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

Abstract

The present invention provides a multi-strand multi-targeting bispecific antigen-binding molecule characterized by comprising a first and a second bispecific entity, each entity comprising a domain that binds to a target, a second domain that binds to an extracellular epitope of human and macaque CD3 epsilon chains wherein the two bispecific entities are linked to each other by a spacer, the spacer separates the first and second bispecific entities. In addition, the invention provides polynucleotides encoding the multi-targeting bispecific antigen binding molecules, vectors comprising such polynucleotides, host cells expressing constructs, and pharmaceutical compositions comprising the antigen binding molecules.

Description

Technical Field

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

[0002] Redirecting T cell activity against tumor cells by bispecific molecules that are independent of T cell receptor specificity is a further developing 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). This new class of protein-based drugs can often bind to two different types of antigens simultaneously. They are known in several structural forms and have been 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 that can be used for immuno-oncology can be antigen-binding polypeptides such as antibodies, for example IgG-like 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 traditional monoclonal antibody (mAb) structure with two Fab arms and one Fc region, except that the two Fab sites bind to different antigens. Non-full-length bispecific antibodies can lack the entire Fc region. These include chemically linked Fabs, consisting only of Fab regions, and various types of bivalent and trivalent single-chain variable fragments (scFv). There are also fusion proteins that simulate the variable domains of two antibodies. An example of this form is a bispecific T cell engager. (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] Exemplary bispecific antibody-derived molecules are The molecule is a recombinant protein construct prepared from two flexibly linked antibody-derived binding domains. One binding domain of the antigen binding molecule is specific for a selected tumor-associated surface antigen on the target cell; the second binding domain is specific for CD3 (a subunit of the T cell receptor complex on T cells). Antigen-binding molecules are uniquely suited to transiently connect T cells to target cells and, at the same time, strongly activate the inherent cytolytic potential of T cells against target cells. The first generation of AMG 103 and AMG 110 were developed to enter clinical trials. An important further development of the antigen binding molecules (see WO 99 / 54440 and WO 2005 / 040220) is the provision of bispecific antigen binding molecules that bind to a context independent epitope at the N-terminus of the CD3ε chain (WO 2008 / 119567). The antigen binding molecules not only show cross-species specificity for CD3ε chains of humans and macaques, or woolly monkeys (Callithrix jacchus), woolly tamarins (Saguinusoedipus) or squirrel monkeys (Saimiri sciureus), but also do not exhibit non-specific activation of T cells to the same extent as observed for previous generation T cell engaging antibodies due to recognition of this specific epitope (rather than the epitope of the CD3 binder in the previously described bispecific T cell engaging molecules). This reduction in T cell activation is associated with less or reduced T cell redistribution in patients, which has been identified as a risk of side effects, for example, in pasotuximab.

[0005] Antibody-based molecules as described in WO 2008 / 119567 are characterized by rapid clearance from the body; therefore, although they can quickly reach most parts of the body, their in vivo applications may be limited by their short persistence in the body. On the other hand, their concentration in the body can be adjusted and fine-tuned in a short time. Because the in vivo half-life of this small single-chain molecule is short, long-term administration by continuous intravenous infusion is used to achieve therapeutic effects. However, bispecific antigen-binding molecules with more favorable pharmacokinetic properties (including longer half-lives as described in WO 2017 / 134140) are available. Increased half-life is typically useful in in vivo applications of immunoglobulins, antibody fragments or constructs of particularly small size, for example, for patient compliance.

[0006] A challenging and persistent problem of antibody-based immuno-oncology is tumor escape. Such tumor escape occurs when the immune system (even when triggered or directed by some antibody-based immunotherapies) is not sufficiently capable of eradicating tumors, which carry accumulated genetic and epigenetic changes and use several mechanisms to win the immune editing process (Keshavarz-Fathi, Mahsa; Rezaei, Nima (2019) "Vaccines for Cancer Immunotherapy"). In general, four mechanisms that interfere with effective anti-tumor immune responses are known: (1) defective tumor antigen processing or presentation, (2) lack of activation mechanisms, (3) inhibition mechanisms and immunosuppressive states, and (4) resistant tumor cells. Especially for the first mechanism, tumor antigens may exist in new forms due to genetic instability, tumor mutations, and escape from the immune system. Epitope-negative tumor cells remain hidden and are therefore resistant to immune rejection. They develop after the elimination of epitope-positive tumor cells, similar to Darwin's theory of natural selection. Therefore, when such tumor cells no longer express their respective antigens due to tumor escape, antibody-based immunotherapy for antigens on tumor cells becomes ineffective. The antigen loss is understood herein as a driving force for tumor escape and can therefore be used interchangeably. Therefore, it is desirable to provide improved antibody-based immuno-oncology that addresses the problem of antigen loss to effectively prevent tumor escape.

[0007] A perhaps more pressing challenge to the widespread application of immuno-oncology relative to T cell-engaging bispecific molecules is the availability of suitable targets (Bacac et al., Clin Cancer Res; 22(13) July 1, 2016). For example, solid tumor targets may be overexpressed on tumor cells but expressed at lower but significant levels in non-malignant primary cells in critical tissues. In nature, according to Bacac et al., T cells can distinguish between high and low antigen-expressing cells by relatively low-affinity T cell receptors (TCRs), which can still achieve high-affinity binding to target cells that express sufficiently high levels of target antigen. Therefore, there is a great need for T cell-engaging bispecific molecules that can facilitate the above and thus maximize the window between killing high and low target-expressing cells. One approach discussed in the art is the use of dual targeting of two antigens, which can improve target selectivity for normal tissues that express only one target antigen or express low levels of both target antigens. This effect is believed to rely on the affinity component mediated by the simultaneous binding of the bsAb to two antigens on the same cell. Therefore, relative to dual targeting itself, some multispecific monoclonal antibodies (mAbs) or other immune constructs are known in the art. WO 2014 / 116846 teaches a multispecific binding protein comprising a first binding site specifically binding to a target cell antigen, a second binding site specifically binding to a cell surface receptor on an immune cell, and a third binding site specifically binding to a cell surface regulator on an immune cell. US2017 / 0022274 discloses a trivalent T cell redirection complex comprising a bispecific antibody, wherein the bispecific antibody has two binding sites for a tumor-associated antigen (TAA) and one binding site for a T cell.

[0008] However, as described above, dual targeting alone may not be sufficient to achieve effective target selectivity in the molecule (Mazor et al., mAbs [monoclonal antibodies] 7: 3, 461-469; May / June 2015). In particular, the configuration of the bsAb binding domain, i.e., monovalent versus bivalent, is a key factor. More importantly, merely providing a bispecific molecule with several valence states may not lead to a clinically suitable treatment method, and the potential risk spectrum in terms of significant immunological side effects (e.g., cytokine release syndrome (CRS)) must also be considered. Therefore, although antibody-based immunotherapy has been successful preclinically and clinically so far, there are still obvious limitations, including differential responses between individuals and cancer types. Dose-limiting toxicity may be a limiting factor in the efficacy of antibody-based immunotherapy, so not all patients respond to treatment at available safe doses. Therefore, it is also necessary to reduce the dose-limiting toxicity of antibody-based immunotherapy so that such therapies can be used for more patients with diverse proliferative diseases.

[0009] Although different multispecific antibodies or antibody fragments are known in the art, some of which address T cells, multitargeting bispecific molecules have not been proposed heretofore that address the need to overcome dose-limiting toxicities in T cell redirecting immunotherapy by increasing the therapeutic window and are stable and readily available therapeutic systems. Summary of the invention

[0010] In view of the above-mentioned various unmet needs, the object of the present invention is to provide a molecule (i.e., a multi-chain molecule) comprising at least two polypeptide chains, which molecule is preferably an antigen binding molecule. The molecule of the present invention is further preferably bispecific, such as a T cell engagement molecule. In addition, the molecule of the present invention is preferably multi-targeted, for example, it is generally capable of immunospecifically binding to at least two antigens on a target cell, which antigens are generally associated with one or more diseases. It is further preferred that the molecule of the present invention is generally capable of immunospecifically binding to two antigens on an effector cell at the same time, preferably for treating the one or more diseases. Thus, the present invention provides a preferred multi-targeting bispecific antigen binding molecule comprising at least one polypeptide, wherein the molecule is characterized in that it comprises at least five different 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 macaque) CD3 chain, wherein the first binding domain and the second binding domain together form a first bispecific entity, (iii.) a spacer that connects but also sufficiently separates the first bispecific entity from the second bispecific entity, the second bispecific entity comprising (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 macaque) CD3 chain. Preferably, these domains are (i.) scFv domains composed of VH and VL domains, respectively, in an amino to carboxyl orientation, wherein a flexible but short peptide linker connects 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, and / or (ii.) Fab domains, which domains comprise a first polypeptide monomer comprising a VL and CL domain, and a second polypeptide monomer comprising a VH and CH domain. Surprisingly, the multi-chain multi-targeted bispecific antigen-binding molecules described herein are generally capable of enabling T cells to distinguish and kill cells that express only one or two targets that are generally associated with a specific disease, thereby opening a therapeutic window and reducing the risk of off-target toxicity and side effects. In addition, the present invention provides polynucleotides encoding multi-targeted bispecific antigen-binding molecules, vectors comprising such polynucleotides, host cells expressing constructs, and pharmaceutical compositions comprising the antigen-binding molecules.

[0011] In a first aspect, it is envisaged in the context of the present invention to provide a multi-chain multi-targeting bispecific molecule comprising at least two polypeptide chains, wherein the molecule comprises

[0012] (i.) a first binding domain that binds to a first target cell surface antigen (TAA1),

[0013] (ii.) a second binding domain which binds to an extracellular epitope of the human and / or macaque CD3 chain,

[0014] (iii.) a third binding domain that binds to a second target cell surface antigen (TAA2), and

[0015] (iv.) a fourth binding domain, which binds to an extracellular epitope of the human and / or macaque CD3 chain,

[0016] wherein 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, and

[0017] Wherein the molecule further comprises a spacer entity selected from:

[0018] (1.) A dimerization domain selected from the group consisting of:

[0019] (a.) An Fc domain comprising a first and a second polypeptide monomer comprising a hinge, a CH2 domain, and a CH3 domain, respectively, wherein the first and the second polypeptide monomers form a heterodimer; wherein the heterodimer is formed by:

[0020] - a charge pair mutation selected from the group consisting of: (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, K370D, and E356K. 0E and K360E, (vii.) D399K, K409D, K392D, E357K, K370E, E356K, and K439E, and (viii.) D399K, K409D, K392D, E357K, K370E, K360E, E356K, and K439D, preferably comprising 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, wherein positions are numbered according to EU; or

[0021] - knobs-into-holes mutations, which preferably comprise T366S, L368A and Y407V mutations in the first polypeptide monomer and T366W mutation in the second monomer, wherein the positions are numbered according to EU;

[0022] (b.) a human serum albumin (HSA) domain comprising a first and a second polypeptide monomer, wherein the first and the second polypeptide monomer correspond to HSA subdomains, respectively, wherein the first and the second polypeptide monomers form a native HSA-like heterodimer; and

[0023] (c.) a Fab comprising a first and a second polypeptide monomer, wherein the first polypeptide monomer comprises a VL and a CL domain, and the second polypeptide monomer comprises a VH and a CH1 domain, wherein the CL and CH1 domains are connected by a disulfide bridge;

[0024] wherein the dimerization domain comprises two N-termini and two C-termini, respectively, wherein at least one N-terminus and one C-terminus are respectively connected to a bispecific entity, wherein any one of the first, second, third and fourth domains can be selected from any form of binding domain, preferably selected from Fab and a single chain domain, and the single chain domain is preferably selected from a single chain Fv (scFv) and scFab;

[0025] (2.) a single chain domain selected from the group consisting of ubiquitin, β2 microglobulin, a VH domain only, a PSI domain from a Met-receptor, a fibronectin type III domain from tenascin, granulocyte-macrophage colony stimulating factor (GM-CSF), interleukin-4, an extracellular domain of CD137L, interleukin-2, a PD-1 binding domain from human programmed cell death 1 ligand 1 (PDL1), Tim-3 (AS24-130), MiniSOG, a programmed cell death protein 1 (PD1) domain, human serum albumin (HSA), or a single chain Fc (scFc) domain comprising two polypeptide monomers each comprising a hinge, a CH2 and CH3 domain, a hinge and additional CH2 and CH3 domains, wherein the two polypeptide monomers are fused to each other via a peptide linker,

[0026] wherein the single chain domain comprises an N-terminus and a C-terminus, which are respectively linked to a bispecific entity, wherein at least one of the first, second, third and fourth binding domains is a double-chain Fab, and any one of the remaining up to at least three binding domains can be selected from any form of binding domain, preferably selected from Fab and a single chain domain, the single chain domain is preferably selected from scFv and scFab;

[0027] The distance between the Cα atom of the first amino acid at the N-terminus and the last amino acid at the C-terminus of the spacer entity is At least wherein the spacer entity separates the first bispecific entity from the second bispecific entity by at least about wherein the indicated distance is preferably understood as the distance between the centers of mass of (i.) the first binding domain and the third binding domain or (ii.) the first bispecific entity and the second bispecific entity, and the spacer entity is located between the first bispecific entity and the second bispecific entity.

[0028] Within said aspects, it is also envisaged in the context of the present invention to provide a multi-chain multi-targeting bispecific antigen binding molecule, wherein when the spacer is a single chain domain, the arrangement of the binding domains in order from amino to carboxyl is selected from the group consisting of:

[0029] (i.) The first and second domains, the spacer, the third and fourth domains

[0030] (ii.) The first and second domains, the spacer, the fourth and third domains

[0031] (iii.) the second and first domains, a spacer, the third and fourth domains, and

[0032] (iv.) The second and first domains, a spacer, and the fourth and third domains.

[0033] Within said aspects, it is also envisaged in the context of the present invention to provide a multi-chain multi-targeting bispecific antigen binding molecule, wherein when the spacer is a single chain domain, the arrangement of the binding domains in order from amino to carboxyl is selected from the group consisting of:

[0034] (i.) a first domain in Fab form, a second domain in scFv form, a spacer, a third domain in Fab form, and a fourth domain in scFv form (e.g. Figure 3 B);

[0035] (ii.) the first domain in Fab form, the second domain in Fab form, a spacer, the third domain in Fab form, and the fourth domain in Fab form (e.g. Figure 3 D)

[0036] (iii.) a first domain in scFv form, a second domain in Fab form, a spacer, a third domain in scFv form, and a fourth domain in Fab form (e.g. Figure 3 H);

[0037] (iv.) a first domain in scFv form, a second domain in scFv form, a spacer, a third domain in scFv form, and a fourth domain in Fab form;

[0038] (v.) a first domain in scFv form, a second domain in scFv form, a spacer, a third domain in Fab form, and a fourth domain in scFv form;

[0039] (vi.) a first domain in Fab form, a second domain in scFv form, a spacer, a third domain in scFv form, and a fourth domain in scFv form; and

[0040] (vii.) a first domain in scFv form, a second domain in Fab form, a spacer, a third domain in scFv form, and a fourth domain in scFv form,

[0041] Each scFv comprises, from amino to carboxyl, VH, a linker and VL or VL, a linker and VH, preferably VH, a linker and VL.

[0042] Within said aspects, it is also envisaged in the context of the present invention to provide a multi-chain multi-targeting bispecific antigen binding molecule, wherein when the spacer is a dimerization domain, the arrangement of the binding domains in order from amino to carboxyl is selected from the group consisting of:

[0043] (i.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that together with the first chain form a Fab, a second domain in scFv form, a first polypeptide monomer of the spacer dimerization domain, and a third chain comprising the second polypeptide monomer of the spacer dimerization domain, a third domain that together with VL and CL of the third domain on the fourth chain form a Fab, and a fourth chain comprising VL and CL of the third domain and a fourth domain in scFv form (e.g. Figure 3 A);

[0044] (ii.) the first domain in Fab form, the second domain in Fab form, a spacer, the third domain in Fab form, and the fourth domain in Fab form (e.g. Figure 3 C);

[0045] (iii.) a first chain comprising the second domain in scFv format, VH and CH1 of the first domain that together with the second chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, a second chain comprising VL and CL of the first domain, a second polypeptide monomer comprising the spacer dimerization domain, a third chain comprising VH and CH1 of the third domain that together with VL and CL of the third domain on the fourth chain form a Fab, and a fourth chain comprising VL and CL of the third domain and the fourth domain in scFv format (e.g. Figure 3 E);

[0046] (iv.) a first chain comprising a second domain in scFv form, VH and CH1 of the first domain that together with the second chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, a second chain comprising VL and CL of the first domain, a second polypeptide monomer comprising the spacer dimerization domain, a fourth domain in scFv form, a third chain comprising VH and CH1 of the third domain that together with VL and CL of the third domain on the fourth chain form a Fab, and a fourth chain comprising VL and CL of the third domain (e.g. Figure 3 F);

[0047] (v.) a first chain comprising a second domain in scFv form, VH and CH1 of the first domain that together with the second chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, a second chain comprising VL and CL of the first domain, a second polypeptide monomer comprising the spacer dimerization domain, a fourth domain in scFv form, a third chain comprising VH and CH1 of the third domain that together with VL and CL of the third domain on the fourth chain form a Fab, and a fourth chain comprising VL and CL of the third domain (e.g. Figure 3 G);

[0048] Each scFv comprises, from amino to carboxyl, VH, a linker and VL or VL, a linker and VH, preferably VH, a linker and VL.

[0049] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein when the spacer is a dimerization domain, the arrangement of these binding domains is in amino to carboxyl order and is selected from the group consisting of:

[0050] (i.) comprising a first domain in scFv form, a first polypeptide monomer of the spacer dimerization domain, a first chain of the third domain in scFv form, and comprising a second domain in scFv form, a second polypeptide monomer of the spacer dimerization domain, a second chain of the fourth domain in scFv form (e.g. Figure 2 A);

[0051] (ii.) comprising a first domain in scFv form, a first polypeptide monomer of the spacer dimerization domain, a first chain of the second domain in scFv form, and comprising a third domain in scFv form, a second polypeptide monomer of the spacer dimerization domain, a second chain of the fourth domain in scFv form (e.g. Figure 2 B);

[0052] (iii.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that together with the first chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, and a third chain comprising the second domain in the form of an scFv, a second polypeptide monomer of the spacer dimerization domain, and a fourth domain in the form of an scFv (e.g. Figure 2 C);

[0053] (iv.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that together with the first chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, a second domain in scFv form, and a third chain comprising a fourth domain in scFv form, a second polypeptide monomer of the spacer dimerization domain, a third domain in scFv form (e.g. Figure 2 D)

[0054] (v.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that together with the first chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, and a third chain comprising the second domain in scFv form, the second polypeptide monomer of the spacer dimerization domain, the third domain in scFv form, and the fourth domain in scFv form (e.g. Figure 2 E);

[0055] (vi.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that together with the first chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, a third domain in scFv form, and a fourth domain in scFv form, and a third chain comprising the second domain in scFv form, a second polypeptide monomer of the spacer dimerization domain (e.g. Figure 2 F);

[0056] (vii.) a first chain comprising VL and CL of the first domain, a first polypeptide monomer comprising VH and CH1 of the first domain that together with the first chain form a Fab, a second chain comprising VH and CH1 of the third domain that together with the third chain form a Fab, a third chain comprising VL and CL of the third domain, a fourth chain comprising VH and CH1 of the second domain that together with the fifth chain form a Fab, 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 (e.g. Figure 2 G);

[0057] (viii.) a first chain comprising VL and CL of the first domain, a first polypeptide monomer comprising VH and CH1 of the first domain that together with the first chain form a Fab, a second chain comprising the second domain in the form of an scFv, a second polypeptide monomer comprising the spacer dimerization domain, a third chain comprising VH and CH1 of the third domain that together with the fourth chain form a Fab, and a fourth chain comprising VL and CL of the third domain (see, e.g., Figure 2 H);

[0058] (ix.) comprising a first domain in scFv format, a first polypeptide monomer of the spacer dimerization domain, a first chain of the third domain in scFv format, VH and CH1 of the second domain that together with the third chain form a Fab, a second polypeptide monomer of the spacer dimerization domain, a second chain of VH and CH1 that together with the fourth chain form a fourth domain of Fab, a third chain comprising VL and CL of the second domain, and a fourth chain comprising VL and CL of the fourth domain (e.g. Figure 2 I);

[0059] (x.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that together with the first chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, and a third chain of the third domain in scFv form, a fourth chain comprising VH and CH1 of the second domain that together with the fourth chain form a Fab, a second polypeptide monomer of the spacer dimerization domain, and a fourth chain comprising VL and CL of the second domain (e.g. Figure 2 J);

[0060] (xi.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that together with the first chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, a second chain comprising VH and CH1 of the fourth domain that together with the fourth chain form a Fab, a second polypeptide monomer of the spacer dimerization domain, a third chain of the third domain that is in the form of an scFv, and a fourth chain comprising VL and CL of the fourth domain (e.g. Figure 2 K);

[0061] (xii.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that together with the first chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, a second polypeptide monomer of the spacer dimerization domain that together with the fourth chain form a Fab, a third domain in scFv form, a third chain of the fourth domain in scFv form, and a fourth chain comprising VL and CL of the second domain (e.g. Figure 2 L);

[0062] (xiii.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that together with the first chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, a third domain in scFv form, a fourth domain in scFv form, a third chain comprising VH and CH1 of the second domain that together with the fourth chain form a Fab, a second polypeptide monomer of the spacer dimerization domain, and a fourth chain comprising VL and CL of the second domain (e.g. Figure 2 M);

[0063] Each scFv comprises VH, linker and VL or VL, linker and VH in the N-to-C orientation, preferably VH, linker and VL.

[0064] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the spacer entity is a globular protein wherein the distance between the Cα atoms of the first amino acid at the N-terminus and the last amino acid at the C-terminus is at least Preferably at least More preferably at least so that the first bispecific entity and the second bispecific entity are effectively spaced apart, preferably at least

[0065] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule, wherein said spacer entity (when the spacer is a single chain) which effectively separates the first bispecific entity and the second bispecific entity is selected from the group consisting of: ubiquitin, β2 microglobulin, a SAND domain, green fluorescent protein (GFP), a VHH antibody lama domain, a PSI domain from a Met receptor, a fibronectin type III domain from tenascin, a granulocyte-macrophage colony stimulating factor (GM-CSF), interleukin-4, a CD137L extracellular domain, an interleukin-2, a PD-1 binding domain from human programmed cell death 1 ligand 1 (PDL1), a Tim-3 (AST) binding domain. 24-130), MiniSOG, a programmed cell death protein 1 (PD1) domain, human serum albumin (HSA) or a derivative of any of the foregoing spacer entities, a multimer of a rigid linker, and an Fc domain or a dimer or trimer thereof, each Fc domain comprising two polypeptide monomers, each polypeptide monomer comprising a hinge, a CH2 and CH3 domain hinge and additional CH2 and CH3 domains, wherein the two polypeptide monomers are fused to each other via a peptide linker, or wherein the two polypeptide monomers are linked together to form a heterodimer via non-covalent CH3-CDH3 interactions and / or covalent disulfide bonds.

[0066] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule, wherein the spacer entity (when a single chain) is at least one Fc domain, preferably one domain or two or three covalently linked domains, which domain or each of these domains comprises, in order from amino to carboxyl:

[0067] Hinge-CH2-CH3-Joint-Hinge-CH2-CH3.

[0068] Within said aspect, it is also envisaged in the context of the present invention to provide an antigen binding molecule, wherein each of said polypeptide monomers in the spacer entity has an amino acid sequence having at least 90% identity with a sequence selected from the group consisting of: SEQ ID NO: 17-24, wherein preferably each of said polypeptide monomers has an amino acid sequence selected from SEQID NO: 17-24.

[0069] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the CH2 domain in the spacer comprises an intradomain cysteine ​​disulfide bridge.

[0070] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule, wherein the single-chain spacer entity comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 13 and 15 to 16 and 25 to 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 lama domain (SEQ ID NO: 1086), PSI domain from Met receptor (SEQ ID NO: 1087), fibronectin type III domain 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), NO: 1092), PD-1 binding domain from human programmed cell death 1 ligand 1 (PDL1) (SEQ ID NO: 1093), Tim-3 (AS24-130) (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 amino acids thereof having at least 90%, preferably 95% or even 98% sequence identity, preferably scFc (SEQ ID NO: 25).

[0072] Within said aspect, it is also envisaged 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, wherein the two peptide monomers preferably form a heterodimer.

[0073] Within said aspects, it is also envisaged that the antigen binding molecule is characterized in that:

[0074] (i) the first domain and the third domain comprise two antibody-derived variable domains, and the second domain and the fourth domain comprise two antibody-derived variable domains;

[0075] (ii) the first domain and the third domain comprise one antibody-derived variable domain, and the second domain and the fourth domain comprise two antibody-derived variable domains;

[0076] (iii) the first domain and the third domain comprise two antibody-derived variable domains, and the second domain and the fourth domain comprise one antibody-derived variable domain; or

[0077] (iv) the first domain comprises an antibody derived variable domain, and the third domain comprises an antibody derived variable domain.

[0078] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule comprising two polypeptide chains, wherein

[0079] The first polypeptide chain comprises the VH of the first domain, the VH of the second domain, preferably a first polypeptide monomer comprising hinge, CH2 and CH3 domains, the VH of the third domain and the VH of the fourth domain; and

[0080] The second polypeptide chain comprises the VL of the first domain, the VL second domain, preferably a first polypeptide monomer comprising hinge, CH2 and CH3 domains, the VL of the third domain and the VL of the fourth domain,

[0081] Preferably, the first and second polypeptide monomers form a heterodimer, thereby connecting the first and second polypeptide chains.

[0082] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule, wherein the first, second, third and fourth binding domains each comprise a VH domain and a VL domain in amino to carboxyl order, wherein the VH and VL within each domain are connected by a peptide linker, preferably a flexible linker comprising serine, glutamine and / or glycine as amino acid structural units, preferably only serine (Ser, S) or glutamine (Gln, Q) and glycine (Gly, G), more preferably (G4S)n or (G4Q)n, even more preferably SEQ ID NO: 1 or 3.

[0083] Within said aspects, it is also envisaged in the context of the present invention to provide a peptide linker, wherein the peptide linker comprises or consists of S(G4X)n and (G4X)n, wherein X is selected from the group consisting of Q, T, N, C, G, A, V, I, L and M, wherein n is an integer selected from the integers 1 to 20, preferably wherein n is 1, 2, 3, 4, 5 or 6, preferably wherein X is Q, wherein preferably the peptide linker is (G4X)n, n is 3, and X is Q.

[0084] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule, wherein 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 flexible linker, which comprises serine, glutamine and / or glycine or glutamic acid, alanine and lysine as amino acid structural units, preferably selected from the group consisting of SEQ ID NO: 1 to 4, 6 to 12 and 1125.

[0085] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule, wherein the peptide linker between the first binding domain or the second binding domain and the spacer, and / or the peptide linker between the third binding domain and the fourth binding domain and the spacer, respectively, is preferably a short linker rich in small amino acids and / or hydrophilic amino acids, preferably glycine, and is preferably SEQ ID NO: 5.

[0086] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule, wherein any one of the first target cell surface antigen and the second target cell surface antigen is selected from the group consisting of: CS1, BCMA, CDH3, FLT3, CD123, CD20, CD22, EpCAM, MSLN and CLL1.

[0087] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the first target cell surface antigen and the second target cell surface antigen are different.

[0088] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the first target cell surface antigen and the second target cell surface antigen are the same.

[0089] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the first binding domain is capable of binding to a first target cell surface antigen and simultaneously the third binding domain is capable of binding to a second target cell surface antigen, preferably wherein the first target cell surface antigen and the second target cell surface antigen are on the same target cell.

[0090] Within said aspect, it is also envisaged in the context of the present invention to provide an antigen binding molecule according to claim 1, wherein the first target cell surface antigen and the second target cell surface antigen are respectively 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.

[0091] Within said aspect, in the context of the present invention it is also envisaged to provide an antigen binding molecule according to claim 1, wherein the first target cell surface antigen and / or the second target cell surface antigen is human MSLN (selected from SEQ ID NO: 1181, 1182 and 1183), and wherein the first and / or third binding domains of the antigen binding molecule of the present invention bind to human MSLN epitope cluster E1 (SEQ ID NO: 1175, according to aa296-346 positions of Kabat) as determined by mouse chimeric sequence analysis as described herein, but preferably do not bind to human MSLN epitope cluster E2 (SEQ ID NO: 1176, according to aa 247-384 positions of Kabat), E3 (SEQ ID NO: 1177, according to aa 385-453 positions of Kabat), E4 (SEQ ID NO: 1178, according to aa 454-501 positions of Kabat) and / or E5 (SEQ ID NO: 1179, according to aa 506-511 positions of Kabat). 502-545 positions).

[0092] Within said aspect, it is also envisaged in the context of the present invention to provide an antigen binding molecule according to 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 wherein the first and / or third binding domain of the antigen binding molecule according to claim 1 binds to human CDH3 epitope clusters D2B (SEQ ID NO: 1171, aa 253-290 positions according to Kabat), D2C (SEQ ID NO: 1172, aa 291-327 positions according to Kabat), D3A (SEQ ID NO: 1173, aa 328-363 positions according to Kabat) and D4B (SEQ ID NO: 1174, aa 476-511 positions according to Kabat), preferably D4B (SEQ ID NO: 1174, aa 476-511 positions according to Kabat), as determined by murine chimeric sequence analysis as described herein.

[0093] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the second and fourth binding domains (CD3 binding domain) both have (i.) an affinity characterized by a KD value of less than about 1.2 x 10-8 M as measured by surface plasmon resonance (SPR), or (ii.) an affinity characterized by a KD value of about 1.2 x 10-8 M as measured by SPR.

[0094] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the second and fourth binding domains (CD3 binding domains) have an affinity characterized by a KD value of about 1.0 x 10-7 to 5.0 x 10-6 M measured by SPR, preferably about 1.0 to 3.0 x 10-6 M, more preferably about 2.5 x 10-6 M measured by SPR.

[0095] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the second and fourth binding domains (CD3 binding domains) have an affinity characterized by a KD value of about 1.0 x 10-7 to 5.0 x 10-6 M measured by SPR, preferably about 1.0 to 3.0 x 10-6 M, more preferably about 2.5 x 10-6 M measured by SPR.

[0096] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the second and fourth binding domains (CD3 binding domains) each individually have at least about 10 times lower activity, preferably at least about 50 times or more preferably at least about 100 times lower activity than one CD3 binding domain comprising VH according to SEQ ID NO 43 and VL according to SEQ ID NO 44 (i.e. in a single targeting environment compared to a dual targeting environment).

[0097] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the second and fourth domains are effector binding domains that bind to the CD3 epsilon chain, these effector binding domains comprising or consisting of a VH region linked to a VL region, wherein

[0098] i) the VH region comprises:

[0099] The CDR-H1 sequence of X1YAX2N, wherein X1 is K, V, S, G, R, T or I; and X2 is M or I;

[0100] The CDR-H2 sequence of RIRSKYNNYATYYADX1VKX2, wherein X1 is S or Q; and X2 is D, G, K, S or E; and

[0101] A CDR-H3 sequence of HX1NFGNSYX2SX3X4AY, wherein X1 is G, R or A; X2 is I, L, V or T; X3 is Y, W or F; and X4 is W, F or Y; and

[0102] ii) wherein the VL region comprises:

[0103] CDR-L1 sequence of X1SSTGAVTX2X3X4YX5N, wherein X1 is G, R or A; X2 is S or T; X3 is G or S; X4 is N or Y; and X5 is P or A;

[0104] X1TX2X3X4X5X6 CDR-L2 sequence; wherein 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; and X6 is P or S; and

[0105] A CDR-L3 sequence of X1LWYSNX2WV, wherein X1 is V, A or T; and X2 is R or L; and

[0106] iii) wherein one or more of the CDR sequences of the VH region of i) and / or the VL region of ii) comprises an amino acid substitution or a combination thereof selected from the group consisting of: X24V and X24F in CDR-H1;

[0107] D15 and X116A in CDR-H2;

[0108] H1, X12E, F4, and N6 in CDR-H3; and

[0109] X11L and W3 in CDR-L3.

[0110] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the second and fourth binding domains comprise a VH region comprising a CDR-H1, CDR-H2 and CDR-H3 selected from the group consisting of SEQ ID NOs 37 to 39, 45 to 47, 53 to 55, 61 to 63, 69 to 71, 436 to 438, 1126 to 1128, 1136 to 1138, 1142 to 1144, 1148 to 1150 and 1217 to 1219, and a VL region comprising a CDR-H1, CDR-H2 and CDR-H3 selected from the group consisting of SEQ ID NOs CDR-L1, CDR-L2 and CDR-L3 of 40 to 42, 48 to 50, 56 to 58, 64 to 66, 72 to 74, 439 to 441, 1129 to 1131, 1139 to 1141, 1145 to 1147, 1151 to 1153, and 1220 to 1222, preferably 61 to 63 and 64 to 66 or 1217 to 1219 and 1220 to 1222.

[0111] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the second and fourth binding domains comprise a VH region selected from SEQ ID NOs 43, 51, 59, 67, 75, 442, 1132 and 1223, preferably 67 or 1223.

[0112] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the second and fourth binding domains comprise a VL region selected from SEQ ID NO 44, 52, 60, 68, 76, 443, 1133 and 1224, preferably 68 or 1224.

[0113] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule, wherein the second and fourth binding domains comprise a VH region selected from the group consisting of SEQ ID NOs 43, 51, 59, 67, 75, 442, 1132 and 1223, preferably 67, and a VL region selected from the group consisting of SEQ ID NOs 44, 52, 60, 68, 76, 443, 1133 and 1224, preferably 68, wherein when the VH region is 1132 and the VL region is 1133, the second binding domain and / or the fourth binding domain as a scFab domain additionally comprises a CH1 domain of SEQ ID NO: 1134 and a CLK domain of SEQ ID NO: 1135, and wherein the VH region and the VL region are connected to each other via a linker preferably selected from the group consisting of SEQ ID NOs 1, 3 and 1125, or wherein the SEQ ID NO of the VH of the VH-CH1 of the second domain and the fourth domain is SEQ ID NO: 1223, the SEQ ID NO of CH1 of VH-CH1 of the second domain and the fourth domain is SEQ ID NO 1224, the SEQ ID NO of VL of VL-CL of the second domain and the fourth domain is SEQ ID NO 1225, and the SEQ ID NO of CL of VL-L of the second domain and the fourth domain is SEQ ID NO1226.Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule, wherein the first and / or third (target) binding domain binds CDH3 and comprises a VH region, the VH region comprising SEQ ID NO: 1154 as CDR-H1, wherein X1 (the number following "X" represents the numerical order of "X" in each amino acid sequence in N- to C-orientation in the sequence listing) is S or N, X2 is Y or S, X3 is P or W, X4 is I or M and X5 is Y, N or H; SEQ ID NO: NO: 1155 as CDR-H2, wherein 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; and X13 is K or Q; and SEQ ID NO:1156 as CDR-H3, wherein 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 wherein the first and / or the third (target) binding domain binds CDH3 and comprises a VL region comprising SEQ ID NO:1158 as CDR-L1, wherein 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; and X9 is A, N or H; SEQ ID NO:1159 as CDR-L2, wherein 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; and X7 is S or E; and SEQ ID NO:1160 as CDR-L3, wherein 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; and X7 is F, Y or L.

[0114] Within said aspects, it is also contemplated in the context of the present invention to provide an antigen binding molecule, wherein the first and / or third (target) binding domain binds to MSLN and comprises a VH region comprising SEQ ID NO: 1162 as CDR-H1, wherein X1 (the number following "X" represents the numerical order of "X" in each amino acid sequence in the N to C orientation of the sequence listing) is S, G or D; X2 is Y, A, G or F; X3 is I, W or M; and X4 is V, S, G, T or H; SEQ ID NO: NO: 1163 as CDR-H2, wherein 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; and X15 is G or S; and SEQ ID NO: 1164 as CDR-H3, wherein 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; and X16 is Y, I or V; and wherein the first and / or the third (target) binding domain binds MSLN and comprises a VL region comprising SEQ ID NO: 1166 as CDR-L1, wherein 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; and X11 is Y or A; SEQ ID NO: 1167 as CDR-L2, wherein 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; and X7 is S or T; and SEQ ID NO: 1168 as CDR-L3, wherein 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; and X8 is V or T.

[0115] Within said aspects, it is also contemplated in the context of the present invention to provide an antigen binding molecule wherein the first and / or third (target) binding domain binds CDH3 and comprises a VH region of SEQ ID NO: 1157, wherein (the number following "X" represents the numerical order of "X" in each amino acid sequence in the N to C orientation in 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 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, or 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; and the VL region of SEQ ID NO 1161, wherein X1 is D or E; X2 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; and X49 is D or E (where all aa in each position are meant to be in the alternative "or" even if not explicitly stated). ;

[0116] Within said aspects, it is also contemplated in the context of the present invention to provide an antigen binding molecule wherein the first and / or third (target) binding domain binds to MSLN and comprises the VH region of SEQ ID NO: 1165, wherein (the number following "X" represents the numerical order of "X" in each amino acid sequence in the N to C orientation in 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 E, A, G, R; X21 is S, T; X22 is V, L; X23 is R, S, K; X24 is V, L; X25 is S, T; X26 is A, K, T; X27 is A, V; X28 is Y, I, F; X29 is S, T; X30 is Y, I, F; X31 is S, T; X32 is A, K, T; X33 is A, V; X34 is Y, I, F; X35 is S, T; X36 is A, K, T; X37 is A, V; X38 is Y, I, F; X39 is S, T; X40 is A, 0 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; and X82 is L, M, T; and SEQ The VL region of ID NO 1169 (the number after "X" indicates the numerical order of "X" in each amino acid sequence in the N to C orientation in the sequence list) 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; X 22 is R, G, S; X23 is D, S, -; X24 is A, S, N, K, T; X25 is Y, WM; 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 is Y, S, A, T; X54 is G, E, Y, H, Q; X55 is A, S; X56 is S, T, F; X57 is P, T; X58 is R, A, L, F; X59 is V, T; X60 is P, C; X61 is V, L; X62 is E, T; X63 is I, V; and X64 is L, K (where all aa in each position are meant to be in the alternative "or", even if not explicitly stated). ;

[0117] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the first and / or third (target) binding domain comprises a VH region comprising a VH region selected from the group consisting of SEQ ID NO: 77 to 79, 86 to 88, 95 to 97, 103 to 105, 111 to 113, 119 to 121, 127 to 129, 135 to 137, 143 to 145, 151 to 153, 159 to 161, 168 to 170, 177 to 179, 185 to 187, 194 to 196, 203 to 205, 212 to 214, 221 to 223, 230 to 232, 238 to 239, 40, 334-336, 356-358, 365-367, 376-378, 385-387, and 194, 432 and 196 CDR-H1, CDR-H2 and CDR-H3, or any combination of CDR-H1, CDR-H2 and CDR-H3 disclosed together in Table 6 of the Sequence Tabulation, preferably 86 to 88, 194 to 196 or 1227 to 1229 and 1237 to 1239.

[0118] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the first and / or third (target) binding domain comprises a VL region comprising a VL region selected from the group consisting of SEQ ID NO: 80 to 82, 89 to 91, 98 to 100, 106 to 108, 114 to 116, 122 to 124, 130 to 132, 138 to 140, 146 to 148, 154 to 156, 162 to 164, 171 to 173, 180 to 182, 188 to 190, 197 to 199, 206 to 208, 215 to 217, 224 to 226, 23 3 to 235, 241 to 243, 337 to 339, 359 to 361, 368 to 370, 379 to 381, 388 to 390, or any combination of CDR-H1, CDR-H2 and CDR-H3 disclosed together in Table 6 of the Sequence Tabulation, preferably 89 to 91 and 197 to 199 or 1230 to 1232 and 1240 1242.

[0119] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the first and / or third (target) binding domain comprises a VH region selected from SEQ ID NO: 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 of Fab) for the first and third binding domain, respectively.

[0120] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the first and / or 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 200 and 93 or 1234+1236 and 1244+1246 (VL and CL in Fab) for the first and third binding domain, respectively.

[0121] Within said aspects, it is also envisaged in the context of the present invention to provide an antigen binding molecule wherein the first and / or third (target) binding domain comprises a VL region selected from SEQ ID NO: 85, 94, 193, 202, 211, 220, 229, 364, 384, 393, preferably 94 and 202, whose stability is increased by a single amino acid exchange (E to I).

[0122] Within said aspects, it is also envisaged in the context of the present invention 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.

[0123] In a second aspect, it is further envisaged in the context of the present invention to provide a polynucleotide encoding an antigen binding molecule of the present invention, preferably selected from SEQ ID NO: 1070 to 1072 and 1074.

[0124] In a third aspect, it is also envisaged in the context of the present invention to provide a vector comprising a polynucleotide of the present invention.

[0125] In a fourth aspect, it is further envisaged in the context of the present invention to provide a host cell transformed or transfected with a polynucleotide or vector of the present invention.

[0126] In a fifth aspect, it is also envisaged in the context of the present invention to provide a method for producing an antigen binding molecule of the present invention, the method comprising culturing the host cell of the present invention under conditions allowing expression of the antigen binding molecule and recovering the produced antigen binding molecule from the culture.

[0127] In a sixth aspect, it is further envisaged in the context of the present invention to provide a pharmaceutical composition comprising an antigen binding molecule of the present invention or an antigen binding molecule produced according to a method of the present invention.

[0128] Within said aspects, it is also envisaged in the context of the present invention that the pharmaceutical composition is stable at about -20°C for at least four weeks.

[0129] It is further envisaged in the context of the present invention to provide an antigen binding molecule of the present invention or an antigen binding molecule produced according to the method of the present invention for use in preventing, treating or ameliorating a disease selected from a proliferative disease, a neoplastic disease, a cancer or an immunological disorder.

[0130] Within said aspects, it is also envisaged in the context of the present invention that the disease is preferably acute myeloid leukemia (AML), non-Hodgkin's lymphoma (NHL), non-small cell lung cancer (NSCLC), pancreatic cancer and colorectal cancer (CRC)]. In a seventh aspect, it is further envisaged in the context of the present invention to provide a method for treating or ameliorating a proliferative disease, the method comprising administering to a subject in need thereof a molecule comprising at least one polypeptide chain, wherein the molecule comprises

[0131] (i.) a first binding domain, which preferably comprises a paratope that specifically binds to a first target cell surface antigen (e.g., TAA1),

[0132] (ii.) a second binding domain, which preferably comprises a paratope that specifically binds to an extracellular epitope of the human (preferably macaque) CD3 epsilon chain,

[0133] (iii.) a third binding domain, which preferably comprises a paratope that specifically binds to a second target cell surface antigen (e.g., TAA2), and

[0134] (iv.) a fourth binding domain, which preferably comprises a paratope that specifically binds to an extracellular epitope of the human (preferably macaque) CD3 epsilon chain,

[0135] wherein 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, and

[0136] wherein the molecule comprises a spacer entity having a molecular weight of at least about greater than about 5 kDa and / or having a length of more than 50 amino acids, wherein the spacer entity separates the first bispecific entity from the second bispecific entity by at least about (the distance between the centers of mass of the first bispecific entity and the second bispecific entity), and the spacer entity is located between the first bispecific entity and the second bispecific entity.

[0137] Within said aspects, it is also contemplated in the context of the present invention to provide a method for addressing disease-associated targets that are significantly co-expressed on pathological physiological tissues and one or more physiological tissues by providing a multi-chain multi-targeting bispecific antigen binding molecule of the form described herein, wherein the molecule addresses (i.) a target expressed on both disease-associated tissues and physiological tissues and (ii.) another target expressed on physiological tissues that are associated with the disease but not under (i.), wherein the method preferably avoids the formation of intra-abdominal adhesions and / or fibrosis when such target is MSLN.

[0138] Within said aspects, it is also envisaged in the context of the present invention that the disease is preferably a neoplastic disease, cancer or an immunological disorder, comprising the step of administering to a subject in need thereof an antigen binding molecule of the present invention or produced according to the method of the present invention, wherein the disease is preferably acute myeloid leukemia, non-Hodgkin's lymphoma, non-small cell lung cancer, pancreatic cancer and / or colorectal cancer.

[0139] Within said aspects, it is also envisaged in the context of the present invention that TAA1 and TAA2 are preferably selected from the group consisting of EpCAM and MSLN, MSLN and EpCAM, MSLN and CDH3, CDH3 and MSLN, FLT3 and CLL1 and CLL1 and FLT3.

[0140] In an eighth aspect, it is also envisaged in the context of the present invention to provide a kit comprising an antigen binding molecule of the present invention or an antigen binding molecule produced according to the method 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

[0141] Figure 1 : Overview of the multi-chain multi-targeted bispecific antigen binding molecules disclosed in the present invention comprising a dimerization domain as a spacer. The black domain is the anti-CD3 domain, and the striped and dotted domains are the target binding domains (see Figure 2 ). The domains in each molecule are arranged as follows: A: the target binding domain and the CD3 binding domain in scFv form, both at the N-terminus and C-terminus, respectively, of the dimerization spacer domain ("cis" orientation, i.e., at least two target binding domains and at least two CD3 binding domains are respectively attached to the same monomer of the dimerization spacer, or in other words, on the same side of the vertical plane defined by the dimerization spacer that cuts the dimerization spacer); B: the first target binding domain is a Fab, the CD3 binding domain and the other target binding domain are in scFv form, both at the N-terminus and C-terminus, respectively, of the spacer dimerization domain (trans orientation, i.e., at least two target binding domains and at least two CD3 binding domains are respectively attached to the same monomer of the dimerization spacer, or in other words, on the same side of the vertical plane defined by the dimerization spacer that cuts the dimerization spacer). D3 binding domains are respectively connected to opposite monomers of the dimerization spacer, or in other words, on opposite sides of the vertical plane defined by the dimerization spacer that cuts the dimerization spacer); C: at least two target binding domains are Fab, and the CD3 binding domain is in the form of scFv, both at the N-terminus and C-terminus of the spacer dimerization domain, respectively (trans); D: at least two target binding domains are Fab, both at the N-terminus of the dimerization spacer domain, and the CD3 binding domain is in the form of scFv; E: target binding domain Fab, at the N-terminus of the spacer dimerization domain, and the CD3 binding domain is Fab, at the C-terminus of the spacer dimerization domain.

[0142] Figure 2 : Figure 2 A to M show examples of multi-chain multi-targeting bispecific antigen-binding molecules of the present invention, wherein the spacer is a dimerized hetero-Fc, and wherein at least three of the four N-termini and C-termini are linked to the target and / or CD3 binding domains, respectively.

[0143] Figure 3 : Figure 3 A to H show examples of multi-chain multi-targeting bispecific antigen-binding molecules of the present invention, wherein the spacer is a dimerized hetero-Fc, and wherein one N-terminus and one C-terminus of the four N-termini and C-termini of the hetero-Fc are connected to the target and / or CD3 binding domains, respectively (A, C, E to G), or wherein the spacer is a scFv (B, D and H).

[0144] Figure 4 : Figure 4 A to I show the cytotoxicity curves of CDH3 T cell engager molecule 1, MSLN T cell engager molecule 1 and MSLN-CDH3 T cell engager molecule 1-7 against parental double positive HCT116 WT cells relative to target knockout HCT116 cells. Effector cells are unstimulated pan T cells.

[0145] Figure 5 : Figure 5 Cytotoxicity curves of CDH3 T cell engager molecule 1, MSLN T cell engager molecule 1, and MSLN-CDH3 T cell engager molecule 1-7 against parental double positive GSU WT cells versus target knockout GSU cells are shown A to I. Effector cells are unstimulated pan T cells.

[0146] Figure 6 : Figure 6 A to C show cytotoxicity curves of single CDH3 T cell engager molecule 1 (A), MSLN T cell engager molecule 1 (B), and MSLN-CDH3 T cell engager molecule 24 comprising two target binders in Fab format against parental double positive GSU WT cells versus target knockout GSU cells. Effector cells are unstimulated pan T cells. DETAILED DESCRIPTION

[0147] In the context of the present invention, a multi-chain multi-targeting bispecific molecule comprising at least five different structural entities is provided, 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 (preferably non-human, such as macaque) CD3ε chain, wherein the first binding domain and the second binding domain together form a first bispecific entity, (iii.) a spacer that connects but separates the first bispecific entity from the second bispecific entity, the second bispecific entity comprising (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 (preferably non-human, such as macaque) CD3ε chain. The molecules of the present invention format typically exhibit advantages characterized by potency and specificity driven by affinity from two targets co-expressed on target cells, which typically results in a reduction in undesirable cytokine release (and associated clinically relevant side effects, such as CRS) while ensuring effective anti-tumor activity, preferably also in solid tumors such as colorectal cancer, non-small cell lung cancer and pancreatic cancer.

[0148] In the context of the present invention, it was surprisingly found that the bispecific (T cell engagement) multi-chain multi-targeting (antigen binding) molecules according to the present invention provide dual avidity effects on the target cell binder and effector cell binder sides due to their specific forms that lead to effective complementary target cell killing. This effect is promoted by the following: molecular forms that specifically target two (different) antigens on a target cell (e.g., a cancer cell) and, by contrast, simultaneously mediate an effective T cell response against the target cell by significantly less targeting of non-target cells. Due to the ability to address two target antigens simultaneously, when two TAAs that are usually associated with disease-related target cells are selected, the possibility of targeting disease-related target cells rather than physiological cells is greatly increased. Therefore, relative to existing bispecific antibodies or antibody-derived constructs that are T cell-engaged, the T cell-engaging multi-chain multi-targeting molecules according to the present invention provide improved efficacy and safety at the same time. The advantageous properties are preferably achieved by the fact that the multi-chain multi-targeting bispecific molecules of the present invention comprise two bispecific entities, each comprising a target binding domain and an effector (CD3) binding domain, which can act in a pathophysiological environment without (e.g., spatially) hindering each other while complementing each other. The interaction of the two bispecific entities within one multi-chain multi-targeting bispecific molecule of the present 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 can interact with their respective binding partners to form a cytolytic synapse between target cells and T cells without interfering 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 effect and thus provide therapeutic function, preferably, the two target binding domains of the first and second bispecific entities must engage their respective targets so that the effector CD3 binding domains of the first and second bispecific entities are fully engaged. In addition, it was unexpectedly found that each of the two bispecific entities must retain function 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 efficacy and implied safety described herein. Particularly surprising is that the two bispecific entities comprising a target binding domain and a CD3 binding domain, respectively, do not need to be on one chain located at the N-terminus and C-terminus of the (central) spacer to be structurally positioned to function as described herein. The target and / or CD3 binder of one or both bispecific entities may be a Fab, i.e. comprising a to chain, respectively. Even more surprisingly, the spacer may also be double-chained, preferably in iso-Fc format.In this case, surprisingly, the same is true for the bispecific entity when the two domains are not on the same chain but are held close together by a four-part spacer (e.g., an iso-Fc), which simultaneously holds the two domains of each bispecific entity in place to act together and separates the two bispecific entities from each other to act without interfering with each other. This arrangement of spacers and domains and the use of two different TAA binding domains and two CD3 binding domains of preferably low affinity (preferably binding to CD3ε) produce a surprising technical effect, namely increasing target cell selectivity and reducing the risk of major side effects (i.e., undesirable cytokine release). At the same time, the multi-chain molecules of the present invention can be produced better in terms of yield and purity.

[0149] As an additional or alternative secondary effect of the increased specificity and thus safety described herein, once target cells (such as cancer cells) have experienced antigen loss and are therefore easily escaping from effective anti-tumor therapy, the probability of multi-chain multi-targeting antigen binding molecules targeting such target cells is greatly increased compared to single targeting molecules, because an effective antigen against the target is retained on cells that have experienced antigen escape. Compared to molecules that contain only one CD3 binder and / or target binder and do not contain two connected but spaced-apart bispecific entities, the effect in terms of increased activity is preferably achieved when both CD3 binders (CD3 binding domains comprising VH and VL of, for example, SEQ ID NOs 67 and 68, respectively, connected by a linker of SEQ ID NO 1 or 3) have low affinity.

[0150] In view of the teachings of the prior art, the above findings based on the present invention are unexpected. For example, antigen binding formats comprising more than one target binding domain and effector binding domain, respectively, are known in the art, such as Adaptir TM However, this format does not provide two bispecific entities that interact with their respective targets and effectors separately and work synergistically at the same time, and therefore cannot achieve dual avidity on the target binder and effector binder sides to effectively provide a large selectivity gap effect for the advantages of multi-chain multi-targeting molecules. According to the present invention, the two bispecific entities must be separated from each other by a certain distance, preferably at least More preferably at least 60, 70, 80, 90 or at least Indicated distance between two bispecific entities In the context of the present invention, it is generally understood as the distance between the center of mass of two bispecific entities, respectively. In general, the center of mass (COM) of the spatial mass distribution (referring to a bispecific entity herein, which comprises a binding domain bound to a target cell surface antigen and a binding domain bound to an extracellular epitope of a human (preferably macaque) CD3ε chain, both binding domains are preferably in the form of Fab or a single domain, preferably selected from scFv and scFab forms and connected by a peptide linker) is understood as the unique point at which the weighted relative position sum of the distribution mass is zero. The distance is generally determined by molecular modeling using a generally accepted modeling program (MD / visualization software), which can identify the COM of a given input structure and is, for example, PyMOL (PyMOL Molecular Graphics System, Version 2.3.3, Schrodinger Company ( LLC.), which is usually based on a collection of snapshot structures from MD simulations. The mass of each atom is usually part of a potential "force field" known in the art. Alternatively and / or additionally, the distance can be determined by crystallography, cryo-electron microscopy or nuclear magnetic resonance analysis techniques.

[0151] The typical method for obtaining distances by molecular modeling provided by the present invention is as follows:

[0152] 1) Obtaining the atomic structure of the complete bispecific antigen binding molecule. The structural source can be selected from the group consisting of:

[0153] a. The resolution is preferably lower than X-ray crystallography of proteins, which enables visualization of the amino acid backbone and side chains;

[0154] b. The resolution is preferably lower than Cryo-electron microscopy (cyo-EM), which enables visualization of amino acid backbones and side chains;

[0155] c. Computer homology modeling of the entire molecule based on a single, highly homologous crystal and / or cro-EM structure (preferably greater than 60% sequence identity);

[0156] d. Involving computer simulation homology modeling connecting 2 or more experimental structures. The structure is preferably identical or highly homologous to the domain found in the complete bispecific antigen binding molecule (preferably higher than 60% sequence identity). In the absence of experimental joint conformations, the model is preferably improved in a clear solvent molecular dynamics (MD) simulation (the simulation length is preferably at least 100ns, unless energy convergence is obtained faster). The simulation is performed using the most advanced software (e.g., Schrodinger, Amber, Gromacs, NAMD or equivalent software), wherein the parameters correspond to room temperature and pressure. No artificial force is applied during the simulation (i.e., preferably excluding methods such as quasi-dynamics or guided molecular dynamics). Similarly, artificial geometric restrictions are preferably not applied to the molecule.

[0157] 2) Identify the center of mass (COM) of the relevant molecular domains. This is usually performed using the MD software or visualization tool used (such as PyMOL or VMD). The center of mass can be defined as the pseudo-atom or non-hydrogen atom that is closest to the true COM. Inter-domain linkers are usually not considered part of the domain.

[0158] 3) Using the same software, report the distance between the two COMs (in As a unit, ). If MD simulations were used to refine homology models (as described in 1d), median distances from multiple simulation snapshots are reported. To further reduce potential inaccuracies in the initial model, at least the first 10% of the simulations were omitted when calculating median distances between COMs and when extracting snapshots for visualizing MD simulations, and preferably up to 50% if the signal changes significantly.

[0159] If not otherwise indicated, the distance in the context of the present invention is is the median distance determined by MD simulations.

[0160] As disclosed herein, the preferred distance between the first and second bispecific entities is promoted by a spacer entity (referred to as spacer) between the two bispecific entities, which separates the two bispecific entities and keeps them in a separate position. The spacer has a certain size, preferably at least greater than 5kDa, more preferably at least about 10, 15, 20, 25, 30, 35, 40, 45 or even at least 50kDa, so as to prevent the undesirable interaction of the two separated bispecific entities. The preferred range of the molecular size of the spacer is about 15 to 200kDa, preferably about 15 to 150kDa, to promote the separation of the two bispecific entities according to the present invention and to maintain the high overall activity of the molecule. Generally, an overly large spacer, for example, greater than about 200kDa, may affect the ability of the two bispecific entities to bind to the two target surface structures on the same target cell, which in turn may reduce the overall activity of the molecule to the target cell. Therefore, with respect to the molecular weight of the spacer, the typical maximum preferred size is about 200kDa, preferably about 150 or 120kDa, and even more preferably about 100kDa. A typical spacer of the largest preferred size is a dual scFc domain as disclosed herein (two scFcs linked to each other to form a larger single chain spacer) of approximately 105.7 kDa.Example sizes of spacers that generally adequately separate the two bispecific entities are the PSI domain of the Met receptor at about 5.3 kDa, ubiquitin at about 8.6 kDa, the fibronectin type III domain from tenascin at about 10.1 kDa, the SAND domain at about 11 kDa, β2 microglobulin at about 11.9 kDa, Tim-3 (aa) at about 12.2 kDa. 24-130), MiniSOG of about 13.3 kDa, SpyCatcher of about 12.1 kDa associated with SpyTag of about 1.7 kDa, preferably linked together by isopeptide bond formation to form a two-chain spacer of about 13.8 kDa, VHH antibody lama 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; T NFSF9) extracellular domain, programmed cell death protein 1 (PD-1) of about 16.6 kDa, green fluorescent protein (GFP) of about 26.3 kDa, single chain Fc region (scFc) as described herein of about 52.8 kDa (about 54.6 kDa with N and C terminal linkers (G4S) 3, respectively), human serum albumin (HSA) of about 66.5 kDa (about 68.3 kDa with N and C terminal linkers (G4S) 3, respectively), and double scFc (two scFcs connected to each other to form a larger single chain spacer) of about 105.7 kDa (about 107.5 kDa with N and C terminal linkers (G4S) 3, respectively). Generally, the more rigid the spacer, the smaller the required intermediate distance, otherwise the intermediate distance must include a safety margin for a flexible spacer.

[0161] Furthermore, preferred spacers in the context of the present invention, such as globular domains, typically have N-terminal and C-terminal ends that are not too close to each other in space, so as to effectively separate the two bispecific entities according to the present invention. In this regard, the spacer typically exhibits a distance between the N-terminal and C-terminal ends that is significantly greater than The distance between the N-terminus and the C-terminus of the spacer is less than or about Thus, the spacer in the context of the present invention is preferably at least 500 nm away from the α-carbon atom of the first amino acid at the N-terminus and the last amino acid at the C-terminus. More preferably at least Even more preferably at least This distance generally ensures that the first and second bispecific entities are separated by at least As described herein. α-Carbon (α-carbon) is understood herein as a term applicable to proteins and amino acids. It is the backbone carbon preceding the carbonyl carbon atom in a molecule. Thus, reading along the backbone of a typical protein would give a sequence such as -[N-Cα-carbonylC]n- (when reading in the N to C direction). The α-carbon is where a different substituent is attached to each different amino acid. That is, the groups hanging from the α-carbon chain give the amino acid diversity. Thus, in the context of the present invention, even if the size of the spacer is at least 5 kDa and the length exceeds 50 aa, if the distance between the α-carbon atom of the first amino acid at the N-terminus and the last amino acid at the C-terminus is too close, that is, if it is only about Such spacers are less preferred. For example, preferred spacers show typical distances between the α-carbon atoms of the first amino acid at the N-terminus and the last amino acid at the C-terminus as follows: scFc (based on 5G4S crystal structure) HSA (based on 5VNW crystal structure): Ubiquitin (based on 1UBQ crystal structure): And SAND (based on 1OQJ crystal structure): In contrast, HSP70-1 (based on the 3JXU crystal structure) shows only a 1:1 distance between the α-carbon atom of the first amino acid at the N-terminus and the last amino acid at the C-terminus. Meanwhile, HSP70-1 in the context of the present invention only provides about The median distance between the COM of the first and second bispecific entities is lower than The median distance threshold is set to be significantly lower than the typical median distance between the COMs of the two bispecific entities. The median distance of (this is facilitated by preferred spacers such as scFc, HSA, ubiquitin and SAND). Among them, scFc (SEQ INNO: 25) is preferred.

[0162] Alternatively, in the context of the present invention, non-globular but rigid linkers can be used as spacers, which separate the two bispecific entities. Such linkers include (PA)25P (SEQ ID NO: 1097) and A(EAAAK)4ALEA(EAAAK)4A (SEQ ID NO: 1096), even though the Mw is lower than 5 kDa (here 4.3 kDa) and the amino acid length is only about or lower than 50 (51 and 46 aa, respectively). However, such spacers are generally less preferred than globular domains, which are preferred to additionally increase half-life.

[0163] As also considered in the context of the present invention, the spacer between the two bispecific entities is a polypeptide generally 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 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 separation of the two bispecific entities according to the present invention. This is to preferably maintain a high overall activity of the entire molecule according to the present invention (not necessarily a single and spaced bispecific entity, which may have a low affinity (and low activity) alone to increase specificity for double positive target cells)), which is generally less than 20pM, preferably less than 5pM, more preferably less than 1pM. Generally, an overly large spacer, for example longer than about 1500 amino acids, may affect the ability of the two bispecific entities to bind to two target surface structures on the same target cell, which in turn may reduce the overall activity of the molecule against the target cell. 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 spacers that adequately separate the two bispecific entities are PD-1 of about (ECD 25-167) 143 aa, scFc of about 484 aa (about 514 aa, with N-terminal and C-terminal linkers (G4S) 3, respectively) as described herein, HSA of about 585 aa (about 615 aa, with N-terminal and C-terminal linkers (G4S) 3, respectively), and bi-scFc of about 968 aa (about 998 aa, with N-terminal and C-terminal linkers (G4S) 3, respectively). Other spacers include ubiquitin of about 76 aa, fibronectin type III domain from tenascin of about 90 aa, SAND domains of about 90 or 97 aa, β2 microglobulin of about 100 aa, Tim-3 (aa 24-130), MiniSOG of about 115 aa, SpyCatcher of about 113 aa associated with SpyTag of about 14 aa, preferably connected together by isopeptide bond formation to form a two-chain spacer of about 127, a VHH antibody lama domain of about 129 aa, a PD-1 binding domain from human programmed cell death 1 ligand (PDL1) of about 126 aa, granulocyte-macrophage colony stimulating factor (GM-CSF) of about 127 aa, interleukin-4 of about 129 aa, interleukin-2 of about 133 aa, CD137L (4-1BBL; TNFSF9) extracellular domain of about 167 aa, and green fluorescent protein (GFP) of about 238 aa.

[0164] The composition and arrangement of the amino acid sequences of the preferred spacers preferably confer a certain rigidity and are not characterized by high flexibility. Rigidity in the context of the present invention is generally present when a spacer of greater than 50 aa and / or a molecular weight exceeding 5 kDa promotes a maximum distance between the centers of mass of the two bispecific entities in a molecule according to the invention that is less than 200% (or 2 times) of the median distance. Thus, a preferred rigid spacer in the context of the present invention extends no more than about 100%, more preferably no more than about 80% of its median length (each calculated as the distance between the centers of mass of the two bispecific entities). Thus, in the context of the present invention, the two bispecific entities are spaced apart by about The preferred rigid spacer (median distance) extends no more than (maximum distance). For example, a typical median distance between the centers of mass of a bispecific entity having a molecule of the invention format comprising a scFc (e.g., SEQ ID NO: 25) as a spacer is about However, the maximum distance in this case is usually about That is, not more than about 100% or even only about 80% of the median distance. In the context of the present invention, such a spacer is considered to be rigid. In contrast, a spacer comprising (G4S) 10 (SEQ ID NO: 8) as a spacer molecule (which is a linear polypeptide without, for example, a globular structure) shows a typical median distance of about The maximum distance is about Therefore, a (G4S) 10Spacers such as show high flexibility rather than rigidity as preferred spacers according to the advantageous features of the present invention. In this regard, the spacer amino acid sequence is generally rich in proline and less rich in serine and glycine. It is particularly envisioned that as a spacer of a folded polypeptide, such as a secondary fold (e.g., a helical structure) or a tertiary fold forms, for example, a three-dimensional protein domain structure, a certain rigidity is ensured by their structure, and further advantageous effects are preferably imparted, such as an in vivo half-life extension of a multi-chain multi-targeted bispecific molecule as a therapeutic agent. Typical domain structures include a hydrophobic core with a hydrophilic surface. In the context of the present invention, proteins with a globular protein structure are preferably used as spacers. In the context of the present invention, globular proteins are understood to be globular ("globular") proteins and are one of the common protein types. Globular proteins in the context of the present invention are characterized by globular protein folding. It is particularly envisioned that a spacer comprises an Fc domain or a portion thereof or multiple, PD-1 or HSA domain. It is also envisioned that a spacer comprising a combination of different globular proteins or portions thereof, which even more preferably comprises an Fc receptor binding function to increase the half-life of a molecule according to the present invention. The formats described herein wherein the two bispecific entities are separated have unique advantages. If there is only one target for the first binding domain to address, then the first domain "uses" only the second domain to engage T cells but not the fourth domain, or alternatively, the third domain uses the fourth domain but not the second domain (or uses it to a lesser extent due to a spacer). If there is only one target, the Kd of the preferred low-affinity CD3 binders disclosed herein prevents effective T cell engagement. Thus, selectivity is increased relative to other (dual) targeting molecules.

[0165] If both targets are present, the multi-chain multi-targeting bispecific T cell engager of the invention will bind more firmly to the target cell (via avidity gain) and both low affinity CD3 binding domains of the invention (such as I2L) can be used to engage T cells (also via avidity gain), e.g., the second domain binds to the CD3 domain on the effector T cell, and the third domain is less likely to bind to the target antigen to form a cytolytic synapse and therefore does not act together as a bispecific entity, which would otherwise result in a less beneficial spectrum of cytotoxic activity. This has the advantage that the first and fourth domains are not left "useless", which would mean that the full effect of the dual affinity generated by the dual binding of the target and effector binding domains, respectively, cannot be utilized. Similarly, the first domain that binds to the target antigen and the fourth domain that binds to the CD3 domain on the effector T cell are prevented from theoretically interacting, which ultimately prevents the second and third domains from forming a cytolytic synapse with the intended "partner domain" in their respective bispecific entities.

[0166] Typically, the favorable avidity effect conferred by the multi-chain multi-targeting bispecific molecules according to the invention is indicated by a differential activity factor or "selectivity gap" between the activity of the molecule on double-positive cells (i.e., target cells carrying (i.) two different targets, the combination of which is overexpressed on the cell type to be targeted and associated with a specific disease, and / or (ii.) one target at an overexpressed level). In either case, the molecule according to the invention that simultaneously targets two (preferably different) targets will preferentially bind to such target cells, and will therefore induce a more pronounced T cell response, compared to non-target cells that express only one of the two targets or one target at a low level. As is preferred for the multi-chain multi-targeting bispecific molecules of the invention, for example, by a lower EC 50 The activity in terms of increased cytotoxicity determined by the value is at least 5 times greater, preferably 10 times, more preferably 30, 50, 80 or even 100 times greater for target cells (e.g., characterized by simultaneous expression of two different targets or a high level of one target) than for non-target cells (e.g., characterized by expression of only one of the two targets or only a low level of one target). The selectivity gap in the context of the present invention is preferably greater than 100 times. In the context of the present invention, it is envisioned that the selectivity gap (which can also be defined as the activity gap) is at least 250, 500, 750 or even 1000 times, which greatly improves the efficacy and safety of the multi-chain multi-targeting bispecific molecules of the present invention compared to various forms of single-targeting bispecific molecules.

[0167] Another aspect envisaged in the context of the present invention is the further support of the dual avidity effect conferred by the format of the multi-chain multi-targeting antigen binding molecule by low affinity (preferably both target antigen binder and CD3 effector binder). In the context of the present invention, preferably the affinity is below KD 1.2 x 10 -8 The activity of a particularly preferred CD3 binder is 10 times lower than the activity of a CD3 binder with a KD of 1.2 x 10-8, more preferably 50 times lower or even more preferably 100 times lower. Without wishing to be bound by theory, the avidity effect is expected to be more pronounced when two binders with relatively balanced affinities, i.e., typically two binders with low to medium-high affinity, preferably low affinity, bind to two targets on the same target cell, compared to binders with mixed affinities or typically higher affinities, which trigger cytolytic activity (this is also the case when only one target on a cell is bound, which cell would, for example, be a physiological non-target cell that should not be targeted to avoid off-target toxicity and related side effects)).

[0168] Therefore, it is preferred that the multi-chain multi-targeted bispecific antigen binding molecules according to the present invention that are combined with two (preferably different) targets on the target cell to show significant cytotoxic activity do show less side effects than the single-targeted bispecific antigen binding molecules that make effector T cells and target cells together.For example, this can be demonstrated by the release of key cytokines IL-2, IL-6, IL-10, TNFa and IFNg significantly reduced, and these cytokines are indicators of clinical stage side effects.For example, relative to corresponding single-targeted bispecific molecules, after using the multi-chain multi-targeted bispecific antigen binding molecules according to the present invention, the release of IL-6 is generally reduced.As is known in the art, interleukin 6 (IL-6) seems to play a key role in CRS pathophysiology, because highly elevated IL-6 levels are observed in CRS patients (Shimabukuro-Vornhagen et al. Journal for ImmunoTherapy of Cancer [Cancer Immunotherapy Magazine] (2018) 6: 56).Since CRS is a serious side effect in immunotherapy, this reduction shows that clinical stage CRS is less.

[0169] In addition, the multi-chain multi-targeted bispecific antigen binding molecules according to the present invention that bind to two (preferably different) targets on target cells to show significant cytotoxic activity preferably do show fewer side effects than single-targeted bispecific antigen binding molecules in terms of toxic tissue damage. It is unexpectedly found that the multi-specific molecules in the form described herein show higher tolerance, i.e., higher doses than the corresponding single-targeted bispecific molecules can be administered without clinical manifestations, such as tissue damage examined by histopathological examination. For example, a 1.5 μg / kg dose of MSLN single-targeted bispecific antigen binding molecules (SEQ ID NO: 1183) cannot be tolerated and causes death, while a 0.1 μg / kg dose is tolerated. In contrast, the multi-chain multi-targeted CDH3-MSLN bispecific molecules (SEQ ID NO: 251) according to the present invention are tolerated at doses up to 1000 μg / kg. The histopathological changes observed with single-targeted molecules at a dose of 1.5 μg / kg are generally more severe than those observed with multi-chain multi-targeted molecules at a dose of 1000 μg / kg. After treatment with multi-chain multi-targeting molecules, there is no adhesion or irreversible fibrotic changes caused by single-targeting molecules. Therefore, the tolerance of the multi-chain multi-targeting molecules according to the present invention is higher than the corresponding single-targeting molecules, for example, 600 (histopathology) to 10.000 (tolerance dose) times, although the efficacy against tumor cells in vitro is equal. Therefore, the multi-chain multi-targeting molecules of the present invention are particularly suitable for the following treatment environments, wherein the addressed targets are not only significantly present in disease-related (pathophysiology) tissues but also mainly present in physiological tissues, however, physiological tissues should not be targeted by cytotoxic immunotherapy. For example, this is the case for MSLN, which is usually expressed in mesothelial cells lining the following body cavities: pleura (pleural cavity around the lungs), peritoneum (abdominal pelvic cavity, including mesentery, omentum, falciform ligament and periuterine membrane) and pericardium (around the heart). There is a risk of serious side effects by addressing targets such as MSLN by cytotoxic immunotherapy, such as intra-abdominal adhesions and / or fibrosis. Intra-abdominal adhesions are understood herein as pathological scars formed between intra-abdominal organs. Adhesions occur in the presence of intra-peritoneal inflammation and cause the peritoneal surfaces to adhere to each other. Adhesions can cause problems if scars restrict the free movement of organs (Mutsaers SE, Prele CM, Pengelly, S., Herrick, SE Mesothelial cells and peritoneal homeostasis. Fertil Steril 2016, 106 (5) 1018-1024).Fibrosis is understood herein as a common pathological consequence of several etiological conditions leading to chronic tissue damage, and is generally defined as excessive deposition of extracellular matrix (ECM) components, leading over time to scar tissue formation and ultimately to organ dysfunction and 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 for addressing disease-associated targets that are significantly co-expressed on pathological physiological tissues and one or more physiological tissues by providing a multi-chain multi-targeting bispecific antigen binding molecule in the form described herein, wherein the molecule addresses (i.) a target expressed on both disease-associated tissues and physiological tissues and (ii.) another target expressed on physiological tissues associated with the disease but not under (i.), wherein the method preferably avoids the formation of intra-abdominal adhesions and / or fibrosis when such a target is MSLN.

[0170] It is envisaged that the bispecific antigen binding molecules according to the invention have cross-reactivity with, for example, cynomolgus monkey tumor-associated antigens such as CDH3, MSLN, CD20, CD22, FLT3, CLL1 and EpCAM. It is particularly envisaged in the context of the present invention that two targets can be addressed simultaneously by one multi-chain multi-targeting bispecific antigen binding molecule.

[0171] Alternatively, in addition to the major advantage of increased selectivity as described herein, dual targeting can mitigate the lack of accessibility of one target when targeting the remaining target can trigger a sufficient residual effect. Examples are (i) the presence of a soluble target, which "masks" the target on target cells by binding the antigen binding molecule without allowing the remaining molecule to have any therapeutic effect, and (ii) antigen loss (reduced target expression on target cells) as a driver of tumor escape.

[0172] For example, the multi-chain multi-targeting antigen-binding molecules according to the present invention, such as constructs against MSLN as TAA1 and CDH3 as TAA2, are suitable for treating, alleviating or preventing cancer, in particular a cancer selected from the group consisting of lung cancer, head and neck cancer, primary or secondary CNS tumors, primary or secondary brain tumors, primary CNS lymphomas, spinal cord axis tumors, brain stem 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, metastatic cancer, cancer of the stomach, gastrointestinal tract (stomach, colorectal, and duodenum), small intestine, biliary tract, urethra, renal cell carcinoma, endometrial cancer, thyroid cancer, testicular cancer, squamous cell carcinoma of the skin, melanoma, stomach cancer, prostate cancer, bladder cancer, osteosarcoma, mesothelioma, Hodgkin's disease, non-Hodgkin's lymphoma, chronic or acute leukemia, chronic myeloid leukemia, lymphocytic lymphoma, multiple myeloma, fibrosarcoma, neuroblastoma, retinoblastoma, and soft tissue sarcoma.

[0173] It is particularly envisaged in the context of the present invention that the multi-chain multitargeting antigen binding molecules preferably address two different target cell surface antigens, thereby being very specific for their target cells and thus preferably safe in their therapeutic use. The efficacy in inhibiting tumor growth has been demonstrated in vivo in a mouse model.

[0174] Preferred target cell surface antigens in the context of the present invention are MSLN, CDH3, FLT3, CLL1, EpCAM, CD20 and CD22. Typically, in the context of the present invention, the target cell surface antigen is a tumor associated antigen (TAA). B lymphocyte antigen CD20 or CD20 Expressed on the surface of all B cells (starting at the pro-B stage (CD45R+, CD117+) and increasing in concentration until maturity). CD22 , or cluster of differentiation-22, is a molecule belonging to the SIGLEC family of lectins. It is found on the surface of mature B cells and, to a lesser extent, on some immature B cells. Fms-like tyrosine kinase 3 ( FLT3 ) is also known as cluster of differentiation 135 (CD135), receptor tyrosine kinase FLT3, or fetal liver kinase 2 (Flk2). FLT3 is a cytokine receptor that belongs to receptor tyrosine kinase III. CD135 is the receptor for the cytokine Flt3 ligand (FLT3L). The FLT3 gene is frequently mutated in acute myeloid leukemia (AML). C-type lectin-like receptor ( CLL1), also known as CLEC12A or MICL. It contains an ITIM motif in the cytoplasmic tail and can associate with the signaling phosphatases SHP-1 and SHP-2. Human MICL is mainly expressed as a monomer in 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 several human tumors. CDH3 ), also known as P-cadherin, is a calcium-dependent cell-cell adhesion glycoprotein composed of five extracellular cadherin repeats, a transmembrane region, and a highly conserved cytoplasmic tail. It is associated with some types of tumors. Epithelial cell adhesion molecule ( EpCAM EpCAM is a transmembrane glycoprotein in epithelia that mediates Ca2+-independent homotypic cell-cell adhesion. EpCAM has oncogenic potential and appears to play a role in tumorigenesis and cancer metastasis.

[0175] Furthermore, in the context of the present invention, it is optionally but advantageously envisaged that the multi-chain multi-targeting antigen binding molecule has a spacer, preferably a globular protein structure, such as a scFc domain or a dimerizing Fc domain, such as an iso-Fc, which also increases the half-life of the molecule and enables intravenous administration only once a week, once every two weeks, once every three weeks or even once every four weeks, or less frequently.

[0176] To determine one or more epitopes of a preferred multi-chain multi-targeted antigen binding molecule according to the invention for, for example, CDH3, MSLN or CD20 epitopes, positioning is performed as described herein. Preferred bispecific antigen binding molecules with CD20 target binders are directed to all epitope clusters E1A, E2B and E2C. An epitope cluster is understood herein as a stretch of amino acids (as disclosed herein and defined according to its position according to Kabat) within a target (as disclosed herein and defined according to its position according to Kabat), if the stretch of amino acids of the human target is replaced by the corresponding stretch of amino acids of the mouse target, then the entire target binder of the multi-chain multi-targeted bispecific antigen binding molecule described herein is essentially no longer bound to the target. Therefore, the epitope cluster method is understood herein as mouse chimera sequence analysis. This method has been described in Münz et al. Cancer Cell International [International Cancer Cell] 2010, 10: 44 and is applied as described in detail in the examples about CDH3 and MSLN.

[0177] Preferred epitope clusters are D4B of CDH3 described herein and E1 of MSLN described herein. As illustrated in the Examples, the selectivity gap of the multi-chain multi-targeting bispecific antigen binding molecules of the invention (relative to comparable single-targeting bispecific antigen binding molecules) is generally even greater, and therefore, more preferably, if the MSLN target binder addresses the E1 epitope cluster, and if the CDH3 target binder addresses the D4B epitope cluster. Although addressing other epitope clusters also results in very high selectivity gaps and associated advantages in efficacy and tolerability / safety, the selectivity gap is particularly high, and therefore molecules comprising target binders addressing E1 and D4B are preferred. Such molecules include, for example, molecules having a MSLN target binder comprising CDR H1-H3 of SEQ ID NOs 774 to 776 and CDR L1-L3 of 777 to 779 (and corresponding VH and VL of 780 and 781), CDR H1-H3 of SEQ ID NOs 782 to 784 and CDR L1-L3 of 785 to 787 (and corresponding VH and VL of 788 and 789), CDR H1-H3 of SEQ ID NOs 806 to 808 and CDR L1-L3 of 809 to 811 (and corresponding VH and VL of 812 and 813), CDR H1-H3 of SEQ ID NOs 838 to 840 and CDR L1-L3 of 841 to 843 (and corresponding VH and VL of 844 and 845), CDR H1-H3 of SEQ ID NOs 862 to 864 and CDR L1-L3 of SEQ ID NOs 865 to 866 H1-H3 and CDR L1-L3 of SEQ ID NOs 865 to 867 (and the corresponding VH and VL of 868 and 869), CDR H1-H3 and CDR L1-L3 of SEQ ID NOs 894 to 896 and CDR L1-L3 of 897 to 899 (and the corresponding VH and VL of 900 and 901), CDR H1-H3 and CDR L1-L3 of SEQ ID NOs 950 to 952 and CDR L1-L3 of 953 to 955 (and the corresponding VH and VL of 956 and 957), CDR H1-H3 and CDR L1-L3 of SEQ ID NOs 1030 to 1032 and CDR L1-L3 of 1033 to 1035 (and the corresponding VH and VL of 1036 and 1037), or CDR H1-H3 of SEQ ID NOs 86 to 88 and CDR L1-L3 of 89 to 91 L1-L3 (and corresponding VH of 92 and VL of 93 or 94). Preferred examples of CDH3 binders that bind to the preferred DB4 epitope cluster comprise CDR H1-H3 of SEQ ID NOs 194, 432 and 196 and CDR L1-L3 of 197 to 199 (and corresponding VH and VL of 433 and 200).Other target binders that preferentially bind to the preferred epitope cluster of D4B are identified herein as, for example, CH3 15-E11 CC and CH3 24-D7 CC.

[0178] It is particularly surprising that the multi-chain multi-targeted antigen binding molecules according to the present invention can bind, preferably simultaneously bind to two different targets. It has been demonstrated herein that multiple targets can be combined simultaneously. However, considering the typical distance between targets, this is surprising. For example, CD20 comprises two small extracellular domains of only 6 amino acids and 47 amino acids. In contrast, CD22 comprises an extracellular domain with 676aa and 7Ig domain length. However, even if the extracellular size and setting are significantly different, the multi-chain multi-targeted antigen binding molecules according to the present invention can successfully address both TAA CD20 and CD22 simultaneously, thereby achieving the benefits of increasing efficacy and reducing toxicity.

[0179] It is envisioned in the context of the present invention that preferred multi-chain multi-targeted antigen binding molecules not only show a favorable ratio of cytotoxicity to affinity, but also show sufficient stability characteristics in addition, so as to contribute to the actual operation of preparing, storing and applying the construct. For example, sufficient stability is characterized by high monomer content (i.e. non-aggregated and / or non-associated natural molecules) after standard preparation, for example, as determined by preparative size exclusion chromatography (SEC) at least 65%, more preferably at least 70% and even more preferably at least 75%. In addition, for example, the turbidity measured as optical absorption at 340nm at a concentration of 2.5mg / ml should preferably be equal to or less than 0.025, more preferably 0.020, for example, so as to draw a conclusion that there is substantially no undesirable aggregate. Advantageously, high monomer content is maintained after incubation under stress conditions (e.g., freezing / thawing) or incubation at 37°C or 40°C. Even more, the multi-chain multi-targeted antigen binding molecules according to the present invention typically have thermal stability, which is at least comparable or even higher than the thermal stability of the bispecific antigen binding molecules (i.e., it is less complex in structure) with only one target binding domain but additionally comprising a CD3 binding domain and an scFc domain with extended half-life. The technician expects that the more complex protein-based molecules in structure are more prone to thermal degradation and other degradations, i.e., poor thermal stability. However, surprisingly, the situation is the opposite, and the multi-chain multi-targeted bispecific antigen binding molecules according to the present invention show at least comparable or even better thermal stability with single-chain molecules. When also testing for long-term storage stability and freeze-thaw stability, the molecules of the present invention advantageously show at least comparable or even better features with single-chain molecules with the same binding domain. Preferably, compared with the corresponding single-chain bispecific antigen binding molecules (i.e., comprising the same target and CD3 binding substances, such as disclosed herein), the molecules of the present invention also show less monomer reduction after storage, and higher protein homogeneity.

[0180] In an embodiment, the present invention provides a multi-chain multi-targeting bispecific antigen binding molecule comprising all four such domains. In a preferred embodiment, the domains in (i.), (ii.), (iii.), and (iv.) are as follows Figure 1 , 2 and arranged as described in 3.

[0181] The term "polypeptide" is understood herein as an organic polymer comprising at least one continuous, unbranched amino acid chain. In the context of the present invention, polypeptides comprising more than one amino acid chain are also envisioned. The polypeptide amino acid chain typically comprises at least 50 amino acids, preferably at least 100, 200, 300, 400 or 500 amino acids. It is also envisioned in the context of the present invention that the amino acid chain of the polymer is linked to an entity that does not consist of amino acids.

[0182] The term "antigen-binding polypeptide" according to the present invention is preferably a polypeptide that is immunospecifically bound to its target or antigen. It typically comprises a heavy chain variable region (VH) and / or a light chain variable region (VL) of an antibody, or comprises a domain derived therefrom. The polypeptide according to the present invention comprises the minimum structural requirements of an antibody that allows immunospecific target binding. This minimum requirement can be defined, for example, by the presence of at least three light chain CDRs (i.e., CDR1, CDR2, and CDR3 in the VL region) and / or three heavy chain CDRs (i.e., CDR1, CDR2, and CDR3 in the VH region), preferably all six CDRs. The antigen binding molecules of the present invention are preferably T cell engagement polypeptides, which are therefore characterized by the presence of three or six CDRs in one or two binding domains, and the technician knows where those CDRs are located in the binding domain (in what order). Preferably, "antigen binding molecules" are understood as "antigen-binding polypeptides" in the context of the present invention. In another embodiment, the antigen-binding polypeptides of the present invention can be aptamers.

[0183] Alternatively, the molecule in the context of the present invention is an antigen-binding polypeptide, which corresponds to an "antibody construct", which typically refers to a molecule in which the structure and / or function is based on the structure and / or function of an antibody (e.g., full-length or complete immunoglobulin molecule). Therefore, the antigen-binding molecule can be extracted in conjunction with its specific target or antigen and / or from the variable heavy chain (VH) and / or variable light chain (VL) domain of an antibody or its fragment. In addition, the domain combined with the binding partner according to the present invention is understood herein as the binding domain of the antigen-binding molecule according to the present invention. Typically, the binding domain according to the present invention includes the minimum structural requirements for the antibody that allows target binding. This minimum requirement can be, for example, defined by at least three light chain CDRs (i.e., CDR1, CDR2, and CDR3 in VL region) and / or three heavy chain CDRs (i.e., CDR1, CDR2, and CDR3 in VH region), preferably all six CDRs. An alternative to defining the minimal structural requirements of an antibody is to define the antibody epitope within a specific target structure, respectively, the protein domain of the target protein constituting the epitope region (epitope cluster) or by reference to a specific antibody that competes with the epitope of the defined antibody. The antibodies on which the construct according to the invention is based include, for example, monoclonal antibodies, recombinant antibodies, chimeric antibodies, deimmunized antibodies, humanized antibodies and human antibodies.

[0184] In the context of the present invention, the polypeptide of the present invention binds to its corresponding target structure in a specific manner. Preferably, each binding domain of the polypeptide according to the present invention comprises a paratope, which "specifically or immunospecifically" binds to its corresponding target structure, "(specifically or immunospecifically) recognizes" its corresponding target structure, or "(specifically or immunospecifically) reacts" with its corresponding target structure. According to the present invention, this means that the polypeptide or its binding domain interacts or (immunospecifically) interacts with a given epitope on a target molecule (antigen) and CD3, respectively. This interaction or binding occurs more frequently and more quickly in an epitope on a specific target than in an alternative substance (non-target molecule), has a longer duration, has a greater affinity, or has some combination of these parameters. However, due to the sequence similarity between homologous proteins in different species, a binding domain that (immunospecifically) specifically binds 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 / immunospecific binding" may include binding of a binding domain to epitopes in more than one species and / or structurally related epitopes. The term "(immuno)selectively binds" does not include binding to structurally related epitopes.

[0185] The binding domain of the antigen binding molecule according to the present invention can for example comprise the CDR group mentioned above. Preferably, those CDRs are contained in the framework of the antibody light chain variable region (VL) and the antibody heavy chain variable region (VH); However, it does not necessarily comprise both. For example, the Fd fragment has two VH regions and generally retains some antigen binding functions of the complete antigen binding domain. Additional examples of antibody fragments, antibody variants, or binding domain formats include (1) a Fab fragment, a monovalent fragment having VL, VH, CL, and CH1 domains; (2) a F(ab')2 fragment, a bivalent fragment having two Fab fragments connected by a disulfide bridge at the hinge region; (3) a Fd fragment having two VH and CH1 domains; (4) an Fv fragment having the VL and VH domains of a single arm of an antibody; (5) a dAb fragment having a VH domain (Ward et al., (1989) Nature 341:544-546); (6) isolated complementarity determining regions (CDRs), and (7) single-chain Fv (scFv), the latter being preferred (e.g., derived from an scFV library). Examples of embodiments of antigen binding molecules according to the invention are described, for example, in WO 00 / 006605, WO 2005 / 040220, WO 2008 / 119567, WO 2010 / 037838, WO 2013 / 026837, WO 2013 / 026833, US2014 / 0308285, US2014 / 0302037, WO 2014 / 144722, WO 2014 / 151910 and WO 2015 / 048272.

[0186] In addition, within the definition of "binding domain" or "domain that binds to..." are fragments of full-length antibodies, such as VH, VHH, VL, (s) dAb, Fv, Fd, Fab, Fab', F(ab')2 or "rIgG" ("half antibody"). The antigen-binding molecules according to the present invention may also include modified antibody fragments, also referred to as antibody variants, such as scFv, two-scFv or two (two)-scFv, scFv-Fc, scFv-zipper, scFab, Fab2, Fab3, double antibodies, single-chain double antibodies, tandem double antibodies (Tandab), tandem two-scFv, tandem three-scFv), "multi-antibodies" (such as three antibodies or four antibodies) and single domain antibodies comprising only one variable domain (which may be VHH, VH or VL, independently of other V regions or domains that specifically bind antigens or epitopes), such as nanobodies or single variable domain antibodies. Typically, the binding domain of the present invention comprises a paratope that promotes binding to its binding partner.

[0187] As used herein, the term "single-chain Fv", "single-chain antibody" or "scFv" refers to a single polypeptide chain antibody fragment comprising variable regions from heavy and light chains, but lacking constant regions. Generally, a single-chain antibody further comprises a polypeptide linker between the VH and VL domains that allows it to form a desired structure that will allow antigen binding. Single-chain antibodies are discussed in detail in Pluckthun, The Pharmacology of Monoclonal Antibodies, Vol. 113, Rosenburg and Moore, ed., Springer-Verlag, New York, pp. 269-315 (1994). Various methods for producing single-chain antibodies are known, including those described in U.S. Pat. Nos. 4,694,778 and 5,260,203; International Patent Application Publication No. WO 88 / 01649; 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 specific embodiments, single-chain antibodies can also be bispecific, multispecific, human and / or humanized and / or synthetic.

[0188] In the context of the present invention, a paratope is understood to be an antigen binding site that is part of a polypeptide as described herein and that recognizes an antigen and binds thereto. A paratope is typically a small region of about at least 5 amino acids. A paratope as understood herein typically comprises portions of antibody-derived heavy chain (VH) and light chain (VL) sequences. Each binding domain of a molecule according to the present invention provides a paratope comprising a set of 6 complementary determining regions (CDR loops) (wherein each three are contained in antibody-derived VH and VL sequences, respectively).

[0189] In addition, the definition of term "antigen binding molecules" includes preferably multivalent (polyvalent / multivalent) constructs, and therefore includes bispecific molecules, wherein bispecific means that it specifically binds to two cell types (i.e., target cells and effector cells) comprising different antigen structures. Because the antigen binding molecules of the present invention are preferably multi-chain multi-targeted, these antigen binding molecules are typically also multivalent (polyvalent / multivalent) molecules, that is, they specifically bind more than two antigen structures, preferably four different binding domains in the context of the present invention, which are two target binding domains and two CD3 binding domains. The term "multi-chain multi-targeting bispecific antigen binding molecules" includes the terms "multi-chain multi-targeting bispecific T cell engager molecules" and "multi-chain multi-targeting bispecific T cell engager polypeptides (MMBiTEP)". Preferably, "multi-chain multi-targeting bispecific antigen binding molecules" are "multi-chain multi-targeting bispecific T cell engager molecules" or "multi-chain multi-targeting bispecific T cell engager polypeptides (MMBiTEP)". The term "multi-chain multi-targeting bispecific T cell engager molecule" is understood to include the term "multi-chain multi-targeting bispecific T cell engager polypeptide". In addition, the definition of the term "antigen binding molecule" includes molecules comprising only one polypeptide chain and molecules consisting of more than one polypeptide chain, which chains can be the same (homodimers, homotrimers or homooligomers) or different (heterodimers, heterotrimers or heterooligomers). Such molecules comprising more than one polypeptide chain (i.e., typically two chains) are typically attached to each other as heterodimers via charge pair binding, for example within a hetero-Fc entity (which serves as a spacer and half-life extension moiety between two bispecific entities described herein). Examples of the above-identified antigen-binding molecules, e.g., 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 Dübel, Antibody Engineering, Springer, 2nd edition 2010, and Little, Recombinant Antibodies for Immunotherapy, Cambridge University Press 2009.

[0190] As used herein, the term "bispecific" refers to an antigen binding molecule that is "at least bispecific", i.e., it addresses two different cell types (i.e., target cells and effector cells), and comprises 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 binding domains bind to two antigens or targets (preferably 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 effector cells (typically T cells). Therefore, the antigen binding molecules according to the present invention are specific for at least two different antigens or targets. For example, the two domains preferably do not bind to the extracellular epitopes of CD3ε of one or more species as described herein.

[0191] The term "target cell surface antigen" refers to an antigen structure expressed by a cell, which is present on the cell surface so that it is accessible to antigen binding molecules as described herein. In the context of the present invention, preferred target cell surface antigens are tumor-associated antigens (TAA). It can be a protein, preferably an extracellular portion of a protein, or a carbohydrate structure, preferably a carbohydrate structure of a protein, such as a glycoprotein. It is preferably a tumor antigen. The term "bispecific antigen binding molecule" of the present invention also encompasses bispecific multi-chain multi-targeted antigen binding molecules, such as three-targeted antigen binding molecules, the latter including three binding domains, or having more than three (e.g., four, five ...) specific constructs.

[0192] Preferred in the context of the present invention are "multi-targeting" molecules, which are understood herein as "typically targeting at least two targets (e.g., TAAs) / molecules of the present invention / target cells". In this regard, multi-targeting molecules, such as antigen binding molecules, have specificity for two typically identical effector structures on effector cells (e.g., CD3, more preferably CD3ε (CD3ε, included whenever "CD3" is mentioned in the present invention)) and at least two target cell surface antigens. The specificity is conferred by corresponding binding domains as defined herein. Typically, "multi-targeting" refers to molecules that are specific for at least two (preferably different) target cell surface antigens (e.g., TAAs), which confer preferred properties to the multi-targeting antigen binding molecules according to the present invention, i.e., mitigating antigen loss and increasing selectivity, i.e., killing target cells that co-express the target for which the molecule of the present invention has a binding domain directed thereto, as well as selectivity for target cells associated with the disease. Therefore, the therapeutic window of the molecules of the present invention is increased relative to single-targeting bispecific molecules, which typically results in higher drug tolerance, as demonstrated herein.

[0193] Antigen binding molecules that engage with T cells, for example, multi-chain polypeptides according to the present invention are preferably bispecific, which are understood herein as generally comprising a domain that binds at least one target antigen and another domain that binds CD3. Therefore, it is not naturally occurring, and its function is significantly different from that of naturally occurring products. Therefore, the polypeptide according to the present invention is an artificial "hybrid" polypeptide comprising at least two different binding domains with different specificities, and is therefore bispecific. Bispecific antigen binding molecules can be produced by a variety of methods (including the fusion of hybridomas or the connection of Fab' fragments). See, for example, Songsivilai and Lachmann, Clin. Exp. Immunol. [Clinical Experimental Immunology] 79: 315-321 (1990).

[0194] At least four binding domains and variable domains (VH / VL) of antigen binding molecules of the present invention typically include peptide linkers (spacer peptides). According to the present invention, term " peptide linker " includes amino acid sequence, and the amino acid sequence of one (variable and / or binding) domain and another (variable and / or binding) domain of antigen binding molecules of the present invention is connected to each other by this amino acid sequence. The peptide linker between the first and second binding domains and the third and fourth domains (wherein the first and third domains are preferably able to combine two targets simultaneously, and these two targets are preferably different targets (such as TAA1 and TAA2), preferably on the same cell) is preferably flexible and limited in length, such as 5,6,7,8,9,10,11,12,13,14,15,16,17 or 18 amino acids. Peptide linkers can also be used for merging spacers with other domains of antigen binding molecules of the present invention. The basic technical feature of this peptide linker is that it does not include any polymerization activity. Suitable peptide linkers are those described in U.S. Patent Nos. 4,751,180 and 4,935,233 or WO 88 / 09344. Peptide linkers can also be used to attach other domains or modules or regions (such as domains of half-life extension) to the antigen binding molecules of the present invention. However, typically the linker between the first target binding domain and the second target binding domain is different from the intra-binder linker connecting VH and VL in the target binding domain. The difference is that the linker between the first binding domain and the second binding domain is one more amino acid than the intra-binder linker in the binding, for example, six and five amino acids, such as SGGGGS vs. GGGGS. This unexpectedly makes the specific antigen binding molecule form as described herein have flexibility and stability at the same time. The spacer (or synonymous spacer entity) between two bispecific entities as described herein is a specific embodiment of a linker, because the spacer also acts as a linker, because it helps to connect two bispecific entities to preferably construct at least one continuous polypeptide chain including four binding domains or parts thereof. However, in addition, the spacer functions as an entity that spatially separates the two bispecific entities. Thus, a spacer in the context of the present invention is a specific embodiment of a linker, which together with two further shorter and flexible linkers at each end helps to connect the two binding domains (of the two different bispecific entities), but first and foremost to space them apart so that the two bispecific entities can advantageously function as described herein, e.g., show an unexpectedly high selectivity gap.

[0195] The antigen binding molecules of the present invention are preferably "antigen binding molecules generated in vitro". This term refers to an antigen binding molecule according to the above definition, wherein all or part of the variable region (e.g., at least one CDR) is generated in non-immune cell selection, such as in vitro phage display, protein chip or any other method that can test the ability of candidate sequences to bind to antigens. Therefore, this term preferably excludes sequences produced only by genomic rearrangement in animal immune cells. "Recombinant antibodies" are antibodies made by using recombinant DNA technology or genetic engineering.

[0196] The term "monoclonal antibody" (mAb) used herein or the monoclonal antibody from which the antigen-binding molecule is derived refers to an antibody obtained from a substantially homogeneous antibody population, that is, except for possible naturally occurring mutations and / or post-translational modifications (e.g., isomerization, amidation) that may be present in small amounts, the individual antibodies constituting the population are identical. Compared to conventional (polyclonal) antibody preparations that typically include different antibodies for different determinants (or epitopes), monoclonal antibodies are highly specific for a single antigen side or determinant on an antigen. In addition to their specificity, monoclonal antibodies are also advantageous in that they are synthesized by hybridoma culture and are therefore not contaminated by other immunoglobulins. The modifier "monoclonal" indicates the characteristics of an antibody obtained from a substantially homogeneous antibody population, and should not be construed as requiring the production of antibodies by any particular method.

[0197] For the preparation of monoclonal antibodies, any technology providing antibodies produced by continuous cell line cultures can be used. For example, the monoclonal antibody to be used can be prepared by the hybridoma method first described by Koehler et al., Nature [Nature], 256:495 (1975), or can be prepared by recombinant DNA methods (see, e.g., U.S. Patent No. 4,816,567). Examples of other technologies for producing human monoclonal antibodies include tri-source hybridoma technology, human B cell hybridoma technology (Kozbor, Immunology Today [Today's Immunology] 4 (1983), 72) and EBV-hybridoma technology (Cole et al., Monoclonal Antibodies and Cancer Therapy [monoclonal antibodies and cancer therapy], Alan R. Liss Company (1985), 77-96).

[0198] The assay can then be performed using standard methods such as enzyme-linked immunosorbent assay (ELISA) and surface plasmon resonance analysis such as Biacore TMHybridomas are screened to identify one or more hybridomas that produce antibodies that specifically bind to a given antigen. Any form of the relevant antigen can be used as an immunogen, such as a recombinant antigen, a naturally occurring form, any variant or fragment thereof, and an antigenic peptide thereof. Surface plasmon resonance, as used in the Biacore system, can be used to increase the efficiency of phage antibodies that bind to epitopes of target cell surface antigens (Schier, Human Antibodies Hybridomas 7 (1996), 97-105; Malmborg, J. Immunol. Methods 183 (1995), 7-13).

[0199] Another exemplary method for preparing monoclonal antibodies includes screening protein expression libraries, such as phage display or ribosome display libraries. Phage display is described, for example, in Ladner et al., U.S. Pat. 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).

[0200] In addition to using display libraries, non-human animals, such as rodents (such as mice, hamsters, rabbits or rats) can also be immunized with relevant antigens. In one embodiment, the non-human animal includes at least a portion of a human immunoglobulin gene. For example, it is possible to use large fragments of human Ig (immunoglobulin) loci to engineer mouse strains that are defective in producing mouse antibodies. Using hybridoma technology, antigen-specific monoclonal antibodies derived from genes with desired specificity can be produced and selected. See, for example, XENOMOUSE TM , Green et al. (1994) Nature Genetics 7: 13-21, US 2003-0070185, WO 96 / 34096 and WO 96 / 33735.

[0201] Monoclonal antibodies can also be obtained from non-human animals and then modified using recombinant DNA techniques known in the art, e.g., humanized, deimmunized, rendered chimeric, etc. 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 mutants with altered effector function(s) (see, e.g., U.S. Pat. No. 5,648,260, Kontermann and Dübel (2010), loc. cit., and Little (2009), loc. cit.).

[0202] In immunology, affinity maturation is a process by which B cells produce antibodies with increased affinity for antigens during an immune response. After repeated exposure to the same antigen, the host will produce antibodies with successively greater affinity. Like 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 are introduced into CDRs using radiation, chemical mutagens, or error-prone PCR. In addition, genetic diversity can be increased by chain shuffling. Two or three rounds of mutations and selection using display methods (such as phage display) typically produce antibody fragments with an affinity in the low nanomolar range.

[0203] The amino acid substitution variation of the preferred type of antigen binding molecule includes replacing one or more hypervariable region residues of the parent antibody (e.g., humanized or human antibody). Generally speaking, one or more obtained variants selected for further development will have improved biological properties relative to the parent antibody that produces them. A convenient way for producing such substitution variants relates to affinity maturation using phage display. In short, several hypervariable region side ends (e.g., 6-7 side ends) are mutated to produce all possible amino acid substitutions at each side end. The antibody variants thus produced are displayed as fusions of the gene III product of the M13 packaged in each particle from filamentous phage particles in a monovalent manner. The biological activity (e.g., binding affinity) of the variants displayed by phage display is then screened as disclosed herein. In order to identify the candidate hypervariable region side ends for modification, alanine scanning mutagenesis can be performed to identify the hypervariable region residues that contribute significantly to antigen binding. Alternatively or additionally, it may be beneficial to analyze the crystal structure of the antigen-antibody complex to identify the 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 according to the techniques described herein. Once such variants are generated, the set of variants is screened as described herein, and antibodies with superior properties in one or more relevant assays can be selected for further development.

[0204] The monoclonal antibodies and antigen-binding molecules of the present invention specifically include "chimeric" antibodies (immunoglobulins) in which a portion of the heavy and / or light chains is identical or homologous to the corresponding sequences in antibodies derived from a particular species or belonging to a particular antibody class or subclass, and the remainder of one or more chains is identical or homologous to the corresponding sequences in antibodies derived from another species or belonging to another antibody class or subclass, as well as fragments of such antibodies, as long as they exhibit the desired biological activity (U.S. Pat. No. 4,816,567; Morrison et al., Proc. Natl. Acad. Sci. USA [Proceedings of the National Academy of Sciences of the United States] 81: 6851-6855 (1984)). The chimeric antibodies of interest herein include "primatized" antibodies, which contain variable domain antigen-binding sequences derived from non-human primates (e.g., Old World monkeys, apes, etc.) and human constant region sequences. A variety of methods for preparing chimeric antibodies have been described. See, e.g., Morrison et al., Proc. Natl. Acad. ScL U. S. A [Proceedings of the National Academy of Sciences of the United States of America] 81:6851, 1985; Takeda et al., Nature [Nature] 314:452, 1985; Cabilly et al., U.S. Pat. No. 4,816,567; Boss et al., U.S. Pat. No. 4,816,397; Tanaguchi et al., EP 0 171 496; EP 0 173 494; and GB 2177096.

[0205] Antibody, antigen binding molecules, antibody fragment or antibody variant can also be modified by specifically lacking human T cell epitopes (called "deimmunization" method) by the method disclosed in, for example, WO 98 / 52976 or WO 00 / 34317. In short, the heavy chain and light chain variable domains of the peptide analysis antibody combined with Class II MHC can be used; These peptides represent potential T cell epitopes (as defined in WO 98 / 52976 and WO 00 / 34317). In order to detect potential T cell epitopes, the computer modeling method called "peptide threading" can be applied, and in addition, for the motif present in VH and VL sequences, the database of human MHC I1 class binding peptides can be searched, as described in WO 98 / 52976 and WO 00 / 34317. These motifs are combined with any one of 18 kinds of main MHC I1 class DR allotypes, and therefore constitute potential T cell epitopes. The potential T cell epitopes detected can be eliminated by replacing a small amount of amino acid residues in the variable domains, or preferably by replacing a single amino acid. Typically, conservative substitutions are made. Generally, but not exclusively, amino acids common to positions in human germline antibody sequences can be used. Human germline sequences are disclosed, for example, in Tomlinson et al. (1992) J. MoI. Biol. [Journal of Molecular Biology] 227: 776-798; Cook, GP et al. (1995) Immunol. Today [Contemporary Immunity] Vol. 16 (5): 237-242; and Tomlinson et al. (1995) EMBO J. [Journal of the European Molecular Biology Association] 14: 14: 4628-4638. The VBASE directory provides a comprehensive directory of human immunoglobulin variable region sequences (compiled by Tomlinson, LA. et al. MRC Centre for Protein Engineering [MRC Protein Engineering Center], Cambridge, UK [Cambridge, UK]). These sequences can be used as a source of human sequences, for example for framework regions and CDRs. Common human framework regions can also be used, for example as described in U.S. Patent No. 6,300,064.

[0206] "Humanized" antibodies, antigen binding molecules, variants or fragments thereof (such as Fv, Fab, Fab', F(ab')2 or other antigen binding subsequences of antibodies) are antibodies or immunoglobulins of most human sequences, which contain one or more minimal sequences derived from non-human immunoglobulins. For the most part, humanized antibodies are human immunoglobulins (receptor antibodies), wherein the residues from the hypervariable regions (also referred to as CDR) of the receptor are replaced by residues from the hypervariable regions of non-human (e.g., rodent) species (donor antibodies) (such as mice, rats, hamsters or rabbits) with desired specificity, affinity and ability. In some cases, the Fv framework region (FR) residues of human immunoglobulins are replaced by corresponding non-human residues. In addition, as used herein, "humanized antibodies" may also include residues not found in both the receptor antibody and the donor antibody. These modifications are performed to further improve and optimize antibody performance. Humanized antibodies may also include at least a portion of an immunoglobulin constant region (Fc), which is typically a human immunoglobulin. For more 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).

[0207] Humanized antibodies or fragments thereof can be produced by replacing the sequences of Fv variable domains that are not directly involved in antigen binding with equivalent sequences of human Fv variable domains. Exemplary methods for producing humanized antibodies or fragments thereof are provided by: Morrison (1985) Science [Science] 229: 1202-1207; Oi et al. (1986) BioTechniques [Biotechnology] 4: 214; and US 5,585,089; US 5,693,761; US ​​5,693,762; US 5,859,205; and US 6,407,213. Those methods include separating, manipulating and expressing nucleic acid sequences encoding all or part of the immunoglobulin Fv variable domains from at least one of the heavy chain or light chain. Such nucleic acids can be obtained from hybridomas that produce antibodies against a predetermined target as described above, as well as other sources. The recombinant DNA encoding the humanized antibody molecule can then be cloned into an appropriate expression vector.

[0208] Humanized antibodies can also be produced using transgenic animals (e.g., mice expressing human heavy and light chain genes but not expressing endogenous mouse immunoglobulin heavy and light chain genes). Winter describes an exemplary CDR transplantation method (U.S. Patent number 5,225,539) that can be used to prepare humanized antibodies as herein described. All CDRs of specific human antibodies can be replaced with at least a portion of non-human CDRs, or only some CDRs can be replaced with non-human CDRs. Only replacement is required for combining humanized antibodies with predetermined antigens with the required CDR quantity.

[0209] Humanized antibodies can be optimized by introducing conservative substitutions, consensus sequence substitutions, germline substitutions and / or back mutations. Such altered immunoglobulin molecules can be prepared by any of several techniques known in the art (e.g., Teng et al., Proc. Natl. Acad. Sci. USA [Proceedings of the National Academy of Sciences of the United States], 80:7308-7312, 1983; Kozbor et al., Immunology Today [Today's Immunology], 4:7279, 1983; Olsson et al., Meth. Enzymol. [Enzymology Methods], 92:3-16, 1982, and EP 239 400).

[0210] The terms "human antibody", "human antigen binding molecules" and "human binding domains" include antibodies, antigen binding molecules and binding domains with antibody regions (such as variable regions and constant regions) or domains substantially corresponding to human germline immunoglobulin sequences known in the art, including those described by, for example, Kabat et al. (1991) (in the above citations). The human antibodies, antigen binding molecules or binding domains of the present invention may include, for example, amino acid residues not encoded by human germline immunoglobulin sequences in CDR and particularly in CDR3 (e.g., mutations introduced by random or side-specific mutagenesis in vitro or by somatic mutations in vivo). Human antibodies, antigen binding molecules or binding domains may have at least one, two, three, four, five or more positions substituted by amino acid residues that are not encoded by human germline immunoglobulin sequences. However, the definitions of human antibodies, antigen binding molecules and binding domains as used herein also encompass "fully human antibodies", which only contain non-artificial and / or genetically altered human antibody sequences, such as those derived using, for example, Xenomouse technology or systems. Preferably, a "fully human antibody" contains no amino acid residues not encoded by human germline immunoglobulin sequences.

[0211] In some embodiments, the antigen binding molecules of the present invention are "isolated" or "substantially pure" antigen binding molecules. When used to describe the antigen binding molecules disclosed herein, "isolated" or "substantially pure" means that the antigen binding molecules have been identified, separated and / or recovered from the components of its production environment. Preferably, the antigen binding molecules do not contain all other components from its production environment or are not substantially related to all other components from its production environment. The contaminating components of its production environment, such as those produced by recombinant transfected cells, are substances that typically interfere with the diagnosis or therapeutic use of polypeptides, and may include enzymes, hormones, and other proteins or non-protein solutes. Antigen binding molecules can, for example, account for at least about 5% or at least about 50% by weight of the total protein in a given sample. It should be understood that, depending on the circumstances, the isolated protein can account for 5% to 99.9% (by weight) of the total protein content. By using an inducible promoter or a high expression promoter, a polypeptide can be prepared at a significantly higher concentration so that it is prepared at an increased concentration level. This definition includes producing antigen binding molecules in a variety of organisms and / or host cells known in the art. In preferred embodiments, the antigen binding molecule is (1) purified to a sufficient extent to obtain at least 15 N-terminal or internal amino acid sequence residues by using a spinning cup sequencer, or (2) purified to homogeneity by SDS-PAGE under non-reducing or reducing conditions using Coomassie blue or preferably silver stain. However, isolated antigen binding molecules are usually prepared by at least one purification step.

[0212] The term "binding domain" in relation to the present invention characterizes a domain that (specifically) binds / interacts / recognizes a given target epitope or a given target flanking end on a target molecule (antigen) (e.g. CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, MSLN, or EpCAM and CD3, respectively). Typically the structure and function of the first and third or second and fourth binding domains (e.g. recognizing CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, MSLN, or EpCAM) and preferably also the structure and / or function of the effector binding domain (typically the second and fourth or first and third binding domain recognizing CD3) are based on the structure and / or function of an antibody (e.g. a full-length or complete immunoglobulin molecule) and / or are extracted from the variable heavy chain (VH) and / or variable light chain (VL) domains of an antibody or a fragment thereof. Preferably, one or more target cell surface antigen binding domains are characterized by the presence of three light chain CDRs (i.e., CDR1, CDR2, and CDR3 in the VL region) and / or three heavy chain CDRs (i.e., CDR1, CDR2, and CDR3 in the VH region). The effector (typically CD3) binding domain also preferably comprises the minimum structural requirements of an antibody that allows target binding. More preferably, the second binding domain comprises at least three light chain CDRs (i.e., CDR1, CDR2, and CDR3 in the VL region) and / or three heavy chain CDRs (i.e., CDR1, CDR2, and CDR3 in the VH region). It is envisioned that the first binding domain and / or the second binding domain are produced or obtainable by phage display or library screening methods, rather than being produced or obtainable by transplanting CDR sequences from pre-existing (monoclonal) antibodies into a scaffold.

[0213] 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 or a chemical cross-linking agent, such as glutaraldehyde). Proteins (including fragments thereof, preferably biologically active fragments, and peptides generally having less than 30 amino acids) comprise two or more amino acids coupled to each other via covalent peptide bonds (producing an amino acid chain).

[0214] As used herein, the term "polypeptide" describes a group of molecules, which are generally composed of more than 30 amino acids. Polypeptides can further form polymers, such as dimers, trimers and higher oligomers, i.e., composed of more than one polypeptide molecule. The polypeptide molecules forming such dimers, trimers, etc. can be identical or different. Therefore, the corresponding high-order structures of such polymers are called homo- or hetero-dimers, homo- or hetero-trimers, etc. An example of a hetero-multimer is an antibody molecule, and its naturally occurring form is composed of two identical polypeptide light chains and two identical polypeptide heavy chains. The terms "peptide", "polypeptide" and "protein" also refer to naturally modified peptides / polypeptides / proteins, wherein the modification is achieved, for example, by post-translational modification (such as glycosylation, acetylation, phosphorylation, etc.). When mentioned in this article, "peptide", "polypeptide" or "protein" can also be chemically modified, such as pegylation. Such modifications are well known in the art and described below.

[0215] Preferably, the binding domain of any one of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN and EpCAM and / or the binding domain of CD3ε is a human binding domain. Antibodies and antigen-binding molecules comprising at least one human binding domain avoid some problems associated with antibodies or antigen-binding molecules with non-human such as rodents (e.g., mice, rats, hamsters or rabbits) variable regions and / or constant regions. The presence of such rodent-derived proteins can result in rapid removal of antibodies or antigen-binding molecules, or can result in the patient generating an immune response to antibodies or antigen-binding molecules. In order to avoid the use of rodent-derived antibodies or antigen-binding molecules, people or fully human antibodies / antigen-binding molecules can be produced by introducing human antibody functions into rodents so that rodents produce fully human antibodies.

[0216] The ability to clone and reconstruct megabase-sized human loci in yeast artificial chromosomes (YACs) and introduce them into the mouse germline provides a powerful approach for elucidating the functional components of very large or roughly located loci and for generating useful human disease models. In addition, using this technology to replace mouse loci with their human equivalents can provide 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.

[0217] The important practical application of this strategy is the "humanization" of the mouse humoral immune system. Human immunoglobulin (Ig) locus is introduced into the mouse in which endogenous Ig gene has been inactivated, and the programmed expression of antibody and the fundamental mechanism of assembly and the opportunity of its effect in B cell development are provided. In addition, this strategy can provide an ideal source for the generation of fully human monoclonal antibodies (mAb)-this is an important milestone that helps to realize the prospect of antibody therapy in human diseases. It is expected that fully human antibodies or antigen binding molecules will minimize the immunogenicity and allergic response inherent to mouse or mouse-derived mAb, and thus increase the effect and safety of the antibody / antigen binding molecules used. It can be expected that the use of fully human antibodies or antigen binding molecules in the treatment of chronic and recurrent human diseases such as inflammation, autoimmunity and cancer that require repeated compound administration provides significant advantages.

[0218] One method to achieve this goal is to engineer mouse strains with large fragments of human Ig loci to produce defective mouse antibodies, which are expected to produce large human antibody repertoires in the absence of mouse antibodies. Large human Ig fragments will maintain large variable gene diversity and appropriate regulation of antibody production and expression. By utilizing mouse mechanisms to achieve antibody diversification and selection and lack of immune tolerance to human proteins, the human antibody repertoire regenerated in these mouse strains should produce high-affinity antibodies for any target antigen (including human antigens). Using hybridoma technology, antigen-specific human mAbs with desired specificity can be easily produced and selected. This general strategy was demonstrated in conjunction with the production of the first XenoMouse mouse strain (see Green et al. Nature Genetics [Natural Genetics] 7: 13-21 (1994)). YAC-engineered XenoMouse strains containing germline configuration fragments of 245kb and 190kb in size, respectively, containing human heavy chain loci and κ light chain loci, contain core variable and constant region sequences. It is proved that YAC containing human Ig is compatible with mouse system to rearrange and express antibodies, and can replace inactivated mouse Ig genes. This is proved by its ability to induce B cell development, produce adult-like human repertoire of fully human antibodies and produce antigen-specific human mAbs. These results also show that the introduction of a larger portion of human Ig loci containing a greater number of V genes, additional regulatory elements and human Ig constant regions can basically reproduce the complete repertoire as a feature of human humoral responses to infection and immunity. The work of Green et al. has recently been extended to the introduction of a human antibody repertoire greater than about 80% by introducing a germline configuration YAC fragment of a megabase-sized human heavy chain locus and a kappa light chain locus, respectively. See Mendez et al. Nature Genetics [Natural Genetics] 15: 146-156 (1997) and U.S. Patent Application Serial No. 08 / 759,620.

[0219] The generation of XenoMouse animals is further discussed and described in the following: U.S. Patent Applications Serial 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 / 46 ...6,649, 08 / 031,801, 08 / 112,848, 08 / 4,582, serial number 08 / 463,191, serial number 08 / 462,837, serial number 08 / 486,853, serial number 08 / 486,857, serial number 08 / 486,859, serial number 08 / 462,513, serial number 08 / 724,752 and serial number 08 / 759,620; and U.S. Patent Nos. 6,162,963; 6,150,584; 6,114,598; 6,075,181 and 5,939,598 and Japanese Patent Nos. 3 068 180 B2, 3 068 506 B2, and 3 068 507 B2. See also Mendez et al. Nature Genetics 15:146-156 (1997) and Green and Jakobovits J. Exp. Med. 188:483-495 (1998), EP 0 463 151 Bl, WO 94 / 02602, WO 96 / 34096, WO 98 / 24893, WO 00 / 76310 and WO 03 / 47336.

[0220] In an alternative approach, other companies, including GenPharm International, Inc., have utilized a "minilocus" approach. In the minilocus approach, an exogenous Ig locus is mimicked by including fragments (individual genes) from the Ig locus. Thus, one or more VH genes, one or more DH genes, one or more JH genes, a mu constant region, and a second constant region (preferably a gamma constant region) are formed into a construct for insertion into an animal. This method is described in: U.S. Pat. Nos. 5,545,807 and 5,545,806 to Surani et al.; 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 (each to Lonberg and Kay), U.S. Pat. Nos. 5,591,669 and 6,023.010 to Krimpenfort and Berns, U.S. Pat. No. 5,612,205 to Berns et al.; 21,367 and 5,789,215, and U.S. Patent No. 5,643,763 to Choi and Dunn, and GenPharm International U.S. Patent Application Serial 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. See also EP 0 546 073 B1, 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 and U.S. Pat. No. 5,981,175. See further 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).

[0221] Kirin has also demonstrated the production of human antibodies from mice into which large segments of chromosomes or entire chromosomes have been introduced by microcell fusion. See European Patent Application Nos. 773 288 and 843 961. Xenerex Biosciences is developing a technology for the potential production of human antibodies. In this technology, SCID mice are reconstituted with human lymphocytes (e.g., B and / or T cells). The mice are then immunized with an antigen and an immune response to the antigen can be generated. See U.S. Patent Nos. 5,476,996; 5,698,767; and 5,958,765.

[0222] Human anti-mouse antibody (HAMA) responses have led the industry to prepare chimeric or other humanized antibodies. However, it is expected that certain human anti-chimeric antibody (HACA) responses will be observed, particularly in long-term or multi-dose use of antibodies. Therefore, it is desirable to provide antigen binding molecules comprising human binding domains for CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN or EpCAM and human binding domains for CD3ε to address the problems and / or effects of HAMA or HACA responses.

[0223] The terms "(specifically) bind to", "(specifically) recognize", "(specifically) target" and "(specifically) react with" mean that according to the present invention, the binding domain interacts or specifically interacts with a given epitope or a given target side on a target molecule (antigen) (here: CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN or EpCAM and CD3ε as effectors, respectively).

[0224] The term "epitope" refers to the side 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 therefore the term epitope is sometimes also referred to herein as an "antigenic structure" or "antigenic determinant". Thus, a binding domain is the "antigen interaction side". The binding / interaction is also understood to define "specific recognition".

[0225] An "epitope" can be formed by consecutive amino acids or non-contiguous amino acids juxtaposed by the tertiary folding of a protein. A "linear epitope" is an epitope in which the primary sequence of amino acids comprises the recognized epitope. A linear epitope typically includes at least 3 or at least 4, and more usually at least 5 or at least 6 or at least 7, e.g., about 8 to about 10 amino acids in a unique sequence.

[0226] In contrast to linear epitopes, a "conformational epitope" is an epitope in which the primary sequence of amino acids constituting the epitope is not the only defining component of the recognized epitope (e.g., an epitope in which the primary sequence of amino acids is not necessarily recognized by a binding domain). Typically, a conformational epitope comprises an increased number of amino acids relative to a linear epitope. With respect to recognition of conformational epitopes, the binding domain recognizes the three-dimensional structure of an antigen, preferably a peptide or protein, or a fragment thereof (in the context of the present invention, the antigenic structure of a binding domain is included in a target cell surface antigen protein). For example, when a protein molecule folds to form a three-dimensional structure, certain amino acids and / or polypeptide backbones that form a conformational epitope are juxtaposed, enabling an antibody to recognize the epitope. Methods for determining epitope conformation include, but are not limited to, x-ray crystallography, two-dimensional nuclear magnetic resonance (2D-NMR) spectroscopy, and site-directed spin labeling and electron paramagnetic resonance (EPR) spectroscopy.

[0227] The following describes a method for epitope mapping: When a region (contiguous amino acid stretch) in a human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM protein is exchanged or replaced with the corresponding region of its 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 animals such as chicken, rat, hamster, rabbit, etc. are also envisioned), reduced binding of the binding domain is expected to occur unless the binding domain is cross-reactive to the non-human, non-primate CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN or EpCAM used. The reduction is preferably at least 10%, 20%, 30%, 40%, or 50%, more preferably at least 60%, 70% or 80%, and most preferably 90%, 95% or even 100% compared to the binding to the corresponding region in human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM protein, thereby setting the binding to the corresponding region in human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM protein to 100%. It is envisioned that the aforementioned 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 chimeras are expressed in CHO cells. It is also envisioned that 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 chimeras are fused to transmembrane domains and / or cytoplasmic domains of different membrane-bound proteins (such as EpCAM).

[0228] In an alternative or additional approach to epitope mapping, several truncated forms of human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM extracellular domains can be generated to determine the specific region recognized by the binding domain. In these truncated forms, different extracellular CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM domains / subdomains or regions are progressively deleted starting from the N-terminus. It is contemplated that truncated forms of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM can be expressed in CHO cells. It is also envisioned that truncated CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM forms can be fused with transmembrane domains and / or cytoplasmic domains of different membrane-bound proteins (e.g., EpCAM). It is also envisioned that truncated CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM forms can contain signal peptide domains at their N-terminals, such as signal peptides derived from mouse IgG heavy chain signal peptides. It is also envisioned that truncated CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM forms can contain v5 domains at their N-terminals (after the signal peptide), which allows verification of its correct expression on the cell surface. It is expected that a reduction or loss of binding will occur for truncated forms of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM that no longer contain the region of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM recognized by the binding domain. The reduction in binding is preferably at least 10%, 20%, 30%, 40%, 50%; more preferably at least 60%, 70%, 80%, and most preferably 90%, 95% or even 100%, thereby setting the binding to the entire human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM protein (or its extracellular region or domain) to 100.

[0229] Another method to determine the contribution of specific residues of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN or EpCAM to the recognition of an antigen binding molecule or binding domain is alanine scanning (see, e.g., Morrison KL and Weiss GA. Cur Opin Chem Biol. 2001 Jun;5(3):302-7), in which each residue to be analyzed is replaced with alanine, e.g., via site-directed mutagenesis. Alanine is used because it has a non-bulky, chemically inert methyl functional group, but still mimics the secondary structure references that many other amino acids have. In situations where the size of the mutated residue is desired to be conservative, bulky amino acids (such as valine or leucine) can sometimes be used. Alanine scanning is a mature technique that has been used for a long time.

[0230] The interaction between a binding domain and an epitope or epitope-containing region means that the binding domain exhibits appreciable affinity for the epitope / epitope-containing region on a particular protein or antigen (here: CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN or EpCAM and CD3, respectively), and generally does not exhibit significant reactivity with proteins or antigens other than CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM or CD3. "Appreciable affinity" includes binding with an affinity of about 10 -6 Preferably, when the binding affinity is about 10 -12 Up to 10 -8 M, 10 -12 Up to 10 -9 M, 10 -12 Up to 10 -10 M, 10 -11 Up to 10 -8 M, preferably about 10 -11 Up to 10 -9M, the binding is considered to be specific. Whether the binding domain specifically reacts or binds to the target can be easily tested in the following manner: compare the reaction of the binding domain with the target protein or antigen with the reaction of the binding domain with proteins or antigens 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 substantially or substantially bind to proteins or antigens other than target cell surface antigens CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM or CD3 (i.e., the first binding domain cannot bind to proteins other than CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM, and the second binding domain cannot bind to proteins other than CD3). The envisioned feature of the antigen binding molecule according to the present invention is that it has excellent affinity characteristics compared to other HLE forms. Therefore, this excellent affinity shows that the half-life in vivo is extended. The longer half-life of the antigen binding molecules according to the present invention can reduce the duration and frequency of administration, which typically helps to improve patient compliance. This is particularly important because the antigen binding molecules of the present invention are particularly beneficial to highly weak or even multi-morbid cancer patients.

[0231] The term "substantially / essentially does not bind" or "cannot bind" means that the binding domain of the present invention does not bind to proteins or antigens other than CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM or CD3 as effectors, that is, it does not show more than 30%, preferably not more than 20%, more preferably not more than 10%, and particularly preferably not more than 9%, 8%, 7%, 6% or 5% reactivity with proteins or antigens other than CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM or CD3 as effectors, thereby setting the binding to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CDH3, MSLN, or EpCAM or CD3 as effectors to 100%, respectively.

[0232] It is believed that specific binding is realized by the specific motif in the amino acid sequence of binding domain and antigen.Therefore, due to the secondary modification of its one-level, two-level and / or three-level structure and described structure, combination is realized.The specific interaction of antigen interaction side end and its specific antigen can cause the simple combination of described side end and antigen.In addition, the specific interaction of antigen interaction side end and its specific antigen can cause the initiation of signal alternatively or additionally, for example due to the variation of inducing antigen conformation, oligomerization of antigen etc.

[0233] The term "variable" refers to the portion of an antibody or immunoglobulin domain that exhibits its sequence variability and is involved in determining the specificity and binding affinity of a particular antibody (i.e., "one or more variable domains"). The pairing of a variable heavy chain (VH) and a variable light chain (VL) together forms a single antigen binding site.

[0234] Variability is not evenly distributed throughout the variable domains of the antibody; it is concentrated in each of the subdomains of the heavy chain variable region and the light chain variable region. These subdomains are referred to as "hypervariable regions" or "complementarity determining regions" (CDRs). The more conservative (i.e., non-hypervariable) parts of the variable domains are referred to as "framework" regions (FRMs or FRs), and provide a scaffold for the six CDRs in three-dimensional space to form an antigen binding surface. The variable domains of naturally occurring heavy and light chains each include four FRM regions (FR1, FR2, FR3, and FR4), which are mainly connected by three hypervariable regions, and these three hypervariable regions form a loop connecting the β-folded structure, and form a part of the β-folded structure in some cases. The hypervariable regions in each chain are closely together by FRMs, and together with the hypervariable regions from another chain, contribute to the formation of the antigen binding side end (see Kabat et al., cited above).

[0235] The term "CDR" and its plural "CDR" refer to the complementarity determining regions of which three constitute the binding characteristics of the light chain variable region (CDR-L1, CDR-L2 and CDR-L3) and three 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 therefore contribute to the functional activity of the antibody molecule: they are the main determinants of antigen specificity.

[0236] Accurately defined CDR boundaries and lengths are subject to different classification and numbering systems. Therefore, CDRs can be referenced by Kabat, Chothia, contact or any other boundary definition (including numbering systems described herein). Despite different boundaries, each of these systems has a certain degree of overlap in constituting the so-called "hypervariable region" in the variable sequence. Therefore, the CDR definitions according to these systems can be different in length and boundary regions relative to adjacent framework regions. See, for example, Kabat (a method based on cross-species sequence variability), Chothia (a method based on crystallographic studies of antigen-antibody complexes) and / or MacCallum (Kabat et al., cited above; Chothia et al., J.MoI.Biol [Journal of Molecular Biology], 1987, 196: 901-917; and MacCallum et al., J.MoI.Biol [Journal of Molecular Biology], 1996, 262: 732). Yet another standard for characterizing the antigen-binding side end is the AbM definition used by the AbM antibody modeling software of Oxford University Molecular. See, for example, Protein Sequence and Structure Analysis of Antibody Variable Domains in: Antibody Engineering Lab Manual (Editors: Duebel, S. and Kontermann, R., Springer-Verlag, Heidelberg). To the extent that the two residue identification techniques define overlapping regions rather than identical regions, they can be combined to define hybrid CDRs. However, numbering according to the so-called Kabat system is preferred.

[0237] Typically, CDR forms a loop structure that can be classified as a canonical structure. The term "canonical structure" refers to the main chain conformation used by the antigen binding (CDR) loop. From comparative structural studies, it has been found that five of the six antigen binding loops have only a limited set of available conformational repertoires. Each canonical structure can be characterized by the torsion angle of the polypeptide backbone. Therefore, the corresponding loops between antibodies can have very similar three-dimensional structures, but most of the loops have high amino acid sequence variability (Chothia and Lesk, J.MoI.Biol. [Molecular Biology Journal], 1987, 196: 901; Chothia et al., Nature [Nature], 1989, 342: 877; Martin and Thornton, J.MoI.Biol [Molecular Biology Journal], 1996, 263: 800). In addition, there is a relationship between the loop structure used and the amino acid sequence around it. The conformation of a specific canonical class is determined by the length of the loop and the amino acid residues located in the key positions within the loop and in the conservative framework (i.e., outside the loop). Therefore, assignment to a specific canonical class can be made based on the presence of these key amino acid residues.

[0238] The term "canonical structure" can also include considerations about the linear sequence of an antibody, for example, as catalogued by Kabat (Kabat et al., cited above). The Kabat numbering scheme (system) is a widely used standard for numbering the amino acid residues of the variable domains of an antibody in a consistent manner, and is a preferred embodiment of the present invention, as also mentioned elsewhere herein. Additional structural considerations can also be used to determine the canonical structure of an antibody. For example, those differences that are not fully reflected by the Kabat numbering can be described by the numbering system of Chothia et al., and / or revealed by other techniques (such as crystallography and two-dimensional or three-dimensional computational modeling). Therefore, a given antibody sequence can be placed in a canonical category that particularly allows identification of an appropriate basic structure (chassis) sequence (for example, based on the expectation of including a variety of canonical structures in a library). The Kabat numbering of an antibody amino acid sequence and the structural considerations as described by Chothia et al., cited above, and its significance to the canonical aspects of interpreting antibody structure are described in the literature. The subunit structure and three-dimensional configuration of immunoglobulins of different classes are well known in the art. For a review of antibody structure, see Antibodies: A Laboratory Manual, Cold Spring Harbor Laboratory, Harlow et al., eds., 1988.

[0239] The CDR3 of light chain and especially the CDR3 of heavy chain can constitute the most important determinant in antigen binding in light chain variable region and heavy chain variable region. In some antigen binding molecules, heavy chain CDR3 seems to constitute the main contact area between antigen and antibody. The in vitro selection scheme that changes CDR3 separately can be used to change the binding characteristics of antibody or determine which residues contribute to the combination of antigen. Therefore, CDR3 is typically the largest source of molecular diversity in antibody binding side end. For example, H3 can be as short as two amino acid residues or more than 26 amino acids.

[0240] In a classical full-length antibody or immunoglobulin, each light (L) chain is linked to a heavy (H) chain by a 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 named CH1. The constant ("C") domain is not directly involved in antigen binding, but exhibits various effector functions, such as antibody-dependent, cell-mediated cytotoxicity and complement activation. The Fc region of an antibody is included in the heavy chain constant domain and is, for example, able to interact with Fc receptors located on the cell surface.

[0241] The sequences of antibody genes after assembly and somatic mutation are highly altered, and it is estimated that these altered genes encode 10 10 Different antibody molecules (Immunoglobulin Genes [immunoglobulin genes], 2nd edition, Jonio et al., ed., Academic Press [Academic Press], San Diego, CA [San Diego, California], 1995). Therefore, the immune system provides an immunoglobulin repertoire. The term "repertoire" refers to at least one nucleotide sequence derived in whole or in part from at least one sequence encoding at least one immunoglobulin. One or more sequences can be produced by in vivo rearrangement of the V, D and J segments of the heavy chain and the V and J segments of the light chain. Alternatively, one or more sequences can be produced from cells in response to rearrangement, such as in vitro stimulation. Alternatively, a portion or all of one or more sequences can be obtained by DNA splicing, nucleotide synthesis, mutagenesis and other methods, see, for example, U.S. Patent No. 5,565,332. The repertoire can include only one sequence or can include multiple sequences, including sequences in a genetic diversity set.

[0242] The term "Fc portion" or "Fc monomer" means, with respect to the present invention, a polypeptide comprising at least one domain having the function of a CH2 domain and at least one domain having the function of a CH3 domain of an immunoglobulin molecule. It is apparent from the term "Fc monomer" that a polypeptide comprising those CH domains is a "polypeptide monomer". An Fc monomer can be a fragment of an immunoglobulin constant region that at least comprises the first constant region immunoglobulin domain (CH1) excluding the heavy chain, but at least retains a functional portion of a CH2 domain and a functional portion of a CH3 domain, wherein the CH2 domain is at the amino terminal end of the CH3 domain. In a preferred aspect of this definition, an Fc monomer can be a polypeptide constant region comprising a portion of an Ig-Fc hinge region, a CH2 region, and a CH3 region, wherein the hinge region is at the amino terminal end of the CH2 domain. It is envisioned that the hinge region of the present invention promotes dimerization. For example, but not limited to, such Fc polypeptide molecules can be obtained by digesting an immunoglobulin region with papain (of course, producing a dimer of two Fc polypeptides). In another aspect of this definition, an Fc monomer can be a polypeptide region comprising a portion of a CH2 region and a CH3 region. For example, but not limited to, such Fc polypeptide molecules can be obtained by digesting immunoglobulin molecules with pepsin. In one embodiment, the polypeptide sequence of the Fc monomer is substantially similar to the following Fc polypeptide sequences: IgG1 Fc region, IgG2 Fc region, IgG3 Fc region, IgG4 Fc region, IgM Fc region, IgA Fc region, IgD Fc region and IgE Fc region. (See, for example, Padlan, Molecular Immunology [Molecular Immunology], 31 (3), 169-217 (1993)). Because there are some changes between immunoglobulins, and only for the sake of clarity, 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 above, the Fc monomer can also include a flexible hinge at the N-terminus of these domains. For IgA and IgM, the Fc monomer can include a J chain. For IgG, the Fc part comprises immunoglobulin domains CH2 and CH3 and the hinge between the first two domains and CH2. Although the border of the Fc part can change, the example of the human IgG heavy chain Fc part comprising functional hinge, CH2 and CH3 domains can be defined as, for example, P476 comprising residue D231 (residue of hinge domain-corresponding to D234 in the following table 1) to the carboxyl terminal of CH3 domain, respectively L476 (for IgG4), wherein numbering is according to Kabat's numbering. Two Fc parts or Fc monomers fused to each other via a peptide linker are preferred examples of the spacer between two bispecific entities of antigen binding molecules of the present invention, which also can be defined as scFc domains.

[0243] In one embodiment of the present invention, it is envisaged that the scFc domains as disclosed herein, ie the Fc monomers respectively fused to each other, are only comprised in the spacer of the antigen binding molecule.

[0244] According to the present invention, the Kabat numbering listed in Table 1 can be used to identify the IgG hinge region by analogy. Consistent with the above, it is envisioned that for the hinge domain / district of the present invention, the minimum requirement comprises the amino acid residues corresponding to the IgG1 sequence segment of D231 D234 to P243 according to the Kabat numbering. It is also envisioned that the hinge domain / district of the present invention comprises the IgG1 hinge sequence DKTHTCPPCP (SEQ ID NO:330) (corresponding to the segment D234 to P243 as shown in Table 1 below-also envisions the variant of the sequence, provided that the hinge region still promotes dimerization) or is composed of the IgG1 hinge sequence. In a preferred embodiment of the present invention, the glycosylation site at the Kabat position 314 of the CH2 domain in the spacer of the antigen binding molecule is removed by N314X substitution, wherein X is any amino acid except Q. The substitution is preferably N314G substitution. In a more preferred embodiment, the CH2 domain further comprises the following substitutions (positions according to Kabat): V321C and R309C (these substitutions introduce intradomain cysteine ​​disulfide bridges at Kabat positions 309 and 321).

[0245] It is also envisioned that the spacer of the antigen binding molecule of the present invention is a scFc domain that may comprise or consist of the following 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. In a preferred embodiment, the peptide linker of the above-mentioned antigen binding molecule 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, wherein x is an integer of 5 or more (e.g. 5, 6, 7, 8, etc. or more), preferably 6 ((Gly4Ser) 6). According to the present invention, Ser can advantageously be replaced by Gln, as disclosed herein. The construct may further comprise the above-mentioned substitution N314X, preferably N314G and / or additional substitutions V321C and R309C. In a preferred embodiment of the antigen binding molecule of the invention as defined above, it is envisaged that the second domain binds to an extracellular epitope of the human and / or macaque CD3ε chain. Table 1: Kabat numbering of amino acid residues in the hinge region

[0246]

[0247] In another embodiment of the present invention, the hinge domain / region comprises or consists of the following: IgG2 subtype hinge sequence ERKCCVECPPCP (SEQ ID NO: 331), IgG3 subtype hinge sequence ELKTPLDTTHTCPRCP (SEQ ID NO: 332) or ELKTPLGDTTHTCPRCP (SEQ ID NO: 333) and / or IgG4 subtype hinge sequence ESKYGPPCPSCP (SEQ ID NO: 444). The IgG1 subtype hinge sequence can be one of the following EPKSCDKTHTCPPCP (as shown in Table 1 and SEQ ID NO: 445). Therefore, these core hinge regions are also contemplated in the context of the present invention.

[0248] The location and sequence of the IgG CH2 and IgG CD3 domains can be identified by analogy using the Kabat numbering listed in Table 2:

[0249] Table 2: Kabat numbering of amino acid residues in the CH2 and CH3 regions of IgG

[0250]

[0251] In one embodiment of the present invention, the amino acid residues highlighted in bold in the CH3 domain of the first or two Fc monomers are deleted.

[0252] The peptide linker to which the polypeptide monomers of the spacer ("Fc part" or "Fc monomer") are fused to each other preferably comprises at least 25 amino acid residues (25, 26, 27, 28, 29, 30, etc.). More preferably, this peptide linker comprises at least 30 amino acid residues (30, 31, 32, 33, 34, 35, etc.). It is also preferred that the linker comprises at most 40 amino acid residues, more preferably at most 35 amino acid residues, most preferably exactly 30 amino acid residues. A preferred embodiment of this 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, wherein x is an integer of 5 or greater (e.g. 6, 7 or 8). Preferably, the integer is 6 or 7, more preferably the integer is 6.

[0253] In the case of using a joint to fuse the first domain with the second domain, and / or the third domain with the fourth domain, and / or the second and third domains with a spacer, the joint preferably has a length and sequence sufficient to ensure that each of the first domain and the second domain can retain its difference binding specificity independently of each other. For connecting the peptide joint of at least two binding domains (or two variable domains) in the antigen binding molecules of the present invention, preferably those peptide joints comprising only a small amount of amino acid residues such as 12 amino acid residues or less. Therefore, the peptide joints of 12,11,10,9,8,7,6 or 5 amino acid residues are preferred. It is envisioned that the peptide joint with less than 5 amino acids comprises 4,3,2 or 1 amino acid, wherein the joint rich in Gly is preferred. The preferred embodiment of the peptide joint for fusing the first domain and the second domain is described in SEQ ID NO:1. The preferred joint embodiment for the peptide joint for fusing the second and third domains to the spacer is (Gly) 4-joint, also referred to as G4-joint.

[0254] In the context of the above-mentioned "peptide linker", a particularly preferred "single" amino acid is Gly. Thus, the peptide linker can be composed of a 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, wherein x is an integer of 1 or greater (e.g., 2 or 3). Preferred linkers are depicted in SEQ ID NOs: 1 to 12. The features of the peptide linker including not promoting secondary structure are known in the art and are described in, for example, Dall'Acqua et al. (Biochem. [Biochemistry] (1998) 37, 9266-9273), Cheadle et al. (Mol Immunol [Molecular Immunology] (1992) 29, 21-30) and Raag and Whitlow (FASEB [Journal of the American Federation of Experimental Biology] (1995) 9 (1), 73-80). In addition, a peptide linker that does not promote any secondary structure is preferred. The connection of the domains to each other can be provided, for example, by genetic engineering, as described in the Examples. Methods for preparing fused 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 or Sambrook et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, 2001).

[0255] 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, diabodies and oligomers of any of these forms.

[0256] According to a particularly preferred embodiment, and as described in the attached examples, the first domain and the second domain of the antigen binding molecule of the present invention are "bispecific single-chain antigen binding molecules", more preferably bispecific "single-chain Fv" (scFv). Although the two domains VL and VH of the Fv fragment are encoded by independent genes, they can be joined by synthetic linkers using recombinant methods, as described above, which enable them to be made into a single protein chain, wherein the VL and VH regions are paired to form a monovalent molecule; see, for example, Huston et al. (1988) Proc. Natl. Acad. Sci USA [Proceedings of the National Academy of Sciences of the United States] 85: 5879-5883. These antibody fragments are obtained using conventional techniques known to those skilled in the art, and the function of the fragments is evaluated in the same manner as for complete or full-length antibodies. Therefore, a single-chain variable fragment (scFv) is a fusion protein of the heavy chain (VH) and light chain (VL) variable regions of an immunoglobulin, typically connected using a short linker peptide of about 10 to about 25 amino acids, preferably about 15 to 20 amino acids. The linker is usually rich in glycine for flexibility and serine or threonine for solubility, and can connect the N-terminus of VH and the C-terminus of VL, or vice versa. Despite the removal of the constant regions and the introduction of the linker, the protein retains the specificity of the original immunoglobulin.

[0257] 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; Blood, (2000), 95, 6, 2098-2103; Brühl, Immunol., (2001), 166, 2420-2426; Kipriyanov, J. Mol. Biol., (1999), 293, 41-56. Techniques described for the production of single-chain antibodies (see, inter alia, U.S. Pat. No. 4,946,778; Kontermann and Dübel (2010), loc. cit., and Little (2009), loc. cit.) can be adapted to produce single-chain antigen-binding molecules that specifically recognize one or more selected targets.

[0258] Bivalent (also referred to as divalent) or bispecific single chain variable fragments (two-scFv or two-scFv of the form (scFv) 2) can be engineered by connecting two scFv molecules (e.g., using a linker as described above). If the two scFv molecules have the same binding specificity, the resulting (scFv) 2 molecule will preferably be referred to as bivalent (i.e., having two valencies for the same target epitope). If the two scFv molecules have different binding specificities, the resulting (scFv) 2 molecule will preferably be referred to as bispecific. Connection can be performed by generating a single peptide chain with two VH regions and two VL regions to generate a tandem scFv (see, e.g., Kufer P. et al., (2004) Trends in Biotechnology [Biological Technology Trends] 22 (5): 238-244). Another possibility is to generate scFv molecules with linker peptides that are too short for the two variable regions to fold together (e.g., about five amino acids), thereby forcing the scFv to dimerize. This type is called a diabody (see, e.g., Hollinger, Philipp et al., (July 1993) Proceedings of the National Academy of Sciences of the United States of America 90(14):6444-8).

[0259] According to the present invention, the first domain, the second domain, the third domain and / or the fourth domain may comprise a single domain antibody, respectively the variable domain or at least the CDR of a single domain antibody. A single domain antibody comprises only one (monomer) antibody variable domain that is capable of selectively binding to a specific antigen independently of other V regions or domains. The first single domain antibodies were engineered from heavy chain antibodies found in camels, and these are called V HH fragment. Cartilaginous fish also have heavy chain antibodies (IgNARs), from which V NAR Single domain antibodies of fragments. An alternative method is to split the dimer variable domains from common immunoglobulins, for example from humans or rodents, into monomers, thereby obtaining VH or VL as single domain Abs. Although most of the research on single domain antibodies is based on the heavy chain variable domains at present, it has also been shown that the nano antibody derived from light chain specifically binds to the target epitope. The example of single domain antibodies is so-called sdAb, nano antibody or single variable domain antibody.

[0260] Thus, (single domain mAb)2 is a monoclonal antigen binding molecule composed of (at least) two single domain monoclonal antibodies selected individually from the group consisting of V H 、V L 、V H H and V NAR The linker is preferably in the form of a peptide linker. Similarly, a "scFv single domain mAb" is a monoclonal antigen binding molecule consisting of at least one single domain antibody as described above and one scFv molecule as described above. Likewise, the linker is preferably in the form of a peptide linker.

[0261] Can determine in competition determination (such as competitive ELISA or cell-based competition determination) whether antigen binding molecule competes with the combination of another given antigen binding molecule.Can also use avidin coupled microparticles (beads).Similar to the ELISA plate coated with avidin, when reacting with biotinylated protein, each of these beads can be used as a substrate that can be measured thereon.Antigen is coated on the beads, and then pre-coated with the first antibody.Add the second antibody and determine any additional combination.Possible means for reading include flow cytometry.

[0262] T cells or T lymphocytes are a type of lymphocyte (itself a type of white blood cell) that plays a central role in cell-mediated immunity. There are several T cell subsets, each with different functions. T cells can be distinguished from other lymphocytes (such as B cells and NK cells) by the presence of T cell receptors (TCRs) on the cell surface. TCRs are responsible for recognizing antigens bound to major histocompatibility complex (MHC) molecules and are composed of two different protein chains. In 95% of T cells, TCRs are composed of alpha (α) and beta (β) chains. When TCRs engage with antigenic peptides and MHC (peptide / MHC complexes), T lymphocytes are activated through a series of biochemical events mediated by related enzymes, co-receptors, specialized adapter molecules, and activated or released transcription factors.

[0263] The CD3 receptor complex is a protein complex and is composed of four chains. In mammals, the complex contains CD3γ (gamma) chain, CD3δ (delta) chain and two CD3ε (ipsilon) chains. These chains associate with T cell receptor (TCR) and so-called ζ (zeta) chain to form T cell receptor CD3 complex and generate activation signals in T lymphocytes. CD3γ (gamma), CD3δ (delta) and CD3ε (ipsilon) chains are highly correlated cell surface proteins of the immunoglobulin superfamily containing a single extracellular immunoglobulin domain. The intracellular tail of the CD3 molecule contains a single conserved motif necessary for the signal transduction ability of TCR, referred to as an activation motif based on immunoreceptor tyrosine or ITAM for short. The CD3ε molecule is a polypeptide, which is encoded by the CD3E gene on chromosome 11 in humans. The most preferred epitope of CD3ε is included in the amino acid residues 1-27 of the human CD3ε extracellular domain. It is envisaged that the antigen binding molecules according to the invention typically and advantageously show less non-specific T cell activation, which is undesirable in specific immunotherapy. This means that the risk of side effects is reduced.

[0264] Redirected lysis of target cells by recruiting T cells via multi-chain multi-targeting, at least bispecific antigen binding molecules involves cytolytic synapse formation and delivery of perforins and granzymes. Engaged T cells are capable of continuous target cell lysis and are not affected by immune escape mechanisms that interfere with peptide antigen processing and presentation or clonal T cell differentiation; see, e.g., WO2007 / 042261.

[0265] The cytotoxicity mediated by the antigen binding molecules of the present invention can be measured in various ways. Effector cells can be (people) CD8 positive T cells or unstimulated (people) peripheral blood mononuclear cells (PBMC) of enrichment such as stimulation. If target cells are macaque origin or macaque CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN or EpCAM transfection expressed or combined with the first domain, then effector cells should also be macaque origin, such as macaque T cell line, such as 4119LnPx. Target cells should express CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN or EpCAM, such as people or macaque CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN or EpCAM (at least the extracellular domain). The target cell can be a cell line (such as CHO) stably or transiently transfected with CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM, such as human or macaque CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM. Generally, EC 50 The value is lower, wherein the target cell line expresses higher levels of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM on the cell surface. The effector cell to target cell (E:T) ratio is generally about 10:1, but can also be varied. The cytotoxic activity of the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM bispecific antigen binding molecules can be 51 The cytotoxicity of the cells can be measured in a Cr-release assay (incubation time of about 18 hours) or in a FACS-based cytotoxicity assay (incubation time of about 48 hours). Modifications to the assay incubation time (cytotoxic response) are also possible. Other methods of measuring cytotoxicity are well known to those skilled in the art and include MTT or MTS assays, ATP-based assays (including bioluminescent assays), sulforhodamine B (SRB) assays, WST assays, clonogenic assays, and ECIS techniques.

[0266] The cytotoxic activity mediated by the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAMxCD3 bispecific antigen binding molecules of the invention is preferably measured in a cell-based cytotoxicity assay. 51 The cytotoxic activity was measured by EC 50The value is expressed as, which corresponds to the half-maximal effective concentration (the concentration of the antigen binding molecule that induces a cytotoxic response between the baseline and the maximum value). Preferably, the EC of the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAMxCD3 bispecific antigen binding molecule is 50 Values ​​are ≤5000pM or ≤4000pM, more preferably ≤3000pM or ≤2000pM, even more preferably ≤1000pM or ≤500pM, even more preferably ≤400pM or ≤300pM, even more preferably ≤200pM, even more preferably ≤100pM, even more preferably ≤50pM, even more preferably ≤20pM or ≤10pM, and most preferably ≤5pM.

[0267] The EC given above 50 The value can be measured in different assays. Those skilled in the art will appreciate that when using stimulated / enriched CD8 + When T cells are used as effector cells, ECs can be expected to be 50 In addition, it can be expected that ECs would be more sensitive to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM when target cells expressed high amounts compared to rats with low target expression. 50 For example, when using stimulated / enriched human CD8 + When T cells are used as effector cells (and cells transfected with CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM, such as CHO cells or CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM-positive human cell lines are used as target cells), the ECs of the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAMxCD3 bispecific antigen binding molecules are 50 The 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 PBMCs are used as effector cells, the EC values ​​of the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAMxCD3 bispecific antigen binding molecules are 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. 50The value is preferably ≤5000pM or ≤4000pM (especially when the target cells are CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM human cell lines), more preferably ≤2000pM (especially when the target cells are CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM transfected cells such as CHO cells), more preferably ≤1000pM or ≤500pM, even more preferably ≤200pM, even more preferably ≤150pM, even more preferably ≤100pM, and most preferably ≤50pM or lower. When a macaque T cell line such as LnPx4119 is used as an effector cell and a macaque CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM-transfected cell line such as CHO cells is used as a target cell line, the EC of the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAMxCD3 bispecific antigen binding molecule is 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, even more preferably ≤ 100 pM, and most preferably ≤ 50 pM.

[0268] Preferably, the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAMxCD3 bispecific antigen binding molecules of the present invention do not induce / mediate lysis or substantially do not induce / mediate lysis of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM-negative cells (e.g., CHO cells). The terms "do not induce lysis", "do not substantially induce lysis", "do not mediate lysis" or "do not substantially mediate lysis" mean that the antigen binding molecules of the present invention induce or mediate no more than 30% lysis of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM negative cells, preferably no more than 20%, more preferably no more than 10%, particularly preferably no more than 9%, 8%, 7%, 6% or 5%, whereby the lysis of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM positive human cell lines is set to 100%. This generally applies to antigen binding molecules at concentrations up to 500 nM. Those skilled in the art know how to measure cell lysis without difficulty. In addition, this specification teaches specific instructions on how to measure cell lysis.

[0269] The difference in cytotoxic activity between monomeric and dimeric isoforms of a single CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAMxCD3 bispecific antigen binding molecule is referred to as the "potency gap". The potency gap can be calculated, for example, as the EC of the monomeric versus dimeric forms of the molecule. 50 The potency gap of the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAMxCD3 bispecific antigen binding molecules of the present invention is preferably ≤5, more preferably ≤4, even more preferably ≤3, even more preferably ≤2, and most preferably ≤1.

[0270] The first, second, third and / or fourth binding domains of the antigen binding molecules of the present invention preferably have cross-species specificity to members of the order Mammalia of primates. Interspecific CD3 binding domains are, for example, those described herein and in WO2008 / 119567. According to one embodiment, the first binding domain and the third binding domain, in addition to being respectively bound to human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM and human CD3, will also be bound to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM / CD3 of primates, including, but not limited to, New World primates (such as woolly monkeys, woolly tamarins or squirrel monkeys), Old World primates (such as baboons and macaques), gibbons, and non-human Homininae.

[0271] 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 macaque CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM, such as CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM of cynomolgus macaque (Macaca fascicularis), and more preferably, binds to macaque CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM expressed on the surface of cells (such as, for example, CHO or 293 cells). The affinity of the first domain to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM, preferably to 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. Affinity can be measured, for example, in 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 macaque CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM is preferably ≤15nM, more preferably ≤10nM, even more preferably ≤5nM, even more preferably ≤1nM, even more preferably ≤0.5nM, even more preferably ≤0.1nM, and most preferably ≤0.05nM or even ≤0.01nM.

[0272] Preferably, the antigen binding molecules according to the invention are for binding to macaque CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM versus human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM [ma CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM:hu CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM] (as determined, for example, by surface plasmon resonance analysis such as BiaCore TM or determined by Scatchard analysis) <100, preferably <20, more preferably <15, further preferably <10, even more preferably <8, more preferably <6 and most preferably <2. The preferred range of the difference in affinity of the antigen binding molecules according to the invention for binding to macaque CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM versus human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM is between 0.1 and 20, more preferably between 0.2 and 10, even more preferably between 0.3 and 6, even more preferably between 0.5 and 3 or between 0.5 and 2.5, and most preferably between 0.5 and 2 or between 0.6 and 2.

[0273] The second and fourth binding domains of the antigen binding molecules of the present invention are generally in combination with human CD3ε and / or macaque CD3ε. In a preferred embodiment to achieve a selectivity gap, the second and fourth binding domains, or alternatively the first and third binding domains further bind to common marmosets, cotton-top tamarins or squirrel monkey CD3ε. Both marmosets and tamarins are New World primates belonging to the subfamily Callitrichidae, while squirrel monkeys are New World primates belonging to the family Cebidae. The binding domain may preferably be selected from sequences identified herein as "I2L" (or synonymously "I2L0"), "I2M" and "I2M2", more preferably "I2L" or "I2L0".

[0274] It is preferred for the antigen-binding molecules of the present invention that the preferred second and fourth binding domains that bind to the extracellular epitopes of human and / or macaque CD3ε chain comprise a VL region comprising CDR-L1, CDR-L2 and CDR-L3 selected from the group consisting of:

[0275] (a) a VL region comprising CDR-L1, CDR-L2 and CDR-L3 selected from SEQ ID NOs 40 to 42, 48 to 50, 56 to 58, 64 to 66, 72 to 74439 to 441, preferably 64 to 66

[0276] (b) CDR-L1 as depicted in SEQ ID NO:27 of WO 2008 / 119567, CDR-L2 as depicted in SEQ ID NO:28 of WO 2008 / 119567 and CDR-L3 as depicted in SEQ ID NO:29 of WO 2008 / 119567;

[0277] (c) CDR-L1 as depicted in SEQ ID NO: 117 of WO 2008 / 119567, CDR-L2 as depicted in SEQ ID NO: 118 of WO 2008 / 119567, and CDR-L3 as depicted in SEQ ID NO: 119 of WO 2008 / 119567;

[0278] (d) CDR-L1 as depicted in SEQ ID NO: 153 of WO 2008 / 119567, CDR-L2 as depicted in SEQ ID NO: 154 of WO 2008 / 119567, and CDR-L3 as depicted in SEQ ID NO: 155 of WO 2008 / 119567; and

[0279] (e) a VL region comprising CDR-L1, CDR-L2 and CDR-L3 of SEQ ID NOs 420 to 422.

[0280] In a further preferred embodiment of the antigen binding molecule of the present invention, the preferred second and fourth binding domains that bind to the extracellular epitope of human and / or macaque CD3ε chain comprise a VH region comprising CDR-H1, CDR-H2 and CDR-H3 selected from the group consisting of:

[0281] (a) a VH region comprising 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;

[0282] (b) CDR-H1 as depicted in SEQ ID NO: 12 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO: 13 of WO 2008 / 119567 and CDR-H3 as depicted in SEQ ID NO: 14 of WO 2008 / 119567;

[0283] (c) CDR-H1 as depicted in SEQ ID NO:30 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO:31 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO:32 of WO 2008 / 119567;

[0284] (d) CDR-H1 as depicted in SEQ ID NO:48 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO:49 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO:50 of WO 2008 / 119567;

[0285] (e) CDR-H1 as depicted in SEQ ID NO:66 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO:67 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO:68 of WO 2008 / 119567;

[0286] (f) CDR-H1 as depicted in SEQ ID NO:84 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO:85 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO:86 of WO 2008 / 119567;

[0287] (g) CDR-H1 as depicted in SEQ ID NO: 102 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO: 103 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO: 104 of WO 2008 / 119567;

[0288] (h) CDR-H1 as depicted in SEQ ID NO: 120 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO: 121 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO: 122 of WO 2008 / 119567;

[0289] (i) CDR-H1 as depicted in SEQ ID NO: 138 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO: 139 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO: 140 of WO 2008 / 119567;

[0290] (j) CDR-H1 as depicted in SEQ ID NO: 156 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO: 157 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO: 158 of WO 2008 / 119567;

[0291] (k) CDR-H1 as depicted in SEQ ID NO: 174 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO: 175 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO: 176 of WO 2008 / 119567; and

[0292] (l) a VH region comprising CDR-H1, CDR-H2 and CDR-H3 of SEQ ID NOs 423 to 425.

[0293] 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 within the third binding domain to form (30) groups, each group comprising CDR-L 1-3 and CDR-H 1-3.

[0294] Preferably, for the antigen binding molecule of the present invention, the third domain that binds to CD3 comprises a VL region selected from the group consisting of: SEQ ID NO: 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 or preferably, SEQ ID NO: 44, 52, 60, 68 and 76 according to the present invention, preferably 68.

[0295] It is also preferred that the third domain binding to CD3 comprises a VH region selected from the group consisting of: SEQ ID NO: 15, 19, 33, 37, 51, 55, 69, 73, 87, 91, 105, 109, 123, 127, 141, 145, 159, 163, 177 or 181 of WO2008 / 119567 or preferably, SEQ ID NO: 43, 51, 59, 67 and 75 according to the present invention, preferably 67.

[0296] More preferably, the antigen binding molecule of the present invention is characterized by binding to the preferred second and fourth domains of CD3 comprising a VL region and a VH region selected from the group consisting of:

[0297] (a) a VL region selected from the group consisting of SEQ ID NOs 44, 52, 60, 68, 76 and 443, and a VH region selected from the group consisting of SEQ ID NOs 43, 51, 59, 67, 75 and 442;

[0298] (b) a VL region as depicted in SEQ ID NO: 17 or 21 of WO 2008 / 119567 and a VH region as depicted in SEQ ID NO: 15 or 19 of WO 2008 / 119567;

[0299] (c) a VL region as depicted in SEQ ID NO: 35 or 39 of WO 2008 / 119567 and a VH region as depicted in SEQ ID NO: 33 or 37 of WO 2008 / 119567;

[0300] (d) a VL region as depicted in SEQ ID NO: 53 or 57 of WO 2008 / 119567 and a VH region as depicted in SEQ ID NO: 51 or 55 of WO 2008 / 119567;

[0301] (e) a VL region as depicted in SEQ ID NO: 71 or 75 of WO 2008 / 119567 and a VH region as depicted in SEQ ID NO: 69 or 73 of WO 2008 / 119567;

[0302] (f) a VL region as depicted in SEQ ID NO: 89 or 93 of WO 2008 / 119567 and a VH region as depicted in SEQ ID NO: 87 or 91 of WO 2008 / 119567;

[0303] (g) a VL region as depicted in SEQ ID NO: 107 or 111 of WO 2008 / 119567 and a VH region as depicted in SEQ ID NO: 105 or 109 of WO 2008 / 119567;

[0304] (h) a VL region as depicted in SEQ ID NO: 125 or 129 of WO 2008 / 119567 and a VH region as depicted in SEQ ID NO: 123 or 127 of WO 2008 / 119567;

[0305] (i) a VL region as depicted in SEQ ID NO: 143 or 147 of WO 2008 / 119567 and a VH region as depicted in SEQ ID NO: 141 or 145 of WO 2008 / 119567;

[0306] (j) a VL region as depicted in SEQ ID NO: 161 or 165 of WO 2008 / 119567 and a VH region as depicted in SEQ ID NO: 159 or 163 of WO 2008 / 119567; and

[0307] (k) a VL region as depicted in SEQ ID NO: 179 or 183 of WO 2008 / 119567 and a VH region as depicted in SEQ ID NO: 177 or 181 of WO 2008 / 119567.

[0308] It is also preferred for the antigen-binding molecule of the present invention that the second domain and the fourth domain that bind to CD3 comprise the VL region as depicted in SEQ ID NO:68 and the VH region as depicted in SEQ ID NO:67.

[0309] 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. Preferably, the VH region is located at the N-terminal end of the joint sequence, and the VL region is located at the C-terminal end of the joint sequence.

[0310] The present invention further provides an antigen binding molecule comprising or having an amino acid sequence 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, 1752 94, 95, 96, 97, 98 or 99% identical to said sequence.

[0311] Covalent modification of antigen binding molecules is also included within the scope of the present invention and is usually, but not always, performed post-translationally. For example, by reacting specific amino acid residues of the antigen binding molecules with an organic derivatizing agent that reacts with selected side chains or N-terminal or C-terminal residues, several types of covalent modification of the antigen binding molecules are introduced into the molecule.

[0312] Cysteinyl residues are most commonly reacted with α-haloacetates (and corresponding amines), such as chloroacetic acid or chloroacetamide, to give carboxymethyl or carboxamidomethyl derivatives. Cysteinyl residues can also be derivatized by reaction with bromotrifluoroacetone, α-bromo-β-(5-imidazolyl)propionic acid, chloroacetyl phosphate, N-alkylmaleimides, 3-nitro-2-pyridyl disulfide, methyl 2-pyridyl disulfide, p-chloromercurybenzoate, 2-chloromercury-4-nitrophenol, or chloro-7-nitrobenzo-2-oxa-1,3-diazole.

[0313] Histidyl residues are derived by reaction with diethylpyrocarbonate at pH 5.5-7.0, as this agent is relatively specific for histidyl side chains. Para-bromophenacyl bromide is also useful; preferably the reaction is carried out in 0.1 M sodium cacodylate at pH 6.0. Lysyl residues and amino terminal residues are reacted with succinic anhydride or other carboxylic anhydrides. Derivatization with these agents has the effect of reversing the charge of lysinyl residues. Other suitable reagents for derivatizing residues containing α-amino groups include imidoesters, such as methyl pyridine imine; pyridoxal phosphate; pyridoxal; chlorine borohydride; trinitrobenzenesulfonic acid; O-methylisourea; 2,4-pentanedione; and transaminase-catalyzed reactions with glyoxylate.

[0314] Arginyl residues are modified by reaction with one or several conventional reagents, among which phenylglyoxal, 2,3-butanedione, 1,2-cyclohexanedione and ninhydrin. Due to the high pKa of the guanidine functional group, the derivatization of arginine residues requires the reaction to be carried out under alkaline conditions. In addition, these reagents can react with lysine groups as well as with the arginine epsilon-amino group.

[0315] Specific modifications can be made to tyrosyl residues, with particular interest in the introduction of spectral labels into tyrosyl residues by reaction with aromatic diazo compounds or tetranitromethane. Most commonly, N-acetylimidazole and tetranitromethane are used to form O-acetyl tyrosyl species and 3-nitro derivatives, respectively. 125 I or 131 I iodinate tyrosyl residues to prepare labeled proteins for radioimmunoassay; the chloramine T method described above is suitable.

[0316] The carboxyl side groups (aspartyl or glutamyl) are selectively modified by reaction with carbodiimide (R'-N=C=N--R'), wherein R and R' are optionally different alkyl groups, such as 1-cyclohexyl-3-(2-morpholinyl-4-ethyl)carbodiimide or 1-ethyl-3-(4-azonia-4,4-dimethylpentyl)carbodiimide. In addition, aspartyl and glutamyl residues are converted to asparaginyl and glutaminyl residues by reaction with ammonium ions.

[0317] Derivatization with bifunctional agents can be used to crosslink the antigen binding molecules of the present invention to a water-insoluble carrier matrix or surface for use in a variety of methods. Commonly used crosslinking agents include, for example, 1,1-bis (diazoacetyl) -2-phenylethane, glutaraldehyde, N-hydroxysuccinimide esters (for example, with esters of 4-azido salicylic acid), homobifunctional imidoesters (including disuccinimide esters, such as 3,3'-dithiobis (succinimidyl propionate)) and difunctional maleimide (such as bis-N-maleimide-1,8-octane). Derivatization agents such as 3-[(p-azidophenyl) dithio] propionimidate methyl esters produce photoactivatable intermediates that can form crosslinks in the presence of light. Alternatively, reactive water-insoluble matrices such as cyanogen bromide activated carbohydrates and reactive substrates, as described in US Pat. Nos. 3,969,287; 3,691,016; 4,195,128; 4,247,642; 4,229,537; and 4,330,440, are used for protein immobilization.

[0318] Glutaminyl residues and asparaginyl residues are usually deamidated to corresponding glutamyl residues and aspartyl residues, respectively. Alternatively, these residues are deamidated under weakly acidic conditions. Any form of these residues falls within the scope of the present invention.

[0319] Other modifications include hydroxylation of proline and lysine, phosphorylation of the hydroxyl groups of seryl or threonyl residues, methylation of the α-amino groups of lysine, arginine, and histidine side chains (TECreighton, Proteins: Structure and Molecular Properties, WH Freeman & Co., San Francisco, 1983, pp. 79-86), acetylation of the N-terminal amine, and amidation of any C-terminal carboxyl group.

[0320] Covalent modification of another type of antigen binding molecules included within the scope of the invention includes changing the glycosylation pattern of the protein. As known in the art, the glycosylation pattern can depend on the sequence of the protein (e.g., the presence or absence of specific glycosylated amino acid residues discussed below) or the host cell or organism in which the protein is produced. Specific expression systems are discussed below.

[0321] Glycosylation of polypeptides is typically N-linked or O-linked. N-linked refers to the carbohydrate moiety being attached to the side chain of an asparagine residue. The tripeptide sequences asparagine-X-serine and asparagine-X-threonine (wherein X is any amino acid except proline) are recognition sequences for enzymatic attachment of the carbohydrate moiety to the asparagine side chain. Therefore, the presence of any of these tripeptide sequences in a polypeptide creates a potential glycosylation site. O-linked glycosylation refers to the attachment of one of the sugars N-acetylgalactosamine, galactose or xylose to a hydroxyamino acid, most commonly serine or threonine, although 5-hydroxyproline or 5-hydroxylysine may also be used.

[0322] Adding glycosylation sites in antigen binding molecules is usually achieved by changing the amino acid sequence so that it contains one or more of the above-mentioned tripeptide sequences (glycosylation sites for N-connections). Changes can also be made by adding or replacing one or more serine or threonine residues to the starting sequence (glycosylation sites for O-connections). For convenience, it is preferred to change the amino acid sequence of the antigen binding molecule by changes in the DNA level, particularly by mutating the DNA encoding the polypeptide at the preselected base so that the codons that will be translated into the desired amino acids are generated.

[0323] Another means of increasing the number of carbohydrate moieties on antigen binding molecules is by chemical or enzymatic coupling of glycosides to proteins. The advantage of these programs is that they do not need to produce proteins in host cells with glycosylation capabilities for N-connections and O-connections. Depending on the coupling mode used, one or more sugars can be attached to (a) arginine and histidine, (b) free carboxyl groups, (c) free sulfhydryl groups, such as those of cysteine, (d) free hydroxyl groups, such as those of serine, threonine or hydroxyproline, (e) aromatic residues, such as those of phenylalanine, tyrosine or tryptophan, or (f) amide groups of glutamine. These methods are described in WO 87 / 05330 and Aplin and Wriston, 1981, CRC Crit.Rev.Biochem. [CRC biochemistry key review], pp. 259-306.

[0324] The removal of carbohydrate moieties present on the starting antigen binding molecule can be accomplished chemically or enzymatically. Chemical deglycosylation requires that the protein be exposed to the compound trifluoromethanesulfonic acid, or an equivalent compound. This treatment results in the cleavage of most or all sugars except the linked sugar (N-acetylglucosamine or N-acetylgalactosamine), while keeping the polypeptide intact. Chemical deglycosylation is described by Hakimuddin et al., 1987, Arch.Biochem.Biophys. [Biochemistry and Biophysics] 259:52 and Edge et al., 1981, Anal.Biochem. [Analytical Biochemistry] 118:131. Enzymatic cleavage of carbohydrate moieties on polypeptides can be achieved by using a variety of endoglycosidases and exoglycosidases, such as by Thotakura et al., 1987, Meth.Enzymol. [Enzymology Methods] 138:350. Glycosylation at potential glycosylation sites can be prevented by using the compound tunicamycin as described by Duskin et al., 1982, J. Biol. Chem. 257: 3105. Tunicamycin blocks the formation of protein-N-glycosidic bonds.

[0325] Other modifications of antigen binding molecules are also contemplated herein.For example, another type of covalent modification of antigen binding molecules includes connecting antigen binding molecules to various non-protein polymers in the manner shown in U.S. Patent Nos. 4,640,835, 4,496,689, 4,301,144, 4,670,417, 4,791,192 or 4,179,337, including but not limited to various polyols, such as polyethylene glycol, polypropylene glycol, polyoxyalkylene or polyethylene glycol and polypropylene glycol copolymers. In addition, as known in the art, amino acid substitutions can be carried out at different positions within antigen binding molecules, such as to promote the addition of polymers such as PEG.

[0326] In some embodiments, the covalent modification of the antigen binding molecules of the present invention includes the addition of one or more labels. The labeling group can be coupled to the antigen binding molecule via spacer arms of various lengths to reduce potential steric hindrance. Various methods for labeling proteins are known in the art and can be used to carry out the present invention. The term "label" or "labeling group" refers to any detectable label. Generally, labels belong to a variety of categories, depending on the assay in which they are to be detected - the following examples include but are not limited to:

[0327] a) Isotopic labels, which may be radioactive isotopes or heavy isotopes, such as 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)

[0328] b) Magnetic labels (e.g. magnetic particles)

[0329] c) Redox-active part

[0330] 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, which can be "small molecule" fluorites or protein fluorites

[0331] e) Enzymatic groups (e.g. horseradish peroxidase, β-galactosidase, luciferase, alkaline phosphatase)

[0332] f) Biotinylation group

[0333] g) a predetermined polypeptide epitope recognized by a second reporter (e.g., a leucine zipper pair sequence, a binding side of a second antibody, a metal binding domain, an epitope tag, etc.)

[0334] "Fluorescent label" means any molecule that can be detected via its intrinsic fluorescent properties. Suitable fluorescent labels include, but are not limited to, fluorescein, rhodamine, tetramethylrhodamine, eosin, erythrosine, coumarin, methyl-coumarin, pyrene, malachite green, stilbene, fluorescent yellow, Cascade Blue J, Texas Red, IAEDANS, EDANS, BODIPY FL, LC Red 640, Cy5, Cy5.5, 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, Oregon).

[00136] In some embodiments, the present invention relates to a plurality of dyes, including fluorophores, comprising fluorescent dyes (fluorophores) and fluorescent dyes (fluorophores). Suitable optical dyes, including fluorophores, are described in the Molecular Probes Handbook by Richard P. Haugland.

[0335] Suitable protein fluorescent markers also include, but are not limited to, green fluorescent proteins, including GFP of Renilla, Ptilosarcus, or Aequorea species (Chalfie et al., 1994, Science 263:802-805), EGFP (Clontech Laboratories, Inc., Genbank Accession No. U55762), blue fluorescent protein (BFP, Quantum Biotechnologies, Inc., 1801 de Maisonneuve Blvd. West, 8th Floor, Montreal, Quebec, Canada H3H 1J9). 1J9); Stauber, 1998, Biotechniques [Biotechnology] 24:462-471; Heim et al., 1996, Curr. Biol. [Current Biology] 6:178-182), enhanced yellow fluorescent protein (EYFP, Crotech Laboratories, Inc.), luciferase (Ichiki et al., 1993, J. Immunol. [Journal of Immunology] 150:5408-5417), β-galactosidase (Nolan et al., 1988, Proc. Natl. Acad. Sci. USA [Proceedings of the National Academy of Sciences of the United States of America] 85:2603-2607) and Renilla (WO 92 / 15673, WO 95 / 07463, WO 98 / 14605, WO 98 / 26277, WO 99 / 49019, U.S. Patent Nos. 5,292,658; 5,418,155; 5,683,888; 5,741,668; 5,777,079; 5,804,387; 5,874,304; 5,876,995; 5,925,558).

[0336] The antigen binding molecules of the present invention can also include other domains, which, for example, contribute to the separation of molecules or to the adaptive pharmacokinetic distribution of the molecule. The domains that contribute to the separation of antigen binding molecules can be selected from peptide motifs or auxiliaryly introduced parts, which can be captured in separation methods (such as separation columns). The non-limiting embodiments of such other domains include peptide motifs called Myc- tags, HAT- tags, HA- tags, TAP- tags, GST- tags, chitin binding domains (CBD- tags), maltose binding protein (MBP- tags), Flag- tags, Strep- tags and variants thereof (such as StrepII- tags) and His- tags. All antigen binding molecules disclosed herein can include His- tag domains, which are generally referred to as continuous His residue repeats of preferably five and more preferably six His residues (six histidines) in the amino acid sequence of the molecule. The His tag can be located, for example, at the N-terminal or C-terminal of the antigen binding molecules, preferably at the C-terminal. Most preferably, hexa-histidine tag (HHHHHH) (SEQ ID NO: 16) is connected to the C-terminus of the antigen binding molecule according to the present invention via a peptide bond. In addition, the conjugate system of PLGA-PEG-PLGA can be combined with poly-histidine tag for sustained release application and improved pharmacokinetic profile.

[0337] The amino acid sequence modification of antigen binding molecules described herein is also contemplated. For example, it may be necessary to improve the binding affinity and / or other biological properties of antigen binding molecules. The amino acid sequence variants of antigen binding molecules are prepared by introducing appropriate nucleotide changes into antigen binding molecule nucleic acids or by peptide synthesis. All amino acid sequence modifications described below should produce antigen binding molecules that still retain the desired biological activity (combined with CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN or EpCAM and CD3) of the unmodified parent molecule.

[0338] The term "amino acid" or "amino acid residue" typically refers to an amino acid having its art-recognized definition, 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 (He 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, synthetic or rare amino acids may be used as desired. In general, amino acids can be grouped as having nonpolar side chains (e.g., Ala, Cys, He, Leu, Met, Phe, Pro, Val); having negatively charged side chains (e.g., Asp, Glu); having positively charged side chains (e.g., Arg, His, Lys); or having uncharged polar side chains (e.g., Asn, Cys, Gln, Gly, His, Met, Phe, Ser, Thr, Trp, and Tyr).

[0339] Amino acid modifications include, for example, deletions and / or insertions and / or substitutions of residues within the amino acid sequence of the antigen binding molecule. Any combination of deletions, insertions and substitutions may be performed to achieve the final construct, provided that the final construct has the desired characteristics. Amino acid changes may also change the post-translational processing of the antigen binding molecule, such as changing the number or position of glycosylation sites.

[0340] For example, 1,2,3,4,5 or 6 amino acids (of course, depending on its length) can be inserted, replaced or lacked in each CDR, and 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20 or 25 amino acids can be inserted, replaced or lacked in each FR.Preferably, the amino acid sequence inset in the antigen binding molecules includes amino acid and / or carboxyl terminal fusions within the length range of 1,2,3,4,5,6,7,8,9 or 10 residues of the polypeptide containing hundreds or more residues, and insets in the sequence of single or multiple amino acid residues.The insertion variants of antigen binding molecules of the present invention include fusions with the N-terminal or C-terminal fusions of the antigen binding molecules of enzyme or with the fusions of polypeptide.

[0341] The sites of greatest interest for substitution mutagenesis include, but are not limited to, CDRs of heavy and / or light chains, particularly hypervariable regions, but FR changes of heavy and / or light chains are also contemplated. Substitutions are preferably conservative substitutions as described herein. Preferably, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10 amino acids can be substituted in CDRs, and 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 or 25 amino acids can be substituted in framework regions (FRs), depending on the length of the CDR or FR. For example, if the CDR sequence encompasses 6 amino acids, it is envisioned that 1, 2 or 3 of these amino acids are substituted. Similarly, if the CDR sequence encompasses 15 amino acids, it is envisioned that 1, 2, 3, 4, 5 or 6 of these amino acids are substituted.

[0342] A useful method for identifying certain residues or regions of an antigen-binding molecule that are preferred locations for mutagenesis is called "alanine scanning mutagenesis," as described by Cunningham and Wells in Science, 244: 1081-1085 (1989). Here, residues or target residue groups within the antigen-binding molecule (e.g., charged residues such as arg, asp, his, lys, and glu) are identified and replaced with neutral or negatively charged amino acids (most preferably alanine or polyalanine) to affect the interaction of the amino acids with the epitope.

[0343] Then by introducing further or other variants at the substitution site or for the substitution site to refine those amino acid positions that show functional sensitivity to substitution.Therefore, although the site or region for introducing amino acid sequence changes is predetermined, the nature of the mutation itself does not need to be predetermined.For example, in order to analyze or optimize the performance of mutation at a given site, alanine scanning or random mutagenesis can be implemented at the target codon or region, and the expressed antigen binding molecule variants are screened to obtain the optimal combination of desired activity.The technology for performing substitution mutations at predetermined sites in DNA with known sequences is well known, for example, M13 primer mutagenesis and PCR mutagenesis.The screening of mutants is carried out using the mensuration of antigen binding activity (such as CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN or EpCAM or CD3 combination).

[0344] In general, if an amino acid is substituted in one or more or all CDRs of the heavy and / or light chain, it is preferred that the "substituted" sequence obtained thereafter is at least 60% or 65%, more preferably 70% or 75%, even more preferably 80% or 85% and particularly preferably 90% or 95% identical to the "initial" CDR sequence. This means that the substitution depends on the length of the CDR and the degree of identity of the "substituted" sequence. For example, a CDR with 5 amino acids is preferably 80% identical to its substituted sequence, so that at least one amino acid is substituted. Thus, the CDRs of an antigen binding molecule may have different degrees of identity with their substituted sequences, for example, CDRL1 may have 80% identity, while CDRL3 may have 90% identity.

[0345] Preferred substitutions (or replacements) are conservative substitutions. However, as long as the antigen binding molecule retains its ability to bind to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM via the first domain and to bind to CD3ε via the second domain and / or its CDR has identity with the sequence replaced thereafter (at least 60% or 65%, more preferably 70% or 75%, even more preferably 80% or 85% and particularly preferably 90% or 95% identical to the "original" CDR sequence), any substitution (including non-conservative substitutions or one or more of the "exemplary substitutions" listed in Table 3 below) is envisioned.

[0346] Conservative substitutions are shown under the heading "Preferred Substitutions" in Table 3. If such substitutions result in a change in biological activity, then more substantial changes denominated "exemplary substitutions" in Table 3, or as described further below with reference to amino acid classes, may be introduced and screened for the desired characteristics.

[0347] Table 3: Amino Acid Substitutions

[0348] Original Exemplary substitutions Preferred substitution Ala(A) val, leu, ile Val Arg(R) lys, gln, asn Lys Asn(N) gln, his, asp, lys, arg Gln Asp(D) glu、asn Glu Cys(C) ser、ala ser Gln(Q) asn、glu asn Glu(E) asp、gln asp Gly(G) Ala ala His(H) asn, gln, lys, arg arg Ile(I) leu, val, met, ala, phe leu Leu(L) norleucine, ile, val, met, ala ile Lys(K) arg, gln, asn arg Met(M) leu, phe, ile leu Phe(F) leu, val, ile, ala, tyr tyr Pro(P) Ala ala Ser(S) Thr thr Thr(T) Ser ser Trp(W) tyr、phe tyr Tyr(Y) trp, phe, thr, ser phe Val(V) ile, leu, met, phe, ala leu

[0349] Substantial modification of the biological properties of the antigen-binding molecules of the invention is accomplished by selecting substitutions that differ significantly in maintaining the following effects: (a) the structure of the polypeptide backbone in the region of the substitution, such as a sheet or helical conformation; (b) the charge or hydrophobicity of the molecule at the target site; or (c) the side chain volume. Naturally occurring residues are grouped based on common side chain properties: (1) hydrophobic: norleucine, met, ala, val, leu, ile; (2) neutral hydrophilic: cys, ser, thr; asn, gln; (3) acidic: asp, glu; (4) basic: his, lys, arg; (5) residues that affect chain orientation: gly, pro; and (6) aromatic: trp, tyr, phe.

[0350] Non-conservative substitutions will require exchanging a member of one of these categories for another. Any cysteine ​​residue that is not involved in maintaining the proper conformation of the antigen-binding molecule can usually be replaced with a serine to improve the oxidative stability of the molecule and prevent abnormal cross-linking. Conversely, one or more cysteine ​​bonds can be added to the antibody to improve its stability (particularly when the antibody is an antibody fragment (such as an Fv fragment)).

[0351] For amino acid sequences, sequence identity and / or similarity are determined using standard techniques known in the art, including, 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 search by similarity method of Pearson and Lipman, 1988, Proc. Nat. Acad. Sci. USA 85:2444, computerized implementations of these algorithms (GAP, BESTFIT, FASTA, and TFASTA in the Wisconsin Genetics Software Package, Genetics Computer Group, 575 Science Drive, Madison, Wis.), Devereux et al., 1984, Nucl. Acids Methods. Res. [Nucleic Acids Research] 12: 387-395, preferably using the default settings, or by inspection. Preferably, the percent identity is calculated by FastDB based on the following parameters: mismatch penalty of 1; gap penalty of 1; gap size penalty of 0.33; and connection 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 Company.

[0352] An example of a useful algorithm is PILEUP. PILEUP creates a multiple sequence alignment from a group of related sequences using progressive pairwise alignments. It can also draw a dendrogram showing the clustering relationships used to create the alignment. PILEUP uses a simplification of the progressive alignment method of Feng and Doolittle, 1987, J. Mol. Evol. [Journal of Molecular Evolution] 35:351-360; this method is similar to the method 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.

[0353] 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. USA 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. The adjustable parameters are set to the following values: overlap interval = 1, overlap fraction = 0.125, word threshold (T) = II. The HSP S and HSP S2 parameters are dynamic values ​​and are established by the program itself based on the composition of the particular sequence and the composition of the particular database against which the sequence of interest is being searched; however, these values ​​can be adjusted to increase sensitivity.

[0354] Another useful algorithm is the gapped BLAST reported by Altschul et al., 1993, Nucl. Acids Res. [Nucleic Acids Research] 25: 3389-3402. Gapped BLAST uses BLOSUM-62 replacement scoring; the threshold T parameter is set to 9; the double-click method triggers non-gap extension, charging a cost of 10+k for a gap length of k; Xu is set to 16, and Xg is set to 40 (for the database search stage) and 67 (for the output stage of the algorithm). Gap alignments are triggered by scores corresponding to about 22 bits.

[0355] Generally speaking, the amino acid homology, similarity or identity between each variant CDR or VH / VL sequence is at least 60% with the sequence depicted herein, and more typically has 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 almost 100% preferably increased homology or identity. In a similar manner, the "percentage (%) of nucleic acid sequence identity" relative to the nucleic acid sequence of the binding protein identified herein is defined as the percentage of nucleotide residues in the candidate sequence that are identical to the nucleotide residues in the coding sequence of the antigen binding molecule. The specific method utilizes the BLASTN module of WU-BLAST-2 set to default parameters, and the overlap interval and overlap score are set to 1 and 0.125, respectively.

[0356] Generally, the nucleic acid sequence homology, similarity or identity between the nucleotide sequence encoding each variant CDR or VH / VL sequence and the nucleotide sequences depicted herein is at least 60%, and more typically has 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 homology or identity of almost 100%. Thus, a "variant CDR" or "variant VH / VL region" has a specified homology, similarity or identity with a parent 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 parent CDR or VH / VL.

[0357] In one embodiment, the antigen binding molecule according to the present invention has a percent identity to the human germline of ≥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 human antibody germline gene products is considered to be an important feature to reduce the risk of therapeutic proteins triggering an immune response to the drug in patients during treatment. Hwang and Foote ("Immunogenicity of engineered antibodies"; Methods 36 (2005) 3-10) showed that the reduction of the non-human part of the drug antigen binding molecule reduces the risk of inducing anti-drug antibodies in patients during treatment. By comparing the antibody drugs evaluated in numerous clinical trials and the corresponding immunogenicity data, the following trend is shown: the humanization of the V region of the antibody makes the protein immunogenicity (average 5.1% of patients) lower than the antibody carrying the unchanged non-human V region (average 23.59% of patients). Therefore, for the protein therapeutic agent based on V region in the form of antigen binding molecules, it is desirable to have a higher degree of identity with human sequences. For the purpose of determining germline identity, the V region of VL can be compared with the amino acid sequence of human germline V segment and J segment (http: / / vbase.mrc-cpe.cam.ac.uk / ) using Vector NTI software and the amino acid sequence (in percentage) is calculated by dividing the same amino acid residues by the total amino acid residues of VL. The same applies to VH segment (http: / / vbase.mrc-cpe.cam.ac.uk / ), except that VH CDR3 can be excluded due to the high diversity of VH CDR3 and the lack of existing human germline VH CDR3 comparison partners. Recombinant technology can then be used to increase the sequence identity with human antibody germline genes.

[0358] In a further embodiment, the bispecific antigen binding molecules of the invention exhibit high monomer yields under standard research scale conditions, such as in a standard two-step purification process. Preferably, the monomer yield of the antigen binding molecules according to the invention is ≥0.25 mg / L supernatant, more preferably ≥0.5 mg / L, even more preferably ≥1 mg / L, and most preferably ≥3 mg / L supernatant.

[0359] Similarly, the yield of the dimeric antigen binding molecule isotype of the antigen binding molecule can be determined, and the monomer percentage (i.e., monomer: (monomer + dimer)) can be determined from this. The productivity of monomer and dimeric antigen binding molecules and the calculated monomer percentage can be obtained, for example, in the SEC purification step of the culture supernatant produced from the standardized research scale in the roller bottle. In one embodiment, the monomer percentage of the antigen binding molecule is ≥80%, more preferably ≥85%, even more preferably ≥90%, and most preferably ≥95%.

[0360] In one embodiment, the preferred plasma stability of the antigen binding molecule (ratio of EC50 with plasma to EC50 without plasma) is ≤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 can be determined by incubating the construct in human plasma at 37°C for 24 hours, followed by incubation at 4 °C for 24 hours. 51 The chromium release cytotoxicity assay is tested by measuring EC50. The effector cells in the cytotoxicity assay can be the enriched human CD8 positive T cells of stimulation. The target cells can be, for example, CHO cells transfected with human CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN or EpCAM. The effector cells and target cells (E: T) ratio can be selected as 10: 1 or 5: 1. The human plasma pool for this purpose is derived from the blood of healthy donors collected by EDTA-coated syringes. Cellular components are removed by centrifugation, and the upper plasma phase is collected and subsequently collected. As a control, the antigen binding molecules are diluted immediately before the cytotoxicity assay in the RPMI-1640 culture medium. The plasma stability is calculated as the ratio of EC50 (after plasma incubation) to EC50 (control).

[0361] In addition, the low conversion rate of monomer to dimer of the preferred antigen binding molecules of the present invention.Conversion can be measured under different conditions and analyzed by high performance size exclusion chromatography.For example, the incubation of the monomer isotype of the antigen binding molecules can be carried out in an incubator for 7 days at 37 ° C and a concentration of, for example, 100 μg / ml or 250 μg / ml.Under these conditions, it is preferred that the antigen binding molecules of the present invention show 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% and most preferably ≤1% or ≤0.5% or even 0%.

[0362] It is also preferred that the bispecific antigen binding molecules of the present invention present very low dimer conversion rate after multiple freeze / thaw cycles. For example, the antigen binding molecule monomer is adjusted to a concentration of 250 μg / ml in, for example, a general formulation buffer, and three freeze / thaw cycles are performed (freezing 30 min at -80 ° C, thawing 30 min at room temperature thereafter), followed by high performance SEC to determine the percentage of the initial monomer antigen binding molecules that have been converted into dimeric antigen binding molecules. Preferably, the dimer percentage of the bispecific antigen binding molecule is ≤5%, 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%, for example, after three freeze / thaw cycles.

[0363] The bispecific antigen binding molecules of the present invention preferably show 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. The thermal stability parameter can be determined as follows according to the antibody aggregation temperature: the antibody solution with a concentration of 250 μg / ml is transferred to a disposable cuvette and placed in a dynamic light scattering (DLS) device. The sample is heated from 40°C to 70°C at a heating rate of 0.5°C / min, and the radius of the measurement is constantly obtained. The aggregation temperature of the antibody is calculated using the increase in radius indicating the melting of proteins and aggregates.

[0364] Alternatively, the temperature melting curve can be determined by differential scanning calorimetry (DSC) to determine the intrinsic biophysical protein stability of antigen binding molecules. These experiments are carried out using MicroCal LLC (Northampton, MA, USA) VP-DSC devices. Compared with samples containing only formulation buffer, the energy absorption of samples containing antigen binding molecules is recorded as 20 ° C to 90 ° C. For example, in SEC running buffer, antigen binding molecules are adjusted to a final concentration of 250 μg / ml. In order to record the corresponding melting curve, the entire sample temperature is gradually increased. At each temperature T, the energy intake of sample and formulation buffer reference is recorded. The energy intake Cp (kcal / mole / ℃) of the sample is deducted from the difference of the reference for the corresponding temperature. Melting temperature is defined as the temperature at the first maximum energy intake.

[0365] It is also envisioned that the CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAMxCD3 bispecific antigen binding molecules of the invention have a turbidity of ≤0.2, preferably ≤0.15, more preferably ≤0.12, even more preferably ≤0.1, and most preferably ≤0.08 (as measured by OD340 after concentrating the purified monomeric antigen binding molecules to 2.5 mg / ml and incubating overnight).

[0366] In a further embodiment, the antigen binding molecules according to the present invention are stable at physiological or slightly low pH, i.e., about pH 7.4 to 6.0. Antigen binding molecules show more tolerance at non-physiological pH such as about pH 6.0, and the recovery of the antigen binding molecules eluted from ion exchange columns relative to the total amount of loaded protein is higher. The recovery of the antigen binding molecules from the ion (e.g., cation) exchange columns of 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%.

[0367] It is further envisaged that the bispecific antigen binding molecules of the invention exhibit therapeutic efficacy or anti-tumor activity. This can be assessed, for example, in studies disclosed in the following generalized example of an advanced human tumor xenograft model:

[0368] On day 1 of the study, 5 x 10 cells of cancer cell lines positive for human target cell antigens (here: CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM) were added. 6 The cells were injected subcutaneously into the right dorsal flank of female NOD / SCID mice. When the average tumor volume reached approximately 100 mm 3 By adding about 2x 10 7 The in vitro expanded human CD3 positive T cells were transplanted into mice by injecting 100 cells into the peritoneal cavity of the animals. The mice in the vehicle control group 1 did not receive effector cells and were used as non-transplanted controls compared with the vehicle control group 2 (receiving effector cells) to monitor the effect of T cells alone on tumor growth. When the average tumor volume reached about 200 mm 3At the beginning of the treatment, the bispecific antigen binding molecules were used for treatment. The average tumor size of each treatment group on the day of treatment should not be statistically different from any other group (analysis of variance). Mice were treated by intravenous push injection of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN or EpCAM and CD3 bispecific antigen binding molecules at 0.5 mg / kg / day for about 15 to 20 days. Tumors were measured by calipers during the study and progress was evaluated by intergroup comparison of tumor volume (TV). Tumor growth inhibition T / C [%] was determined by calculating TV as T / C%=100×(median TV of analysis group) / (median TV of control group 2).

[0369] Those skilled in the art know how to modify or adjust certain parameters of the study, such as the number of tumor cells injected, the injection site, the number of transplanted human T cells, the amount of bispecific antigen binding molecules to be administered, and the timeline, while still obtaining meaningful and reproducible results. Preferably, the tumor growth inhibition T / C [%] ≤ 70 or ≤ 60, more preferably ≤ 50 or ≤ 40, even more preferably ≤ 30 or ≤ 20 and most preferably ≤ 10 or ≤ 5 or even ≤ 2.5. Tumor growth inhibition is preferably close to 100%.

[0370] In a preferred embodiment of the antigen-binding molecule of the present invention, the antigen-binding molecule is a single-chain antigen-binding molecule.

[0371] Furthermore, in a preferred embodiment of the antigen-binding molecule of the present invention, the spacer comprises, in order from amino to carboxyl groups:

[0372] Hinge-CH2-CH3-Joint-Hinge-CH2-CH3.

[0373] 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 NO: 17-24. In a preferred embodiment of the present invention, each of the polypeptide monomers has an amino acid sequence selected from SEQ ID NO: 17-24.

[0374] In addition, in one embodiment of the present invention, the CH2 domain of one or preferably each (two) polypeptide monomer of the spacer comprises an intradomain cysteine ​​disulfide bridge. As known in the art, the term "cysteine ​​disulfide bridge" refers to a functional group having a general structure R-S-S-R. This connection is also referred to as an SS bond or a disulfide bond and is derived by coupling two thiol groups of a cysteine ​​residue. For the antigen-binding molecules of the present invention, it is particularly preferred that the cysteine ​​that forms the cysteine ​​disulfide bridge in the mature antigen-binding molecule is introduced into the amino acid sequence of the CH2 domain corresponding to 309 and 321 (Kabat numbering).

[0375] In one embodiment of the invention, the glycosylation site in Kabat position 314 of the CH2 domain is removed. Preferably, the removal of the glycosylation site is achieved by a N314X substitution, wherein X is any amino acid except Q. Preferably, the substitution is N314G. In a more preferred embodiment, the CH2 domain further comprises the following substitutions (according to Kabat positions): V321C and R309C (these substitutions introduce intradomain cysteine ​​disulfide bridges at Kabat positions 309 and 321).

[0376] It is assumed that preferred features of the antigen binding molecules of the present invention compared to bispecific heteroFc antigen binding molecules known in the art may for example relate inter alia to the introduction of the above modifications in the CH2 domain. Thus, it is preferred for the constructs of the present invention that the CH2 domain in the spacer of the antigen binding molecule of the present invention comprises intradomain cysteine ​​disulfide bridges at Kabat positions 309 and 321 and / or that the glycosylation site at Kabat position 314 is removed, preferably by N314G substitution.

[0377] In a further preferred embodiment of the present invention, the CH2 domain in the spacer of the antigen-binding molecule of the present invention comprises 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.

[0378] In one embodiment, the present invention provides an antigen binding molecule, wherein:

[0379] (i) the first domain comprises two antibody variable domains, and the second domain comprises two antibody variable domains;

[0380] (ii) the first domain comprises one antibody variable domain, and the second domain comprises two antibody variable domains;

[0381] (iii) the first domain comprises two antibody variable domains, and the second domain comprises one antibody variable domain; or

[0382] (iv) the first domain comprises an antibody variable domain, and the second domain comprises an antibody variable domain.

[0383] The first domain and the second domain can be binding domains that each comprises two antibody variable domains (such as VH and VL domains). Examples of such binding domains that comprise two antibody variable domains are described above and include, for example, Fv fragments, scFv fragments or Fab fragments described above. Alternatively, one or both of these binding domains can only comprise a single variable domain. Examples of such single domain binding domains are described above and include, for example, nano antibodies or single variable domain antibodies that only comprise a variable domain, which can be VHH, VH or VL that specifically binds to antigens or epi-positions independently of other V regions or domains.

[0384] In a preferred embodiment of the antigen binding molecules of the present invention, the second and third binding domains are fused to the spacer via a peptide linker.Preferred peptide linkers have been described above and are characterized in that the amino acid sequence Gly-Gly-Gly-Gly-Ser, i.e. Gly4Ser (SEQ ID NO:7), or its polymer, i.e. (Gly4Ser) x, wherein x is 1 or a greater integer (e.g. 2 or 3).The particularly preferred linker for the first domain and the second domain to be fused to the spacer is described in SEQ ID NO:7.

[0385] The antigen binding molecules of the present invention include a first domain that binds to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM, preferably to one or more extracellular domains (ECD) of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM. It should be understood that in the context of the present invention, the term "binding to the extracellular domain of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM" means 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, when CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM is expressed by a naturally expressing cell or cell line and / or by a cell or cell line transformed or (stably / transiently) transfected therewith, the first domain according to the present invention preferably binds to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM. In a preferred embodiment, 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 (e.g., BIAcore or Scatchard), the first binding domain also binds to CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM. A "target cell" can 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 human or animal body, such as a cancer or tumor cell that expresses a specific CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM.

[0386] 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 further 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 human and macaque CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM / CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAMECD. "CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM extracellular domain" or "CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM ECD" refers to a CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM region or sequence that is substantially free of the transmembrane and cytoplasmic domains of CS1, BCMA, CD20, CD22, FLT3, CD123, CLL1, CHD3, MSLN, or EpCAM. It will be appreciated 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 invention are identified according to criteria routinely used in the art for identifying the type of hydrophobic domain. The exact boundaries of the transmembrane domain may vary, but most likely will vary by no more than about 5 amino acids at either end of the domains specifically mentioned herein.

[0387] Preferred binding domains that bind to CD3 are disclosed in WO 2010 / 037836 and WO 2011 / 121110. Any binding domain for CD3 described in these applications may be used in the context of the present invention.

[0388] The present invention further provides polynucleotides / nucleic acid molecules encoding the antigen binding molecules of the present invention. Polynucleotides are biopolymers consisting of 13 or more nucleotide monomers covalently bonded in a chain. DNA (such as cDNA) and RNA (such as mRNA) are examples of polynucleotides with different biological functions. Nucleotides are organic molecules that serve as monomers or subunits of nucleic acid molecules such as DNA or RNA. Nucleic acid molecules or polynucleotides can be double-stranded and single-stranded, linear and circular. It is preferably contained in a vector, and the vector is preferably contained in a host cell. For example, the host cell can express antigen binding molecules after transformation or transfection with the vector or polynucleotide of the present invention. For this purpose, polynucleotides or nucleic acid molecules are operably connected to control sequences.

[0389] The genetic code is a set of rules for translating 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 to carry the amino acids and read out the mRNA three nucleotides at a time. The code defines how the sequence of these nucleotide triplets, called codons, specifies which amino acid will be added next during protein synthesis. With some exceptions, a three-nucleotide codon in a nucleic acid sequence specifies a single amino acid. Because the vast majority of genes are encoded using the exact same code, this particular code is often referred to as the canonical or standard genetic code. While the genetic code determines the protein sequence for a given coding region, other genomic regions may influence when and where these proteins are produced.

[0390] In addition, the present invention provides vectors, which include polynucleotides / nucleic acid molecules of the present invention. A vector is a nucleic acid molecule used as a medium for transferring (exogenous) genetic material into a cell. The term "vector" encompasses but is not limited to plasmids, viruses, cosmids, and artificial chromosomes. Generally speaking, engineered vectors include a replication origin, a multiple cloning site, and a selective marker. The vector itself is typically a nucleotide sequence, which is typically a DNA sequence of a larger sequence that includes an insert (transgene) and a "skeleton" that serves as a vector. In addition to transgenic inserts and skeletons, modern vectors can include other features: promoters, genetic markers, antibiotic resistance, reporter genes, targeting sequences, protein purification tags. A vector referred to as an expression vector (expression construct) is particularly useful for expressing a transgene in a target cell, and typically has a control sequence.

[0391] The term "control sequence" refers to DNA sequences necessary for the expression of an operably linked coding sequence in a particular host organism. For example, control sequences suitable for prokaryotes include a promoter, optionally an operator sequence, and a ribosome binding side. Eukaryotic cells are known to utilize promoters, polyadenylation signals, and enhancers.

[0392] When a nucleic acid is in a functional relationship with another nucleic acid sequence, the nucleic acid is "operably linked". For example, if the DNA of a presequence or secretory leader sequence is expressed as a preprotein that participates in the secretion of a polypeptide, the DNA of the presequence or secretory leader sequence is operably linked to the DNA of the polypeptide; if a promoter or enhancer affects the transcription of a coding sequence, the promoter or enhancer is operably linked to the sequence; or if the ribosome binding side is positioned so as to facilitate translation, the ribosome binding side is operably linked to the coding sequence. Generally speaking, "operably linked" means that the DNA sequence connected is continuous, and in the case of a secretory leader sequence, it is continuous and in the reading phase. However, enhancers do not have to be continuous. Connection is accomplished by joining at convenient restriction sites. If such sites are not present, synthetic oligonucleotide adapters or joints are used according to conventional practice.

[0393] "Transfection" is the process of intentionally introducing nucleic acid molecules or polynucleotides (including vectors) into target cells. The term is primarily used for non-viral methods in eukaryotic cells. Transduction is often used to describe the transfer of nucleic acid molecules or polynucleotides mediated by viruses. Transfection of animal cells typically involves opening transient pores or "holes" in the cell membrane to allow uptake of the substance. Transfection can be performed using calcium phosphate, by electroporation, by cell extrusion, or by mixing cationic lipids with the substance to produce liposomes, which fuse with the cell membrane and deposit their cargo inside.

[0394] The term "transformation" is used to describe the non-viral transfer of nucleic acid molecules or polynucleotides (including vectors) into bacteria, as well as into non-animal eukaryotic cells (including plant cells). Thus, transformation is a genetic change in a bacterial or non-animal eukaryotic cell that results from the direct uptake of exogenous genetic material (nucleic acid molecules) from its surroundings through one or more cell membranes and the subsequent incorporation of such material. Transformation can be achieved by artificial means. For transformation to occur, the cell or bacterium must be in a competent state, which may occur as a time-limited response to environmental conditions such as starvation and cell density.

[0395] In addition, the present invention provides host cells, which are transformed or transfected with polynucleotides / nucleic acid molecules or vectors of the present invention. As used herein, the term "host cell" or "receptor cell" is intended to include a receptor that may be or has been a carrier, an exogenous nucleic acid molecule, and a polynucleotide encoding the antigen binding molecules of the present invention; and / or any single cell or cell culture of the receptor of the antigen binding molecules themselves. The corresponding substance is introduced into the cell by means of transformation, transfection, etc. The term "host cell" is also intended to include the offspring or potential offspring of a unicellular cell. Because some modifications may occur in offspring due to natural, unexpected or intentional mutations or due to environmental influences, this offspring may in fact be not completely identical with the parental cell (in morphology or genome or entire DNA complement), but is still included in the scope of the term as used herein. Suitable host cells include prokaryotic cells 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 mice, rats, macaques or people.

[0396] Antigen binding molecules of the present invention can be produced in bacteria. After expression, antigen binding molecules of the present invention are separated from Escherichia coli cell paste with soluble fractions, and can be purified by, for example, affinity chromatography and / or size exclusion. Final purification can be carried out by methods similar to those for purifying antibodies expressed, for example, in CHO cells.

[0397] In addition to prokaryotes, eukaryotic microorganisms (such as filamentous fungi or yeast) are suitable cloning or expression hosts for the antigen binding molecules of the present invention. Saccharomyces cerevisiae or common baker's yeast is the most commonly used of the lower eukaryotic host microorganisms. However, many other genera, species and strains are generally available and can be used herein, such as Schizosaccharomyces pombe; Kluyveromyce 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 (EP 402 226); Pichia pastoris (EP 183070); Candida; Trichoderma reesei (EP 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.

[0398] Suitable host cells for expressing the glycosylated antigen binding molecules of the present invention are derived from multicellular organisms. Examples of invertebrate cells include plant cells and insect cells. Many baculovirus strains and variants and corresponding permissive insect host cells from hosts such as Spodopterafrugiperda (caterpillars), Aedes aegypti (mosquitoes), Aedes albopictus (mosquitoes), Drosophila melanogaster (fruit flies) and Bombyx mori (Bombyx mori) have been identified. Multiple virus strains for transfection are publicly available, such as the L-1 variant of Autographa californica NPV and the Bm-5 strain of Bombyx mori NPV, and according to the present invention, such viruses can be used as viruses herein, particularly for transfecting Spodoptera frugiperda cells.

[0399] Plant cell cultures of cotton, corn, potato, soybean, petunia, tomato, Arabidopsis, and tobacco can also be used as hosts. Cloning and expression vectors useful for producing proteins in plant cell cultures 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.

[0400] However, the greatest interest has been in vertebrate cells, and propagation of vertebrate cells in culture (tissue culture) has become routine procedure. Examples of useful mammalian host cell lines are monkey kidney CV1 line transformed by SV40 (COS-7, ATCC CRL1651); human embryonic kidney line (293 cells or 293 cells subcloned for growth in suspension culture, Graham et al., J. Gen Virol. [Journal of General Virology] 36:59 (1977)); baby hamster kidney cells (BHK, ATCC CCL10); Chinese hamster ovary cells / -DHFR (CHO, Urlaub et al., Proc. Natl. Acad. Sci. USA [Proceedings of the National Academy of Sciences of the United States of America] 77:4216 (1980)); mouse Sertoli cells (TM4, Mather, Biol. Reprod. [Biology of Reproduction] 23:243-251 (1980)); monkey kidney cells (CVI ATCC CCL 70); African green monkey kidney cells (VERO-76, ATCC CRL1587); human cervical carcinoma cells (HELA, ATCC CCL 2); canine 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, 14138065); mouse mammary tumor (MMT 060562, ATCC CCL5 1); TRI cells (Mather et al., Annals N.Y Acad. Sci. [Annals of the New York Academy of Sciences] (1982) 383:44-68); MRC 5 cells; FS4 cells; and human hepatoma cell line (Hep G2).

[0401] In a further embodiment, the present invention provides a method for producing an antigen-binding molecule of the present invention, the method comprising culturing the host cell of the present invention under conditions allowing expression of the antigen-binding molecule of the present invention, and recovering the produced antigen-binding molecule from the culture.

[0402] As used herein, the term "culture" refers to the in vitro maintenance, differentiation, growth, proliferation and / or propagation of cells in a culture medium under suitable conditions. The term "expression" includes any step involved in producing the antigen binding molecules of the present invention, including but not limited to transcription, post-transcriptional modification, translation, post-translational modification and secretion.

[0403] When using recombinant technology, antigen binding molecules can be produced in the cell, in the periplasmic space or directly secreted into the culture medium. If antigen binding molecules are produced in the cell, then as a first step, the granular fragments of the host cell or dissolved fragments are removed, for example, by centrifugation or ultrafiltration. Carter et al., Bio / Technology [biological / technology] 10: 163-167 (1992) describes the procedure for separating the antibodies secreted into the periplasmic space of Escherichia coli. In short, the cell paste is thawed in the presence of sodium acetate (pH 3.5), EDTA and phenylmethylsulfonyl fluoride (PMSF) for about 30 minutes. Cell debris can be removed by centrifugation. In the case of secreting antibodies into the culture medium, commercially available protein concentration filters are usually first used, such as Amicon or Millipore Pellicon ultrafiltration units to concentrate the supernatant from such expression systems. Protease inhibitors (such as PMSF) can be included in any of the aforementioned steps to inhibit proteolysis, and antibiotics can be included to prevent the growth of foreign contaminants.

[0404] The antigen binding molecules of the present invention prepared by host cells can be recovered or purified using, for example, hydroxyapatite chromatography, gel electrophoresis, dialysis, and affinity chromatography. Depending on the antibody to be recovered, other techniques for protein purification may also be used, such as fractionation on an ion exchange column, ethanol precipitation, reversed-phase HPLC, chromatography on silica, chromatography on heparin, SEPHAROSE on an anion or cation exchange resin (e.g., polyaspartic acid column), TM Chromatography, chromatofocusing, SDS-PAGE, and ammonium sulfate precipitation. Where the antigen binding molecules of the invention comprise a CH3 domain, Bakerbond ABX resin (JT Baker, Phillipsburg, NJ) can be used for purification.

[0405] Affinity chromatography is the preferred purification technique. The matrix to which the affinity ligand is attached is most often agarose, but other matrices are also available. Mechanically stable matrices such as controlled pore glass or poly(styrenedivinyl)benzene allow faster flow rates and shorter processing times than can be achieved with agarose.

[0406] In addition, the present invention provides a pharmaceutical composition comprising an antigen binding molecule of the present invention or an antigen binding molecule produced according to the process of the present invention. For a pharmaceutical composition of the present invention, the homogeneity of the preferred antigen binding molecule is ≥80%, more preferably ≥81%, ≥82%, ≥83%, ≥84% or ≥85%, further preferably ≥86%, ≥87%, ≥88%, ≥89% or ≥90%, more preferably ≥91%, ≥92%, ≥93%, ≥94% or ≥95% and most preferably ≥96%, ≥97%, ≥98% or ≥99%.

[0407] 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 present invention preferably comprise one or more antigen binding molecules of the present invention 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 acceptable ingredients of the composition are preferably nontoxic to the receptor at the dosage and concentration used. The pharmaceutical composition of the present invention includes, but is not limited to, liquid, frozen and lyophilized compositions.

[0408] The compositions of the present invention may include a pharmaceutically acceptable carrier. Generally speaking, as used herein, "pharmaceutically acceptable carrier" means any and all aqueous and non-aqueous solutions, sterile solutions, solvents, buffers (e.g., phosphate buffered saline (PBS) solutions), water, suspensions, emulsions (e.g., oil / water emulsions), various types of wetting agents, liposomes, dispersion media, and coatings compatible with drug administration, particularly parenteral administration. The use of such media and medicaments in pharmaceutical compositions is well known in the art, and compositions comprising such carriers can be formulated by well-known conventional methods.

[0409] Some embodiments provide pharmaceutical compositions comprising antigen binding molecules of the present invention and one or more other excipients (such as those excipients described illustratively in this chapter and elsewhere herein). In this regard, excipients can be used for multiple purposes in the present invention, such as adjusting the physical, chemical or biological properties of the preparation, such as adjusting viscosity and / or the method of the present invention to improve effectiveness and / or stabilize such preparations and methods, to resist the degradation and corruption caused by, for example, pressure occurring during preparation, administration and subsequent use before manufacture, transportation, storage, use.

[0410] In certain embodiments, pharmaceutical compositions may contain formulation materials for the purpose of altering, maintaining or preserving aspects of the composition such as pH, osmotic pressure, viscosity, clarity, color, isotonicity, odor, sterility, stability, dissolution or release rate, adsorption or permeability (see REMINGTON'S PHARMACEUTICAL SCIENCES, 18th edition, (A. R. Genrmo, ed.), 1990, Mack Publishing Company). In such embodiments, suitable formulation materials may include, but are not limited to:

[0411] Amino acids, such as glycine, alanine, glutamine, asparagine, threonine, proline, 2-phenylalanine, including charged amino acids, preferably lysine, lysine acetate, arginine, glutamate and / or histidine

[0412] Antimicrobial agents, such as antibacterial and antifungal agents

[0413] Antioxidants, such as ascorbic acid, methionine, sodium sulfite or sodium bisulfite;

[0414] Buffers, buffer systems and buffering agents for maintaining the composition at physiological pH or slightly lower pH, preferably at a lower pH of 4.0 to 6.5; examples of buffers are borate, bicarbonate, Tris-HCl, citrate, phosphate or other organic acid, succinate, phosphate and histidine; for example Tris buffer at about pH 7.0-8.5;

[0415] Non-aqueous solvents such as propylene glycol, polyethylene glycol, vegetable oils such as olive oil, and injectable organic esters such as ethyl oleate;

[0416] Aqueous carriers include water, alcoholic / aqueous solutions, emulsions or suspensions, including saline and buffered media;

[0417] Biodegradable polymers, such as polyesters;

[0418] · bulking agents, such as mannitol or glycine;

[0419] Chelating agents, such as ethylenediaminetetraacetic acid (EDTA);

[0420] Isotonicity agents and absorption delaying agents;

[0421] Complexing agents, such as caffeine, polyvinylpyrrolidone, β-cyclodextrin or hydroxypropyl-β-cyclodextrin;

[0422] Fillers

[0423] Monosaccharides; disaccharides; and other carbohydrates (such as glucose, mannose or dextrin); the carbohydrate may be a non-reducing sugar, preferably trehalose, sucrose, octasulfate, sorbitol or xylitol;

[0424] (Low molecular weight) proteins, polypeptides or protein carriers, such as human or bovine serum albumin, gelatin or immunoglobulins, preferably of human origin;

[0425] Coloring and flavoring agents;

[0426] Sulfur-containing reducing agents, such as glutathione, lipoic acid, sodium thioacetate, thioglycerol, [α]-monothioglycerol, and sodium thiosulfate

[0427] Diluent;

[0428] Emulsifiers;

[0429] Hydrophilic polymers such as polyvinylpyrrolidone

[0430] Salt-forming counterions, such as sodium;

[0431] preservatives, such as antimicrobials, antioxidants, chelating agents, inert gases, etc.; examples are: benzalkonium chloride, benzoic acid, salicylic acid, thimerosal, phenethyl alcohol, methylparaben, propylparaben, chlorhexidine, sorbic acid or hydrogen peroxide;

[0432] Metal complexes, such as Zn-protein complexes;

[0433] Solvents and co-solvents (such as glycerol, propylene glycol or polyethylene glycol);

[0434] Sugars and sugar alcohols, such as trehalose, sucrose, octasulfate, mannitol, sorbitol or xylitol, stachyose, mannose, sorbitol, xylose, ribose, myoinisitose, galactose, lactitol, ribitol, myoinisitol, dulcitol, glycerol, cyclitols (e.g., inositol), polyethylene glycol; and polyols;

[0435] Suspension concentrate;

[0436] Surfactants or wetting agents such as pluronics, PEG, sorbitan esters, polysorbates (such as polysorbate 20, polysorbate 50), triton, tromethamine, lecithin, cholesterol, tyloxapal; the surfactant may be a detergent, preferably with a molecular weight > 1.2 KD, and / or a polyether, preferably with 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;

[0437] Stability enhancers, such as sucrose or sorbitol;

[0438] Tonicity enhancing agents, such as alkali metal halides, preferably sodium chloride or potassium chloride; mannitol sorbitol;

[0439] Parenteral delivery vehicles include sodium chloride solution, Ringer's dextrose, dextrose and sodium chloride, lactated Ringer's, or fixed oils;

[0440] • Intravenous delivery vehicles, including fluid and nutrient replenishers, electrolyte replenishers (such as those based on Ringer's dextrose).

[0441] In the context of the present invention, the pharmaceutical composition, which is preferably a liquid composition or may be a solid composition obtained by lyophilization or may be a reconstituted liquid composition, comprises

[0442] (a) an antigen-binding molecule comprising at least four binding domains, wherein:

[0443] The first and third domains bind to target cell surface antigens and have an isoelectric point (pI) in the range of 4 to 9,5;

[0444] The second and fourth domains bind to CD3; and have a pi in the range of 8 to 10, preferably 8.5 to 9.0; and

[0445] A spacer, which preferably comprises two polypeptide monomers, each polypeptide monomer comprising a hinge, a CH2 domain and a CH3 domain, wherein the two polypeptide monomers are fused to each other via a peptide linker;

[0446] (b) at least one buffer;

[0447] (c) at least one sugar; and

[0448] (d) at least one surfactant;

[0449] And wherein the pH of the pharmaceutical composition is in the range of 3.5 to 6.

[0450] It is further envisaged in the context of the present invention that the at least one buffer is present in a concentration range of 5 to 200 mM, more preferably in a concentration range of 10 to 50 mM. It is envisaged in the context of the present invention that the at least one sugar is selected from the group consisting of a monosaccharide, a disaccharide, a cyclic polysaccharide, a sugar alcohol, a linear branched glucan or a linear unbranched glucan. It is also envisaged in the context of the present invention that the disaccharide is selected from the group consisting of sucrose, trehalose and mannitol, sorbitol and combinations thereof. It is further envisaged in the context of the present invention that the sugar alcohol is sorbitol. It is envisaged in the context of the present invention that the at least one sugar is present in a concentration range of 1% to 15% (m / V), preferably in a concentration range of 9% to 12% (m / V).

[0451] It is also envisaged in the context of the present invention that at least one surfactant is selected from the group consisting of polysorbate 20, polysorbate 40, polysorbate 60, polysorbate 80, poloxamer 188, pluronic F68, Triton X-100, polyoxyethylen, PEG 3350, PEG 4000 and combinations thereof. It is further envisaged in the context of the present invention that at least one surfactant is present in a concentration range of 0.004% to 0.5% (m / V), preferably in a range of 0.001% to 0.01% (m / V). It is envisaged in the context of the present invention that the pH of the composition is in the range of 4.0 to 5.0, preferably 4.2. It is also envisaged in the context of the present invention that the pharmaceutical composition has an osmotic pressure in the range of 150 to 500 mOsm. It is further envisaged in the context of the present invention that the pharmaceutical composition further comprises an excipient selected from the group consisting of one or more polyols and one or more amino acids. In the context of the present invention it is envisaged that the one or more excipients are present in a concentration range of 0.1% to 15% (w / V).

[0452] It is also envisaged in the context of the present invention that the pharmaceutical composition comprises

[0453] (a) an antigen-binding molecule as described above,

[0454] (b) 10 mM glutamate or acetate,

[0455] (c) 9% (m / V) sucrose or 6% (m / V) sucrose and 6% (m / V) hydroxypropyl-β-cyclodextrin,

[0456] (d) 0.01% (m / V) Polysorbate 80

[0457] And wherein the pH of the liquid pharmaceutical composition is 4.2.

[0458] In the context of the present invention it is further envisaged that the antigen-binding molecule is present in a concentration range of from 0.1 to 8 mg / ml, preferably from 0.2 - 2.5 mg / ml, more preferably from 0.25 - 1.0 mg / ml.

[0459] It will be apparent to those skilled in the art that, for example, the different components of the pharmaceutical composition (such as those listed above) can have different effects, and amino acids can act as buffers, stabilizers and / or antioxidants; mannitol can act as a bulking agent and / or tonicity enhancer; sodium chloride can act as a delivery vehicle and / or tonicity enhancer; and so on.

[0460] It is envisaged that, in addition to the polypeptides of the invention as defined herein, the compositions of the invention can comprise additional bioactive agents, depending on the intended use of the composition. Such agents can be drugs acting on the gastrointestinal system, drugs acting as cytostatic agents, drugs for preventing hyperuricemia, drugs for inhibiting the immune response (such as corticosteroids), drugs for modulating the inflammatory response, drugs acting on the circulatory system and / or agents known in the art such as cytokines. It is also envisaged that the antigen-binding molecules of the invention are applied in co-therapy, i.e. in combination with another anti-cancer drug.

[0461] In certain embodiments, the optimal pharmaceutical composition can affect the physical state, stability, in vivo release rate and in vivo clearance rate of the antigen-binding molecule of the invention. In certain embodiments, the major vehicle or carrier in the pharmaceutical composition can be aqueous or non-aqueous in nature. For example, suitable vehicles or carriers can be water for injection, physiological saline solution or artificial cerebrospinal fluid, possibly supplemented with other substances common in compositions for parenteral administration. Neutral buffered saline or saline mixed with serum albumin are additional exemplary vehicles. In certain embodiments, the antigen-binding molecule of the invention can be prepared for storage by mixing the selected components of the desired purity with optional formulations (REMINGTON'S PHARMACEUTICAL SCIENCES, supra) in the form of a lyophilized cake or an aqueous solution. In addition, in certain embodiments, the antigen-binding molecule of the invention can be formulated as a lyophilizate using a suitable excipient such as sucrose.

[0462] When considering parenteral administration, the therapeutic composition of the present invention can be provided in a pharmaceutically acceptable vehicle in the form of a pyrogen-free, parenterally acceptable aqueous solution comprising the required antigen binding molecules of the present invention. The particularly suitable vehicle for parenteral injection is sterile distilled water, in which the antigen binding molecules of the present invention are mixed with the sterile isotonic solution of suitable preservation. In certain embodiments, the preparation can be related to the preparation of desired molecules with the control or sustained release of a product (the product can be delivered via reservoir injection) provided (such as injectable microspheres, biologically erodible particles, polymer compounds (such as polylactic acid or polyglycolic acid), beads or liposomes). In certain embodiments, hyaluronic acid can also be used, and the hyaluronic acid has the effect of promoting the duration of circulation. In certain embodiments, implantable drug delivery devices can be used to introduce desired antigen binding molecules.

[0463] Other pharmaceutical compositions will be apparent to those skilled in the art, including the preparations related to the antigen binding molecules of the present invention being made into continuous or controlled delivery / release preparations. It is also known to those skilled in the art to prepare various other continuous or controlled delivery methods (such as liposome carriers, bioerodible microparticles or porous beads and reservoir injections). Refer to, for example, International Patent Application No. PCT / US93 / 00829, which describes the controlled release of porous polymer microparticles for delivering pharmaceutical compositions. Sustained release formulations may include a semipermeable polymer matrix in the form of a molded article (such as a film or microcapsule). Sustained release matrices may include polyesters, hydrogels, polylactides (as disclosed in U.S. Pat. No. 3,773,919 and European Patent Application Publication No. EP 058481), copolymers of L-glutamic acid and γ-ethyl-L-glutamic acid (Sidman et al., 1983, Biopolymers [Biopolymers] 2: 547-556), poly (2-hydroxyethyl methacrylate) (Langer et al., 1981, J. Biomed. Mater. Res. [Journal of Biomedical Materials Research] 15: 167-277 and Langer, 1982 Chem. Tech. [Chemical Technology] 12: 98-105), ethylene vinyl acetate (Langer et al., 1981, supra) or poly-D (-) -3-hydroxybutyric acid (European Patent Application Publication No. EP 133,988). Sustained release composition...

Claims

1. A molecule comprising at least two polypeptide chains, wherein the molecule comprises (i.) a first binding domain that binds to a first target cell surface antigen (TAA1), (ii.) a second binding domain which binds to an extracellular epitope of the human and / or macaque CD3 chain, (iii.) a third binding domain that binds to a second target cell surface antigen (TAA2), and (iv.) a fourth binding domain, which binds to an extracellular epitope of the human and / or macaque CD3 chain, wherein 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, and wherein the molecule further comprises a spacer entity selected from: (1.) A dimerization domain selected from the group consisting of: (a.) An Fc domain comprising a first and a second polypeptide monomer comprising a hinge, a CH2 domain, and a CH3 domain, respectively, wherein the first and the second polypeptide monomers form a heterodimer; wherein the heterodimer is formed by: - a charge pair mutation selected from the group consisting of: (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, K370D, and E356K. 0E and K360E, (vii.) D399K, K409D, K392D, E357K, K370E, E356K, and K439E, and (viii.) D399K, K409D, K392D, E357K, K370E, K360E, E356K, and K439D, preferably comprising 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, wherein positions are numbered according to EU; or - a knob-in-hole mutation, which preferably comprises T366S, L368A and Y407V mutations in the first polypeptide monomer and T366W mutation in the second monomer, wherein the positions are numbered according to EU; (b.) a human serum albumin (HSA) domain comprising a first and a second polypeptide monomer, wherein the first and the second polypeptide monomer correspond to HSA subdomains, respectively, wherein the first and the second polypeptide monomers form a native HSA-like heterodimer; and (c.) a Fab comprising a first and a second polypeptide monomer, wherein preferably the first polypeptide monomer comprises a VL and a CL domain, and the second polypeptide monomer comprises a VH and a CH1 domain, wherein the CL and CH1 domains are connected by a disulfide bridge; wherein the dimerization domain comprises two N-termini and two C-termini, respectively, wherein at least one N-terminus and one C-terminus are respectively connected to a bispecific entity, wherein any one of the first, second, third and fourth domains can be selected from any form of binding domain, preferably selected from Fab and a single chain domain, and the single chain domain is preferably selected from a single chain Fv (scFv) and scFab; (2.) a single chain domain selected from the group consisting of ubiquitin, β2 microglobulin, a VH domain only, a PSI domain from a Met-receptor, a fibronectin type III domain from tenascin, granulocyte-macrophage colony stimulating factor (GM-CSF), interleukin-4, interleukin-2, a PD-1 binding domain from human programmed cell death 1 ligand 1 (PDL1), Tim-3 (AS24-130), MiniSOG, a programmed cell death protein 1 (PD1) domain, human serum albumin (HSA), or a plurality of sequences each comprising a hinge, a CH2 and a C A single-chain Fc (scFc) domain of two polypeptide monomers comprising an H3 domain, a hinge and additional CH2 and CH3 domains, wherein the two polypeptide monomers are fused to each other via a peptide linker, wherein the single-chain domain comprises an N-terminus and a C-terminus, which are respectively connected to a bispecific entity, wherein at least one of the first, second, third and fourth binding domains is 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 a single-chain domain, the single-chain domain is preferably selected from scFv and scFab; The distance between the Cα atom of the first amino acid at the N-terminus and the last amino acid at the C-terminus of the spacer entity is At least wherein the spacer entity separates the first bispecific entity from the second bispecific entity by at least about wherein the indicated distance is preferably understood as the distance between the centers of mass of (i.) the first binding domain and the third binding domain or (ii.) the first bispecific entity and the second bispecific entity, and the spacer entity is located between the first bispecific entity and the second bispecific entity.

2. The molecule according to claim 1, which is an antigen binding molecule, preferably a bispecific antigen binding molecule, more preferably a multi-chain multi-targeting bispecific antigen binding molecule.

3. The antigen binding molecule according to claim 2, wherein when the spacer is a single-chain domain, the arrangement of the binding domains from amino to carboxyl sequence is selected from the group consisting of: (i.) The first and second domains, the spacer, the third and fourth domains (ii.) The first and second domains, the spacer, the fourth and third domains (iii.) the second and first domains, a spacer, the third and fourth domains, and (iv.) The second and first domains, a spacer, and the fourth and third domains.

4. The antigen binding molecule according to any one of the preceding claims, wherein when the spacer is a single-chain domain, the arrangement of the binding domains in order from amino to carboxyl is selected from the group consisting of: (i.) a first domain in the form of a Fab, a second domain preferably in the form of a single chain domain of an scFv, a spacer, a third domain in the form of a Fab, and a fourth domain preferably in the form of a single chain domain of an scFv; (ii.) the first domain in Fab form, the second domain in Fab form, a spacer, the third domain in Fab form, and the fourth domain in Fab form; (iii.) preferably a first domain in the form of a single chain domain 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.) preferably a first domain in the form of a single chain domain 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, preferably in the form of a single chain domain of an scFv, a second domain, preferably in the form of a single chain domain of an scFv, a spacer, a third domain in the form of a Fab, and a fourth domain, preferably in the form of a single chain domain of an scFv; (vi.) a first domain in the form of a Fab, a second domain preferably in the form of a single chain domain of an scFv, a spacer, a third domain preferably in the form of a single chain domain of an scFv, and a fourth domain preferably in the form of a single chain domain of an scFv; and (vii.) a first domain, preferably in the form of a single chain domain of an scFv, a second domain, preferably in the form of a Fab, a spacer, a third domain, preferably in the form of a single chain domain of an scFv, and a fourth domain, preferably in the form of a single chain domain of an scFv, Each scFv comprises, from amino to carboxyl, VH, a linker and VL or VL, a linker and VH, preferably VH, a linker and VL.

5. The antigen binding molecule according to any one of the preceding claims, wherein when the spacer is a dimerization domain, the arrangement of the binding domains in order from amino to carboxyl is selected from the group consisting of: (i.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that together with the first chain form a Fab, a second domain in scFv form, a first polypeptide monomer of the spacer dimerization domain, and a third chain comprising the second polypeptide monomer of the spacer dimerization domain, a third domain comprising VH and CH1 of the third domain that together with VL and CL of the third domain on the fourth chain form a Fab, and a fourth chain comprising VL and CL of the third domain and a fourth domain in scFv form; (ii.) a first domain in Fab form, a second domain in Fab form, a spacer comprising a first and a second polypeptide monomer of the spacer dimerization domain, a third domain in Fab form, and a fourth domain in Fab form; (iii.) a first chain comprising the second domain in scFv format, VH and CH1 of the first domain that together with the second chain form a Fab, a 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, VH and CH1 of the third domain that together with VL and CL of the third domain on the fourth chain form a Fab, and a fourth chain comprising VL and CL of the third domain and the fourth domain in scFv format; (iv.) a first chain comprising a second domain in scFv format, VH and CH1 of the first domain that together with the second chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, a second chain comprising VL and CL of the first domain, a second polypeptide monomer comprising the spacer dimerization domain, a fourth domain in scFv format, a third chain comprising VH and CH1 of the third domain that together with VL and CL of the third domain on the fourth chain form a Fab, and a fourth chain comprising VL and CL of the third domain; (v.) a first chain comprising a second domain in scFv format, VH and CH1 of the first domain that together with the second chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, a second chain comprising VL and CL of the first domain, a second polypeptide monomer comprising the spacer dimerization domain, a fourth domain in scFv format, a third chain comprising VH and CH1 of the third domain that together with VL and CL of the third domain on the fourth chain form a Fab, and a fourth chain comprising VL and CL of the third domain; Each scFv comprises, from amino to carboxyl, VH, a linker and VL or VL, a linker and VH, preferably VH, a linker and VL.

6. The antigen binding molecule according to any one of the preceding claims, wherein when the spacer is a dimerization domain, the arrangement of the binding domains is in amino to carboxyl order and is selected from the group consisting of: (i.) comprising a first domain in scFv form, a first polypeptide monomer of the spacer dimerization domain, a first chain of the third domain in scFv form, and comprising a second domain in scFv form, a second polypeptide monomer of the spacer dimerization domain, and a second chain of the fourth domain in scFv form; (ii.) comprising a first domain in scFv form, a first polypeptide monomer of the spacer dimerization domain, a first chain of a second domain in scFv form, and comprising a third domain in scFv form, a second polypeptide monomer of the spacer dimerization domain, and a second chain of a fourth domain in scFv form; (iii.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that together with the first chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, and a third chain comprising the second domain in the scFv form, a second polypeptide monomer of the spacer dimerization domain, and a fourth domain in the scFv form; (iv.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that together with the first chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, a second domain in scFv form, and a third chain comprising a fourth domain in scFv form, a second polypeptide monomer of the spacer dimerization domain, and a third domain in scFv form; (v.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that together with the first chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, and a third chain comprising the second domain in scFv form, the second polypeptide monomer of the spacer dimerization domain, the third domain in scFv form, and the fourth domain in scFv form; (vi.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that together with the first chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, a third domain in scFv form, and a fourth domain in scFv form, and a third chain comprising the second domain in scFv form, a second polypeptide monomer of the spacer dimerization domain; (vii.) a first chain comprising VL and CL of the first domain, a first polypeptide monomer comprising VH and CH1 of the first domain that together with the first chain form a Fab, a second chain comprising VH and CH1 of the third domain that together with the third chain form a Fab, a third chain comprising VL and CL of the third domain, a fourth chain comprising VH and CH1 of the second domain that together with the fifth chain form a Fab, 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; (viii.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that together with the first chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, a second chain comprising a fourth domain in the form of an scFv, a second polypeptide monomer of the spacer dimerization domain, a third chain comprising VH and CH1 of the third domain that together with the fourth chain form a Fab, and a fourth chain comprising VL and CL of the third domain; (ix.) comprising a first domain in scFv format, a first polypeptide monomer of the spacer dimerization domain, a first chain of the third domain in scFv format, VH and CH1 that together with the third chain form the second domain of Fab, a second polypeptide monomer of the spacer dimerization domain, a second chain of VH and CH1 that together with the fourth chain form the fourth domain of Fab, a third chain comprising VL and CL of the second domain, and a fourth chain comprising 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 that together with the first chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, and a third chain of the third domain in the form of an scFv, a fourth chain comprising VH and CH1 of the second domain that together with the fourth chain form a Fab, a second polypeptide monomer of the spacer dimerization domain, 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 that together with the first chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, and a second chain of the second domain in scFv form, a third chain comprising VH and CH1 of the fourth domain that together with the fourth chain form a Fab, a second polypeptide monomer of the spacer dimerization domain, and a third chain of the third domain in scFv form, and a fourth chain comprising VL and CL of the fourth domain; (xii.) a first chain comprising VL and CL of the first domain, a second chain comprising VH and CH1 of the first domain that together with the first chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, a second polypeptide monomer of the spacer dimerization domain that together with the fourth chain form a Fab, a third domain in scFv form, a third chain of the fourth domain in scFv form, 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 that together with the first chain form a Fab, a first polypeptide monomer of the spacer dimerization domain, a third domain in scFv form, a fourth domain in scFv form, a third chain comprising VH and CH1 of the second domain that together with the fourth chain form a Fab, a second polypeptide monomer of the spacer dimerization domain, and a fourth chain comprising VL and CL of the second domain; Each scFv comprises VH, linker and VL or VL, linker and VH in the N-to-C orientation, preferably VH, linker and VL.

7. An antigen binding molecule according to any one of the preceding claims, wherein the spacer entity is a globular protein, wherein the distance between the Cα atoms of the first amino acid at the N-terminus and the last amino acid at the C-terminus is at least Preferably at least More preferably at least so that the first bispecific entity and the second bispecific entity are effectively spaced apart, preferably at least 8. An antigen binding molecule according to any of the preceding claims, wherein the spacer entity that sufficiently separates the first bispecific entity and the second bispecific entity and is a single chain is selected from the group consisting of: ubiquitin, β2 microglobulin, SAND domain, green fluorescent protein (GFP), VHH antibody lama domain, PSI domain from Met receptor, fibronectin type III domain from tenascin, granulocyte-macrophage colony stimulating factor (GM-CSF), interleukin-4, CD137L extracellular domain, interleukin-2, PD-1 domain from human programmed cell death 1 ligand 1 (PDL1), A binding domain, Tim-3 (AS24-130), MiniSOG, a programmed cell death protein 1 (PD1) domain, human serum albumin (HSA) or a derivative of any of the foregoing spacer entities, a multimer of a rigid linker, and an Fc domain or a dimer or trimer thereof, each Fc domain comprising two polypeptide monomers, each polypeptide monomer comprising a hinge, a CH2 and CH3 domains, a hinge and additional CH2 and CH3 domains, wherein the two polypeptide monomers are fused to each other via a peptide linker, or wherein the two polypeptide monomers are linked together to form a heterodimer via non-covalent CH3-CH3 interactions and / or covalent disulfide bonds.

9. An antigen binding molecule according to any one of the preceding claims, wherein when single chain, the spacer entity is at least one Fc domain, preferably one domain or two or three covalently linked domains, each of which, in order from amino to carboxyl, comprises: Hinge-CH2-CH3-Joint-Hinge-CH2-CH3.

10. The antigen binding molecule according to any one of the preceding claims, wherein each of the polypeptide monomers of 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 NO: 17-24, wherein preferably each of the polypeptide monomers has an amino acid sequence selected from SEQ ID NO: 17-24.

11. The antigen binding molecule according to any one of the preceding claims, wherein the CH2 domain in the spacer comprises an intradomain cysteine ​​disulfide bridge.

12. The antigen binding molecule according to any one of the preceding claims, wherein the spacer entity comprises an amino acid sequence selected from the group consisting of: SEQ ID NO: 13 and 15 to 16 and 25 to 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 lama domain (SEQ ID NO: 1086), PSI domain from Met receptor (SEQ ID NO: 1087), fibronectin type III domain 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 from human programmed cell death 1 ligand 1 (PDL1) (SEQ ID NO: 1093), NO:1093), Tim-3 (AS24-130) (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 thereof having at least 90%, preferably 95% or even 98% sequence identity, preferably scFc (SEQ ID NO:25).

13. An antigen binding molecule according to any one of the preceding claims, wherein the first peptide monomer of the first peptide chain in the dimerization spacer is SEQ ID NO 35 and the second peptide monomer of the second peptide chain in the dimerization spacer is SEQ ID NO 36, wherein the two peptide monomers preferably form a heterodimer.

14. An antigen binding molecule according to any one of the preceding claims, wherein the antigen binding molecule is characterized in that (i) the first domain and the third domain comprise two antibody-derived variable domains, and the second domain and the fourth domain comprise two antibody-derived variable domains; (ii) the first domain and the third domain comprise one antibody-derived variable domain, and the second domain and the fourth domain comprise two antibody-derived variable domains; (iii) the first domain and the third domain comprise two antibody-derived variable domains, and the second domain and the fourth domain comprise one antibody-derived variable domain; or (iv) the first domain comprises an antibody derived variable domain, and the third domain comprises an antibody derived variable domain.

15. An antigen binding molecule according to any one of the preceding claims 1 to 7, wherein the antigen binding molecule comprises two polypeptide chains, wherein The first polypeptide chain comprises the VH of the first domain, the VH of the second domain, preferably a first polypeptide monomer comprising hinge, CH2 and CH3 domains, the VH of the third domain and the VH of the fourth domain; and The second polypeptide chain comprises the VL of the first domain, the VL second domain, preferably a first polypeptide monomer comprising hinge, CH2 and CH3 domains, the VL of the third domain and the VL of the fourth domain, Preferably, the first and second polypeptide monomers form a heterodimer, thereby connecting the first and second polypeptide chains.

16. An antigen binding molecule according to any one of the preceding claims, wherein the antigen binding molecule, wherein the first, second, third and fourth binding domains each comprise a VH domain and a VL domain in an amino to carboxyl order, wherein the VH and VL within each domain are connected by a peptide linker, preferably a flexible linker comprising serine, glutamine and / or glycine as amino acid structural units, preferably only serine (Ser, S) or glutamine (Gln, Q) and glycine (Gly, G), more preferably (G4S)n or (G4Q)n, even more preferably SEQ ID NO: 1 or 3.

17. An antigen binding molecule according to any one of the preceding claims, wherein the peptide linker comprises S(G4X)n and (G4X)n or consists of, wherein X is selected from the group consisting of Q, T, N, C, G, A, V, I, L and M, and wherein n is an integer selected from integers 1 to 20, preferably wherein n is 1, 2, 3, 4, 5 or 6, preferably wherein X is Q, wherein preferably the peptide linker is (G4X)n, n is 3, and X is Q.

18. An antigen binding molecule according to any one of the preceding claims, wherein the peptide linker between the first binding domain and the second binding domain and between the third binding domain and the fourth binding domain is preferably a flexible linker, which comprises serine, glutamine and / or glycine or glutamic acid, alanine and lysine as amino acid structural units, preferably selected from the group consisting of SEQ ID NO: 1 to 4, 6 to 12 and 1125.

19. An antigen binding molecule according to any one of the preceding claims, wherein the peptide linker between the first binding domain or the second binding domain and the spacer, and / or the peptide linker between the third binding domain and the fourth binding domain and the spacer are preferably short linkers rich in small amino acids and / or hydrophilic amino acids, preferably glycine, respectively and are preferably SEQ ID NO:

5.

20. The antigen binding molecule according to any one of the preceding claims, wherein any one of the first target cell surface antigen and the second target cell surface antigen is selected from the group consisting of: CS1, BCMA, CDH3, FLT3, CD123, CD20, CD22, EpCAM, MSLN and CLL1.

21. The antigen binding molecule according to any one of the preceding claims, wherein the first target cell surface antigen and the second target cell surface antigen are not the same.

22. The antigen binding molecule according to any one of the preceding claims 1 to 20, wherein the first target cell surface antigen and the second target cell surface antigen are the same.

23. An antigen binding molecule according to any one of the preceding claims, wherein the first binding domain is capable of binding to the first target cell surface antigen and simultaneously the third binding domain is capable of binding to the second target cell surface antigen, preferably wherein the first target cell surface antigen and the second target cell surface antigen are on the same target cell.

24. according to the antigen binding molecule described in any one of the preceding claims, wherein the first target cell surface antigen and the second target cell surface antigen are respectively 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.

25. An antigen binding molecule according to any one of the preceding claims, 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 wherein the first and / or third binding domains of the antigen binding molecule of the present invention bind to human MSLN epitope cluster E1 (SEQ ID NO: 1175, according to aa 296-346 positions of Kabat) as determined by mouse chimeric sequence analysis as described herein, but preferably do not bind to human MSLN epitope cluster E2 (SEQ ID NO: 1176, according to aa 247-384 positions of Kabat), E3 (SEQ ID NO: 1177, according to aa 385-453 positions of Kabat), E4 (SEQ ID NO: 1178, according to aa 454-501 positions of Kabat) and / or E5 (SEQ ID NO: 1179, according to aa 502-545 positions of Kabat).

26. An antigen binding molecule according to any one of the preceding claims, wherein the first target cell surface antigen and / or the second target cell surface antigen is human CDH3 (SEQ ID NO: 1170), and wherein the first and / or third binding domain of the antigen binding molecule according to claim 1 binds to human CDH3 epitope clusters D2B (SEQ ID NO: 1171, aa 253-290 positions according to Kabat), D2C (SEQ ID NO: 1172, aa 291-327 positions according to Kabat), D3A (SEQ ID NO: 1173, aa 328-363 positions according to Kabat) and D4B (SEQ ID NO: 1174, aa 476-511 positions according to Kabat), preferably D4B (SEQ ID NO: 1174, aa 476-511 positions according to Kabat), as determined by murine chimeric sequence analysis as described herein.

27. An antigen binding molecule according to any of the preceding claims, wherein the second binding domain and the fourth binding domain (CD3 binding domain) both have (i.) an affinity characterized by a KD value of less than about 1.2 x 10-8M as measured by surface plasmon resonance (SPR), or (ii.) an affinity characterized by a KD value of about 1.2 x 10-8M as measured by SPR.

28. An antigen binding molecule according to any of the preceding claims, wherein the second binding domain and the fourth binding domain (CD3 binding domain) have an affinity characterized by a KD value of about 1.0 x 10-7 to 5.0 x 10-6 M measured by SPR, preferably about 1.0 to 3.0 x 10-6 M, more preferably about 2.5 x 10-6 M measured by SPR.

29. An antigen binding molecule according to any of the preceding claims, wherein the second binding domain and the fourth binding domain (CD3 binding domain) have an affinity characterized by a KD value of about 1.0 x 10-7 to 5.0 x 10-6 M measured by SPR, preferably about 1.0 to 3.0 x 10-6 M, more preferably about 2.5 x 10-6 M measured by SPR.

30. An antigen binding molecule according to any of the preceding claims, wherein the second binding domain and the fourth binding domain (CD3 binding domain) each individually have an activity that is at least about 10 times, preferably at least about 50 times or more preferably at least about 100 times lower than a CD3 binding domain comprising VH according to SEQ ID NO 43 and VL according to SEQ ID NO 44 (i.e. in a single targeting environment compared to a dual targeting environment).

31. according to the antigen binding molecule in any one of the preceding claims, wherein the second binding domain and the fourth binding domain comprise a VH region and a VL region, the VH region comprises a CDR-H1, a CDR-H2 and a CDR-H3 selected from SEQ ID NOs 37 to 39, 45 to 47, 53 to 55, 61 to 63, 69 to 71, 436 to 438, 1126 to 1128, 1136 to 1138, 1142 to 1144, 1148 to 1150 and 1217 to 1219, and the VL region comprises a CDR-H2, a CDR-H3 selected from SEQ ID NOs 40 to 42, 48 to 50, 56 to 58, 64 to 66, 72 to 74, 439 to 441, 1129 to 1131, 1139 to 1141, 1145 to 1147, 1151 to 1153, and 1220 to 1222, preferably CDR-L1, CDR-L2, and CDR-L3 of 61 to 63 and 64 to 66 or 1217 to 1219 and 1220 to 1222.

32. An antigen binding molecule according to any one of the preceding claims, 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 or 1223.

33. An antigen binding molecule according to any one of the preceding claims, wherein the second binding domain and the fourth binding domain comprise a VL region selected from SEQ ID NO 44, 52, 60, 68, 76, 443, 1133 and 1224, preferably 68 or 1224.

34. according to the antigen binding molecule described in any one of the preceding claims, wherein the second binding domain and the fourth binding domain comprise a VH region selected from SEQ ID NO 43, 51, 59, 67, 75, 442, 1132 and 1223, preferably 67, and a VL region selected from SEQ ID NO 44, 52, 60, 68, 76, 443, 1133 and 1224, preferably 68, wherein when the VH region is 1132 and the VL region is 1133, the second binding domain and / or the fourth binding domain additionally comprise a CH1 domain of SEQ ID NO: 1134 and a CLK domain of SEQ ID NO: 1135 as a scFab domain, and wherein the VH region and the VL region are connected to each other by a linker preferably selected from SEQ ID NO 1, 3 and 1125, or wherein the SEQ ID NO of the VH of the VH-CH1 of the second domain and the fourth domain is SEQ ID NO: 1223, the SEQ ID NO of CH1 of VH-CH1 of the second domain and the fourth domain is SEQ ID NO 1224, the SEQ ID NO of VL of VL-CL of the second domain and the fourth domain is SEQ ID NO 1225, and the SEQ ID NO of CL of VL-L of the second domain and the fourth domain is SEQ ID NO 1226.

35. according to the antigen binding molecule described in any one of the preceding claims, wherein the first and / or the third (target) binding domain binds CDH3 and comprises a VH region, the VH region comprises SEQ ID NO:1154 as CDR-H1, wherein X1 (the number following "X" represents the numerical order of "X" in each amino acid sequence of N- to C- orientation in the sequence list) is S or N, X2 is Y or S, X3 is P or W, X4 is I or M and X5 is Y, N or H; SEQ ID NO: 1155 as CDR-H2, wherein 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; and X13 is K or Q; and SEQ ID NO:1156 as CDR-H3, wherein 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 wherein the first and / or the third (target) binding domain binds CDH3 and comprises a VL region comprising SEQ ID NO:1158 as CDR-L1, wherein 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; and X9 is A, N or H; SEQ ID NO:1159 as CDR-L2, wherein X1 is Y, G, W, or 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; and X7 is S or E; and SEQ ID NO:1160 as CDR-L3, wherein 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; and X7 is F, Y or L.

36. according to the antigen binding molecules described in any one of the preceding claims, wherein the first and / or the third (target) binding domain binds MSLN and comprises a VH district, the VH district comprises SEQ ID NO:1162 as CDR-H1, wherein X1 (the numerical order of "X" in each amino acid sequence of N to C orientation in the sequence table after "X") is S, G or D; X2 is Y, A, G or F; X3 is I, W or M; and X4 is V, S, G, T or H; SEQ ID NO: 1163 as CDR-H2, wherein 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; and X15 is G or S; and SEQ ID NO: 1164 as CDR-H3, wherein 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; and X16 is Y, I or V; and wherein the first and / or the third (target) binding domain binds MSLN and comprises a VL region comprising SEQ ID NO: 1166 as CDR-L1, wherein 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; and X11 is Y or A; SEQ ID NO: 1167 as CDR-L2, wherein 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; and X7 is S or T; and SEQ ID NO: 1168 as CDR-L3, wherein 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; and X8 is V or T.

37. An antigen binding molecule according to any one of the preceding claims, wherein the first and / or the third (target) binding domain binds CDH3 and comprises a VH region of SEQ ID NO: 1157, wherein (the number following "X" represents the numerical order of "X" in each amino acid sequence in the N to C orientation in the sequence list) 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, X 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; and the VL region of SEQ ID NO 1161, wherein X1 is D or E; X2 Q or V; X3 is L, M; X4 is A, S or D; X5 is F, S or T; X6 is A or S; X7 is A or V; X8 is P, V or L; X9 is D or E; X10 is A or V; X11 is I or L; X12 is T, S or N; X13 is K or R; X14 is A, S; or X15 is Q, D, S, G or E; X16 is S, D or N; X17 is V, I or L; X18 is K, Y, S or H; X19 is S or N; X20 is F, L or M; X21 is A, N or H; X22 is K or Q; X23 is A, P or V; X24 is K or R; X25 is I or V; X26 is Y, G, W or N; X27 is T or A; X28 is S or K; X29 is T, N or R;X30 is L or R; X31 is E, A, V or H; X32 is S or E; X33 is A, S, V or D; X34 is D or E; X35 is T or K; X36 is S or R; X37 is A, S or P; X38 is F or V; X39 is A, G; X40 is T or V; X41 is Q or V; X42 is Q, N, H; X43 is F, L, Y, W, N or H; X44 is A, D, Y, S or N; X45 is Q, R, S, G, W or M; X46 is F, Y or T; X47 is F, Y or L; X48 is V or L; and X49 is D or E (wherein all aa at each position preferably mean in an alternative "or" manner, even if not explicitly stated). ; 38. according to the antigen binding molecule described in any one of the preceding claims, wherein the first and / or the third (target) binding domain binds MSLN and comprises the VH district of SEQ ID NO:1165, wherein (the number following "X" represents the numerical order of "X" in each amino acid sequence of N to C orientation in the sequence list) X1 is E or Q; X2 is V, L or Q; X3 is E or Q; X4 is A, G or P; X5 is E or G; X6 is V or L; X7 is V or K; X8 is K or Q; X9 is G or S; X10 is E, A, G or R; X11 is S or T; X12 is V or L; X13 is R, S or K; X14 is V or L; X15 is S or T; X16 is A, K or T; X17 is A or V; X18 is Y, I or F; X19 is S or T; X2 0 is S or F; X21 is S or T; X22 is D, G or S; X23 is Y, G, A or F; X24 is I, W or M; X25 is G, S, V, T or H; X26 is I or V; X27 is A or P; X28 is M, K or Q; X29 is G or C; X30 is I, M, V or L; X31 is A, G or S; X32 is A, S, N, W, Y or V; X33 is Y, S or N; X34 is Y, G, P or S; X35 is D, H, S or N; X36 is G or S; X37 is E, G or S; X38 is G, S, N, F, T or Q; X39 is S, K, W, D, I or T; X40 is Y or N; X41 is A or N; X42 is A, P, N, E, D, I or Q; X43 is D, A, S or K; X44 is V, L or F; X45 is K, Q; X46 is G or S; X47 is V or F; X48 is I or M; X49 is S or T; X50 is R or V; X51 is N or T; X52 is A or S; X53 is I or K; X54 is S or N; X55 is S, T or Q; X56 is A, L or F; X57 is Y, S or F; X58 is L or M; X59 is E, K or Q; X60 is M or L; X61 is S or N; X62 is R or S ; X63 is V or L; X64 is R or T; X65 is A or S; X66 is D, A or E; X67 is R or K; X68 is D, E, V or L; X69 is E, R, G or P; X70 is R, A, N or Y; X71 is G, S, Y, V or H; X72 is A, P, F, D or Y; X73 is R or G; X74 is M, R, S or D; X75 is E or G; X76 is Y or L; X77 is Y or F; X78 is Y, S or F; X79 is A, G, S, T or H; X80 is L, M or F; X81 is Y, I or V; and X82 is L, M or T; and SEQ VL region of ID NO 1169 (the number after "X" indicates the numerical order of "X" in each amino acid sequence in the N to C orientation in the sequence list) X1 is E, S or D; X2 is Y, I or L; X3 is E, V or T; X4 is V, L or M;X5 is P or S; X6 is G or S; X7 is S or T; X8 is V or L; X9 is A, V or L; X10 is P or V; X11 is E, Q or D; X12 is R or T; X13 is A or V; X14 is S or T; X15 is I or L; X16 is S or T; X17 is A or S; X18 is G or S; X19 is E or Q; X20 is G, S or K; X21 is I, V, L or F; X X22 is R, G or S; X23 is D or S; X24 is A, S, N, K or T; X25 is Y, W or M; X26 is V or L; X27 is Y or A; X28 is K or Q; X29 is A, S or V; X30 is R, V or K; X31 is V or L; X32 is A, G or Q; X33 is A or S; X34 is S or T; X35 is G, S, K, I or T; X36 is R or L; X37 is Y or A; X38 is K or Q; X39 is A or S; X40 is R or L; X41 is V or L; X42 is A, G or Q; X43 is A or S; X44 is S or T; X45 is G, S, K, I or T; X46 is R or L; X47 is Y or A; X48 is K or Q; X49 is A, S or V; X50 is R, V or K; X51 is V or L; X52 is A, G or Q; X53 is A or S; X54 is S or T; X55 is G, S, K, I or T; X56 is R or L; X57 is Y or A; X58 is K or Q; X59 is K or I; X60 is R or L; X61 is K or I; X62 is R or L; X63 is K or I; X64 is K or I; X65 is K or I; X66 is R or L; X67 is Y or A; X68 is K or I; X69 is K or I; X70 is K or I; X71 is K or I; X72 is K or I 7 is A, P or Q; X38 is S or T; X39 is I or V; X40 is E, S or D; X41 is G or N; X42 is N or T; X43 is D or T; X44 is A or F; X45 is R, G or S; X46 is L or T; X47 is E or Q; X48 is A or P; X49 is E or M; X50 is E or F; X51 is D, V or T; X52 is A or Q; X53 is Y, S, A or T; X54 is G, E, Y, H or Q; X55 is A or S; X56 is S, T or F; X57 is P or T; X58 is R, A, L or F; X59 is V or T; X60 is P or C; X61 is V or L; X62 is E or T; X63 is I or V; and X64 is L or K (wherein all aa at each position preferably mean in an alternative "or" manner, even if not explicitly stated). ; 39. An antigen binding molecule according to any one of the preceding claims, wherein the first and / or the third (target) binding domain comprises a VH region comprising a CDR-H1, CDR-H2 and CDR-H3 selected from the group consisting of: SEQ ID NO:77 to 79, 86 to 88, 95 to 97, 103 to 105, 111 to 113, 119 to 121, 127 to 129, 135 to 137, 143 to 145, 151 to 153, 159 to 161, 168 to 170, 177 to 179, 185 to 187, 194 to 196, 203 to 205, 212 to 214, 221 to 223, 230 to 232, 238 to 240, 334 to 336, 356 to 358, 365 to 367, 376 to 378, 385 to 387 and 194, 432 and 196, 446 to 448, 454 to 45 6, 462 to 464, 470 to 472, 478 to 480, 486 to 488, 494 to 496, 502 to 504, 510 to 512, 518 to 520, 526 to 528, 534 to 536, 542 to 544, 550 to 552, 558 to 560, 566 to 568, 574 to 576, 582 to 584, 590 to 592, 598 to 600, 606 to 608, 614 to 616, 622 to 624, 630 to 632, 638 to 640, 646 to 648, 654 to 656, 662 to 664, 670 to 672, 678 to 68 0, 686 to 688, 694 to 696, 702 to 704, 710 to 712, 718 to 720, 726 to 728, 734 to 736, 742 to 744, 750 to 752, 758 to 760, 766 to 768, 774 to 776, 782 to 784, 790 to 792, 798 to 800, 806 to 808, 814 to 816, 822 to 826, 830 to 832, 838 to 840, 846 to 848, 854 to 856, 862 to 864, 870 to 872, 878 to 880, 886 to 888, 894 to 896, 902 to 911 to 1024, 1030 to 1032, 1038 to 1040, 1046 to 1048, 1054 to 1056, and 1062 to 1064, or any combination of CDR-H1, CDR-H2 and CDR-H3 as disclosed together in Table 6 of the Sequence Tab.,86 to 88 and 194, 432 and 196 are preferred for the first binding domain and the third binding domain, respectively, 194, 432 and 196 are more preferred for the first binding domain, 86 to 88 are more preferred for the third binding domain, or 1227 to 1229 and 1237 to 1239 are more preferred for the first binding domain and the third binding domain. , 40. An antigen binding molecule according to any one of the preceding claims, wherein the first and / or the third (target) binding domain comprises a VL region comprising a CDR-L1, CDR-L2 and CDR-L3 selected from the group consisting of: SEQ ID NO: 80 to 82, 89 to 91, 98 to 100, 106 to 108, 114 to 116, 122 to 124, 130 to 132, 138 to 140, 146 to 148, 154 to 156, 162 to 164, 171 to 173, 180 to 182, 188 to 190, 197 to 199, 206 to 208, 215 to 217, 224 to 226, 233 to 235, 241 to 243, 337 to 339, 359 to 361, 368 to 370, 379 to 381, 388 to 390, 449 to 451, 457 to 459, 465 to 467 , 473 to 475, 481 to 483, 489 to 491, 497 to 499, 505 to 507, 513 to 515, 521 to 523, 529 to 531, 537 to 539, 545 to 547, 553 to 555, 561 to 563, 569 to 571, 577 to 579, 585 to 587, 593 to 595, 601 to 603, 609 to 611, 617 to 619, 625 to 627, 633 to 635, 641 to 643, 649 to 651, 657 to 659, 665 to 667, 673 to 675, 681 to 683, 689 to 691, 697 to 699, 705 to 707, 713 to 715, 721 to 723, 729 to 731, 737 to 739, 745 to 747, 753 to 755, 761 to 763, 769 to 771, 777 to 779, 785 to 787, 793 to 795, 801 to 803, 809 to 811, 817 to 819, 825 to 829, 833 to 835, 841 to 843, 849 to 851, 857 to 859, 865 to 867, 873 to 875, 881 to 883, 889 to 891, 897 to 899, 905 to 907, to 1065, or any combination of CDR-L1, CDR-L2 and CDR-L3 as disclosed together in Table 6 of the Sequence Listing,Preferably 89 to 91 and 197 to 199 for the first binding domain and the third binding domain, respectively, more preferably 197 to 199 for the first binding domain, more preferably 89 to 91 for the third binding domain, or more preferably 1230 to 1232 and 1240 1242 for the first binding domain and the third binding domain. , 41. An antigen binding molecule according to any one of the preceding claims, wherein the first and / or the third (target) binding domain comprises a VH region selected from the group consisting of SEQ ID NO: 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, 5 24, 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, 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 as in Table 52 of the sequence listing, preferably 433 and 92 for the first binding domain and the third binding domain, respectively, more preferably 433 for the first binding domain, more preferably 92 for the third binding domain, or more preferably 1233+1235 and 1243+1245 for the first binding domain and the third binding domain (VH and CH1 in Fab).

42. An antigen binding molecule according to any one of the preceding claims, wherein the first and / or the third (target) binding domain comprises a VL region selected from the group consisting of 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, 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, 81 3, 821, 829, 837, 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 any VL as disclosed together in Table 52 of the Sequence Listing, preferably 200 and 93 for the first binding domain and the third binding domain, respectively, more preferably 200 for the first binding domain, more preferably 93 for the third binding domain, or more preferably 1234+1236 and 1244+1246 for the first binding domain and the third binding domain (VL and CL in Fab).

43. An antigen binding molecule according to any one of the preceding claims, wherein the first and / or third (target) binding domain comprises a VL region selected from SEQ ID NO: 85, 94, 193, 202, 211, 220, 229, 364, 384, 393, preferably 94 and 202 with increased stability by a single amino acid exchange (E to I).

44. An antigen binding molecule according to any one of the preceding claims, 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, or any other full-length multi-targeted bispecific antigen binding molecule disclosed in Table 6 of the Sequence Table.

45. A polynucleotide encoding the antigen-binding molecule according to any one of claims 1 to 44.

46. ​​A vector comprising the polynucleotide according to claim 45.

47. A host cell transformed or transfected with the polynucleotide according to claim 45 or the vector according to claim 46.

48. A method for producing an antigen binding molecule according to any one of claims 1 to 44, the method comprising culturing the host cell of the present invention under conditions allowing expression of the antigen binding molecule and recovering the produced antigen binding molecule from the culture.

49. A pharmaceutical composition comprising an antigen binding molecule according to any one of claims 1 to 38 or produced according to the method of claim 48.

50. The pharmaceutical composition of claim 49, which is stable at about -20°C for at least four weeks.

51. An antigen binding molecule according to claims 1 to 44 or produced according to the method of claim 48, for use in the prevention, treatment or alleviation of a disease selected from a proliferative disease, a neoplastic disease, a cancer or an immunological disorder.

52. The antigen binding molecule according to claim 51, wherein the disease is preferably acute myeloid leukemia (AML), non-Hodgkin's lymphoma (NHL), non-small cell lung cancer (NSCLC), pancreatic cancer and colorectal cancer (CRC).

53. A method for treating or ameliorating a proliferative disease, the method comprising administering to a subject in need thereof a molecule comprising at least one polypeptide chain, wherein the molecule comprises (i.) a first binding domain that binds to a first target cell surface antigen (TAA1), (ii.) a second binding domain which binds to an extracellular epitope of the human and / or macaque CD3 chain, (iii.) a third binding domain that binds to a second target cell surface antigen (TAA2), and (iv.) a fourth binding domain, which binds to an extracellular epitope of the human and / or macaque CD3 chain, wherein 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, and Wherein the molecule further comprises a spacer entity selected from: (1.) A dimerization domain selected from the group consisting of: (a.) An Fc domain comprising a first and a second polypeptide monomer comprising a hinge, a CH2 domain, and a CH3 domain, respectively, wherein the first and the second polypeptide monomers form a heterodimer; wherein the heterodimer is formed by: - a charge pair mutation selected from the group consisting of: (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, K370D, and E356K. 0E and K360E, (vii.) D399K, K409D, K392D, E357K, K370E, E356K, and K439E, and (viii.) D399K, K409D, K392D, E357K, K370E, K360E, E356K, and K439D, preferably comprising 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, wherein positions are numbered according to EU; or - a knob-in-hole mutation, which preferably comprises T366S, L368A and Y407V mutations in the first polypeptide monomer and T366W mutation in the second monomer, wherein the positions are numbered according to EU; (b.) a human serum albumin (HSA) domain, comprising a first and a second polypeptide monomer, wherein the first and the second polypeptide monomer correspond to HSA subdomains, respectively, wherein the first and the second polypeptide monomers form a native HSA-like heterodimer; and (c.) a Fab comprising a first and a second polypeptide monomer, wherein preferably the first polypeptide monomer comprises a VL and a CL domain, and the second polypeptide monomer comprises a VH and a CH1 domain, wherein the CL and CH1 domains are connected by a disulfide bridge; wherein the dimerization domain comprises two N-termini and two C-termini, respectively, wherein at least one N-terminus and one C-terminus are respectively connected to a bispecific entity, wherein any one of the first, second, third and fourth domains can be selected from any form of binding domain, preferably selected from Fab and a single chain domain, and the single chain domain is preferably selected from a single chain Fv (scFv) and scFab; (2.) a single chain domain selected from the group consisting of ubiquitin, β2 microglobulin, a VH domain only, a PSI domain from a Met-receptor, a fibronectin type III domain from tenascin, granulocyte-macrophage colony stimulating factor (GM-CSF), interleukin-4, interleukin-2, a PD-1 binding domain from human programmed cell death 1 ligand 1 (PDL1), Tim-3 (AS24-130), MiniSOG, a programmed cell death protein 1 (PD1) domain, human serum albumin (HSA), or a single chain Fc (scFc) domain comprising two polypeptide monomers each comprising a hinge, a CH2 and CH3 domain, a hinge and additional CH2 and CH3 domains, wherein the two polypeptide monomers are fused to each other via a peptide linker, wherein the single chain domain comprises an N-terminus and a C-terminus, which are respectively connected to a bispecific entity, wherein at least one of the first, second, third and fourth binding domains is 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 a single chain domain, and the single chain domain is preferably selected from scFv and scFab; The distance between the Cα atom of the first amino acid at the N-terminus and the last amino acid at the C-terminus of the spacer entity is At least wherein the spacer entity separates the first bispecific entity from the second bispecific entity by at least about , wherein the indicated distance is preferably understood as the distance between the centers of mass of (i.) the first binding domain and the third binding domain or (ii.) the first bispecific entity and the second bispecific entity, and the spacer entity is located between the first bispecific entity and the second bispecific entity, comprising the step of administering an antigen binding molecule of the invention or an antigen binding molecule produced according to the method of the invention to a subject in need thereof, wherein the disease is preferably acute myeloid leukemia, non-Hodgkin's lymphoma, non-small cell lung cancer, pancreatic cancer and / or colorectal cancer.

54. A method according to claim 53, wherein the method comprises addressing a disease-associated target that is significantly co-expressed on a pathological physiological tissue and one or more physiological tissues by providing a multi-targeted, bispecific antigen binding molecule in the form described herein, wherein the molecule addresses (i.) a target expressed on both the disease-associated tissue and the physiological tissue and (ii.) another target expressed on a physiological tissue associated with the disease but not under (i.), wherein the method preferably avoids the formation of intra-abdominal adhesions and / or fibrosis when such a target is MSLN.

55. The method of claim 53, wherein the disease is a neoplastic disease, cancer or an immunological disorder.

56. The method according to claim 55, wherein the disease is preferably acute myeloid leukemia, non-Hodgkin's lymphoma, non-small cell lung cancer, pancreatic cancer and / or colorectal cancer.

57. The method according to claim 47, wherein the 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.

58. A kit comprising an antigen binding molecule according to any one of claims 1 to 44 or produced according to the method of claim 48, a polynucleotide according to claim 45, a vector according to claim 46 and / or a host cell according to claim 47.

Images