During the test, tumors are measured with calipers, and progression is evaluated by comparing the tumor volume (TV) between groups. Tumor growth inhibition T / C [%] is determined by calculating TV as T / C% = 100×(median TV of the analysis group) / (median TV of control group 2).

[0247] One of ordinary skill in the art knows how to obtain meaningful and reproducible results while changing or adapting certain parameters of this test, such as the number of tumor cells to inject, the injection site, the number of human T cells to transplant, the amount of bispecific antigen-binding molecule to administer, and the schedule. Preferably, the tumor growth inhibition T / C [%] is ≤70 or ≤60, more preferably ≤50 or ≤40, even more preferably ≤30 or ≤20, and most preferably ≤10 or ≤5 or even ≤2.5. It is preferred that the tumor growth inhibition is close to 100%.

[0248] In a preferred embodiment of the antigen-binding molecule of the present invention, the antigen-binding polypeptide is a single-chain antigen-binding molecule.

[0249] Also, in a preferred embodiment of the antigen-binding molecule of the present invention, the spacer, in the order from amino to carboxyl, hinge-CH2-CH3-linker-hinge-CH2-CH3 is included.

[0250] In one embodiment of the present invention, each of the polypeptide monomers of the spacer has an amino acid sequence that is at least 90% identical to a sequence selected from the group consisting of SEQ ID NOs: 17 - 24. In a preferred embodiment or the present invention, each of the polypeptide monomers has an amino acid sequence selected from SEQ ID NOs: 17 - 24.

[0251] In one embodiment of the present invention, the CH2 domain of one or preferably each (both) of the polypeptide monomers of the spacer contains an intradomain cysteine disulfide bridge. As is known in the art, the term "cysteine disulfide bridge" refers to a functional group having the general structure R-S-S-R. This linkage is also called an SS bond or a disulfide bridge and is obtained by the bonding of two thiol groups of cysteine residues. With respect to the antigen-binding molecule of the present invention, it is particularly preferred that the cysteines that form cysteine disulfide bridges within the mature antigen-binding molecule are introduced into the amino acid sequence of the CH2 domain corresponding to positions 309 and 321 (Kabat numbering).

[0252] In one embodiment of the present invention, the glycosylation site at Kabat position 314 of the CH2 domain is removed. Removal of this glycosylation site is preferably achieved by N314X substitution (X is any amino acid other than Q). The substitution is preferably N314G. In a more preferred embodiment, the CH2 domain further comprises the following substitutions (positions according to Kabat), V321C and R309C (by these substitutions, an intradomain cysteine disulfide bridge is introduced at Kabat positions 309 and 321).

[0253] For example, the preferred features of the antigen-binding molecule of the present invention compared to bispecific heterologous Fc antigen-binding molecules known in the art may be related, inter alia, to the introduction of the modifications described above in the CH2 domain. Thus, with respect to the construct of the present invention, it is preferred that the CH2 domain in the spacer of the antigen-binding molecule of the present invention contains an intradomain cysteine disulfide bridge at Kabat positions 309 and 321 and / or the glycosylation site at Kabat position 314 is removed, preferably by N314G substitution.

[0254] In a further preferred embodiment of the present invention, the CH2 domain in the spacer of the antigen-binding molecule of the present invention contains an intradomain cysteine disulfide bridge at Kabat positions 309 and 321, and the glycosylation site at Kabat position 314 is removed by N314G substitution. Most preferably, the polypeptide monomer of the spacer of the antigen-binding molecule of the present invention has an amino acid sequence selected from the group consisting of SEQ ID NOs: 17 and 18.

[0255] In one embodiment, the present invention provides an antigen-binding molecule, (i) wherein the first domain comprises two antibody variable domains and the second domain comprises two antibody variable domains; (ii) wherein the first domain comprises one antibody variable domain and the second domain comprises two antibody variable domains; (iii) wherein the first domain comprises two antibody variable domains and the second domain comprises one antibody variable domain; or (iv) wherein the first domain comprises one antibody variable domain and the second domain comprises one antibody variable domain. The present invention provides an antigen-binding molecule.

[0256] Thus, the first domain and the second domain can be binding domains each comprising two antibody variable domains such as VH domains and VL domains. Examples of such binding domains comprising two antibody variable domains described above herein include, for example, Fv fragments, scFv fragments, or Fab fragments described above herein. Alternatively, either or both of those binding domains can comprise only a single variable domain. Examples of such single-domain-binding domains described above herein include, for example, nanobodies or single variable domain antibodies comprising only one variable domain that can specifically bind an antigen or epitope independently of other V regions or domains, such as VHH, VH, or VL.

[0257] In a preferred embodiment of the antigen-binding molecule of the present invention, the second binding domain and the third binding domain are fused to the spacer via a peptide linker. Preferred peptide linkers are described above herein and are characterized by the amino acid sequence Gly-Gly-Gly-Gly-Ser, i.e., Gly4Ser (SEQ ID NO: 7) or a polymer thereof, i.e., (Gly4Ser)x where x is an integer of 1 or more (e.g., 2 or 3). A particularly preferred linker for fusing the first domain and the second domain to the spacer is shown in SEQ ID NO: 7.

[0258] The antigen-binding molecule of the present invention comprises a first domain, which binds to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM, preferably to the extracellular domain (ECD) of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM. In the context of the present invention, the term "binds to the extracellular domain of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM" is understood to imply that the binding domain binds to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM expressed on the surface of target cells. Thus, the first domain according to the present invention preferably binds to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM when CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM is expressed by a cell or cell line expressing it naturally and / or when CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM is expressed by a cell or cell line transformed or (stably / transiently) transfected with CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM. In a preferred embodiment, the first binding domain also binds to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM when CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM is used as a "target" or "ligand" molecule in an in vitro binding assay such as BIAcore or Scatchard.The "target cell" may be any prokaryotic or eukaryotic cell that expresses CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM on its surface. Preferably, the target cell is a cell that is part of the body of a human or animal. For example, it is a cancer cell or tumor cell that specifically expresses CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM.

[0259] Preferably, the first binding domain binds to human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM / CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM ECD. In a more preferred embodiment, the first binding domain binds to macaque CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM / CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM ECD. According to the most preferred embodiment, the first binding domain binds to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM / CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM ECD of both human and macaque. The "extracellular domain of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM", namely the "ECD of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM", refers to the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM region or sequence that essentially does not contain the transmembrane domain and cytoplasmic domain of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM. It will be understood by those skilled in the art that the transmembrane domains identified for the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM polypeptides of the present invention are identified according to the criteria routinely used in the art to identify this type of hydrophobic domain. Although the exact boundaries of the transmembrane domain may vary, it is very likely to be about 5 or fewer amino acids at either end of any of the domains specifically mentioned herein.

[0260] Preferred binding domains that bind to CD3 are disclosed in WO 2010 / 037836 and WO 2011 / 121110. Any binding domain of CD3 described in these applications may be used in the context of the present invention.

[0261] The present invention further provides a polynucleotide / nucleic acid molecule encoding the antigen-binding molecule of the present invention. A polynucleotide is a biopolymer composed of 13 or more nucleotide monomers covalently linked within a chain. DNA (e.g., cDNA) and RNA (e.g., mRNA) are examples of polynucleotides having different biological functions. A nucleotide is an organic molecule that functions as a monomer or subunit of a nucleic acid molecule such as DNA or RNA. The nucleic acid molecule or polynucleotide may be double-stranded or single-stranded, and may be linear or circular. The nucleic acid molecule or polynucleotide is preferably contained within a vector, and the vector is preferably contained within a host cell. The host cell is capable of expressing the antigen-binding molecule, for example, after being transformed or transfected with the vector or polynucleotide of the present invention. Therefore, the polynucleotide or nucleic acid molecule is operably linked to a control sequence.

[0262] The genetic code is a set of rules for translating the information encoded within genetic material (nucleic acids) into proteins. Biological decoding in living cells is accomplished by ribosomes that link amino acids in the order specified by mRNA, using tRNA molecules that carry amino acids and read three nucleotides in the mRNA at a time. This code defines how a three-nucleotide sequence of nucleotides called a codon specifies the amino acid to be added next during protein synthesis. With a few exceptions, a codon of three nucleotides in a nucleic acid sequence specifies one amino acid. Since most genes are encoded by exactly the same code, this particular code is often referred to as the standard genetic code or the canonical genetic code. While the genetic code determines the protein sequence of a given coding region, other genomic regions can affect when and where these proteins are produced.

[0263] Furthermore, the present invention provides a vector comprising the polynucleotide / nucleic acid molecule of the present invention. A vector is a nucleic acid molecule used as a vehicle for introducing (foreign) genetic material into a cell. The term "vector" includes, but is not limited to, plasmids, viruses, cosmids, and artificial chromosomes. Generally, a modified vector contains an origin of replication, a multiple cloning site, and a selectable marker. A vector itself is generally a nucleotide sequence (generally a DNA sequence) composed of an insert (transgene) and a larger sequence that functions as the "backbone" of the vector. Up-to-date vectors can include additional features in addition to the transgene insert and backbone: promoters, gene markers, antibiotic resistance, reporter genes, targeting sequences, protein purification tags. A vector called an expression vector (expression construct) is specifically for expressing a transgene in a target cell and generally has regulatory sequences.

[0264] The term "control array" refers to the DNA array necessary for the expression of an operably linked code array in a particular host organism. Suitable control arrays for prokaryotes include, for example, a promoter, optionally an operator array, and a ribosome binding site. Eukaryotic cells are known to utilize a promoter, a polyadenylation signal, and an enhancer.

[0265] Nucleic acids are "operably linked" when they are in a functional relationship with another nucleic acid array. For example, the DNA of a pre-array or a secretion leader is operably linked to the DNA of a polypeptide when expressed as a protein precursor involved in the secretion of the polypeptide; a promoter or enhancer is operably linked to a coding array when it affects the transcription of the array; or a ribosome binding site is operably linked to a coding array when arranged to facilitate translation. Generally, "operably linked" means that the linked DNA arrays are continuous, and in the case of a secretion leader, it means continuous and within the reading frame. However, an enhancer does not need to be continuous. This linkage is achieved by ligation at convenient restriction sites. If such sites do not exist, synthetic oligonucleotide adapters or linkers are used according to conventional practices.

[0266] "Transfection" is the process of intentionally introducing a nucleic acid molecule or polynucleotide (including a vector) into a target cell. This term is mainly used for non-viral methods in eukaryotic cells. Transduction is often used to describe the virus-mediated introduction of a nucleic acid molecule or polynucleotide. Transfection of animal cells typically involves creating transient pores or "holes" in the cell membrane to allow the uptake of materials. Transfection can be carried out using calcium phosphate, by electroporation, by cell squeezing, or by mixing a cationic lipid with the material to form liposomes, which are then fused with the cell membrane to deposit the internal cargo.

[0267] The term "transformation" is used to describe the non-viral introduction of nucleic acid molecules or polynucleotides (including vectors) into bacteria and non-animal eukaryotic cells, including plant cells. Thus, transformation is a genetic modification of bacteria or non-animal eukaryotic cells that results from direct uptake from its surroundings through the cell membrane and subsequent integration of foreign genetic material (nucleic acid molecules). Transformation can be caused by artificial means. To cause transformation, the cells or bacteria must be in a competent state in which transformation can occur as a time-limited response to environmental conditions such as starvation and cell density.

[0268] Furthermore, the present invention provides host cells transformed or transfected with the polynucleotide / nucleic acid molecule or vector of the present invention. As used herein, the term "host cell" or "recipient cell" is intended to include any individual cell or cell culture that can or was a recipient of a vector encoding an antigen-binding molecule of the present invention, a foreign nucleic acid molecule, and a polynucleotide; and / or a recipient of the antigen-binding molecule itself. The introduction of each substance into the cell is effected by transformation, transfection, etc. The term "host cell" is also intended to include the progeny or potential progeny of a single cell. In subsequent generations, such progeny may not actually be (morphologically, or with respect to the genome or total DNA complement) identical to the parent cell because certain modifications may occur due to spontaneous, accidental, or intentional mutations, or due to environmental influences, but this is still included within the scope of this term as used herein. Suitable host cells include prokaryotic or eukaryotic cells, and also include, but are not limited to, bacteria, yeast cells, fungal cells, plant cells and animal cells, such as insect cells and mammalian cells, such as mouse, rat, macaque or human.

[0269] The antigen-binding molecules of the present invention can be produced within bacteria. After expression, the antigen-binding molecules of the present invention can be isolated from E. coli cell paste in the soluble fraction and purified, for example, by affinity chromatography and / or size exclusion chromatography. Final purification can be carried out in the same manner as the process for purifying an antibody expressed in CHO cells, for example.

[0270] In addition to prokaryotes, eukaryotic microorganisms such as filamentous fungi or yeasts are suitable cloning hosts or expression hosts for the antigen-binding molecules of the present invention. Saccharomyces cerevisiae or common baker's yeast is the most commonly used among lower eukaryotic host microorganisms. However, several other genera, species, and strains are generally available and useful herein, for example, Schizosaccharomyces pombe, Kluyveromyces hosts such as K. lactis, K. fragilis (ATCC 12424), K. bulgaricus (ATCC 16045), K. wickeramii (ATCC 24178), K. waltii (ATCC 56500), K. drosophilarum (ATCC 36906), K. thermotolerans, and K. marxianus; Yarrowia (European Patent No. 402226); Pichia pastoris (European Patent No. 183070); Candida; Trichoderma reesia (European Patent No. 244234); Neurospora crassa; Schwanniomyces such as Schwanniomyces occidentalis; and filamentous fungi such as Neurospora, Penicillium, Tolypocladium, and Aspergillus hosts such as A. nidulans and A. niger.

[0271] Suitable host cells for expressing the glycosylated antigen-binding molecules of the present invention are those derived from multicellular organisms. Examples of invertebrate cells include plant cells and insect cells. Many baculovirus strains and variants and the corresponding permissive insect host cells derived from the host (e.g., Spodoptera frugiperda (armyworm), Aedes aegypti (mosquito), Aedes albopictus (mosquito), Drosophila melanogaster (fruit fly), and Bombyx mori (silkworm)) have been identified. Various virus strains for transfection, such as the L-1 variant of Autographa californica NPV and the Bm-5 strain of Bombyx mori NPV, are publicly available, and such viruses may be used for transfection of Spodoptera frugiperda cells, in particular, as the viruses of the present specification according to the present invention.

[0272] Plant cell cultures of cotton, corn, potato, soybean, petunia, tomato, Arabidopsis thaliana, and tobacco can also be used as hosts. Cloning and expression vectors useful for protein production in plant cell culture are known to those skilled in the art. See, for example, Hiatt et al., Nature (1989) 342:76-78, Owen et al. (1992) Bio / Technology 10:790-794, Artsaenko et al. (1995) The Plant J 8:745-750, and Fecker et al. (1996) Plant Mol Biol 32:979-986.

[0273] However, there is the greatest interest in vertebrate cells, and it has become a common procedure to grow vertebrate cells in a culture (tissue culture) medium. Examples of useful mammalian host cell lines are monkey kidney CV1 line transformed by SV40 (COS-7, ATCC CRL 1651); human embryonic kidney line (293 cells or 293 cells subcloned for growth in suspension culture, Graham et al., J. Gen Virol. 36:59 (1977)); baby hamster kidney cells (BHK, ATCC CCL 10); Chinese hamster ovary cells / -DHFR (CHO, Urlaub et al., Proc. Natl. Acad. Sci. USA 77:4216 (1980)); mouse Sertoli cells (TM4, Mather, Biol. Reprod. 23:243-251 (1980)); monkey kidney cells (CVI ATCC CCL 70); African green monkey kidney cells (VERO-76, ATCC CRL 1587); human cervical carcinoma cells (HELA, ATCC CCL 2); dog kidney cells (MDCK, ATCC CCL 34); buffalo rat liver cells (BRL 3A, ATCC CRL 1442); human lung cells (W138, ATCC CCL 75); human liver cells (Hep G2, 1413 8065); mouse mammary tumor (MMT 060562, ATCC CCL5 1); TRI cells (Mather et al., Annals N.Y Acad. Sci. (1982) 383:44-68); MRC 5 cells; FS4 cells; and human hepatoma line (Hep G2).

[0274] In a further embodiment, the invention provides a process for producing an antigen-binding molecule of the invention, the process comprising culturing a host cell of the invention under conditions that permit expression of the antigen-binding molecule of the invention and recovering the produced antigen-binding molecule from the culture.

[0275] As used herein, the term "culturing" refers to the in vitro maintenance, differentiation, growth, proliferation, and / or propagation of cells under suitable conditions in a medium. The term "expression" includes any step involved in the production of an antigen-binding molecule of the present invention, such as, but not limited to, transcription, post-transcriptional modification, translation, post-translational modification, and secretion.

[0276] When recombinant techniques are used, the antigen-binding molecule can be produced in the periplasmic space within the cell or secreted directly into the medium. When the antigen-binding molecule is produced intracellularly, as a first step, host cells or lysed fragments that are particulate debris are removed, for example, by centrifugation or ultrafiltration. Carter et al., Bio / Technology 10:163-167 (1992) describes a procedure for isolating antibodies secreted into the periplasmic space of Escherichia coli (E. coli). Briefly, the cell paste is thawed in the presence of sodium acetate (pH 3.5), EDTA, and phenylmethylsulfonyl fluoride (PMSF) for about 30 minutes. The cell debris can be removed by centrifugation. When the antibody is secreted into the medium, the supernatant from such an expression system is generally first concentrated using a commercially available protein concentration filter (e.g., an Amicon or Millipore Pellicon ultrafiltration unit). To inhibit proteolysis, a protease inhibitor such as PMSF may be included in any of the foregoing steps, and an antibiotic may be included to prevent the growth of exogenous contaminating bacteria.

[0277] The antigen-binding molecules of the present invention prepared from host cells can be recovered or purified, for example, using hydroxyapatite chromatography, gel electrophoresis, dialysis, and affinity chromatography. Depending on the antibody to be recovered, other protein purification techniques can also be used, such as fractionation on an ion exchange column, ethanol precipitation, reverse phase HPLC, chromatography on silica, chromatography on heparin SEPHAROSE (trademark), chromatography on an anion or cation exchange resin (e.g., polyaspartic acid column), chromatofocusing, SDS-PAGE, and ammonium sulfate precipitation. When the antigen-binding molecule of the present invention contains a CH3 domain, Bakerbond ABX resin (J.T. Baker, Phillipsburg, NJ) is useful for purification.

[0278] A preferred purification technique is affinity chromatography. Although the matrix to which the affinity ligand binds is most often agarose, other matrices are also available. Mechanically stable matrices, such as controlled pore glass or poly(styrene divinyl) benzene, allow for faster flow rates and shorter processing times than those achievable using agarose.

[0279] Furthermore, the present invention provides a pharmaceutical composition comprising the antigen-binding molecule of the present invention or an antigen-binding molecule produced according to the method of the present invention. In the pharmaceutical composition of the present invention, the homogeneity of the antigen-binding molecule is preferably ≧80%, more preferably ≧81%, ≧82%, ≧83%, ≧84%, or ≧85%, even more preferably ≧86%, ≧87%, ≧88%, ≧89%, or ≧90%, still even more preferably ≧91%, ≧92%, ≧93%, ≧94%, or ≧95%, and most preferably ≧96%, ≧97%, ≧98%, or ≧99%.

[0280] As used herein, the term "pharmaceutical composition" relates to a composition suitable for administration to a patient (preferably a human patient). Particularly preferred pharmaceutical compositions of the invention comprise one or more antigen-binding molecules of the invention, preferably in a therapeutically effective amount. Preferably, the pharmaceutical composition further comprises a suitable formulation of one or more (pharmaceutically effective) carriers, stabilizers, excipients, diluents, solubilizers, surfactants, emulsifiers, preservatives and / or adjuvants. The components of the acceptable composition are preferably non-toxic to the recipient at the dosages and concentrations employed. Pharmaceutical compositions of the invention include, but are not limited to, liquid compositions, frozen compositions, and lyophilized compositions.

[0281] The compositions of the invention may include a pharmaceutically acceptable carrier. Generally, as used herein, "pharmaceutically acceptable carrier" means any aqueous and non-aqueous solution, sterile solution, solvent, buffer, such as phosphate buffered saline (PBS) solution, water, suspension, emulsion such as oil / water emulsion, various kinds of wetting agents, liposomes, dispersion media and coatings, which are compatible with pharmaceutical administration, particularly parenteral administration. The use of such media and agents in pharmaceutical compositions is well known in the art, and compositions containing such carriers can be formulated by well-known conventional methods.

[0282] Certain embodiments provide a pharmaceutical composition comprising an antigen-binding molecule of the invention and further one or more excipients as exemplified elsewhere in this section and in this specification. Excipients can be used for a wide variety of purposes such as adjusting the physical, chemical or biological properties of the formulation such as adjusting viscosity, and / or for improving efficacy and / or stabilizing such formulations, and for processes of the invention, e.g., for degradation and damage due to stresses occurring during manufacture, transport, storage, preparation before use, administration and after these, and can be used in the invention in view of such processes.

[0283] In certain embodiments, the pharmaceutical composition may contain formulation materials for the purpose of altering, sustaining, or protecting, for example, the pH, osmolality, viscosity, clarity, color, isotonicity, odor, sterility, stability, dissolution or release rate, adsorption, or penetration of the composition (see REMINGTON’S PHARMACEUTICAL SCIENCES, 18” Edition, (A.R. Genrmo, ed.), 1990, Mack Publishing Company). In such embodiments, suitable formulation materials may include, but are not limited to: · Amino acids, such as glycine, alanine, glutamine, asparagine, threonine, proline, 2-phenylalanine, such as charged amino acids, preferably lysine, lysine acetate, arginine, glutamate, and / or histidine · Antimicrobial agents, such as antibacterial drugs and antifungal agents · Antioxidants, such as ascorbic acid, methionine, sodium sulfite, or sodium bisulfite; · Buffers, buffer systems, and buffering agents used to maintain the composition at a physiological pH or a slightly lower pH, preferably a lower pH of 4.0 - 6.5; examples of buffers are borates, bicarbonates, Tris-HCl, citrates, phosphates, or other organic acids, succinates, phosphates, and histidine; for example, Tris buffer at about pH 7.0 - 8.5; · Nonaqueous solvents, such as propylene glycol, polyethylene glycol, vegetable oils such as olive oil, and injectable organic esters such as ethyl oleate; · Aqueous carriers such as water, alcoholic / aqueous solutions, emulsions, or suspensions, such as physiological saline and buffered media; · Biodegradable polymers such as polyesters; · Fillers such as mannitol or glycine; · Chelating agents such as ethylenediaminetetraacetic acid (EDTA); · Isotonic agents and absorption delaying agents; · Complexing agents, such as caffeine, polyvinylpyrrolidone, β-cyclodextrin, or hydroxypropyl-β-cyclodextrin · Bulking agent; · Monosaccharides; disaccharides; and other carbohydrates (e.g., glucose, mannose, or dextrin); the carbohydrate may be a non-reducing sugar and preferably may be trehalose, sucrose, octasulfate, sorbitol, or xylitol; · (Low molecular weight) proteins, polypeptides, or proteinaceous carriers such as human or bovine serum albumin, gelatin, or preferably immunoglobulins of human origin; · Colorants and flavorants; · Sulfur-containing reducing agents such as glutathione, thioctic acid, sodium thioglycolate, thioglycerol, [alpha]-monothioglycerol, and sodium thiosulfate · Diluents; · Emulsifiers; · Hydrophilic polymers such as polyvinylpyrrolidone; · Salt-forming counterions such as sodium; · Preservatives such as antibacterial agents, antioxidants, chelating agents, inert gases, etc.; examples include benzalkonium chloride, benzoic acid, salicylic acid, thimerosal, phenethyl alcohol, methylparaben, propylparaben, chlorhexidine, sorbic acid, or hydrogen peroxide; · Metal complexes such as Zn-protein complexes; · Solvents and co-solvents (e.g., glycerin, propylene glycol, or polyethylene glycol); · Sugars and sugar alcohols such as trehalose, sucrose, octasulfate, mannitol, sorbitol or xylitol, stachyose, mannose, sorbose, xylose, ribose, myo-inositol, galactose, lactitol, ribitol, myo-inositol, galactitol, glycerol, cyclitols (e.g., inositol), polyethylene glycol; and polyhydric sugar alcohols; · Suspending agents; · Surfactants or wetting agents, such as Pluronic, PEG, sorbitan esters, polysorbates, such as polysorbate 20, polysorbate, Triton, tromethamine, lecithin, cholesterol, tyloxapol; the surfactant can preferably be a detergent having a molecular weight > 1.2 KD and / or a polyether preferably having a molecular weight > 3 KD; non-limiting examples of preferred detergents are Tween 20, Tween 40, Tween 60, Tween 80, and Tween 85; non-limiting examples of preferred polyethers are PEG 3000, PEG 3350, PEG 4000, and PEG 5000; · Stability enhancers such as sucrose or sorbitol; · Isotonicity enhancers, such as alkali metal halides, preferably sodium chloride or potassium chloride, mannitol, sorbitol; · Parenteral delivery vehicles, such as sodium chloride solution, Ringer's dextrose, dextrose, and sodium chloride, lactated Ringer's solution, or fixed oil; · Intravenous delivery vehicles, such as body fluids, and nutritional supplements, electrolyte supplements (e.g., those based on Ringer's dextrose).

[0284] In the context of the present invention, a pharmaceutical composition, which can preferably be a liquid composition, or a solid composition obtained by lyophilization, or a reconstituted liquid composition, is (a) an antigen-binding molecule comprising at least four binding domains, wherein · the first domain and the third domain bind to a target cell surface antigen and have an isoelectric point (pI) in the range of 4 to 9.5; · the second domain and the fourth domain bind to CD3; and have a pI in the range of 8 to 10, preferably 8.5 to 9.0; · the spacer preferably comprises two polypeptide monomers, each comprising a hinge, CH2 domain, and CH3 domain, and the two polypeptide monomers are fused to each other via a peptide linker, the antigen-binding molecule; (b) at least one buffer; (c) at least one sugar; and (d) at least one surfactant comprising the pH of the pharmaceutical composition is in the range of 3.5 to 6.

[0285] Furthermore, in connection with the present invention, it is contemplated that at least one buffering agent is present in a concentration range of 5 to 200 mM, more preferably in a concentration range of 10 to 50 mM. It is contemplated in connection with the present invention that at least one saccharide is selected from the group consisting of monosaccharides, disaccharides, cyclic polysaccharides, sugar alcohols, linear or branched dextrans or linear or unbranched dextrans. In connection with the present invention, it is also contemplated that disaccharides are selected from the group consisting of sucrose, trehalose, and mannitol, sorbitol, and combinations thereof. Furthermore, it is contemplated in connection with the present invention that the sugar alcohol is sorbitol. It is contemplated in connection with the present invention that at least one saccharide is present at a concentration in the range of 1 to 15% (m / V), preferably in the concentration range of 9 to 12% (m / V).

[0286] At least one surfactant is also contemplated in the context of the present invention to be selected from the group consisting of polysorbate 20, polysorbate 40, polysorbate 60, polysorbate 80, poloxamer 188, pluronic F68, Triton X-100, polyoxyethylene, PEG 3350, PEG 4000, and combinations thereof. Further, it is contemplated in the context of the present invention that at least one surfactant is present at a concentration in the range of 0.004 - 0.5% (m / V), preferably in the range of 0.001 - 0.01% (m / V). The pH of the composition is contemplated in the context of the present invention to be in the range of 4.0 - 5.0, preferably 4.2. The pharmaceutical composition is also contemplated in the context of the present invention to have a molar osmotic concentration in the range of 150 - 500 mOsm. The pharmaceutical composition is further contemplated in the context of the present invention to further comprise an excipient selected from the group consisting of one or more polyols and one or more amino acids. The one or more excipients are contemplated in the context of the present invention to be present in a concentration range of 0.1 - 15% (w / V).

[0287] The pharmaceutical composition comprises (a) an antigen-binding molecule as discussed above, (b) 10 mM glutamate or acetate, (c) 9% (m / V) sucrose or 6% (m / V) sucrose, and 6% (m / V) hydroxypropyl-β-cyclodextrin, (d) 0.01% (m / V) polysorbate 80 and it is contemplated in the context of the present invention that the pH of the liquid pharmaceutical composition is 4.2 This is also contemplated in the context of the present invention.

[0288] Furthermore, it is contemplated in the context of the present ...

Claims

1. A molecule comprising at least two polypeptide chains, wherein the molecule is (i.) A first binding domain that binds to the first target cell surface antigen (TAA1), (ii.) A second binding domain that binds to the extracellular epitope of the human and / or maca CD3 chain, (iii.) A third binding domain that binds to the second target cell surface antigen (TAA2), and (iv.) A fourth binding domain that binds to the extracellular epitope of the human and / or maca CD3 chain. Includes, The first binding domain and the second binding domain form a first dual specificity entity, the third binding domain and the fourth binding domain form a second dual specificity entity, and the molecule further comprises a spacer entity. The original spacer is, An Fc domain comprising a first polypeptide monomer and a second polypeptide monomer, each comprising a hinge, a CH2 domain, and a CH3 domain, respectively, wherein the first polypeptide monomer and the second polypeptide monomer form a heterodimer; The aforementioned heterodimer, (a.) D399K, K409D, K392D, and E356K, (ii.) D399K, K409D, K392D, E357K, K370D, and E356K, (iii.) D399K, K409D, K392D, E356K, and K439D, (iv.) D399K, K409D, and K392D, (v.) D399K, K409D, K392D, E357K, and E370K, (vi.) D399K, K409D, K392D, E357K, K370E, and K360E, (vii.) D399K, K409D, K3 A charge pair mutation selected from 92D, E357K, K370E, E356K, and K439E, and (viii.) D399K, K409D, K392D, E357K, K370E, K360E, E356K, and K439D, preferably comprising the K392D, K409D and / or K439D mutations in the CH3 domain of the first polypeptide monomer and the E356K and / or D399K mutation in the CH3 domain of the second polypeptide monomer, with the position determined by EU numbering; or (b) Preferably, the first polypeptide monomer contains the T366S, L368A, and Y407V mutations, and the second monomer contains the T366W mutation, with the position determined by EU numbering, which is a knob-into-hole mutation. Formed by, The dimerization domain comprises two N-terminants and two C-terminants, each of which at least one N-terminant and one C-terminant is linked to a bispecific entity, and any of the first, second, third, and fourth domains can be selected from any form of binding domain, preferably selected from Fab and a single-stranded domain, and the single-stranded domain is preferably selected from single-stranded Fv (scFv) and scFab. The distance between the Cα atom of the first amino acid located at the N-terminus of the spacer entity and the Cα atom of the last amino acid located at the C-terminus is at least 30 Å, and the spacer entity separates the first bispecific entity and the second bispecific entity by a distance of at least about 50 Å, the indicated distance being preferably understood as (i) the distance between the first binding domain and the third binding domain, or (ii) the distance between the mass centers of the first bispecific entity and the second bispecific entity, and the spacer entity is positioned between the first bispecific entity and the second bispecific entity. The first target cell surface antigen and the second target cell surface antigen are not identical, preferably the first binding domain can bind to the first target cell surface antigen and the third binding domain can simultaneously bind to the second target cell surface antigen, and preferably the first target cell surface antigen and the second target cell surface antigen are located on the same target cell. molecule.

2. If the spacer is a single-chain domain, the arrangement of the binding domains in the order of amino to carboxyl is, (i.) The first domain and the second domain, the spacer, the third domain and the fourth domain (ii.) The first domain and the second domain, the spacer, the fourth domain and the third domain (iii.) The second domain and the first domain, spacer, the third domain and the fourth domain, and (iv.) The second domain and the first domain, spacer, the fourth domain and the third domain Selected from the group consisting of, Preferably, when the spacer is a single-chain domain, the arrangement of the binding domains in the order of amino to carboxyl is (i.) A first domain in the form of Fab, a second domain in the form of a single-stranded domain, preferably in the form of an scFv, a spacer, a third domain in the form of a Fab, and a fourth domain in the form of a single-stranded domain, preferably in the form of an scFv; (ii.) A first domain in the form of a Fab, a second domain in the form of a Fab, a spacer, a third domain in the form of a Fab, and a fourth domain in the form of a Fab; (iii.) A single-stranded domain, preferably a first domain in the form of scFv, a second domain in the form of Fab, a spacer, a third domain in the form of scFv, and a fourth domain in the form of Fab; (iv.) A single-stranded domain, preferably a first domain in the form of scFv, a second domain in the form of scFv, a spacer, a third domain in the form of scFv, and a fourth domain in the form of Fab; (v.) A first domain in the form of a single-stranded domain, preferably in the form of scFv; a second domain in the form of a single-stranded domain, preferably in the form of scFv; a third domain in the form of a spacer, Fab; and a fourth domain in the form of a single-stranded domain, preferably in the form of scFv; (vi.) A first domain in the form of Fab, a second domain in the form of a single-stranded domain, preferably in the form of scFv, a spacer, a third domain in the form of a single-stranded domain, preferably in the form of scFv, and a fourth domain in the form of a single-stranded domain, preferably in the form of scFv; and (vii.) A first domain in the form of a single-stranded domain, preferably in the form of an scFv; a second domain in the form of a Fab; a spacer; a third domain in the form of a single-stranded domain, preferably in the form of an scFv; and a fourth domain in the form of a single-stranded domain, preferably in the form of an scFv. Selected from the group consisting of, Each scFv comprises VH, linker and VL, or VL, linker and VH, preferably VH, linker and VL, in the order of amino to carboxyl. The antigen-binding molecule according to claim 1.

3. If the spacer is a dimerizing domain, the arrangement of the binding domains in the order of amino to carboxyl is, (i.) A first chain comprising the VL and CL of the first domain; a second chain comprising the VH and CH1 of the first domain, a second domain in the form of scFv, and the first polypeptide monomer of the spacer dimerization domain, which together form a Fab with the first chain; a third chain comprising the second polypeptide monomer of the spacer dimerization domain, and the third domain comprising the VH and CH1 of the third domain, which together form a Fab with the VL and CL of the third domain of the fourth chain; and a fourth chain comprising the VL and CL of the third domain, and the fourth domain in the form of scFv; (ii.) A first domain in the form of Fab, a second domain in the form of Fab, a spacer comprising the first and second polypeptide monomers of the spacer dimerization domain, a third domain in the form of Fab, and a fourth domain in the form of Fab; (iii.) A first chain comprising the second domain in the form of scFv, VH and CH1 of the first domain which together with the second chain form a Fab, and the first polypeptide monomer of the spacer dimerization domain; a second chain comprising VL and CL of the first domain; a third chain comprising the second polypeptide monomer of the spacer dimerization domain, and the third domain which together with the VL and CL of the third domain which form a Fab, and the fourth chain comprising VL and CL of the third domain and the fourth domain in the form of scFv; (iv.) A first chain comprising the second domain in the form of scFv, VH and CH1 of the first domain which together with the second chain form a Fab, and the first polypeptide monomer of the spacer dimerization domain; a second chain comprising VL and CL of the first domain; a third chain comprising the second polypeptide monomer of the spacer dimerization domain, the fourth domain in the form of scFv, and the third domain which together with VL and CL of the third domain which form a Fab; and a fourth chain comprising VL and CL of the third domain; (v.) A first chain comprising the second domain in the form of scFv, VH and CH1 of the first domain which together with the second chain form Fab, and the first polypeptide monomer of the spacer dimerization domain; a second chain comprising VL and CL of the first domain; a third chain comprising the second polypeptide monomer of the spacer dimerization domain, the fourth domain in the form of scFv, and the third domain which together with the VL and CL of the third domain which form Fab; and a fourth chain comprising VL and CL of the third domain. Selected from the group consisting of; Each scFv comprises VH, linker and VL, or VL, linker and VH, preferably VH, linker and VL, in the order of amino to carboxyl. The antigen-binding molecule according to any one of claims 1 to 2.

4. If the spacer is a dimerizing domain, the arrangement of the binding domains in the order of amino to carboxyl is, (i.) A first chain comprising the first domain in the form of scFv, a first polypeptide monomer of the spacer dimerization domain, and a third domain in the form of scFv; and a second chain comprising the second domain in the form of scFv, a second polypeptide monomer of the spacer dimerization domain, and a fourth domain in the form of scFv; (ii.) A first chain comprising the first domain in the form of scFv, a first polypeptide monomer of the spacer dimerization domain, and a second domain in the form of scFv; and a second chain comprising the third domain in the form of scFv, a second polypeptide monomer of the spacer dimerization domain, and a fourth domain in the form of scFv; (iii.) A first chain comprising VL and CL of the first domain; a second chain comprising VH and CH1 of the first domain, the first polypeptide monomer of the spacer dimerization domain, and a third domain in the form of scFv, which together form Fab with the first chain; and a third chain comprising the second domain in the form of scFv, the second polypeptide monomer of the spacer dimerization domain, and a fourth domain in the form of scFv; (iv.) A first chain comprising VL and CL of the first domain; a second chain comprising VH and CH1 of the first domain, a first polypeptide monomer of the spacer dimerization domain, and a second domain in the form of scFv, which together form Fab with the first chain; and a third chain comprising a fourth domain in the form of scFv, a second polypeptide monomer of the spacer dimerization domain, and a third domain in the form of scFv; (v.) A first chain comprising VL and CL of the first domain; a second chain comprising VH and CH1 of the first domain, which together form Fab with the first chain, and a first polypeptide monomer of the spacer dimerization domain; and a third chain comprising the second domain in the form of scFv, a second polypeptide monomer of the spacer dimerization domain, a third domain in the form of scFv, and a fourth domain in the form of scFv; (vi.) A first chain comprising VL and CL of the first domain; a second chain comprising VH and CH1 of the first domain, a first polypeptide monomer of the spacer dimerization domain, a third domain in the form of scFv, and a fourth domain in the form of scFv, which together form Fab with the first chain; and a third chain comprising the second domain in the form of scFv and a second polypeptide monomer of the spacer dimerization domain; (vii.) A first chain comprising VL and CL of the first domain; a second chain comprising VH and CH1 of the first domain forming a Fab together with the first chain, a first polypeptide monomer of the spacer dimerization domain; a third chain comprising VL and CL of the third domain; a fourth chain comprising VH and CH1 of the third domain forming a Fab together with the fifth chain, a second polypeptide monomer of the spacer dimerization domain; a fifth chain comprising VL and CL of the second domain; and a sixth chain comprising VL and CL of the fourth domain; (viiii.) A first chain comprising VL and CL of the first domain; a second chain comprising VH and CH1 of the first domain, a first polypeptide monomer of the spacer dimerization domain, and a second domain in the form of scFv, which together form Fab with the first chain; a third chain comprising a fourth domain in the form of scFv, a second polypeptide monomer of the spacer dimerization domain, and VH and CH1 of the third domain, which together form Fab with the fourth chain; and a fourth chain comprising VL and CL of the third domain; A first chain comprising the first domain in the form of (ix.) scFv, the first polypeptide monomer of the spacer dimerization domain, and the third domain in the form of scFv; a second chain comprising the VH and CH1 of the second domain which together with the third chain form a Fab, the second polypeptide monomer of the spacer dimerization domain, and the VH and CH1 of the fourth domain which together with the fourth chain form a Fab; a third chain comprising the VL and CL of the second domain; and a fourth chain comprising the VL and CL of the fourth domain; (x.) A first chain comprising VL and CL of the first domain; a second chain comprising VH and CH1 of the first domain, the first polypeptide monomer of the spacer dimerization domain, and the third domain in the form of scFv, which together form Fab with the first chain; a third chain comprising VH and CH1 of the second domain, the second polypeptide monomer of the spacer dimerization domain, and the fourth domain in the form of scFv, which together form Fab with the fourth chain; and a fourth chain comprising VL and CL of the second domain; (xi.) A first chain comprising VL and CL of the first domain; a second chain comprising VH and CH1 of the first domain, the first polypeptide monomer of the spacer dimerization domain, and the second domain in the form of scFv, which together form Fab with the first chain; a third chain comprising VH and CH1 of the fourth domain, the second polypeptide monomer of the spacer dimerization domain, and the third domain in the form of scFv, which together form Fab with the fourth chain; and a fourth chain comprising VL and CL of the fourth domain; (xi.) A first chain comprising VL and CL of the first domain; a second chain comprising VH and CH1 of the first domain, which together form a Fab with the first chain, and a first polypeptide monomer of the spacer dimerization domain; a third chain comprising VH and CH1 of the second domain, which together form a Fab with the fourth chain, a second polypeptide monomer of the spacer dimerization domain, the third domain in the form of scFv, and the fourth domain in the form of scFv; and a fourth chain comprising VL and CL of the second domain; (xiii.) A first chain comprising VL and CL of the first domain; a second chain comprising VH and CH1 of the first domain which together form a Fab with the first chain, a first polypeptide monomer of the spacer dimerization domain, a third domain in the form of scFv, and a fourth domain in the form of scFv; a third chain comprising VH and CH1 of the second domain which together form a Fab with the fourth chain, and a second polypeptide monomer of the spacer dimerization domain; and a fourth chain comprising VL and CL of the second domain. Selected from the group consisting of; Each scFv includes VH, linker and VL, or VL, linker and VH, preferably VH, linker and VL, in the direction from N to C. The antigen-binding molecule according to claim 1.

5. When the single strand consists of at least one Fc domain, the spacer entity is preferably one domain or two or three covalently linked domains, each of which is in the order of amino to carboxyl: Hinge - CH2 - CH3 - Linker - Hinge - CH2 - CH3 Includes, Preferably, each of the polypeptide monomers in the spacer entity has an amino acid sequence that is at least 90% identical to a sequence selected from the group consisting of SEQ ID NOs: 17 to 24, Preferably, the CH2 domain in the spacer includes intradomain cysteine ​​disulfide bridges, or The antigen-binding molecule according to claim 1, wherein the spacer entity comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 13, 15-16, and 25-34.

6. The antigen-binding molecule according to claim 1, wherein the first peptide monomer of the first peptide chain in the dimerization spacer is SEQ ID NO: 35, the second peptide monomer of the second peptide chain in the dimerization spacer is SEQ ID NO: 36, and the two peptide monomers preferably form a heterodimer.

7. The aforementioned antigen-binding molecule (i) The first domain and the third domain each comprise two antibody-derived variable domains, and the second domain and the fourth domain each comprise two antibody-derived variable domains; (ii) The first domain and the third domain each comprise a variable domain derived from one antibody, and the second domain and the fourth domain each comprise a variable domain derived from two antibodies; (iii) The first domain and the third domain each contain variable domains derived from two antibodies, and the second domain and the fourth domain each contain variable domains derived from one antibody; or (iv) The first domain comprises a variable domain derived from one antibody, and the third domain comprises a variable domain derived from one antibody. The antigen-binding molecule according to claim 1, characterized in that...

8. The antigen-binding molecule comprises two polypeptide chains, The first polypeptide chain comprises the VH of the first domain, the VH of the second domain, the first polypeptide monomer which preferably includes hinge, CH2 and CH3 domains, the VH of the third domain, and the VH of the fourth domain; The second polypeptide chain comprises the VL of the first domain, the VL of the second domain, the first polypeptide monomer which preferably includes hinge, CH2 and CH3 domains, the VL of the third domain, and the VL of the fourth domain. Preferably, the first polypeptide monomer and the second polypeptide monomer form a heterodimer, thereby linking the first polypeptide chain and the second polypeptide chain. The antigen-binding molecule according to claim 1.

9. The first, second, third, and fourth binding domains each contain a VH domain and a VL domain in the order of amino to carboxyl, and the VH and VL within each domain are linked by a peptide linker, preferably a mobile linker containing serine, glutamine and / or glycine as amino acid components, preferably only serine (Ser, S) or glutamine (Gln, Q) and glycine (Gly, G), more preferably (G4S)n or (G4Q)n, and even more preferably SEQ ID NO: 1 or 3. The peptide linker preferably includes or consists of S(G4X)n and (G4X)n (where X is selected from the group consisting of Q, T, N, C, G, A, V, I, L, and M, and n is an integer selected from integers 1 to 20, preferably n is 1, 2, 3, 4, 5, or 6, and preferably X is Q), and preferably the peptide linker is (G4X)n (where n is 3 and X is Q), The peptide linker between the first binding domain and the second binding domain and the third binding domain and the fourth binding domain is preferably a mobile linker selected from the group consisting of SEQ ID NOs: 1-4, 6-12 and 1125, comprising serine, glutamine and / or glycine or glutamic acid, alanine and lysine as amino acid components. The antigen-binding molecule according to claim 1, wherein the peptide linker between the first binding domain or the second binding domain and the spacer, and / or between the third binding domain and the fourth binding domain and the spacer, is preferably a short linker rich in low molecular weight and / or hydrophilic amino acids, preferably glycine, and preferably SEQ ID NO:

5.

10. Either the first target cell surface antigen or the second target cell surface antigen is selected from the group consisting of CS1, BCMA, CDH3, FLT3, CD123, CD20, CD22, EpCAM, MSLN, and CLL1. Preferably, the antigen-binding molecule according to claim 1, wherein the first target cell surface antigen and the second target cell surface antigen are selected from the group consisting of CS1 and BCMA, BCMA and CS1, FLT3 and CD123, CD123 and FLT3, CD20 and CD22, CD22 and CD20, EpCAM and MSLN, MSLN and EpCAM, MSLN and CDH3, CDH3 and MSLN, FLT3 and CLL1, and CLL1 and FLT3, respectively.

11. The second and fourth binding domains include a VH region containing CDR-H1, CDR-H2, and CDR-H3 selected from sequence numbers 37-39, 45-47, 53-55, 61-63, 69-71, 436-438, 1126-1128, 1136-1138, 1142-1144, 1148-1150, and 1217-1219, and sequence numbers 40-4 2. A VL region including CDR-L1, CDR-L2, and CDR-L3 selected from 48-50, 56-58, 64-66, 72-74, 439-441, 1129-1131, 1139-1141, 1145-1147, 1151-1153, and 1220-1222, preferably 61-63 and 64-66 or 1217-1219 and 1220-1222. Preferably, the second binding domain and the fourth binding domain include a VH region selected from SEQ ID NOs: 43, 51, 59, 67, 75, 442, 1132 and 1223, preferably 67 or 1223. Preferably, the second binding domain and the fourth binding domain include a VL region selected from sequence numbers 44, 52, 60, 68, 76, 443, 1133 and 1224, preferably 68 or 1224. Preferably, the second and fourth binding domains include a VH region selected from SEQ ID NOs: 43, 51, 59, 67, 75, 442, 1132, and 1223, preferably 67, and a VL region selected from SEQ ID NOs: 44, 52, 60, 68, 76, 443, 1133, and 1224, preferably 68, wherein the VH region is 1132 and the VL region is 1133, and the second and / or fourth binding domains include, as scFab domains, the CH1 domain of SEQ ID NO: 1134 and the CLK domain of SEQ ID NO: 1135 The antigen-binding molecule according to claim 1, comprising a , wherein the VH and VL regions are linked to each other by a linker preferably selected from SEQ ID NOs: 1, 3 and 1125, or the VH of the VH-CH1 of the second and fourth domains is SEQ ID NO: 1223, the CH1 of the VH-CH1 of the second and fourth domains is SEQ ID NO: 1224, the VL of the VL-CL of the second and fourth domains is SEQ ID NO: 1225, and the CL of the VL-CL of the second and fourth domains is SEQ ID NO: 1226.

12. The first (target) binding domain and / or the third (target) binding domain are sequence numbers 77-79, 86-88, 95-97, 103-105, 111-113, 119-121, 127-129, 135-137, 143-145, 151-153, 159-161, 168-170, 177-179, 185-187, 194-196, 203-205, 212-214, 221-223, 230-232, 238-240, 334-336, 356-358, 365-367, 376-378, 385-387 and 194, 432 and 19 6, 446-448, 454-456, 462-464, 470-472, 478-480, 486-488, 494-496, 502-504, 510-512, 518-520, 526-528, 534-536, 542-544, 550-552, 558-560, 56 6-568, 574-576, 582-584, 590-592, 598-600, 606-608, 614-616, 622-624, 630-632, 638-640, 646-648, 654-656, 662-664, 670-672, 678-680, 686-68 8, 694-696, 702-704, 710-712, 718-720, 726-728, 734-736, 742-744, 750-752, 758-760, 766-768, 774-776, 782-784, 790-792, 798-800, 806-808, 81 4-816, 822-826, 830-832, 838-840, 846-848, 854-856, 862-864, 870-872, 878-880, 886-888, 894-896, 902-904, 910-912, 918-920, 926-928, 934-93 6, CDR-H1, CDR-H2 and CDR-H3 selected from 942-944, 950-952, 958-960, 966-968, 974-976, 982-984, 990-992, 998-1000, 1006-1008, 1014-1016, 1022-1024, 1030-1032, 1038-1040, 1046-1048, 1054-1056 and 1062-1064, or preferably sequence listing 6, preferably 86-88 and 194, 432 and 196 with respect to the first and third binding domains, respectively.More preferably, the VH region includes any combination of CDR-H1, CDR-H2, and CDR-H3 as disclosed in 194, 432, and 196 with respect to the first binding domain and 86 to 88 with respect to the third binding domain or in 1227 to 1229 and 1237 to 1239 with respect to the first and third binding domains. Preferably, the first (target) binding domain and / or the third (target) binding domain are sequence numbers 80-82, 89-91, 98-100, 106-108, 114-116, 122-124, 130-132, 138-140, 146-148, 154-156, 162-164, 171-173, 180-182, 188-190, 197-199, 206-208, 215-217, 224-226, 233-235, 241-243, 337-339, 359-361, 368-370, 379-381, 388-390, 449 ~451, 457~459, 465~467, 473~475, 481~483, 489~491, 497~499, 505~507, 513~515, 521~523, 529~531, 537~539, 545~547, 553~555, 561~563, 569~571, 577~579, 585~587, 593~595, 601~603, 609~611, 617~619, 625~627, 633~635, 641~643, 649~651, 657~659, 665~667, 673~675, 681~683, 689~691, 697-699, 705-707, 713-715, 721-723, 729-731, 737-739, 745-747, 753-755, 761-763, 769-771, 777-779, 785-787, 793-795, 801-803, 809-811, 81 7-819, 825-829, 833-835, 841-843, 849-851, 857-859, 865-867, 873-875, 881-883, 889-891, 897-899, 905-907, 913-915, 921-923, 929-931, 937-9 CDR-L1, CDR-L2, and CDR-L3 selected from 39, 945-947, 953-955, 961-963, 969-971, 977-979, 985-987, 993-995, 1001-1003, 1009-1011, 1017-1019, 1025-1027, 1033-1035, 1041-1043, 1049-1051, 1057-1059, and 1065-1067, or preferably sequence listing 6, preferably 89-91 and 197-199 with respect to the first and third binding domains, respectively.The antigen-binding molecule according to claim 1, more preferably comprising a VL region including any combination of CDR-L1, CDR-L2, and CDR-L3, as disclosed in 197-199 with respect to the first binding domain and 89-91 with respect to the third binding domain, or in 1230-1232 and 1240-1242 with respect to the first and third binding domains.

13. The first (target) binding domain and / or the third (target) binding domain are sequence numbers 83, 92, 101, 109, 117, 125, 133, 141, 149, 157, 165, 174, 183, 191, 200, 209, 218, 227, 236, 244, 340, 362, 371, 382, ​​391, and 433, 452, 460, 468, 476, 484, 492 ,500,508,516,524,532,540,548,556,564,572,580,588,596,604,612,620,628,636,644,652,660,668,676,684,692,700,708,716,724,732,740,748,756,764,772,780,788,796,804,812,820,828 , a VH region selected from 836, 844, 852, 860, 868, 876, 884, 892, 900, 908, 916, 924, 932, 940, 948, 956, 964, 972, 980, 988, 996, 1004, 1012, 1020, 1028, 1036, 1044, 1052, 1060, and 1068, or preferably any VH disclosed together in sequence listing 52, preferably 433 and 92 with respect to the first binding domain and the third binding domain, more preferably 433 with respect to the first binding domain and 92 with respect to the third binding domain, or 1233+1235 and 1243+1245 (VH and CH1 in Fab) with respect to the first binding domain and the third binding domain, The first (target) binding domain and / or the third (target) binding domain are sequence numbers 84, 93, 102, 110, 118, 126, 134, 142, 150, 158, 166, 175, 184, 192, 201, 210, 219, 228, 237, 245, 341, 363, 372, 383, 392, 453, 461, 469, 477, 485, 493, 501 ,509,517,525,533,541,549,557,565,573,581,589,597,605,613,621,629,637,645,653,661,669,677,685,693,701,709,717,725,733,741,749,757,765,773,781,789,797,805,813,821,829,83 VL regions selected from 7, 845, 853, 861, 869, 877, 885, 893, 901, 909, 917, 925, 933, 941, 949, 957, 965, 973, 981, 989, 997, 1005, 1013, 1021, 1029, 1037, 1045, 1053, 1061, and 1069, or preferably including any VLs disclosed together in sequence listing 52, preferably 200 and 93 with respect to the first and third binding domains, more preferably 200 with respect to the first binding domain and 93 with respect to the third binding domain, or 1234+1236 and 1244+1246 (VLs and CLs in Fab) with respect to the first and third binding domains, respectively. Preferably, the antigen-binding molecule according to claim 1, wherein the first (target)-binding domain and / or the third (target)-binding domain comprises a VL region whose stability is increased by a single amino acid exchange (E to I) selected from SEQ ID NOs: 85, 94, 193, 202, 211, 220, 229, 364, 384, 393, preferably 94 and 202.

14. The antigen-binding molecule according to claim 1, comprising a combination of amino acid sequences selected from the group consisting of SEQ ID NOs: 1259 and 1251, 1247 and 1248, 1249 and 1250, 1254, 1255 and 1253, 1252, 1257, 1253 and 1256, and 1254, 1258, 1253 and 1256.

15. A pharmaceutical composition comprising the antigen-binding molecule described in claim 1.

16. An antigen-binding molecule according to claim 1, for use in the prevention, treatment, or improvement of a disease selected from proliferative disorders, neoplastic disorders, cancer, or immunodeficiency, The aforementioned disease is preferably acute myeloid leukemia (AML), non-Hodgkin lymphoma (NHL), non-small cell lung cancer (NSCLC), pancreatic cancer, and colorectal cancer (CRC), and is an antigen-binding molecule.