Antibodies that bind EGFR and cMET

Abstract

The present application relates to antibodies that bind to EGFR and cMET. The invention, as disclosed herein, relates to bispecific antibodies comprising a first variable domain that can bind an extracellular portion of epidermal growth factor receptor (EGFR) and a second variable domain that can bind an extracellular portion of MET prooncogene receptor tyrosine kinase (cMET). The antibodies may comprise a common light chain, which may be human antibodies. The antibody may be a full-length antibody. In some embodiments, the bispecific antibody is an antibody in the form of IgG1 that is anti-EGFR, anti-cMET stoichiometric 1: 1. In some embodiments, the antibodies have a variable domain that can bind to EGFR and a variable domain that can bind to cMET.

Description

[0001] This application is a divisional application of the Chinese patent application with application number 201880059748.5 and invention name “Antibodies binding to EGFR and cMET”. The original application is the PCT international application PCT / NL2018 / 050537 filed on August 9, 2018, which entered the Chinese national phase on March 13, 2020. Technical Field

[0002] The present invention relates to the field of antibodies. In particular, the present invention relates to the field of therapeutic antibodies (including human antibodies) for treating diseases involving abnormal cells. In addition, the present invention relates to antibodies that bind to EGFR and cMET, including multispecific antibodies, and their use in binding to EGFR and cMET positive cells (particularly tumor cells). Background Art

[0003] Epidermal growth factor (EGF) receptor (EGFR) is a cell surface receptor for members of the epidermal growth factor family (EGF family) of extracellular protein ligands. EGFR is also called the ErbB-1 receptor. The receptor has been given a variety of names in the past (EGFR; ERBB; ERBB1; HER1; PIG61; mENA). In the present invention, the names of ErbB-1, EGFR or HER1 in humans will be used interchangeably. EGFR is a member of the ErbB receptor family, which is a subfamily of four closely related receptor tyrosine kinases: ErbB-1 (EGFR), ErbB-2 (HER2 / c-neu; Her2), ErbB-3 (Her 3) and ErbB-4 (Her 4).

[0004] EGFR is present on the cell surface and can be activated by the combination of its specific ligands (including epidermal growth factor and transforming growth factor α (transforming growth factor α, TGF α)). Once activated by its growth factor ligand, the receptor can undergo a conversion from an inactive majority of monomeric forms to an active homodimer. In addition to forming a homodimer after ligand binding, EGFR can also be paired with another member of the ErbB receptor family (e.g., ErbB2) to produce activated heterodimers. Dimers can also be formed in the absence of ligand binding, and can form activated EGFR clusters after ligand binding.

[0005] EGFR dimerization stimulates intrinsic intracellular protein-tyrosine kinase (PTK) activity. This activity induces several signal transduction cascades that lead to cell proliferation and differentiation. The kinase domain of EGFR can cross-phosphorylate tyrosine residues of other receptors in complex with it and can activate itself in this way.

[0006] Mutations involving EGFR have been identified in several types of cancer. It is the target of an expanding class of anticancer treatments. Such treatments include EGFR tyrosine kinase inhibitors (EGFR-TKIs), such as gefitinib and erlotinib for lung cancer, and antibodies, such as cetuximab and panitumumab for colon cancer and head and neck cancer.

[0007] Cetuximab and panitumumab are monoclonal antibodies that inhibit the receptor. Other monoclonal antibodies in clinical development are zalutumumab, nimotuzumab, and matuzumab. The purpose of monoclonal antibodies is to block extracellular ligand-induced receptor activation, primarily by blocking the binding of the ligand to the receptor. With the binding site blocked, the signal-inducing molecule cannot effectively attach and therefore cannot activate downstream signaling. Ligand-induced receptor activation can also be inhibited by stabilization of the inactive receptor conformation (matuzumab).

[0008] To date, EGFR targeted therapy has been associated with the development of treatment resistance over time. A variety of mechanisms for EGFR-TKI resistance have been described. In patients with advanced non-small cell lung cancer (NSCLC), resistance mechanisms include the occurrence of secondary mutations (e.g., T790M, C797S), activation of alternative signaling (e.g., Met, HGF, AXL, Hh, IGF-1R), abnormal downstream pathways (e.g., AKT mutations, PTEN deletions), EGFR-TKI-mediated apoptosis pathway damage (e.g., BCL2-like 11 / BIM deletion polymorphisms) and histological transformation. Although some resistance mechanisms have been identified, others remain to be identified. Similarly, colorectal cancer patients treated with EGFR antibodies also develop resistance over time. This can occur through the emergence of KRAS mutations. For those without KRAS mutations; Amplification of the MET proto-oncogene may be associated with acquired resistance during anti-EGFR therapy (Bardelli et al., 2013; Cancer Discov. Jun; 3(6): 658-73. doi: 10.1158 / 2159-8290. CD-12-0558). Tumors may be ab initio resistant or develop resistance during treatment. Resistance to EGFR targeted therapy is seen in many EGFR-positive cancers, and more effective EGFR cancer treatments have been demonstrated in the art that improve the standard of care and have excellent capabilities in dealing with EGFR targeted therapy resistance.

[0009] The dysregulation of MET proto-oncogene, receptor tyrosine kinase (cMET) and hepatocyte growth factor (HGF) has been reported in a variety of tumors. Ligand-driven cMET activation has been observed in several cancers. Elevated serum and intratumoral HGF has been observed in lung cancer, breast cancer and multiple myeloma (JM Siegfried et al., Ann Thorac Surg 66, 1915 (1998); PC Ma et al., Anticancer Res 23, 49 (2003); BE Elliott et al. Can J Physiol Pharmacol 80, 91 (2002); C. Seidel, et al, Med Oncol 15, 145 (1998)). Overexpression, amplification or mutation of cMET has been reported in a variety of cancers (e.g., colorectal cancer, lung cancer, gastric cancer, and renal cancer) and can drive ligand-independent receptor activation (C. Birchmeier et al, Nat Rev Mol Cell Biol 4, 915 (2003); G. Maulik et al., Cytokine Growth Factor Rev 13, 41 (2002)). HGF expression is also associated with activation of the HGF / cMET signaling pathway and is also one of the escape mechanisms of tumors selected by EGFR targeted therapy.

[0010] The cMET receptor is formed by proteolytic processing of a common precursor into a single-pass, disulfide-linked α / β heterodimer. The extracellular portion of cMET is composed of three domain types. The N-terminal region folds to form a large semaphorin (Sema) domain, which encompasses the entire α subunit and part of the β subunit. The plexin–semaphorin–integrin (PSI) domain follows the Sema domain and contains four disulfide bonds. This domain is connected to the transmembrane helices by four immunoglobulin–plexin–transcription (IPT) domains that are associated with immunoglobulin-like domains. Intracellularly, the cMET receptor contains a tyrosine kinase catalytic domain flanked by unique membrane-proximal and carboxyl-terminal sequences (Organ and Tsao. Therapeutic advances in medical oncology 3.1_suppl (2011): S7-S19, which is incorporated herein by reference in its entirety).

[0011] The ligand of cMET, hepatocyte growth factor (HGF; also called scatter factor) and its splice isoforms (NK1, NK2) are known ligands for the cMET receptor. HGF was identified as a potent mitogen / morphogen in 1991. The HGF / cMET signaling pathway plays an important role in the development and progression of a variety of cancers. Dysregulation and / or hyperactivation of HGF or cMET in human cancers is associated with a poor prognosis. cMET can be activated by overexpression, amplification or mutation. Activation can promote the development, progression, invasive growth and metastasis of cancer. cMET can be activated in an HGF-dependent and HGF-independent manner. HGF-independent activation occurs in the case of cMET overexpression. Abundant cMET can also trigger (hetero)dimerization and intracellular signaling in the absence of ligand. Other ligands do not appear to affect the function of such cMET-overexpressing cells. cMET amplification is associated with cMET overexpression and has become a biomarker for tumor subtypes.

[0012] HGF is ubiquitously expressed throughout the body, indicating that this growth factor is a systemically available cytokine and is derived from the tumor stroma. Positive paracrine and / or autocrine loops of cMET activation may lead to further expression of cMET. The HGF-specific antibody Rilotumumab (AMG102) is being developed for gastric cancer. Phase I and II trials appeared promising, but a Phase III study (RILOMET-2) with cisplatin and capecitabine as first-line treatment in gastric cancer was terminated following a safety review of Study 20070622 by a preplanned data monitoring committee.

[0013] The relevance of cMET / HGF signaling to resistance to EGFR-targeted therapies has stimulated the development of approaches to counteract resistance. To date, antibody-based approaches include anti-HGF antibodies; anti-cMET or cMET antibodies and anti-cMET / EGFR, which have not been clinically effective (reviewed in Lee et al., 2015; Immunotargets and Therapy 4:35–44). The cMET antibody Onartuzumab (MetMab TM ) and Emibetuzumab (LY-2875358) have been evaluated in Phase II clinical trials. Among them, Onartuzumab appears to be effective against colorectal cancer in combination therapy with the EGFR inhibitor erlotinib. However, these results may not be repeated in randomized Phase III clinical trials. MetMAb is a monovalent monoclonal antibody (mAb) against cMET that blocks the binding of HGF to cMET and subsequent pathway activation (Jin et al., 2008 Cancer Research Vol. 68: pp 4360-68).

[0014] To overcome the problems of anti-EGFR, cMET and HGF immunotherapy, the present invention provides a novel bispecific antibody comprising a first variable domain that can bind to the extracellular portion of the epidermal growth factor receptor (EGFR) and a second variable domain that can bind to the extracellular portion of the cMET proto-oncogene receptor tyrosine kinase (cMET).

[0015] To date, some bispecific EGFR x cMET antibodies have been described in the art. Castoldi R. et al. (2013) described a bispecific EGFR x cMET antibody named MetHer1, which has the cMET binding site of antibody 5D5 (or MetMab) and the EGFR binding site of cetuximab. The fixed EGFR and cMET binding stoichiometry of the bispecific antibody is 2:1 (see Supplementary Figure).

[0016] US20140378664 describes a cMET x EGFR bispecific antibody among other bispecific antibodies. The complete bispecific antibody was produced as a single protein and subsequently cleaved by proteolysis. The two VH / VL domains were produced as single-chain Fv fragments. Binding of the antibody induced cMET degradation and Akt phosphorylation in gastric cancer cell lines. Moores et al (2016) described a bispecific cMET x EGFR antibody designated JNJ-61186372 generated by controlled Fab-arm exchange (cFAE) with mutations at positions 405 and 409 according to EU numbering, which may be potentially immunogenic. JNJ-61186372 was shown to be active in vivo using a xenograft model of the tumor cell line H1975 expressing the cMET ligand HGF. It is known that this tumor model will depend on the ADCC activity of the antibody (Ahmed et al., 2015). JNJ-61186372 has been reported to have an affinity imbalance with approximately 40-fold higher affinity for cMET than for EGFR (Moores et al. (2016)), and the anti-EGFR arm derived from zaluzumab is known to cause infusion-related reactions, skin diseases, and other issues.

[0017] LY3164530 is a bispecific cMET x EGFR antibody comprising the EGFR binding domain of cetuximab as a single chain Fv fragment fused to the heavy chain variable domain of the cMET binding antibody LY2875358 (Emibetuzumab; Kim and Kim 2017). It is a so-called dual variable domain antibody comprising two binding sites for each antigen. No data on HGF inhibition of the antibody were provided. The antibody reportedly binds to cMET and EGFR and internalizes them without agonistic activity. The authors reviewed a variety of cMET, EGFR, and cMET x EGFR targeted therapies and concluded that none of these inhibitors have shown significant efficacy in clinical trials to date.

[0018] Therefore, there is a need for new bispecific cMET x EGFR antibodies, including those that may have the superior characteristics as described herein. Summary of the invention

[0019] In one aspect, the present invention provides a bispecific antibody comprising a first variable domain that can bind to the extracellular portion of human epidermal growth factor receptor (EGFR) and a second variable domain that can bind to the extracellular portion of human MET proto-oncogene receptor tyrosine kinase (cMET).

[0020] Bispecific antibodies may comprise a common light chain. The first and second variable domains preferably comprise the same or substantially the same (common) light chain variable region. The common light chain variable region may be those known to be well paired with the diverse human variable region gene segments undergoing recombination. More preferably, the common light chain is preferably a variable region encoded by a germline Vk gene segment, preferably an O12 / IgVκ1-39*01 variable region gene segment. The preferred light chain variable region comprises rearranged IgVκ1-39*01 / IGJκ1*01 or IgVκ1-39*01 / IGJκ5*01. The light chain of the cMET binding arm and the light chain of the EGFR binding arm are preferably the same (common) light chain. The common light chain is preferably a rearranged κ light chain IgVκ1-39*01 / IGJκ1*01 or IgVκ1-39*01 / IGJκ5*01 connected to a human light chain constant region. The bispecific antibody may be a human antibody. The bispecific antibody can be a full-length antibody. It can have a variable domain that can bind to EGFR and a variable domain that can bind to cMET. In one aspect, the variable domain that can bind to human EGFR can also beneficially bind to mouse EGFR and / or cynomolgus EGFR. The variable domain that can bind to human EGFR can bind to domain III of human EGFR. The variable domain that can bind to cMET can block the binding of antibody 5D5 to cMET. The variable domain that can bind to cMET can block the binding of HGF to cMET. The Kd of the cMET antibody can be at least 10 times smaller than the Kd of the EGFR antibody. The amino acids at positions 405 and 409 in one CH3 domain can be the same as the amino acids at the corresponding positions in another CH3 domain (EU numbering).

[0021] The first variable domain may comprise a heavy chain variable region having a CDR1 sequence of SYGIS, a CDR2 sequence of WISAYX1X2NTNYAQKLQG and a CDR3 comprising the sequence X3X4X5X6HWWLX7A

[0022] wherein X1=N or S; X2=A or G; X3=D or G; X4=R, S or Y; X5=H, L or Y; X6=D or W and X7=D or G; and there are 0 to 5 amino acid insertions, deletions, substitutions, additions or a combination thereof at positions other than X1 to X7.

[0023] The second variable domain may comprise a heavy chain variable region having an amino acid sequence of one of the sequences of SEQ ID NOs: 1 to 23 with 0 to 10, preferably 0 to 5, amino acid insertions, deletions, substitutions, additions or a combination thereof.

[0024] A bispecific antibody is described wherein

[0025] X1=N; X2=G; X3=D; X4=S; X5=Y; X6=W and X7=G;

[0026] X1=N; X2=A; X3=D; X4=S; X5=Y; X6=W and X7=G;

[0027] X1=S; X2=G; X3=D; X4=S; X5=Y; X6=W and X7=G;

[0028] X1=N; X2=G; X3=D; X4=R; X5=H; X6=W and X7=D;

[0029] X1=N; X2=A; X3=D; X4=R; X5=H; X6=W and X7=D;

[0030] X1=S; X2=G; X3=D; X4=R; X5=H; X6=W and X7=D;

[0031] X1=N; X2=G; X3=G; X4=Y; X5=L; X6=D and X7=G;

[0032] X1=N; X2=A; X3=G; X4=Y; X5=L; X6=D and X7=G; or

[0033] X1=S; X2=G; X3=G; X4=Y; X5=L; X6=D and X7=G. In some embodiments

[0034] X1=N; X2=G; X3=D; X4=R; X5=H; X6=W and X7=D;

[0035] X1=N; X2=A; X3=D; X4=R; X5=H; X6=W and X7=D; or

[0036] X1=S; X2=G; X3=D; X4=R; X5=H; X6=W and X7=D.

[0037] In a preferred embodiment, X3 to X7 =DRHWD, and X1 and X2 are NG; SG or NA.

[0038] Bispecific antibodies are described wherein the heavy chain variable region of the second variable domain comprises an amino acid sequence of one of the sequences of SEQ ID NO: 1 to 3, 7, 8, 10, 13, 15, 16, 17, 21, 22 or 23 having 0 to 10, preferably 0 to 5, amino acid insertions, deletions, substitutions, additions or a combination thereof.

[0039] The present invention also provides a method for treating a subject having a tumor, the method comprising administering a bispecific antibody as described herein to an individual in need thereof. Typically, the individual is one having a disease involving abnormal cells, for example, the individual may have a tumor.

[0040] The present invention also provides a bispecific antibody comprising a first variable domain that can bind to the extracellular portion of an epidermal growth factor receptor (EGFR) and a second variable domain that can bind to the extracellular portion of a MET proto-oncogene receptor tyrosine kinase (cMET), wherein the first variable domain comprises a heavy chain variable region having a CDR1 sequence of SYGIS, a CDR2 sequence of WISAYX1X2NTNYAQKLQG, and a CDR3 comprising a sequence of X3X4X5X6HWWLX7A, wherein

[0041] X1=N or S; X2=A or G; X3=D or G; X4=R, S or Y; X5=H, L or Y; X6=D or W and X7=D or G; there are 0 to 5 amino acid insertions, deletions, substitutions, additions or a combination thereof at positions other than X1 to X7, and wherein the second variable domain comprises a heavy chain variable region having an amino acid sequence of one of the sequences of SEQ ID NOs: 1 to 23 with 0 to 10, preferably 0 to 5, amino acid insertions, deletions, substitutions, additions or a combination thereof.

[0042] The first variable domain preferably comprises a heavy chain variable region having CDR1 sequence SYGIS, CDR2 sequence WISAYNGNTNYAQKLQG and a CDR3 comprising the sequence DRHWHWWLDAFDY, and the second variable domain preferably comprises a heavy chain variable region having CDR1 sequence SYSMN, CDR2 sequence WINTYTGDPTYAQGFTG and CDR3 sequence ETYYYDRGGYPFDP.

[0043] The present invention also provides a bispecific antibody of the present invention as disclosed herein for use in treating a subject suffering from a disease involving abnormal cells, such as a tumor.

[0044] Also provided is the use of a bispecific antibody of the invention as disclosed herein in the preparation of a medicament for treating a disease involving abnormal cells, such as a tumor or cancer.

[0045] Also provided are methods for treating a subject having a tumor, preferably an EGFR-positive tumor, a cMET-positive tumor, or an EGFR and cMET-positive tumor, comprising administering a bispecific antibody to an individual in need thereof.

[0046] The antibodies of the invention as disclosed herein preferably inhibit the migration of HGF-induced EBC1 cells in a wound healing assay. Preferably, the inhibition is better than the combination of Cetuximab and MetMab. For example, inhibition is preferably achieved by preventing wound closure in the presence of HGF with or without EGF (HGF is present at 15 ng / ml, and EGF is present in an amount of 12.5 ng / ml when present).

[0047] When used in combination with a TKI, the antibodies of the invention disclosed herein inhibit the HGF and EGF / HGF-induced growth of the EGFR TKI-resistant tumor cell lines PC-9 and HCC827. The TKI is preferably gefitinib.

[0048] The antibodies of the invention disclosed herein inhibit HGF-induced growth of HGF-responsive cells, preferably EGFR TKI-resistant tumor cell lines PC-9 or HCC827.

[0049] The antibodies of the invention disclosed herein inhibit EGF-induced growth of EGF-responsive cells without inducing toxicities (e.g., rash and diarrhea) associated with high affinity bivalent EGFR antibodies. This makes the antibodies ideally suited for combination with TKIs, which have their own toxicity profiles.

[0050] The present invention also includes pharmaceutical compositions comprising the bispecific antibodies disclosed herein.

[0051] The antibodies of the invention can be used to treat tumors that are resistant to treatment with EGFR tyrosine kinase inhibitors, such as erlotinib, gefitinib, or afatinib; analogs of erlotinib, gefitinib, or afatinib; or a combination of one or more of their corresponding compounds and / or analogs.

[0052] Treatment according to the invention may further comprise treatment with an EGFR tyrosine kinase inhibitor, for example wherein the EGFR tyrosine kinase inhibitor is erlotinib.

[0053] Therefore, the bispecific antibodies of the present invention can be administered simultaneously, sequentially or separately with the EGFR tyrosine kinase inhibitor.

[0054] The present invention also includes a nucleic acid molecule or a set of nucleic acid molecules, which encodes, alone or together, a heavy chain or a heavy chain variable region of a bispecific antibody or variant thereof disclosed herein. A nucleic acid molecule or a set of nucleic acid molecules is also provided, which encodes an antibody disclosed herein.

[0055] In a preferred embodiment, the heavy chain comprises an IgG1 antibody, preferably a constant region of a human IgG1 antibody. The CH2 region of the IgG1 constant region can be modified to change the ADCC and / or CDC activity of the antibody, or the ADCC and / or CDC activity of the antibody is not changed. In a preferred embodiment, the change results in enhanced ADCC and / or CDC activity. In a preferred embodiment, the CH3 region of the antibody is modified to promote heterodimerization of the heavy chain, the heavy chain comprising a first heavy chain that binds to EGFR and a second heavy chain that binds to cMET.

[0056] The present invention also includes cells comprising one or more nucleic acid molecules, which encode bispecific antibodies or variants thereof as disclosed herein, either alone or together. A method for using the cells to produce bispecific antibodies or variants thereof disclosed herein is also provided, preferably, the method also includes harvesting the bispecific antibodies or variants thereof from the cell culture.

[0057] The present invention also includes a cell system comprising the bispecific antibodies or variants thereof disclosed herein.

[0058] The present invention also provides a cell expressing a bispecific antibody and / or comprising a nucleic acid molecule encoding the bispecific antibody.

[0059] The present invention also encompasses a bispecific antibody as disclosed herein further comprising a label, preferably a label for in vivo imaging. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 : The amino acid sequence of the heavy chain variable region of the variable domain mentioned in this application.

[0061] Figure 2 : Functionality of anti-EGFR cLC bivalent antibodies in inhibiting EGF-induced A431 cell death. The Y axis (count) shows the fluorescence readout of the assay, indicating the number of metabolically active cells as a function of the antibody concentration used (X axis). PG3370 is able to inhibit EGF-induced cell death and therefore shows enhanced cell growth with increasing antibody concentrations. Molecules with the variable region amino acid sequence of cetuximab / Erbitux are referred to herein as cetuximab or reference antibody cetuximab is used as an internal standard (black dots) in the experiment.

[0062] Figure 3 : Effects of cMET x EGFR bispecific antibody on wound healing in H385 cells (Panel A) and EBC-1 cells (Panel B).

[0063] Cells were incubated without (mock) or with 12.5 ng / ml EGF or 15 ng / ml HGF or a combination of HGF and EGF (15 ng / ml and 12.5 ng / ml) with the addition of 5 separate cMETxEGFR bispecific antibodies. As a control, cetuximab in combination with 2994 Fab was included. The Y axis depicts the percentage of wound closure measured by time series microscopy.

[0064] Figure 4 : FACS analysis for EGFR and cMET expression analysis in TKI-resistant NSCLC cells HCC827 and PC-9 cells.

[0065] (A) Expression of EGFR (x-axis) and cMET (y-axis) in these two cell lines was characterized using fluorescently labeled antibodies. All HCC827 cells showed EGFR expression and could be subdivided into EGFR high cMET pos EGFR pos cMET neg PC-9 cells contain a small population of EGFR high and cMET pos cells, and the smallest group of EGFR pos and cMET neg cell.

[0066] (B) Graph representing the distribution of different cell populations in PC-9 and HCC827 cells.

[0067] Figure 5 : Examples of the effects of PB8532 and PB8388 on HGF-induced resistance to TKI inhibitors in PC-9 (Panel A) and HCC827 (Panel B) cells.

[0068] Cells were pre-treated with bispecific PB8532, PB8388 or a cetuximab / 5D5 Fab cocktail and incubated with HGF and / or EGF in combination with TKI inhibitors, after which proliferation was measured. PB8532 inhibited HGF-mediated and EGF-mediated gefitinib resistance in PC-9 cells and HCC827 cells.

[0069] Figure 6 : Effect of treatment with the indicated antibodies on HGF-induced cMET phosphorylation or EGF-induced EGFR phosphorylation on PC-9 and HCC827 cells. Antibodies (100 nM) were incubated at 37°C for 15 minutes before cell extracts were generated and applied to Western blot analysis for detection of (p) EGFR and (p) cMET. Anti-vinculin antibody was included as a protein loading control.

[0070] Figure 7 : MF3370 and its variants. The CDR1, CDR2 and CDR3 sequences in MF8226 are underlined from left to right. The CDRs in other sequences are in corresponding positions.

[0071] Figure 8 : MF4356 and its variants. The CDR1, CDR2 and CDR3 sequences in MF4356 are underlined from left to right. The CDRs in other sequences are in corresponding positions.

[0072] Figure 9: Common light chain used in monospecific and bispecific IgGs.

[0073] Fig. 9A : Common light chain amino acid sequence. Fig. 9B : Common light chain variable domain DNA sequence and translation (IGKV1-39 / jk1). Fig. 9C : Common light chain constant region DNA sequence and translation. Fig.9D : IGKV1-39 / jk5 common light chain variable domain translation. Fig.9E : V zone IGKV1-39A.

[0074] Figure 10: IgG heavy chains used to generate bispecific molecules. Fig. 10A : CH1 area. Fig. 10B Figure A: Hinge region. Figure C: CH2 region. Fig. 10D : CH2 containing L235G and G236R silent substitutions. Fig.10E : CH3 domain containing L351K and T366K substitutions (KK). Fig.10F : CH3 domain containing L351D and L368E substitutions (DE).

[0075] Fig.11 : Inhibition of EGF binding to recombinant EGFR in ELISA.

[0076] Biotinylated EGF was allowed to bind to coated EGFR in the presence of serial dilutions of IgG. Cetuximab was used as a positive control, and PG2708 was used as a negative control antibody (Neg ctrl Ab). EGF binding was detected by streptavidin HRP.

[0077] Fig.12: Cynomolgus EGFR cross reactivity was determined by FACS analysis. CHO-K1 cells were transfected with human EGFR or cynomolgus EGFR constructs. Antibodies were allowed to bind to transfected cells and CHO-K1 cells at 5 μg / ml. Cetuximab was used as a positive control, and PG2708 was used as a negative control antibody (Neg ctrl Ab). Bound antibodies were detected by PE-conjugated antibodies.

[0078] Fig.13 : Mouse EGFR and cMET cross-reactivity determined by ELISA. Above ;Antibodies at a fixed concentration (5 μg / ml) were tested in serial titrations in microtiter plates coated with mouse EGFR and human EGFR. Anti-EGFR antibodies and PG2708 (neg Ctrl Ab) were allowed to bind and detected by HRP-conjugated antibodies. Below Serial titrations of antibodies were allowed to bind to coated human and mouse cMET. Human / mouse cross-reactive antibody BAF527 was included as a positive control antibody, and PG2708 was added as a negative control antibody (Neg Ctrl Ab). Bound antibodies were detected by streptavidin HRP.

[0079] Figure 14: Inhibition of ligand-dependent N87 proliferation. Serial antibody titrations were incubated with N87 cells in the presence of HGF (A), EGF (B) or EGF / HGF (C). Cell proliferation was measured by Alamar Blue. An equimolar concentration of Fab 5D5 / cetuximab was included as a positive control antibody. The Y-axis represents the fluorescence intensity as an indicator of cell proliferation. The X-axis represents the different concentrations of the tested antibodies.

[0080] Figure 15: Example of ADCC activity of cMETxEGFR bispecific antibodies in N87 cells (A) and MKN-45 cells (B) using a high affinity FcγRIIIa ADCC reporter assay. The X-axis represents increasing antibody concentrations. The Y-axis represents luminescence (RLU) as a readout of ADCC activity. The anti-EGFR antibody cetuximab was included as a positive control antibody.

[0081] Figure 16: Effect of HGF on the potency of TKI erlotinib and gefitinib in PC-9 (A) and HCC827 (B) cells. Cells were incubated with increasing concentrations of HGF (0 to 120 ng / mL) in combination with 300 nM erlotinib or gefitinib, after which cell proliferation was measured. In both cell lines, HGF induced a dose-dependent resistance to TKIs.

[0082] Figure 17: Affinity binding test of ADCC-enhanced c-MET x EGFR variants. CHO-K1 cells stably expressing EGFR (A) or MKN-45 cells endogenously expressing c-MET (B) were cultured at 2×10 5 Cells / well were incubated with increasing concentrations of antibodies as specified. After washing, binding was detected with anti-human IgG-PE (3 μg / ml). The stained cells were analyzed on the iQue system and the mean fluorescence intensity (MFI) was calculated. The control antibodies were MF1337xMF1337; (TTxTT negative control; black triangles at the bottom) and MF4356xMF3770 (PB8532p04; c-METxEGFR positive control for c-MET; black triangles). TT represents tetanus toxoid. ADCC represents antibodies with enhanced ADCC function by co-transfection with DNA encoding the RMD enzyme that removes fucose residues from the Fc region of IgG1.

[0083] Fig.18 :The results of the ADCC reporter assay for confirming the enhancement of ADCC effector function. The BxPC-3 cells expressing EGFR (left) or the MKN-45 cells expressing c-MET (right) were mixed with ADCC effector cells at an E:T ratio of 15:1 and incubated in the presence of a titration solution (0.01 to 10 μg / ml) of the test antibody. After 6 hours, Bio-Glo reagent was added and luminescence was measured using a microplate reader. The higher the level of luminescence, the higher the degree of interaction between the target cell and the effector cell induced by the test antibody. The upper figure shows the results of the high affinity assay, and the lower figure shows the results of the low affinity assay. The negative control antibody was PG1337p218 (anti-TT, light triangle at bottom); other control antibodies were 3178x4280 (HER3 x EGFR, ADCC enhanced, light closed circle (upper in left upper panel); 3178x4280 (HER3 x EGFR, non-ADCC enhanced, black cross, at bottom); 4356x3370 (c-MET x EGFR, ADCC enhanced, open light circle); 3370x4356 (EGFR x c-MET, non-ADCC enhanced, black cross and dashed line); and cetuximab (anti-EGFR, small black circle).

[0084] Fig.19 : Erlotinib induces antitumor responses in NGS-hHGFki mice transplanted with HCC827 cells for as long as the mice were treated. The black arrow indicates the start of treatment.

[0085] Fig. 20: PB8532 alone and in combination with erlotinib induces anti-tumor responses in NGS-hHGFki mice transplanted with HCC827 cells. Black arrows indicate the start of treatment; gray arrows in the X-axis indicate weekly antibody treatments.

[0086] Fig.21 : Even after treatment cessation, the anti-tumor response induced by PB8532 alone and in combination with erlotinib was superior to that induced by erlotinib. The black arrows indicate the start of treatment; the grey arrows in the X-axis indicate weekly antibody treatments.

[0087] Fig. 22 : The anti-tumor response induced by PB8532 alone and in combination with erlotinib is superior to the anti-tumor response induced by erlotinib. The black arrow indicates the start of treatment; the gray arrows in the X-axis indicate weekly antibody treatment. Treatment with cMET antibody LY2875358, whether combined with erlotinib treatment or not, was less effective than PB8532 even without erlotinib treatment.

[0088] Fig.23 : The anti-tumor response induced by the cMETxEGFR bispecific antibody PB19478 was also effective when the tumor was resistant to erlotinib. The black arrows indicate the start of erlotinib treatment and the start of PB19478 treatment. DETAILED DESCRIPTION

[0089] EGFR is a member of a family of four receptor tyrosine kinases (RTKs) called Her- or cErbB-1, -2, -3 and -4. EGFR has an extracellular domain (ECD) consisting of four subdomains, two of which are involved in ligand binding, and one of which is involved in homodimerization and heterodimerization (Ferguson (2008)). The reference numbers used in this section refer to the reference numbers in the list entitled "Citations in this specification", each of which is incorporated by reference. EGFR integrates extracellular signals from a variety of ligands to produce a variety of intracellular responses (Yarden at al. 2001; and Jorrisen et al. 2003). EGFR is involved in several human epithelial malignancies, particularly breast cancer, bladder cancer, non-small cell lung cancer, lung cancer, colon cancer, ovarian cancer, head and neck cancer, and brain cancer. Activating mutations in the gene, as well as overexpression of the receptor and its ligand, have been found, leading to an autocrine activation loop (for a review, see Robertson et al. 2000). Therefore, this RTK has been widely used as a target for cancer therapy. Both small molecule inhibitors targeting RTK and monoclonal antibodies (mAbs) against the extracellular ligand binding domain have been developed and have shown several clinical successes to date, although most are for selected groups of patients. The database accession number for the human EGFR protein and its encoding gene is (GenBank NM_005228.3). Other database identifiers for genes and / or proteins are HGNC: 3236; Entrez Gene: 1956; Ensembl: ENSG00000146648; OMIM: 131550 and UniProtKB: P00533. The accession number is given mainly to provide another method for identifying the EGFR protein as a target, and the actual sequence of the EGFR protein bound by the antibody may vary, for example, because of mutations in the encoding gene (e.g., mutations occurring in some cancers, etc.). Unless otherwise indicated, when EGFR is referred to herein, the reference is to human EGFR.Antigen binding sites that bind EGFR bind EGFR and its various variants, such as those expressed on some EGFR-positive tumors.

[0090] The term "EGFR ligand" as used herein refers to a polypeptide that binds to and activates EGFR. Some examples of EGFR ligands include, but are not limited to, EGF, TGF-α, HB-EGF, amphiregulin, betacellulin, and epiregulin (reviewed by Olayioye MA et al.; EMBO J (2000) Vol 19: pp 3159-3167). The term includes biologically active fragments and / or variants of naturally occurring polypeptides.

[0091] cMET, also known as tyrosine protein kinase MET or hepatocyte growth factor receptor (HGFR), is a protein encoded by the MET gene in humans. The protein has tyrosine kinase activity. The primary single-chain precursor protein is cleaved after translation to produce α and β subunits, which are linked via disulfide bonds to form the mature receptor.

[0092] Abnormally activated cMET can induce tumor growth, the formation of new blood vessels that supply nutrients to the tumor (angiogenesis), and the spread of cancer to other organs (metastasis). cMET is deregulated in many types of human malignancies, including kidney cancer, liver cancer, stomach cancer, breast cancer, and brain cancer. The cMET gene is known by many different names, such as MET proto-oncogene receptor tyrosine kinase; hepatocyte growth factor receptor; tyrosine protein kinase Met; discrete factor receptor; proto-oncogene C-Met; HGF / SF receptor; HGF receptor; SF receptor; EC 2.7.10.1; Met proto-oncogene; EC 2.7.10; DFNB97; AUTS9; RCCP2; C-Met; MET; HGFR; The external ID of cMET is HGNC: 7029; Entrez gene: 4233; Ensembl: ENSG00000105976; OMIM: 164860 and UniProtKB: P08581. The accession number is given primarily to provide another method of identifying the cMET protein as a target, the actual sequence of the cMET protein bound by the antibody may vary, for example because of mutations in the encoding gene (such as those that occur in some cancers, etc.). Unless otherwise indicated, when cMET is referred to herein, the reference is to human cMET.Antigen binding sites that bind cMET bind cMET and its various variants, such as those expressed on some cMET-positive tumors.

[0093] Antibodies typically recognize only a portion of an antigen. Antigens are often, but not necessarily, proteins. The recognition or binding site on an antigen bound by an antibody is called an epitope, where an epitope can be linear or conformational. The binding of an antibody to an antigen is typically specific. The "specificity" of an antibody refers to its selectivity for a particular epitope, while "affinity" refers to the strength of the interaction between the antigen binding site of an antibody and the epitope to which it binds.

[0094] Exemplary antibodies of the invention disclosed herein bind to EGFR and cMET (preferably human EGFR and human cMET). The EGFR / cMET bispecific antibodies of the invention disclosed herein bind to EGFR, and under otherwise identical conditions, bind to the cognate receptors ErbB-2 and ErbB-4 of the same species at least 100 times lower. The EGFR / cMET bispecific antibodies of the invention disclosed herein bind to cMET, and under otherwise identical conditions, bind to the receptors ErbB-2 and ErbB-4 of the same species at least 100 times lower. Considering that the receptor is a cell surface receptor, the binding can be evaluated on cells expressing the receptor. The bispecific antibodies of the invention disclosed herein preferably bind to human EGFR, cynomolgus monkey EGFR and / or mouse EGFR.

[0095] Antibodies that bind to EGFR and cMET may also bind to other proteins if such other proteins contain the same epitope. Therefore, the term "binding" does not exclude the binding of an antibody to another protein or more proteins containing the same epitope. Such binding is generally referred to as cross-reactivity. EGFR / cMET bispecific antibodies generally do not bind to other proteins except EGFR and / or cMET on the cell membrane in postnatal humans (preferably adults). The antibodies according to the present invention disclosed herein are generally capable of binding to EGFR with a binding affinity (i.e., equilibrium dissociation constant Kd) of at least 1×10e-6M, as outlined in more detail below.

[0096] The term "antibody" as used herein means a protein molecule that preferably belongs to the class of immunoglobulins. Antibodies generally comprise two variable domains that bind to an epitope on an antigen. Such domains are derived from or have sequence homology with the variable domains of an antibody. The bispecific antibodies of the present invention as disclosed herein preferably comprise two variable domains. Antibodies for therapeutic use are preferably as close as possible to the natural antibodies of the subject to be treated (e.g., human antibodies for human subjects). Antibody binding can be expressed in terms of specificity and affinity. Specificity determines which antigen or epitope thereof is specifically bound by the binding domain. Generally speaking, antibodies for therapeutic applications may have an affinity as high as 1×10e-10M or higher. Antibodies of the present invention disclosed herein, such as bispecific antibodies, preferably comprise the constant domains (Fc portions) of natural antibodies. The antibodies of the present invention disclosed herein are generally bispecific full-length antibodies of the preferred human IgG subclass. Preferably, the antibodies of the present invention are of the human IgG1 subclass. Such antibodies of the present invention as disclosed herein can have good ADCC properties, have good half-life after administration to the human body, and there are CH3 transformation technologies that can provide modified heavy chains that preferentially form heterodimers compared to homodimers after co-expression in clonal cells. The ADCC activity of antibodies can also be improved by techniques known to those skilled in the art.

[0097] The antibody of the present invention as disclosed herein is preferably a "full-length" antibody. The term "full-length" according to the present invention as disclosed herein is defined as comprising a substantially complete antibody, however, it is not necessary to have all the functions of a complete antibody. For the avoidance of doubt, a full-length antibody comprises two heavy chains and two light chains. Each chain comprises a constant (C) and a variable (V) region, which can be decomposed into domains designated as CH1, CH2, CH3, VH and CL, VL. Generally speaking, antibodies bind to antigens through the variable domains contained in the Fab portion, and after binding, can interact with molecules and cells of the immune system through the constant domains (mainly through the Fc portion). The full-length antibodies according to the present invention disclosed herein encompass antibodies in which mutations providing desired characteristics may be present. The term "full-length antibody" includes antibodies in which one or several amino acid residues are missing and the specificity and / or affinity characteristics of the resulting antibody are not substantially changed. For example, an IgG antibody may have 1 to 20 amino acid residues inserted, deleted or replaced or a combination thereof in the constant region.

[0098] The antibodies of the present invention as disclosed herein are preferably bispecific IgG antibodies, preferably bispecific full-length IgG1 antibodies, and more preferably human IgG1. Full-length IgG antibodies are preferred because they generally have a good half-life, and for reasons of immunogenicity it is desirable to maintain complete proximity to autologous (human) molecules. In some embodiments, the antibodies of the present invention are full-length IgG1, full-length IgG2, full-length IgG3, or full-length IgG4 antibodies.

[0099] The invention disclosed herein includes bispecific antibodies comprising a first variable domain that can bind to the extracellular portion of EGFR and a second variable domain that can bind to the extracellular portion of cMET, wherein the first variable domain binds to EGFR with an affinity less than the Kd of cetuximab, which is 0.39 nM (Kim et al 2008). The first variable domain preferably binds to EGFR with a Kd of 10e-6M to 10e-9M. The Kd is preferably 10e-7M to 10e-9M, preferably 10e-8M to 10e-9M. The second variable domain preferably binds to cMET with a Kd of 10e-7M or less. The Kd is preferably 10e-7M to 10e-11M. The affinity of the second variable domain for cMET is preferably higher than the affinity of the first variable domain for EGFR. In other words, in this preferred embodiment, the Kd of the antibody against cMET is less than the Kd of the antibody against EGFR. In a preferred embodiment, the Kd of the antibody against cMET is at least 5 times smaller than the Kd of the antibody against EGFR, and preferably at least 10 times smaller. In this embodiment, the Kd values ​​for the corresponding antigens are preferably as shown in this paragraph. This appropriate imbalance in affinity allows the bispecific antibodies of the invention disclosed herein to dock on cells, preferably by binding to EGFR, and block the binding of the ligand HGF to cMET.

[0100] The variable domains that can be combined with EGFR are preferably the variable domains that suppress EGF-induced A431 cell death in the case of bivalent monospecific antibodies.Preferably, the inhibition of EGF-induced cell death is measured under the concentration of 10nM EGF and 10μg / ml antibody.The inhibition of EGF-induced cell death can be detected by comparing the cell number with and without antibody after A431 is cultivated for 3 to 7 days under the condition of allowing (but except EGF) A431 cell growth.Not bound by theory, it is believed that the combination of antibody and EGFR blocks the combination of EGF and EGFR.The variable domains that can be combined with EGFR are preferably the variable domains that suppress the proliferation of EGF-induced BxPC3 or BxPC3-luc2 cells in the case of bivalent monospecific antibodies.

[0101] The antibodies of the invention as disclosed herein preferably inhibit HGF-induced migration of EBC1 cells in a wound healing assay. The wound healing assay is preferably an assay as described in the Examples. The inhibition of wound healing is superior to the combination of Cetuximab and MetMab. The inhibition is usually not 100%. Some wound healing also occurs in the presence of inhibitory antibodies.

[0102] When used in combination with a TKI, the antibodies of the invention as disclosed herein inhibit the HGF and EGF / HGF-induced growth of the EGFR TKI-resistant tumor cell lines PC-9 and HCC827. The TKI is preferably gefitinib.

[0103] The antibodies of the invention as disclosed herein inhibit HGF-induced growth of HGF-responsive cells, preferably EGFR TKI-resistant tumor cell lines PC-9 or HCC827.

[0104] The antibodies of the invention as disclosed herein inhibit EGF-induced growth of EGF-responsive cells without inducing significant common toxicities associated with high affinity bivalent EGFR antibodies, such as rash and diarrhea, etc. This makes the antibodies ideally suited for combination with TKIs, which have their own toxicity profile.

[0105] Induced growth is preferably measured using an assay as described in the Examples. Inhibition is usually not 100%. Some growth also occurs in the case of inhibitory antibodies.

[0106] The variable domain that can bind to EGFR and comprises the amino acid sequence of MF3370 or its variant as shown in this article preferably binds to EGFR domain III (see Table 4 of International Patent Application PCT / NL2015 / 050124; WO2015 / 130172, which is incorporated herein by reference). The binding of the variable domain to EGFR can be inhibited by cetuximab. The variable domain binds to an epitope that is different from the epitope recognized by cetuximab and zalumuzumab. For example, the variable domain binds to mouse EGFR, while cetuximab and zalumuzumab do not bind, indicating that one or more residues that differ between mouse and human EGFR domain III play a role in the binding of cetuximab and zalumuzumab, but do not play a role in the antibodies of the invention described herein. One advantage of the bispecific antibodies of the invention described herein that have cross-reactivity to human, mouse, and cynomolgus monkey EGFR is that they allow xenograft studies to be used with human cancer models, which can be more predictive in terms of efficacy and toxicity because the antibodies also bind to normal mouse cells that have receptors, while also being able to be used in cynomolgus monkey toxicology studies. In one aspect, the present invention provides a bispecific antibody comprising a first variable domain that can bind to the extracellular portion of human epidermal growth factor receptor (EGFR) and a second variable domain that can bind to the extracellular portion of human MET proto-oncogene receptor tyrosine kinase (cMET), wherein the first variable domain can also bind to mouse EGFR, cynomolgus monkey EGFR, or both.

[0107] The cMET variable domain preferably comprises the amino acid sequence of MF4356 or a variant thereof as shown herein, and preferably blocks the binding of the antibody MetMab to cMET. The variable domain preferably blocks the binding of the ligand HGF to cMET. When the binding of MetMab to cMET under half-maximal binding conditions is reduced by at least 40%, and preferably at least 60%, in the presence of a saturating amount of the variable domain, the variable domain blocks the binding of the antibody MetMab to cMET. The variable domain is preferably provided in the case of a bivalent monospecific antibody. The cMET variable domain may preferably bind to the sema domain of cMET. The cMET variable domain of the present invention may compete with 5D5 for binding to cMET, or may not compete with a reported anti-cMET reference antibody, such as 5D5. See Table 2.

[0108] The variable domain of the present invention disclosed herein can bind to EGFR (first variable domain), and preferably comprises a heavy chain variable region having a CDR1 sequence SYGIS, a CDR2 sequence W1SAYX1X2NTNYAQKLQG and a CDR3 comprising the sequence X3X4X5X6HWWLX7A, wherein X1=N or S; X2=A or G; X3=D or G; X4=R, S or Y; X5=H, L or Y; X6=D or W and X7=D or G.

[0109] X 1至7 Preferably:

[0110] X1=N; X2=G; X3=D; X4=S; X5=Y; X6=W and X7=G;

[0111] X1=N; X2=A; X3=D; X4=S; X5=Y; X6=W and X7=G;

[0112] X1=S; X2=G; X3=D; X4=S; X5=Y; X6=W and X7=G;

[0113] X1=N; X2=G; X3=D; X4=R; X5=H; X6=W and X7=D;

[0114] X1=N; X2=A; X3=D; X4=R; X5=H; X6=W and X7=D;

[0115] X1=S; X2=G; X3=D; X4=R; X5=H; X6=W and X7=D;

[0116] X1=N; X2=G; X3=G; X4=Y; X5=L; X6=D and X7=G;

[0117] X1=N; X2=A; X3=G; X4=Y; X5=L; X6=D and X7=G; or

[0118] X1=S; X2=G; X3=G; X4=Y; X5=L; X6=D and X7=G.

[0119] In a preferred embodiment

[0120] X1=N; X2=G; X3=D; X4=R; X5=H; X6=W and X7=D;

[0121] X1=N; X2=A; X3=D; X4=R; X5=H; X6=W and X7=D; or

[0122] X1=S; X2=G; X3=D; X4=R; X5=H; X6=W and X7=D.

[0123] Preferably X1=N; X2=G; X3=D; X4=R; X5=H; X6=W and X7=D.

[0124] The amino acid after the amino acid A in the sequence X9X4X5X6HWWLX7A in the CDR3 sequence of the first variable domain may vary. The amino acid sequence after the sequence X3X4X5X6HWWLX7A may be FDY. The CDR3 of the first variable domain preferably comprises the sequence X3X4X5X6HWWLX7AF, preferably X3X4X5X6HWWLX7AFD, more preferably X3X4X5X6HWWLX7AFDY.

[0125] The first variable domain preferably comprises a heavy chain variable region having a CDR1 sequence of SYGIS, a CDR2 sequence of WISAYNGNTNYAQKLQG and a CDR3 sequence of X3X4X5X6HWWLX7A.

[0126] The first variable domain preferably comprises a heavy chain variable region having a CDR1 sequence SYGIS, a CDR2 sequence WISAYNGNTNYAQKLQG and a CDR3 comprising the sequence DRHWHWWLDA. The amino acid after the LDA sequence in the CDR3 sequence of the first variable domain can vary. The amino acid sequence after the sequence LDA can be FDY. The CDR3 of the first variable domain preferably comprises the sequence DRHWHWWLDAF, preferably DRHWHWWLDAFD), more preferably DRHWHWWLDAFDY.

[0127] The first variable domain preferably comprises a heavy chain variable region having Figure 7 The following amino acid sequence as shown in: MF3353, MF8229, MF8228, MF3370, MF8233, MF8232, MF3393, MF8227 or MF8226, said amino acid sequence having at most 10, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or combinations thereof relative to the sequence shown. In a preferred embodiment, the first variable domain comprises a heavy chain variable region having Figure 7 The following amino acid sequences are shown in: MF3353, MF8229, MF8228, MF3370, MF8233, MF8232, MF3393, MF8227 or MF8226.

[0128] The variable domain (second variable domain) that can bind to cMET preferably comprises a heavy chain variable region comprising an amino acid sequence of one of the sequences of SEQ ID NOs: 1 to 23 with 0 to 10, preferably 0 to 5 amino acid insertions, deletions, substitutions, additions or a combination thereof. The heavy chain variable region of the second variable domain preferably comprises an amino acid sequence of one of the sequences of SEQ ID NOs: 1 to 3, 7, 8, 10, 13, 15, 16, 17, 21, 22 or 23 with 0 to 10, preferably 0 to 5 amino acid insertions, deletions, substitutions, additions or a combination thereof. The heavy chain variable region of the second variable domain preferably comprises an amino acid sequence of one of the sequences of SEQ ID NOs: 2, 7, 8, 10, 13 or 23 with 0 to 10, preferably 0 to 5 amino acid insertions, deletions, substitutions, additions or a combination thereof. The heavy chain variable region of the second variable domain preferably comprises the amino acid sequence of SEQ ID NO: 13 or SEQ ID NO: 23 having 0 to 10, preferably 0 to 5, amino acid insertions, deletions, substitutions, additions or a combination thereof.

[0129] In a preferred embodiment, the first variable domain comprises a heavy chain variable region having a CDR1 sequence SYGIS, a CDR2 sequence WISAYNGNTNYAQKLQG and a CDR3 comprising the sequence DRHWHWWLDA, preferably DRHWHWWLDAFDY, and wherein the second variable domain comprises a heavy chain variable region having a CDR1 sequence SYSMN, a CDR2 sequence WINTYTGDPTYAQGFTG and a CDR3 sequence ETYYYDRGGYPFDP. The CDR1, CDR2 and CDR3 of the light chain of the first variable domain and the second variable domain preferably comprise the amino acid sequences CDR1-QSISSY, CDR2-AAS, CDR3-QQSYSTP, i.e., the CDRs of IGKV1-39 (according to IMGT), respectively.

[0130] In a preferred embodiment, the first variable domain comprises a heavy chain variable region having a CDR1 sequence SYGIS, a CDR2 sequence WISAYNGNTNYAQKLQG, and a CDR3 comprising the sequence DRHWHWWLDA, and wherein the second variable domain comprises a heavy chain variable region having a CDR1 sequence TYSMN, a CDR2 sequence WINTYTGDPTYAQGFTG, and a CDR3 comprising the sequence ETYFYDRGGYPFDP. The CDR1, CDR2, and CDR3 of the light chain of the first and second variable domains preferably comprise the amino acid sequences CDR1-QSISSY, CDR2-AAS, CDR3-QQSYSTP, i.e., the CDRs of IGKV1-39 (according to IMGT), respectively.

[0131] A bispecific antibody comprising a first variable domain that can bind to the extracellular portion of EGFR and a second variable domain that can bind to the extracellular portion of cMET, wherein the first variable domain comprises a heavy chain variable region having a CDR1 sequence of SYGIS, a CDR2 sequence of WISAYNANTNYAQKLQG, and a CDR3 comprising the sequence DRHWHWWLDA, and wherein the second variable domain comprises a heavy chain variable region having a CDR1 sequence of SYSMN, a CDR2 sequence of WINTYTGDPTYAQGFTG, and a CDR3 sequence of ETYYYDRGGYPFDP. The CDR1, CDR2, and CDR3 of the light chain of the first variable domain and the second variable domain preferably comprise the amino acid sequences CDR1-QSISSY, CDR2-AAS, CDR3-QQSYSTP, i.e., the CDRs of IGKV1-39 (according to IMGT), respectively.

[0132] A bispecific antibody comprising a first variable domain that can bind to the extracellular portion of EGFR and a second variable domain that can bind to the extracellular portion of cMET, wherein the first variable domain comprises a heavy chain variable region having a CDR1 sequence of SYGIS, a CDR2 sequence of WISAYNANTNYAQKLQG, and a CDR3 comprising the sequence DRHWHWWLDA, and wherein the second variable domain comprises a heavy chain variable region having a CDR1 sequence of TYSMN, a CDR2 sequence of WINTYTGDPTYAQGFTG, and a CDR3 comprising the sequence ETYFYDRGGYPFDP. The CDR1, CDR2, and CDR3 of the light chain of the first variable domain and the second variable domain preferably comprise the amino acid sequences CDR1-QSISSY, CDR2-AAS, CDR3-QQSYSTP, i.e., the CDRs of IGKV1-39 (according to IMGT), respectively.

[0133] A bispecific antibody comprising a first variable domain that can bind to the extracellular portion of EGFR and a second variable domain that can bind to the extracellular portion of cMET, wherein the first variable domain comprises a heavy chain variable region having a CDR1 sequence of SYGIS, a CDR2 sequence of WISAYSGNTNYAQKLQG, and a CDR3 comprising the sequence DRHWHWWLDA, and wherein the second variable domain comprises a heavy chain variable region having a CDR1 sequence of SYSMN, a CDR2 sequence of WINTYTGDPTYAQGFTG, and a CDR3 sequence of ETYYYDRGGYPFDP. The CDR1, CDR2, and CDR3 of the light chain of the first variable domain and the second variable domain preferably comprise the amino acid sequences CDR1-QSISSY, CDR2-AAS, CDR3-QQSYSTP, i.e., the CDRs of IGKV1-39 (according to IMGT), respectively.

[0134] A bispecific antibody comprising a first variable domain that can bind to the extracellular portion of EGFR and a second variable domain that can bind to the extracellular portion of cMET, wherein the first variable domain comprises a heavy chain variable region having a CDR1 sequence of SYGIS, a CDR2 sequence of WISAYSGNTNYAQKLQG, and a CDR3 comprising the sequence DRHWHWWLDA, and wherein the second variable domain comprises a heavy chain variable region having a CDR1 sequence of TYSMN, a CDR2 sequence of WINTYTGDPTYAQGFTG, and a CDR3 comprising the sequence ETYFYDRGGYPFDP. The CDR1, CDR2, and CDR3 of the light chain of the first variable domain and the second variable domain preferably comprise the amino acid sequences CDR1-QSISSY, CDR2-AAS, CDR3-QQSYSTP, i.e., the CDRs of IGKV1-39 (according to IMGT), respectively.

[0135] As described herein, the CDR1, CDR2 and CDR3 of the light chain of the first variable domain and the second variable domain preferably comprise the amino acid sequences CDR1-QSISSY, CDR2-AAS, CDR3-QQSYSTP, i.e., the CDRs of IGKV1-39 (according to IMGT). In some embodiments of the bispecific antibodies as described herein, the first variable domain and the second variable domain comprise a common light chain, preferably Fig. 9B of light chain.

[0136] In another preferred embodiment, the EGFR / cMET bispecific antibody comprises an extracellular portion that can bind to human EGFR and comprises Figure 1 The first variable domain of CDR1, CDR2 and CDR3 of the heavy chain variable region of MF3755 shown in FIG, and the extracellular portion that can bind to human cMET and comprises Figure 1 The second variable domain of CDR1, CDR2 and CDR3 of the heavy chain variable region of MF4297 shown in . The light chain variable region in the first variable domain and the second variable domain is preferably a common light chain variable region as described herein. The CDR1, CDR2 and CDR3 of the light chain of the first variable domain and the second variable domain preferably comprise the amino acid sequence CDR1-QSISSY, CDR2-AAS, CDR3-QQSYSTP, i.e., the CDRs of IGKV1-39 (according to IMGT), respectively. In a preferred embodiment, the antibody comprises a heavy chain variable region having as Figure 1The amino acid sequence of MF3755 shown in , said amino acid sequence has at most 10, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably has 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the specified sequence. In a preferred embodiment, the first variable domain comprises a heavy chain variable region having Figure 1 The amino acid sequence of MF3755 is shown in . The variable domain (second variable domain) that can bind to cMET preferably comprises a heavy chain variable region, wherein the heavy chain variable region comprises a heavy chain variable region having 0 to 10, preferably 0 to 5 amino acid insertions, deletions, substitutions, additions or a combination thereof. Figure 1 The amino acid sequence of MF4297 is shown in . The heavy chain variable region of the second variable domain preferably comprises Figure 1 The amino acid sequence of MF4297 is shown in .

[0137] In the context of the present invention, the term "bispecific" (bs) means that the antibody is able to bind to two different targets or two epitopes on the same target, for example, wherein one variable domain of the antibody (as defined above) binds to an epitope on EGFR and the second variable domain binds to an epitope on cMET. Depending on the expression level, (sub)cellular localization and stoichiometry of the two antigens recognized by the bispecific antibody, the two Fab arms of the antibody may or may not bind their epitopes simultaneously. One arm of a bispecific antibody generally comprises the variable domain of one antibody and the other arm comprises the variable domain of another antibody (i.e., one arm of the bispecific antibody is formed by a heavy chain paired with a light chain, and the other arm is formed by a different heavy chain paired with a light chain). Therefore, the stoichiometry of the preferred bispecific antibodies of the present invention disclosed herein is 1:1 EGFR:cMET binding.

[0138] The heavy chain variable regions of the bispecific antibodies of the present invention as disclosed herein are usually different from each other, while the light chain variable regions are preferably the same. The bispecific antibodies in which different heavy chain variable regions are associated with the same light chain variable region are also referred to as bispecific antibodies with a common light chain variable region (common light chain variable region, cLcv). Preferably, the light chain constant region is also the same. Such bispecific antibodies are referred to as having a common light chain (cLc). Therefore, a bispecific antibody according to the present invention as disclosed herein is also provided, wherein the two arms comprise a common light chain.

[0139] The term "common light chain" according to the present invention disclosed herein refers to two or more light chains in a bispecific antibody, and the light chain can be identical or have some amino acid sequence differences, while the binding specificity of the full-length antibody is not affected. For example, within the scope of the definition of the common light chain as used herein, it is possible to prepare or find light chains that are not identical but still functionally equivalent, such as by introducing and testing conservative amino acid changes, which are amino acid changes that do not or only partially promote binding specificity in the region when paired with the heavy chain, etc. The terms "common light chain", "common LC", "cLC", "single light chain" are used interchangeably herein whether or not the term "rearranged" is added. The terms "common light chain variable region", "common VL", "common LCv", "cLCv", "single VL" are used interchangeably herein whether or not the term "rearranged" is added. A preferred aspect of the present invention is that the bispecific antibody has a common light chain (variable region) that can be combined with at least two and preferably multiple heavy chains (variable regions) with different binding specificities to form an antibody (e.g., WO2009 / 157771) with a functional antigen binding domain. The common light chain (variable region) is preferably a human light chain (variable region). The common light chain (variable region) preferably has a germline sequence. The preferred germline sequence is a light chain variable region with good thermodynamic stability, yield and solubility. The preferred germline light chain is O12. The common light chain preferably comprises a light chain encoded by a germline human Vk gene segment, and is preferably a rearranged germline human κ light chain IgVκ1-39*01 / IGJκ1*01 ( Fig. 9A ). The common light chain variable region is preferably the variable region of a rearranged germline human κ light chain IgVκ1-39*01 / IGJκ1*01. The common light chain preferably comprises a sequence having 0 to 5 amino acid insertions, deletions, substitutions, additions or a combination thereof. Fig. 9B Or the light chain variable region shown in 9D. The common light chain preferably also comprises a light chain constant region, preferably a kappa light chain constant region. The nucleic acid encoding the common light chain can be codon optimized for the cell system used to express the common light chain protein. The encoding nucleic acid can deviate from the germline nucleic acid sequence.

[0140] In a preferred embodiment, the light chain comprises a light chain region comprising a light chain region having 0 to 10, preferably 0 to 5 amino acid insertions, deletions, substitutions, additions or a combination thereof. Fig. 9A The amino acid sequence of the O12 / IgVκ1-39*01 gene segment shown in . Throughout the specification, the phrase "O12 light chain" will be used as "comprising a light chain having 0 to 10, preferably 0 to 5 amino acid insertions, deletions, substitutions, additions or a combination thereof". Fig. 9AThe amino acid sequence of the O12 / IgVκ1-39*01 gene segment shown in FIG. 1 is an abbreviation for "light chain variable region of the light chain". IgVκ1-39 is an abbreviation for immunoglobulin variable κ1-39 gene. The gene is also called immunoglobulin κ variable 1-39; IGKV139; IGKV1-39; O12a or O12. The external ID of the gene is HGNC: 5740; Entrez gene: 28930; Ensembl: ENSG00000242371. A preferred amino acid sequence of IgVκ1-39 is Fig.9E . It lists the sequences of the V regions. The V regions can be combined with one of the five J regions. Fig. 9B and 9D Two preferred sequences of IgVκ1-39 in combination with the J region are described. The linked sequences are denoted IGKV1-39 / jk1 and IGKV1-39 / jk5; alternative names are IgVκ1-39*01 / IGJκ1*01 or IgVκ1-39*01 / IGJκ5*01 (according to the nomenclature of the IMGT database World Wide Web at imgt.org).

[0141] Preferably, the light chain variable region comprising O12 / IgVκ1-39*01 is a germline sequence. Also preferably, the light chain variable region comprising IGJκ1*01 or / IGJκ5*01 is a germline sequence. In a preferred embodiment, the IGKV1-39 / jk1 or IGKV1-39 / jk5 light chain variable region is a germline sequence.

[0142] In a preferred embodiment, the light chain variable region comprises germline O12 / IgVκ1-39*01. In a preferred embodiment, the light chain variable region comprises a kappa light chain IgVκ1-39*01 / IGJκ1*01 or IgVκ1-39*01 / IGJκ5*01. In a preferred embodiment, IgVκ1-39*01 / IGJκ1*01. The light chain variable region preferably comprises a germline kappa light chain IgVκ1-39*01 / IGJκ1*01 or a germline kappa light chain IgVκ1-39*01 / IGJκ5*01, preferably a germline IgVκ1-39*01 / IGJκ1*01.

[0143] Mature B cells that produce antibodies with O12 light chains typically produce light chains that have undergone one or more mutations relative to the germline sequence (i.e., the normal sequence in the non-lymphoid cells of an organism). The process that leads to these mutations is generally called somatic (hyper) mutation. The resulting light chain is called an affinity matured light chain. Such a light chain is an O12-derived light chain when derived from an O12 germline sequence. In this specification, the phrase "common light chain" will include "common light chain-derived light chain", and the phrase "O12 light chain" will include an O12-derived light chain. Mutations introduced by somatic hypermutation can also be artificially introduced in the laboratory. In the laboratory, other mutations can also be introduced without affecting the properties of the light chain (type, not necessarily in quantity). If the light chain contains 0 to 10, preferably 0 to 5 amino acid insertions, deletions, substitutions, additions, or a combination thereof, such as Fig. 9A , Fig. 9B , Fig.9D or Fig.9E In a preferred embodiment, the O12 light chain is a light chain comprising 0 to 9, 0 to 8, 0 to 7, 0 to 6, 0 to 5, 0 to 4 amino acid insertions, deletions, substitutions, additions or a combination thereof. Fig. 9A , 9B , 9D or 9E. In a preferred embodiment, the O12 light chain is a light chain comprising 0 to 5, preferably 0 to 4, more preferably 0 to 3 amino acid insertions, deletions, substitutions, additions or a combination thereof. Fig. 9A , Fig. 9B , Fig.9D or Fig.9E In a preferred embodiment, the O12 light chain is a light chain having a sequence as shown in . In a preferred embodiment, the O12 light chain is a light chain having 0 to 2, more preferably 0 to 1, most preferably 0 amino acid insertions, deletions, substitutions, additions or a combination thereof as shown in Fig. 9A , Fig. 9B , Fig.9D or Fig.9E In a preferred embodiment, the O12 light chain is a light chain comprising a sequence as shown in Fig. 9A or Fig. 9B In a preferred embodiment, the light chain comprises Fig. 9A In a preferred embodiment, the light chain variable region comprises Fig. 9B The 1, 2, 3, 4 or 5 amino acid substitutions mentioned are preferably conservative amino acid substitutions, and insertions, deletions, substitutions or combinations thereof are preferably not in the CDR3 region of the VL chain, preferably not in the CDR1, CDR2 or CDR3 region or FR4 region of the VL chain.

[0144] The common light chain may have a lambda light chain, and therefore this is also provided in the context of the present invention as disclosed herein, but a kappa light chain is preferred. The constant portion of the common light chain of the present invention as disclosed herein may be a constant region of a kappa or lambda light chain. It is preferably a constant region of a kappa light chain, preferably wherein the common light chain is a germline light chain, preferably a rearranged germline human kappa light chain comprising an IgVK1-39 gene segment, most preferably a rearranged germline human kappa light chain IgVK1-39*01 / IGJK1*01 (Fig. 9). The terms rearranged germline human kappa light chain IgVκ1-39*01 / IGJκ1*01, IGKV1-39 / IGKJ1, huVκ1-39 light chain or huVκ1-39 for short, or simply 1-39 are used interchangeably throughout this application.

[0145] The common light chain producing cell can produce, for example, the rearranged germline human kappa light chain IgVκ1-39*01 / IGJκ1*01 and a light chain comprising the variable region of the mentioned light chain fused to a lambda constant region.

[0146] In a preferred embodiment, the light chain variable region comprises an amino acid sequence having 0 to 10, preferably 0 to 5, amino acid insertions, deletions, substitutions, additions or a combination thereof.

[0147]

[0148] In a preferred embodiment, the light chain variable region comprises 0 to 9, 0 to 8, 0 to 7, 0 to 6, 0 to 5, 0 to 4, preferably 0 to 3, preferably 0 to 2, preferably 0 to 1, and preferably 0 amino acid insertions, deletions, substitutions, additions or combinations thereof relative to the specified amino acid sequence. A combination of insertions, deletions, additions or substitutions is a claimed combination if the aligned sequences do not differ at more than 5 positions. In a preferred embodiment, the light chain variable region comprises the amino acid sequence

[0149]

[0150] In a preferred embodiment, the light chain variable region comprises the amino acid sequence

[0151]

[0152] In another preferred embodiment, the light chain variable region comprises the amino acid sequence

[0153]

[0154] Amino acid insertion, deletion, replacement, addition or its combination is preferably not in the CDR3 region of the light chain variable region, preferably not in the CDR1 or CDR2 region of the light chain variable region. In a preferred embodiment, the light chain variable region does not comprise deletion, addition or insertion relative to the specified sequence. In this embodiment, the heavy chain variable region may have 0 to 5 amino acid replacements relative to the specified amino acid sequence. The amino acid replacement is preferably a conservative amino acid replacement. The CDR1, CDR2 and CDR3 of the light chain of the antibody of the present invention preferably comprise the amino acid sequence CDR1-QSISSY, CDR2-AAS, CDR3-QQSYSTP, i.e., the CDR of IGKV1-39 (according to IMGT), respectively.

[0155] The bispecific antibodies as described herein preferably have a heavy chain variable region / light chain variable region (VH / VL) combination that binds to the extracellular portion of EGFR and a second VH / VL combination that binds to the extracellular portion of cMET. In a preferred embodiment, the VL in the first VH / VL combination is similar to the VL in the second VH / VL combination. In a more preferred embodiment, the VL in the first and second VH / VL combinations are the same. In a preferred embodiment, the bispecific antibody is a full-length antibody having a heavy / light (H / L) chain combination that binds to the extracellular portion of EGFR and a H / L chain combination that binds to the extracellular portion of cMET. In a preferred embodiment, the light chain in the first H / L chain combination is similar to the light chain in the second H / L chain combination. In a more preferred embodiment, the light chains in the first and second H / L chain combinations are the same.

[0156] Several methods for producing host cells whose expression is conducive to the production of bispecific antibodies or, conversely, monospecific antibodies have been published. In the present invention, preferably, relative to the production of corresponding monospecific antibodies, the cell expression of antibody molecules is more conducive to the production of bispecific antibodies. This is usually achieved by modifying the constant region of the heavy chain so that it is more conducive to heterodimerization (i.e., heavy chain dimerization combined with another heavy chain / light chain) than homodimerization. In a preferred embodiment, the bispecific antibody of the present invention as disclosed herein comprises two different immunoglobulin heavy chains with a compatibility heterodimerization domain. A variety of compatibility heterodimerization domains have been described in the art. The compatibility heterodimerization domain is preferably a compatibility immunoglobulin heavy chain CH3 heterodimerization domain. When using a wild-type CH3 domain, the co-expression of two different heavy chains (A and B) and a common light chain will produce three different antibody substances: AA, AB and BB. AA and BB are representations of two monospecific bivalent antibodies, and AB is a representation of a bispecific antibody. In order to increase the percentage of the desired bispecific product (AB), CH3 engineering can be used, or in other words, heavy chains with compatible heterodimerization domains can be used, as defined below. The art describes a variety of ways in which such heterodimerization of heavy chains can be achieved. One way is to generate a "knob into hole" bispecific antibody.

[0157] As used herein, the term "compatible heterodimerization domain" refers to a protein domain that is engineered such that the engineered domain A' will preferentially form a heterodimer with the engineered domain B', and vice versa, homodimerization between A'-A' and B'-B' is reduced.

[0158] In US13 / 866,747 (now authorized as US 9,248,181), US14 / 081,848 (now authorized as US 9,358,286) and PCT / NL2013 / 050294 (published as WO2013 / 157954, incorporated herein by reference), methods and means for producing bispecific antibodies using compatible heterodimerization domains are disclosed. These means and methods may also be advantageously used in the present invention. Specifically, the bispecific antibodies of the present invention as disclosed herein preferably comprise mutations to substantially produce expression of bispecific full-length IgG molecules in a host cell. Preferred mutations are amino acid substitutions L351K and T366K ("KK variant" heavy chain) in the first CH3 domain and amino acid substitutions L351D and L368E ("DE variant" heavy chain) in the second domain, or vice versa. US 9,248,181 and US 9,358,286 patents and WO2013 / 157954 PCT application (which are incorporated herein by reference) demonstrate that DE variants preferentially pair with KK variants to form heterodimers (so-called "DEKK" bispecific molecules). Due to the repulsion between the charged residues in the CH3-CH3 interface between the same heavy chains, homodimerization of the DE variant heavy chains is not favorable (DEDE homodimers).

[0159] Bispecific antibodies can be generated by (transient) transfection of a plasmid encoding a light chain and two different heavy chains, which are transformed through CH3 to ensure effective heterodimerization and the formation of bispecific antibodies. The generation of these chains in a single cell results in a formation that is more conducive to bispecific antibodies relative to the formation of monospecific antibodies. The preferred mutation that substantially only produces bispecific full-length IgG1 molecules is an amino acid replacement at positions 351 and 366 in the first CH3 domain, such as L351K and T366K (numbered according to EU numbering) ("KK variant" heavy chain), and an amino acid replacement at positions 351 and 368 in the second CH3 domain, such as L351D and L368E ("DE variant" heavy chain), or vice versa (see, e.g., Fig.10E and 10F ).

[0160] In one embodiment, the heavy chain / light chain combination comprising a variable domain that binds to EGFR comprises a DE variant of the heavy chain. In this embodiment, the heavy chain / light chain combination comprising a variable domain that can bind to cMET comprises a KK variant of the heavy chain. The KK variant of the heavy chain that binds to cMET does not produce homodimers, thereby making the observed effect of the bispecific antibody on HGF-induced cMET activation inhibition very precise. It avoids the activation (excitement) of cMET that is sometimes observed with a bivalent cMET antibody.

[0161] The Fc region mediates the effector functions of antibodies, such as complement-dependent cytotoxicity (CDC), antibody-dependent cellular cytotoxicity (ADCC) and antibody-dependent cell phagocytosis (ADCP). Depending on the application of therapeutic antibodies or Fc fusion proteins, it is expected to reduce or increase effector functions. In some embodiments of the present invention as disclosed herein, when an immune response is to be activated, enhanced or stimulated, reduced effector functions can be expected. Antibodies with reduced effector functions can be used for targeting cell surface molecules of immune cells, etc.

[0162] The antibody with reduced effector function is preferably an IgG antibody comprising a modified CH2 / lower hinge region, for example to reduce Fc-receptor interactions or reduce C1q binding. In some embodiments, the antibody of the present invention is an IgG antibody with a mutated CH2 and / or lower hinge domain, such that the interaction of the bispecific IgG antibody with the Fc-γ receptor is reduced. The antibody comprising a mutated CH2 region is preferably an IgG1 antibody. Such a mutated IgG1 CH2 and / or lower hinge domain preferably comprises an amino acid substitution at position 235 and / or position 236 (EU numbering), preferably L235G and / or G236R substitution ( Fig. 10D ).

[0163] The antibodies of the present invention as disclosed herein preferably have effector functions. The bispecific antibodies of the present invention as disclosed herein preferably comprise antibody-dependent cell-mediated cytotoxicity (ADCC). Antibodies can be modified to enhance ADCC activity (for a review, see Cancer Sci. 2009 Sep; 100 (9): 1566-72). The modified therapeutic antibodies have improved effector functions (Kubota T, Niwa R, Satoh M, Akinaga S, Shitara K, Hanai N). There are several in vitro methods for determining the efficacy of antibodies or effector cells in inducing ADCC. Among them are chromium 51 [Cr51] release assays, europium [Eu] release assays, and sulfur 35 [S35] release assays. Generally speaking, a labeled target cell line expressing a surface-exposed antigen is incubated with an antibody specific for the antigen. After washing, effector cells expressing Fc receptor CD16 are co-incubated with antibody-labeled target cells. Target cell lysis is then measured by a scintillation counter or spectrophotometry by releasing intracellular markers. In one aspect, the bispecific antibodies of the present invention as disclosed herein exhibit ADCC activity. In such an aspect, the bispecific antibodies may have improved ADCC activity. In another aspect, the bispecific antibodies of the present invention as disclosed herein do not exhibit ADCC activity. In such an aspect, the antibody may have reduced ADCC by one or more CH2 mutations as described elsewhere herein and by techniques known in the art. A technique for enhancing the ADCC of an antibody is non-fucosylation. (See, e.g., Junttila, TT, K. Parsons, et al. (2010). "Superior In vivo Efficacy of Afucosylated Trastuzumab in the Treatment of HER2-Amplified Breast Cancer." Cancer Research 70 (11): 4481-4489). Therefore, non-fucosylated bispecific antibodies according to the present invention as disclosed herein are also provided. Alternatively or additionally, ADCC enhancement may be achieved using a variety of other strategies, including, for example, glycoengineering (Kyowa Hakko / Biowa, GlycArt (Roche) and Eureka Therapeutics) and mutagenesis, all of which attempt to improve Fc binding to the low affinity activating FcγRIIIa, and / or reduce binding to the low affinity inhibitory FcγRIIb. The bispecific antibodies of the invention as disclosed herein are preferably afucosylated in order to enhance ADCC activity.The bispecific antibodies of the invention as disclosed herein preferably comprise reduced amounts of fucosylated N-linked carbohydrate structures in the Fc region when compared to the same antibody produced in normal CHO cells.

[0164] Variants of antibodies or bispecific antibodies as described herein include functional parts, derivatives and / or analogs of antibodies or bispecific antibodies. Variants maintain the binding specificity of (bispecific) antibodies. Functional parts, derivatives and / or analogs maintain the binding specificity of (bispecific) antibodies. As described herein, the binding specificity is defined by the ability to bind the extracellular part of the first membrane protein and the second membrane protein.

[0165] The bispecific antibodies of the invention disclosed herein are preferably used for humans. A preferred antibody of the present invention is a human or humanized antibody. The constant region of the bispecific antibodies of the invention disclosed herein is preferably a human constant region. The constant region may contain one or more, preferably no more than 10, preferably no more than 5 amino acid differences from the constant region of a naturally occurring human antibody. Preferably, the constant portion is derived entirely from a naturally occurring human antibody. The various antibodies produced herein are derived from a human antibody variable domain library. Therefore, these variable domains are human. The unique CDR region may be derived from a human, be synthetic, or be derived from another organism. When the variable region has an amino acid sequence identical to the amino acid sequence of the variable region of a naturally occurring human antibody (except for the CDR region), it is considered to be a humanized variable region. In some such embodiments, the VH of the variable domain of the antibody binding to EGFR or cMET of the present invention as disclosed herein may contain one or more, preferably no more than 10, preferably no more than 5 amino acid differences from the variable region of a naturally occurring human antibody, excluding possible differences in the amino acid sequence of the CDR region. The light chain variable region of the EGFR binding domain and / or cMET binding domain in the antibodies of the invention as disclosed herein may contain one or more, preferably no more than 10, preferably no more than 5 amino acid differences from the variable region of a naturally occurring human antibody, not counting possible differences in the amino acid sequence of the CDR region. The light chain in the antibodies of the invention as disclosed herein may contain one or more, preferably no more than 10, preferably no more than 5 amino acid differences from the variable region of a naturally occurring human antibody, not counting possible differences in the amino acid sequence of the CDR region. Such mutations also occur naturally in the context of somatic hypermutation.

[0166] At least for the heavy chain variable region, the antibody can be derived from a variety of animal species. Common practice is humanization, such as humanizing the murine heavy chain variable region. There are a variety of ways in which humanization can be achieved, including grafting the CDRs into a human heavy chain variable region having a 3D structure that matches the 3D structure of the murine heavy chain variable region; deimmunization of the murine heavy chain variable region, preferably by removing known or suspected T cell or B cell epitopes from the murine heavy chain variable region. The removal is usually performed by replacing one or more amino acids in the epitope with another (usually conservative) amino acid, so that the sequence of the epitope is modified so that it is no longer a T cell or B cell epitope.

[0167] The immunogenicity of the deimmunized murine heavy chain variable region in humans is lower than that of the original murine heavy chain variable region. Preferably, the variable region or domain of the invention as disclosed herein is further humanized, for example as veneer. By using veneer technology, external residues that are easily encountered by the immune system are selectively replaced by human residues to provide a hybrid molecule comprising a veneered surface that is weakly immunogenic or substantially non-immunogenic. The animal used in the invention as disclosed herein is preferably a mammal, more preferably a primate, and most preferably a human.

[0168] The bispecific antibody according to the present invention as disclosed herein preferably comprises the constant region of a human antibody. According to the difference of its heavy chain constant domain, antibodies are divided into five categories or isotypes: IgG, IgA, IgM, IgD and IgE. These categories or isotypes include at least one of the heavy chains named with corresponding Greek letters. A preferred embodiment includes antibodies, wherein the constant region is selected from IgG, IgA, IgM, IgD and IgE constant regions, more preferably, the constant region comprises an IgG constant region, i.e. selected from IgG1, IgG2, IgG3 and IgG4. Preferably, the constant region is an IgG1 or IgG4 constant region, more preferably a mutated IgG1 constant region. Some changes in the IgG1 constant region occur naturally and / or are allowed without changing the immunological properties of the resulting antibody. Changes can also be artificially introduced to place certain preferred features on antibodies or parts thereof. Such features are described, for example, in the context of CH2 and CH3 herein. Generally speaking, about 1 to 10 amino acid insertions, deletions, substitutions or combinations thereof are allowed in the constant region.

[0169] Figure 1 , 7 or 8 VH chain relative to Figure 1 , 7 or the VH chain shown in 8 preferably has up to 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10 amino acid insertions, deletions, substitutions or a combination thereof; relative to Figure 1 , 7or the VH chain shown in 8 preferably has 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof; relative to Figure 1 , 7 The VH chain shown in 8 preferably has 0, 1, 2, 3 or 4 insertions, deletions, substitutions or combinations thereof, preferably 0, 1, 2 or 3 insertions, deletions, substitutions or combinations thereof, more preferably 0, 1 or 2 insertions, deletions, substitutions or combinations thereof, and preferably 0 or 1 insertion, deletion, substitution or combinations thereof. One or more amino acid insertions, deletions, substitutions or combinations thereof are preferably not in the CDR1, CDR2 and / or CDR3 regions of the VH chain. It is also preferably not present in the Fr4 region. Amino acid substitutions are preferably conservative amino acid substitutions.

[0170] Rational approaches have been developed towards minimizing the content of non-human residues in the human environment. A variety of methods can be used to successfully graft the antigen binding properties of an antibody onto another antibody. The binding properties of an antibody may be primarily present in the exact sequence of the CDR3 region, which is generally supported by the sequences of the CDR1 and CDR2 regions in the variable domain in combination with the appropriate structure of the variable domain as a whole. The amino acid sequence of the CDR region as described herein is determined using the Kabat definition. Currently, a variety of methods are available for grafting the CDR region onto the appropriate variable domain of another antibody. Some of these methods are reviewed in JC Almagro1 and J. Fransson (2008) Frontiers in Bioscience 13, 1619-1633, which is included herein by reference. Therefore, the present invention as disclosed herein also provides a human or humanized bispecific antibody comprising a first antigen binding site that binds to EGFR and a second antigen binding site that binds to cMET, wherein the variable domain comprising the EGFR binding site comprises such as Figure 1 The VH CDR3 sequence for MF3370 is shown in , and the variable domain containing the cMET binding site comprises Figure 1 The VH CDR3 region for MF4356 is shown in FIG. 1 . The VH variable region comprising the EGFR binding site preferably comprises Figure 1 The VH variable region containing the cMET binding site preferably comprises the following: Figure 1 The sequences of the CDR1, CDR2 and CDR3 regions of the VH chain shown for MF4356 are shown in FIG. CDR grafting can also be used to generate Figure 1The different framework may be the framework of another human VH, or the framework of a different mammal. Therefore, the present invention as disclosed herein also provides a human or humanized bispecific antibody comprising a first antigen binding site that binds to EGFR and a second antigen binding site that binds to cMET, wherein the variable domain comprising the EGFR binding site comprises Figure 7 The VH CDR3 sequence for MF8233 is shown in , and the variable domain comprising the cMET binding site comprises Figure 8 The VH CDR3 region for MF8230 is shown in FIG. The VH variable region comprising the EGFR binding site preferably comprises Figure 7 The VH variable region comprising the cMET binding site preferably comprises the following: Figure 8 The sequences of the CDR1, CDR2 and CDR3 regions of the VH chain shown for MF8230 are shown in FIG. CDR grafting can also be used to generate Figure 7 or Figure 8 The CDR regions of a VH of the same human, but with a VH chain of a different framework. The different framework may be the framework of another human VH, or the framework of a different mammal.

[0171] Therefore, the present invention as disclosed herein also provides a human or humanized bispecific antibody comprising a first antigen binding site that binds to EGFR and a second antigen binding site that binds to cMET, wherein the variable domain comprising the EGFR binding site comprises Figure 1 The VH CDR3 sequence for MF3370 is shown in , and the variable domain containing the cMET binding site comprises Figure 8 The VH CDR3 region for MF8230 is shown in FIG. The VH variable region comprising the EGFR binding site preferably comprises Figure 1 The VH variable region containing the cMET binding site preferably comprises the following: Figure 8 The sequences of the CDR1, CDR2 and CDR3 regions of the VH chain shown for MF8230 are shown in FIG. CDR grafting can also be used to generate Figure 7 or Figure 8 The CDR regions of a VH of the same human, but with a VH chain of a different framework. The different framework may be the framework of another human VH, or the framework of a different mammal.

[0172] Therefore, the present invention as disclosed herein also provides a human or humanized bispecific antibody comprising a first antigen binding site that binds to EGFR and a second antigen binding site that binds to cMET, wherein the variable domain comprising the EGFR binding site comprises Figure 7 The VH CDR3 sequence for MF8233 is shown in , and the variable domain comprising the cMET binding site comprises Figure 8 The VH CDR3 region for MF4356 is shown in FIG. 1 . The VH variable region comprising the EGFR binding site preferably comprises Figure 7 The VH variable region comprising the cMET binding site preferably comprises the following: Figure 8 The sequences of the CDR1, CDR2 and CDR3 regions of the VH chain shown for MF4356 are shown in FIG. CDR grafting can also be used to generate Figure 7 or Figure 8 The CDR regions of a VH of the same human, but with a VH chain of a different framework. The different framework may be the framework of another human VH, or the framework of a different mammal.

[0173] Therefore, the present invention as disclosed herein also provides a human or humanized bispecific antibody comprising a first antigen binding site that binds to EGFR and a second antigen binding site that binds to cMET, wherein the variable domain comprising the EGFR binding site comprises Figure 7 The VH CDR3 sequence for MF8232 is shown in , and the variable domain comprising the cMET binding site comprises Figure 8 The VH CDR3 region for MF8230 is shown in FIG. The VH variable region comprising the EGFR binding site preferably comprises Figure 7 The VH variable region containing the cMET binding site preferably comprises the following: Figure 8 The sequences of the CDR1, CDR2 and CDR3 regions of the VH chain shown for MF8230 are shown in FIG. CDR grafting can also be used to generate Figure 7 or Figure 8 The CDR regions of a VH of the same human, but with a VH chain of a different framework. The different framework may be the framework of another human VH, or the framework of a different mammal.

[0174] Therefore, the present invention as disclosed herein also provides a human or humanized bispecific antibody comprising a first antigen binding site that binds to EGFR and a second antigen binding site that binds to cMET, wherein the variable domain comprising the EGFR binding site comprises Figure 7The VH CDR3 sequence for MF8232 is shown in , and the variable domain comprising the cMET binding site comprises Figure 8 The VH CDR3 region for MF4356 is shown in FIG. 1 . The VH variable region comprising the EGFR binding site preferably comprises Figure 7 The VH variable region containing the cMET binding site preferably comprises the following: Figure 8 The sequences of the CDR1, CDR2 and CDR3 regions of the VH chain shown for MF4356 are shown in FIG. CDR grafting can also be used to generate Figure 7 or Figure 8 The CDR regions of a VH of the same human, but with a VH chain of a different framework. The different framework may be the framework of another human VH, or the framework of a different mammal.

[0175] The present invention also provides a human or humanized bispecific antibody comprising a first variable domain that binds to EGFR and a second variable domain that binds to cMET, wherein the first variable domain comprises a first variable domain that has at most 10, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably has 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof. Figure 7 The heavy chain variable region of the amino acid sequence of MF3370 shown in , and wherein the second variable domain comprises a heavy chain variable region having 0 to 10, preferably 0 to 5, amino acid insertions, deletions, substitutions, additions or a combination thereof. Figure 8 The present invention also provides a human or humanized bispecific antibody comprising a first variable domain that binds to EGFR and a second variable domain that binds to cMET, wherein the first variable domain comprises ... Figure 7 The heavy chain variable region of the amino acid sequence of MF8233 shown in , and wherein the second variable domain comprises a sequence having 0 to 10, preferably 0 to 5, amino acid insertions, deletions, substitutions, additions or a combination thereof. Figure 8 The heavy chain variable region of MF8230 (SEQ ID NO: 13) has the amino acid sequence shown in FIG.

[0176] The present invention also provides a human or humanized bispecific antibody comprising a first variable domain that binds to EGFR and a second variable domain that binds to cMET, wherein the first variable domain comprises a first variable domain that has at most 10, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably has 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof. Figure 7 The heavy chain variable region of the amino acid sequence of MF3370 shown in , and wherein the second variable domain comprises a heavy chain variable region having 0 to 10, preferably 0 to 5, amino acid insertions, deletions, substitutions, additions or a combination thereof. Figure 8 The heavy chain variable region of MF8230 (SEQ ID NO: 13) has the amino acid sequence shown in FIG.

[0177] The present invention also provides a human or humanized bispecific antibody comprising a first variable domain that binds to EGFR and a second variable domain that binds to cMET, wherein the first variable domain comprises a first variable domain that has at most 10, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably has 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof. Figure 7 The heavy chain variable region of the amino acid sequence of MF8233 shown in , and wherein the second variable domain comprises a sequence having 0 to 10, preferably 0 to 5, amino acid insertions, deletions, substitutions, additions or a combination thereof. Figure 8 The heavy chain variable region of MF4356 (SEQ ID NO: 23) is shown in FIG.

[0178] The present invention also provides a human or humanized bispecific antibody comprising a first variable domain that binds to EGFR and a second variable domain that binds to cMET, wherein the first variable domain comprises a first variable domain that has at most 10, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably has 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof. Figure 7 The heavy chain variable region of the amino acid sequence of MF8232 shown in , and wherein the second variable domain comprises a heavy chain variable region comprising 0 to 10, preferably 0 to 5, amino acid insertions, deletions, substitutions, additions or a combination thereof. Figure 8 The heavy chain variable region of MF4356 (SEQ ID NO: 23) is shown in FIG.

[0179] The present invention also provides a human or humanized bispecific antibody comprising a first variable domain that binds to EGFR and a second variable domain that binds to cMET, wherein the first variable domain comprises a first variable domain that has at most 10, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably has 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof. Figure 7 The heavy chain variable region of the amino acid sequence of MF8232 shown in , and wherein the second variable domain comprises a heavy chain variable region comprising 0 to 10, preferably 0 to 5, amino acid insertions, deletions, substitutions, additions or a combination thereof. Figure 8 The heavy chain variable region of MF8230 (SEQ ID NO: 23) is shown in FIG.

[0180] The up to 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5, amino acid substitutions mentioned are preferably conservative amino acid substitutions, insertions, deletions, substitutions or a combination thereof, preferably not in the CDR3 region of the VH chain, preferably not in the CDR1, CDR2 or CDR3 regions of the VH chain, and preferably not in the FR4 region.

[0181] Bispecific antibodies can be produced in a variety of ways. One method involves expressing two different heavy chains and two different light chains in cells, and collecting antibodies produced by cells. Antibodies produced in this way will generally include a collection of antibodies with different combinations of heavy chains and light chains, some of which are desired bispecific antibodies. The bispecific antibodies can then be purified from the collection. The ratio of bispecific antibodies produced by cells to other antibodies can be increased in a variety of ways. In a preferred embodiment, the ratio is increased by expressing a common light chain instead of expressing two different light chains in the cell. When a common light chain is expressed with two different heavy chains, the ratio of bispecific antibodies produced by cells to other antibodies is significantly increased relative to the expression of two different light chains. The ratio of bispecific antibodies produced by cells can be further increased by stimulating two different heavy chains to pair with each other relative to the pairing of two identical heavy chains. Methods and means for producing bispecific antibodies (from single cells) are disclosed, thereby providing means that are more conducive to the formation of bispecific antibodies relative to the formation of monospecific antibodies. These methods can also be advantageously used in the present invention. Therefore, the present invention as disclosed herein provides in one aspect a method for producing a bispecific antibody from a single cell, wherein the bispecific antibody comprises two CH3 domains capable of forming an interface, the method comprising providing in the cell: a) a first nucleic acid molecule encoding a heavy chain containing a first CH3 domain, b) a second nucleic acid molecule encoding a heavy chain containing a second CH3 domain, wherein the nucleic acid molecule provides for preferential pairing of the heavy chain containing the first CH3 domain with the heavy chain containing the second CH3 domain, the method further comprising culturing the host cell and allowing the expression of the two nucleic acid molecules and harvesting the bispecific antibody from the culture. The first and second nucleic acid molecules may be part of the same nucleic acid molecule, vector or gene delivery vector, and may be integrated at the same site of the host cell genome. Alternatively, the first and second nucleic acid molecules are provided to the cell separately.

[0182] A preferred embodiment provides a method for producing a bispecific antibody according to the present invention as disclosed herein from a single cell, wherein the bispecific antibody comprises two CH3 domains capable of forming an interface, the method comprising providing:

[0183] - a cell having: a) a first nucleic acid molecule encoding a heavy chain comprising an antigen binding site that can bind to EGFR and comprising a first CH3 domain; and b) a second nucleic acid molecule encoding a heavy chain comprising an antigen binding site that binds to ErbB-3 and comprising a second CH3 domain, wherein the nucleic acid molecules provide for preferential pairing of the first and second CH3 domains,

[0184] The method also includes culturing the cell and allowing expression of the protein encoded by the two nucleic acid molecules, and harvesting the bispecific IgG antibody from the culture. In a particularly preferred embodiment, the cell also has a third nucleic acid molecule encoding a common light chain. The first, second and third nucleic acid molecules can be part of the same nucleic acid molecule, vector or gene delivery vector, and can be integrated in the same site of the host cell genome. Alternatively, the first, second and third nucleic acid molecules are provided to the cell, respectively. As described above, the preferred common light chain is based on O12, which is preferably rearranged germline human κ light chain IgVκ1 39*01 / IGJκ1*01. The means for preferential pairing of the first CH3 domain and the second CH3 domain are preferably corresponding mutations in the CH3 domain of the heavy chain coding region. Preferred mutations for preferentially producing bispecific antibodies are amino acid replacements L351K and T366K (EU numbering) in the first CH3 domain and amino acid replacements L351D and L368E in the second CH3 domain, or vice versa. Therefore, a method for producing a bispecific antibody according to the present invention as disclosed herein is also provided, wherein the first CH3 domain comprises amino acid replacements L351K and T366K (EU numbering), and wherein the second CH3 domain comprises amino acid replacements L351D and L368E, and the method further comprises culturing the cell and allowing expression of the protein encoded by the nucleic acid molecule, and harvesting the bispecific antibody from the culture. A method for producing a bispecific antibody according to the present invention as disclosed herein is also provided, wherein the first CH3 domain comprises amino acid replacements L351D and L368E (EU numbering), and wherein the second CH3 domain comprises amino acid replacements L351K and T366K, and the method further comprises culturing the cell and allowing expression of the nucleic acid molecule, and harvesting the bispecific antibody from the culture. Antibodies that can be produced by these methods are also part of the present invention. The CH3 heterodimerization domain is preferably an IgG1 heterodimerization domain. The heavy chain constant region comprising the CH3 heterodimerization domain is preferably an IgG1 constant region.

[0185] In one embodiment of the present invention, a nucleic acid molecule encoding an antibody heavy chain variable region is included. The nucleic acid molecule (usually an in vitro, isolated or recombinant nucleic acid molecule) preferably encodes Figure 7 or Figure 8 or a heavy chain variable region having 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof. Figure 7 or Figure 8 In a preferred embodiment, the nucleic acid molecule comprises Figure 7 or Figure 8The codon-optimized amino acid sequence of the amino acid sequence shown in . The codon optimization is optimized for the type and / or cell type of antibody-producing cells. For example, for CHO production, the nucleic acid sequence of the molecule is codon-optimized for use in Chinese hamster cells. The present invention also provides a coding Figure 7 or Figure 8 A nucleic acid molecule of the heavy chain.

[0186] The nucleic acid molecules used in the present invention as disclosed herein are generally, but not absolutely, ribonucleic acid (RNA) or deoxyribonucleic acid (DNA). Alternative nucleic acids can be used by those skilled in the art. The nucleic acid according to the present invention as disclosed herein is, for example, contained in a cell. When the nucleic acid is expressed in the cell, the cell can produce an antibody according to the present invention as disclosed herein. Therefore, in one embodiment of the present invention, a cell comprising an antibody according to the present invention as disclosed herein and / or a nucleic acid according to the present invention as disclosed herein is included. The cell is preferably an animal cell, more preferably a mammalian cell, more preferably a primate cell, and most preferably a human cell. Suitable cells are any cells that can contain and preferably produce an antibody according to the present invention as disclosed herein and / or a nucleic acid according to the present invention as disclosed herein.

[0187] The present invention as disclosed herein also provides cells comprising antibodies according to the present invention as disclosed herein. Preferably, the cells (typically in vitro, isolated or recombinant cells) produce the antibodies. The cells may also be storage cells capable of producing the antibodies when taken out of storage and cultured. In a preferred embodiment, the cells are hybridoma cells, Chinese hamster ovary (CHO) cells, NS0 cells or PER-C6TM cells. In a particularly preferred embodiment, the cells are CHO cells. Cell cultures comprising cells according to the present invention as disclosed herein are also provided. Several institutions and companies have developed cell lines for large-scale production of antibodies, for example for clinical applications. Some non-limiting examples of such cell lines are CHO cells, NS0 cells or PER.C6 TM Cells. These cells are also used for other purposes, such as the production of proteins. Cell lines developed for industrial-scale production of proteins and antibodies are also referred to herein as industrial cell lines. Therefore, a preferred embodiment includes the use of cell lines developed for large-scale production of antibodies for the production of antibodies of the invention as disclosed herein, the cell lines preferably comprising cells for the production of antibodies, comprising a gene encoding a protein such as Figure 7 or Figure 8 Nucleic acid molecules of the VH, VL and / or heavy chains shown in .

[0188] The present invention also provides a method for producing an antibody, comprising culturing a cell of the present invention as disclosed herein and harvesting the antibody from the culture. Preferably, the cell is cultured in a serum-free medium. Preferably, the cell is suitable for growth in suspension. Also provided is an antibody obtainable by a method for producing an antibody according to the present invention as disclosed herein. Preferably, the antibody is purified from the culture medium. Preferably, the antibody is affinity purified.

[0189] Cells of the invention as disclosed herein are, for example, hybridoma cell lines, CHO cells, 293F cells, NS0 cells or another cell type known for their suitability for antibody production for clinical purposes. In a particularly preferred embodiment, the cells are human cells. Preferably, cells transformed by the adenovirus E1 region or its functional equivalent. A preferred example of such a cell line is the PER.C6TM cell line or its equivalent. In a particularly preferred embodiment, the cells are CHO cells or variants thereof. Preferably, variants of antibodies are expressed using a glutamine synthetase (GS) vector system.

[0190] After transient transfection in suspension 293F cells, the antibodies of the invention as disclosed herein can be produced at levels of >50 mg / L. The bispecific antibodies can be purified to a purity greater than 98% with a yield of >70%. Analytical characterization studies have shown that the bispecific IgG1 antibody spectrum is comparable to the bivalent monospecific IgG1. In terms of functional activity, the bispecific antibodies of the invention as disclosed herein can show superior efficacy compared to cetuximab in vitro and in vivo.

[0191] The present invention also provides a pharmaceutical composition comprising an antibody according to the invention as disclosed herein.The pharmaceutical composition preferably comprises an excipient or carrier, which is preferably pharmaceutically acceptable.

[0192] The antibody may comprise a label, preferably a label for in vivo imaging. Such a label is generally optional for therapeutic applications. For example, in a diagnostic context, a label may be helpful. For example, it can be used to visualize target cells in vivo. A variety of labels are suitable, and many labels are well known in the art. In a preferred embodiment, the label is a radioactive label for detection. In another preferred embodiment, the label is an infrared label. Preferably, the infrared label is suitable for in vivo imaging. A variety of infrared labels are available to those skilled in the art. Some preferred infrared labels are, for example, IRDye 800; IRDye 680RD; IRDye 680LT; IRDye 750; IRDye 700DX; IRDye 800RSIRDye 650; IRDye 700 phosphoramidite; IRDye 800 phosphoramidite (LI-COR USA; 4647Superior Street; Lincoln, Nebraska).

[0193] The present invention also provides a method for treating a subject having a tumor or being at risk of having the tumor, the method comprising administering an antibody or pharmaceutical composition according to the present invention as disclosed herein to a subject in need thereof. The tumor is preferably an EGFR, cMET or EGFR / cMET positive tumor. Before starting the treatment, the method preferably also includes determining whether the subject suffers from such an EGFR, cMET or EGFR / cMET positive tumor. The present invention also provides an antibody or pharmaceutical composition of the present invention as disclosed herein for treating a subject suffering from an EGFR, cMET or EGFR / cMET positive tumor or being at risk of suffering from the same.

[0194] To confirm whether a tumor is EGFR positive, a technician can, for example, determine EGFR amplification and / or immunohistochemical staining. At least 10% of the tumor cells in the biopsy should be positive. The biopsy may also contain 20%, 30%, 40%, 50%, 60%, 70% or more positive cells. To confirm whether a tumor is cMET positive, a technician can, for example, determine cMET amplification and / or immunohistochemical staining. At least 10% of the tumor cells in the biopsy should be positive. The biopsy may also contain 20%, 30%, 40%, 50%, 60%, 70% or more positive cells.

[0195] The present invention as disclosed herein can be applied to a wide range of cancers, such as breast cancer; colon cancer; pancreatic cancer; gastric cancer; ovarian cancer; colorectal cancer; head and neck cancer; lung cancer, including non-small cell lung cancer; bladder cancer, etc. The tumor can be an EGFR, cMET or EGFR / cMET positive cancer. One embodiment of the present invention can preferably treat breast cancer, such as positive cancer of early breast cancer. In another embodiment of the present invention, EGFR, cMET or EGFR / cMET positive cancer of colorectal cancer can be preferably treated. The present invention as disclosed herein can be applied to a wide range of EGFR, cMET or EGFR / cMET positive cancers, such as breast cancer; colon cancer; pancreatic cancer; gastric cancer; ovarian cancer; colorectal cancer; head and neck cancer; lung cancer, including non-small cell lung cancer; bladder cancer, etc. The subject is preferably a human subject. The subject is preferably a subject suitable for antibody therapy using an EGFR-specific antibody, such as cetuximab. In a preferred embodiment, the present invention can preferably treat a subject comprising a tumor, preferably an EGFR / cMET positive cancer, preferably a tumor / cancer having an EGFR RTK resistance phenotype, an EGFR monoclonal antibody resistance phenotype, or a combination thereof.

[0196] The amount of the antibody to be administered to the patient is generally within the therapeutic window, meaning that a sufficient amount is used to obtain a therapeutic effect without exceeding the threshold value that causes an unacceptable degree of side effects. The lower the amount of antibody required to obtain the desired therapeutic effect, the larger the therapeutic window is generally. Therefore, antibodies according to the present invention as disclosed herein that exert sufficient therapeutic effects at low doses are preferred. The dosage may be within the range of the dosing regimen of cetuximab. The dosage may also be lower.

[0197] Compared with cetuximab under otherwise similar conditions, the bispecific antibodies according to the present invention as disclosed herein preferably induce less skin toxicity. Compared with cetuximab under otherwise similar conditions, the bispecific antibodies according to the present invention as disclosed herein preferably produce less proinflammatory chemokines, preferably CXCL14. Compared with cetuximab under otherwise similar conditions, the bispecific antibodies according to the present invention as disclosed herein preferably induce less damage to antimicrobial RNases (RNAse) (preferably RNase 7).

[0198] The present invention particularly describes targeting EGFR and cMET receptors, and causing antibodies that potently inhibit proliferation of in vitro cancer cell lines and inhibit tumor growth in vivo. The bispecific antibodies of the present invention as disclosed herein can combine low toxicity spectrum with high efficacy. The antibodies of the present invention as disclosed herein can be used for EGFR targeted therapy of various types and lines. When compared with antibodies that bind to the same antigen with two arms, the antibodies of the present invention as disclosed herein can have an improved therapeutic window. When compared with cetuximab antibodies, the bispecific antibodies of the present invention as disclosed herein can show better growth inhibition in vitro, in vivo, or a combination thereof.

[0199] The present invention also provides a bispecific antibody of the present invention as disclosed herein for use in treating a subject that may suffer from one or more of a variety of different types of tumors. The tumor may be an EGFR-positive tumor, a cMET-positive tumor, or an EGFR and cMET-positive tumor. The tumor may be breast cancer; colon cancer; pancreatic cancer; gastric cancer; ovarian cancer; colorectal cancer; head and neck cancer; lung cancer, including non-small cell lung cancer; or bladder cancer. The tumor may be resistant to treatment with an EGFR tyrosine kinase inhibitor. The EGFR tyrosine kinase inhibitor is preferably erlotinib, gefitinib or afatinib; an analog of erlotinib, gefitinib or afatinib; or a combination of one or more of their corresponding compounds and / or analogs. The treatment preferably also includes treatment with an EGFR tyrosine kinase inhibitor. When co-treated with an EGFR tyrosine kinase inhibitor, the tumor may be resistant to treatment with an EGFR tyrosine kinase inhibitor. Co-treatment at least partially restores the sensitivity of the tumor to the tyrosine kinase inhibitor. The EGFR tyrosine kinase inhibitor may be a first generation EGFR tyrosine kinase inhibitor. Some examples of clinically relevant first generation EGFR tyrosine kinase inhibitors are erlotinib and gefitinib.In this and other embodiments, the tumor may be a HGF-associated tumor.

[0200] EGFR-positive tumors are typically tumors with EGFR activating mutations. EGFR activating mutations are EGFR mutations that lead to activation of the EGF / EGFR signaling pathway. EGFR activating mutations may be important for the cancerous state of the tumor. One of the ways in which such tumors can become insensitive to EGFR targeted therapy is by activating the HGF / cMET signaling pathway. The tumor may be a tumor associated with HGF. Activation of the cMET / HGF signaling pathway is one of the ways in which EGFR-positive tumors can escape treatment with EGFR targeted therapy. The cMET / HGF pathway can be activated in a variety of ways. A variety of methods of activation are described in the art, some of which are described in detail herein. The antibodies of the present invention as disclosed herein are particularly suitable for treating tumors in which activation of the cMET / HGF signaling pathway is associated with the presence or excess of HGF. Such cMET-positive tumors are referred to as HGF-associated tumors or HGF-dependent tumors. The antibodies of the present invention as disclosed herein may also be used to at least partially inhibit this possible escape mechanism of EGFR-positive tumors. In addition, such tumors can escape EGFR targeted therapy by selected outgrowth of tumor cells in which the cMET / HGF signaling pathway is activated. Such cells may be present at the beginning of EGFR targeted therapy. Such cells have a selective growth advantage compared to HGF / cMET signaling negative tumor cells. The tumor may be a tumor in which the HGF / cMET signaling pathway is activated. The tumor may be a tumor associated with elevated levels of hepatocyte growth factor (HGF) or overexpression of the HGF receptor c-Met. The tumor may be a tumor in which growth is driven by EGF and / or HGF. If the signaling pathway in the tumor cell is activated in response to the presence of a growth factor, and the removal of the growth factor results in inhibition of tumor cell growth, it indicates that the tumor is driven by a certain growth factor. The reduction can be measured by reduced cell division and / or induced cell killing such as apoptosis. In the presence of HGF, if the tumor grows or grows faster under conditions that otherwise allow tumor growth, the tumor is an HGF-related tumor.

[0201] EGFR targeted therapy for various tumors has been reviewed by Vecchione et al., EGFR-targeted therapy. "Experimental cell research Vol 317 (2011): 2765-2771. In general, EGFR targeted therapy is a treatment with molecules that interact with EGFR and inhibit EGFR-mediated signal transduction in cells.

[0202] The therapeutic method or antibodies for treatment as specified herein preferably also include the step of determining whether the tumor is an HGF-associated tumor.

[0203] The antibodies of the invention as disclosed herein can inhibit the growth of tumors associated with HGF.

[0204] In some embodiments, where the cMET binding variable domain is described as having a CDR2 sequence of "WINTYTGDPTYAQGFTG", the CDR2 sequence may also be "WINTYTGDPTYAQGFT".

[0205] When a range is given herein as a number from 1 to 2, the range includes the number 1 and the number 2. For example, a range of 2 to 5 includes the numbers 2 and 5.

[0206] When an affinity is referred to herein as being higher than another affinity, Kd = lower than another Kd. For the avoidance of doubt, a Kd of 10e-9 M is lower than a Kd of 10e-8 M. An antibody with a Kd of 10e-9 M has a higher affinity for a target than when the Kd is 10e-8 M.

[0207] Citation of a patent document or other matter cited herein is not to be taken as an admission that the document or matter is known or that the information it contains was part of the common general knowledge as of the priority date of any claim.

[0208] For purposes of clarity and concise description, features are described herein as part of the same or separate embodiments, however, it should be understood that the scope of the invention as disclosed herein may include embodiments having a combination of all or some of the described features.

[0209] Example

[0210] As used herein, "MFXXXX" wherein X is independently a number from 0 to 9 refers to a Fab comprising a variable domain wherein VH has an amino acid sequence identified by 4 numbers. Unless otherwise specified, the light chain variable region of the variable domain generally has Fig. 9A ,generally Fig. 9B "MFXXXX VH" refers to the amino acid sequence of a VH identified by 4 numbers. MF also contains the constant region of the light chain and the constant region of the heavy chain that usually interacts with the constant region of the light chain. PG refers to a monospecific antibody comprising the same heavy chain and light chain. PB refers to a bispecific antibody with two different heavy chains. The VH variable regions of the heavy chains are different, and usually the CH3 regions are also different, with one heavy chain having a KK mutation in its CH3 domain and the other heavy chain having a complementary DE mutation in its CH3 domain (reference PCT / NL2013 / 050294 (published as WO2013 / 157954)).

[0211] Example 1: Materials and Methods

[0212] Cell lines:

[0213] Buy EBC-1[JCRB0820]、PC-9[RCB0446]、H358[ CRL-5807 TM ]、HCC827[ CRL-2868 TM ]、MKN-45[DSMZ ACC 409]N87[ CRL-5822 TM ] and A431[ CRL-1555 TM ] cell line and were routinely maintained in growth medium supplemented with 10% heat-inactivated fetal bovine serum (FBS). HEK293F Freestyle cells were obtained from Invitrogen and were routinely maintained in 293 FreeStyle medium.

[0214] cDNA constructs:

[0215] Generation of cMET and EGFR expression vectors for stable cell lines (cMET and EGFR) and for immunization (cMET)

[0216] The full-length cDNAs of each target, containing unique restriction sites for cloning and Kozak consensus sequences for efficient translation, were synthesized or obtained by PCR amplification of commercially available expression constructs containing the target cDNAs with specific primers that introduced unique restriction sites for cloning and Kozak consensus sequences for efficient translation. The full-length cDNAs of each target were cloned into eukaryotic expression constructs such as pcDNA3.1, while the extracellular domains were cloned into pVAX1 and pDisplay. The insert sequences were verified by comparison with the NCBI reference amino acid sequences.

[0217] Amino acid sequence for expression on the cell surface Full-length human EGFR insert sequence (same as GenBank: NP_00533):

[0218]

[0219] in:

[0220] -MRPSGTAGAALLALLAALCPASR: signal peptide.

[0221]

[0222] ECD of human EGFR.

[0223] -IATGMVGALLLLLVVALGIGLFM: predicted TM area.

[0224]

[0225] Intracellular tail region.

[0226] Naturally occurring EGFR variant VAR_066493 caused by in-frame deletion of exons 2 to 7 [Ji H., Zhao X; PNAS 103:7817-7822 (2006)], amino acid sequence of the extracellular domain of human EGFR varIII. The _ below indicates the position of the deletion of amino acids 30 to 297

[0227]

[0228] in:

[0229] -MRPSGTAGAALLALLAALCPASR: signal peptide.

[0230]

[0231] ECD of EGFRvarIII.

[0232] Amino acid sequence chimeric macaque (Macaca mulatta) extracellular EGFR domain hybridized with human EGFR transmembrane and intracellular domains for expression on the cell surface (identical to GenBank: XP_014988922.1). The human EGFR sequence is underlined in the following examples.

[0233]

[0234] in:

[0235] -MGPSGTAGAALLALLAALCPASR: signal peptide.

[0236]

[0237] ECD of cyEGFR.

[0238] Amino acid sequence for expression on the cell surface Full-length human cMET insertion sequence (identical to GenBank: P08581-2) This sequence differs from the reference sequence in that it has the following insertion at position 755-755: S→STWWKEPLNIVSFLFCFAS.

[0239]

[0240]

[0241] in:

[0242] -MKAPAVLAPGILVLLFTLVQRSNG: signal peptide

[0243]

[0244] :ECD of human cMET

[0245] -GLIAGVVSISTALLLLLGFFLWL: transmembrane region

[0246]

[0247] Intracellular region.

[0248] Reference Antibodies

[0249] Anti-cMET antibodies are known in the art (Table 1). Monospecific bivalent cMET antibodies were constructed according to published information and expressed in 293F Freestyle cells. Table 1 shows relevant public information. Monospecific bivalent antibodies against cMET were constructed according to published information and expressed in 293F Freestyle cells. For HGF ligand blocking assays, the VH- and VL-encoding gene segments of the patented anti-cMET antibodies were re-cloned in phage display vectors for display on filamentous phage.

[0250] Reference antibody Cetuximab (Erbitux) was used as the reference antibody for the EGFR Fab group.

[0251] 2994Fab protein was generated from purified PG2994 IgG by papain digestion. Therefore, PG2994 was incubated with papain (Pierce #44985) coupled to beads and allowed to digest for 5.5 hours at 37°C under rotation. Fab fragments were purified from the digestion mixture by filtering on MabSelectSure LX. The fractions containing Fab protein were concentrated to 3 ml using vivaspin20 of 10 kDa and further purified by gel filtration using a superdex75 16 / 600 column in PBS.

[0252] Example 2

[0253] Generation of bivalent monoclonal antibodies and antibody characterization

[0254] The VH gene of a unique antibody judged by the VH gene sequence and some of its sequence variants was cloned into a backbone IgG1 vector. 293F Freestyle cells suitable for suspension were cultured in T125 flasks on a shaking platform until the density reached 3.0 × 10 6 cells / ml. 6 to 0.5×10 6 The density of 100 viable cells / ml was inoculated in each well of a 24-well plate. Cells were transiently transfected with a separate sterile DNA:PEI mixture and further cultured. 7 days after transfection, supernatants were harvested and filtered through 0.22 μM (Sartorius) and purified on protein A beads using batch purification, followed by buffer exchange with PBS.

[0255] Inhibition of EGF-mediated apoptosis

[0256] High concentrations (10 nM) of EGF induce (apoptotic) cell death of A431 cells [(Gulli et al., 1996)]. This effect can be dose-dependently restored by the addition of ligand-blocking anti-EGFR antibodies (such as cetuximab).

[0257] To test the efficacy of bivalent anti-EGFR IgG in inhibiting EGF-induced A431 cell death, the antibody was incubated in a series of titrations (from 10 μg / ml) in the presence of 10 nM EGF. Each assay plate contained a series of dilutions of a negative antibody (Ctrl Ab; PG2708) and a positive control antibody (cetuximab) used as a reference control. On the third day, Alamar blue (Invitrogen, #DAL1100) (20 μl per well) was added and fluorescence was measured using 560 nm excitation and 590 nm readout on a Biotek Synergy 2 multi-mode microplate reader after incubation for 6 hours (at 37 ° C) with Alamar blue. Figure 2 The activity of cLC EGFR antibodies compared to that of cetuximab and a control antibody is shown. Antibodies PG4280, 3755 and 3752 were more potent than cetuximab, while antibodies PG4281 and PG3370 showed lower potency.

[0258] EGF blocking ELISA

[0259] EGFR-specific phage were tested for binding to recombinant EGFR in the absence and presence of molar excess of ligand (EGF). Therefore, 5 μg / ml of goat anti-human IgG was coated onto MAXISORP TMThe ELISA plate was overnight at 4°C. The wells of the ELISA plate were blocked with PBS (pH 7.2) containing 2% ELK while shaking (700 rpm) at RT for 1 h. Next, 5 μg / ml recombinant human EGFR-Fc was allowed to incubate for 1 h at RT. At the same time, IgG was mixed with human biotinylated EGF in a series of titrations at RT for 1 h. After washing away unbound human EGFR-Fc, the antibody / EGF mixture was added and allowed to bind at RT for 1 h. The bound EGF was detected by HRP-streptavidin for 1 h at RT. As a control, the procedure was performed simultaneously with an antibody specific for the coated antigen (not shown) and a negative control phage (Neg Ctrl Ab). The bound secondary antibody was visualized by TMB / H2O2 staining and was visualized by OD 450 nm The staining was quantified. Fig.11 described that the PG3370 antibody, which is less potent than cetuximab in inhibiting EGF-mediated apoptosis, showed similar EGF blocking activity.

[0260] Cross-reactivity test between cynomolgus monkey EGFR and mouse EGFR

[0261] To test whether anti-EGFR IgG reacts with cynomolgus EGFR, constructs encoding full-length human EGFR and expression constructs encoding cynomolgus ECD fused to intracellular human EGFR were transfected into (antigen-negative) CHO cells, and the cells were then stained with 5 μg / ml of anti-EGFR antibody and finally analyzed by FACS. As a positive control for staining, the clinically used antibody cetuximab was used because it is known that the antibody cross-reacts with cynomolgus EGFR. PG3370, PG3752, PG4280, and PG4281 were shown to react with cynomolgus EGFR because the staining of cells expressing human EGFR was almost indistinguishable from that of cells expressing the chimeric receptor ( Fig.12 ).

[0262] To test the anti-EGFR IgG for its cross-reactivity with murine EGFR, an ELISA was performed. Serial titrations of recombinant mouse EGFR ECD-Fc starting at 5 μg / ml and diluted to 0.038 μg / ml were coated onto MAXISORP TMThe ELISA plates were incubated overnight at 4°C. Binding of anti-EGFR IgG to this antigen was tested at a fixed concentration of 5 μg / ml and allowed to bind for 1H at RT. As a positive control for antibody immunoreactivity, the same ELISA setup was performed using human EGFR ECD-Fc fusion protein as antigen (R&D systems). Next, goat anti-mouse IgG HRP conjugate (BD Bioscience) was added and allowed to bind for 2 hours at RT. Bound IgG was detected by OD450 nm measurement. Antibody PG3370 has been shown to recognize mouse EGFR as well as human EGFR with similar affinity ( Fig.13 - top). Cetuximab does not recognize mouse EGFR (125084 Erbitux Pharmacology Review Part 2 – FDA). Therefore, PG3370 and cetuximab do not recognize the same epitope on human EGFR.

[0263] Mouse cMET Cross-Reactivity Test

[0264] To test PG3342 for its cross-reactivity with mouse cMET, an ELISA was performed. A fixed concentration of mouse HGF R / c-MET Fc (R&D systems) was diluted to 2.5 μg / ml in PBS and coated onto MAXISORP TM The ELISA plates were incubated overnight at 4°C. Binding of anti-cMET IgG to this antigen was tested in a semi-log titration (starting at 10 μg / ml). The antibodies were allowed to bind for 1H at RT. As a positive control for antibody immunoreactivity, the same ELISA setup was performed using human HGF R / c-MET Fc fusion protein as antigen (R&D systems). Next, goat anti-mouse IgG HRP conjugate (BD Bioscience) was added and allowed to bind for 2 hours at RT. Bound IgG was detected by OD450 nm measurement. Antigen affinity purified polyclonal goat IgG BAF527 against mouse cMET conjugated to biotin was included as a positive control antibody. No cross-reactivity with mouse cMET was observed with the PG3342 antibody ( Fig.13 below).

[0265] Cross-blocking assay for cMET antibodies

[0266] The competition of cMET-specific phage with cMET reference antibody was tested in ELISA. Therefore, 2.5 μg / ml of cMET-Fc fusion protein was coated onto MAXISORP TMThe ELISA plate was kept overnight at 4°C. The wells of the ELISA plate were blocked with PBS (pH 7.2) containing 2% ELK for 1H at RT while shaking (700rpm). Next, reference or negative control IgG was added at a concentration of 5μg / ml and allowed to bind at 700rpm for 15 minutes at RT. Next, 5μl of PEG-precipitated phage was added and allowed to bind at 700rpm for 1H at RT. The bound phage was detected with an HRP-labeled anti-M13 antibody at 700rpm for 1H at RT. As a control, the procedure was performed simultaneously with antibodies specific for the coated antigen and negative control phage. The bound secondary antibody was visualized by TMB / H2O2 staining and was visualized by OD 450nm The staining was quantified by measurement. Table 2 shows that MF4040 and MF4356 showed competition with the 5D5 reference antibody. MF4297 competed with 13.3.2 and C8H241 to a lesser extent. The positive control phages all showed complete competition with the corresponding IgG, while the no antibody control did not affect the competition assay.

[0267] Generation of bispecific antibodies

[0268] Bispecific antibodies are generated by transient co-transfection of two plasmids encoding IgG with different VH domains, using patented CH3 modification technology to ensure efficient heterodimerization and bispecific antibody formation. A common light chain on the same plasmid or on another plasmid is also co-transfected in the same cell. In our co-pending applications (e.g., WO2013 / 157954 and WO2013 / 157953; incorporated herein by reference), we disclose methods and means for producing bispecific antibodies from single cells, thereby providing means that are more conducive to the formation of bispecific antibodies relative to the formation of monospecific antibodies. These methods can also be advantageously used in the present invention. Specifically, preferred mutations that substantially only produce bispecific full-length IgG molecules are amino acid replacements at positions 351 and 366 in the first CH3 domain, such as L351K and T366K (numbered according to EU numbering) ("KK variant" heavy chain), and amino acid replacements at positions 351 and 368 in the second CH3 domain, such as L351D and 10L368E ("DE variant" heavy chain), or vice versa. Previously, it has been demonstrated in our co-pending applications that negatively charged DE variant heavy chains and positively charged KK variant heavy chains preferentially pair to form heterodimers (so-called "DEKK" bispecific molecules). Due to the strong repulsion between charged residues in the CH3-CH3 interface between the same heavy chains, homodimerization of DE variant heavy chains (DE-DE homodimers) or homodimerization of KK variant heavy chains (KK-KK homodimers) is unfavorable.

[0269] cMET and EGFR Fab arms were cloned in suitable KK and DE vectors (Table 3). After production, bispecific IgG was purified by protein A batch purification, and the buffer was exchanged for PBS. Successful production produced a minimum concentration of 0.1 mg / ml IgG1 full-length antibody, and a unique number (PBnnnnn; wherein nnnnn represents a randomly generated number) was assigned to identify the specific combination of 2 different target-binding Fab fragments. The combination of the successfully produced bispecific IgG and its corresponding target was tested in ELISA.

[0270] Example 3

[0271] Screening of c-MET x EGFR bispecific antibodies in EGF / HGF and HGF and EGF proliferation assays

[0272] The efficacy of the cMET x EGFR bispecific antibody panel was tested in N87 cells using HGF / EGF, HGF and EGF assays. The N87 cell line (formally known as NCI-N87) is a gastric cancer cell line derived from a metastatic site and has high EGFR expression levels and moderate cMET expression levels (Zhang et al, 2010). The antibodies were tested in 8 steps of semi-log titration (10 μg / ml to 3.16 ng / ml). Each antibody was tested in duplicate. The anti-RSV-G antibody PG2708 was used as a negative control. The reference antibody 2994 Fab was used as a positive control for the HGF assay, and the reference antibody cetuximab was used as a positive control for the EGF assay.

[0273] Equimolar 1:1 cetuximab / 5D5 Fab was used as a positive control for the EGF, HGF, and EGF / HGF assays.

[0274] Wells with one ligand, or a combination of ligands, and a culture medium control are included to determine the assay window. Antibodies are diluted in chemically defined starvation medium (CDS: RPMI1640 culture medium, containing 80U penicillin and 80μg streptomycin / ml, 0.05% (w / v) BSA and 10μg / ml holo-transferrin), and 50μl of the diluted antibodies are added to the wells of a 96-well blank well transparent bottom plate (Costar). Add ligand (50μl / well stock solution, the stock solution contains 400ng / ml HGF and 4ng / ml EGF diluted in CDS, and EGF / HGF at a concentration of 4ng / ml EGF / 400ng / ml HGF: R&D systems, cat.nr.396-HB and 236-EG). N87 cells are trypsinized, harvested and counted, and then 8000 cells in 100μl CDS are added to each well of the plate. To avoid edge effects, the plates were placed at RT for one hour and then placed in a container in a 37°C cell culture incubator for three days. On the fourth day, Alamar Blue (Invitrogen, #DAL1100) was added (20 μl per well) and fluorescence was measured using 560 nm excitation and 590 nm readout on a Biotek Synergy 2 multi-mode microplate reader after 6 hours of incubation with Alamar Blue (at 37°C). Fluorescence values ​​were normalized to uninhibited growth (no antibody, but with the addition of both ligands). An example of HGF, EGF, and EGF / HGF proliferation assays is shown in Figure 14 (respectively Fig.14A , B and C).

[0275] The results of various experiments are listed in Table 4. In the N87 HGF / EGF assay, 14 different cMETxEGFR bispecific antibodies were identified with comparable potency to the reference monospecific antibody (an equimolar mixture of cetuximab and 5D5 Fab): PB7679, PB7686, PB8218, PB8244, PB8292, PB8316, PB8340, PB8364, PB8388, PB8511, PB8535, PB8583, PB8607, and PB8640.

[0276] In the N87 EGF assay, 11 different cMETxEGFR bispecific antibodies were identified with comparable potency to monospecific cetuximab: PB7679, PB8244, PB8292, PB8340, PB8364, PB8388, PB8511, PB8535, PB8583, PB8607, and PB8640. They all contain the EGFR Fab arm MF3755. In the HGF N87 assay, 9 bispecific antibodies were identified that showed higher potency compared to the monospecific 5D5 Fab reference antibody: PB8218, PB8388, PB8511, PB8532, PB8535, PB8545, PB8583, PB8639, and PB8640. They contain six different cMET Fab arms: MF4040, MF4297, MF4301, MF4356, MF4491 and MF4506.

[0277] ADCC activity

[0278] 24 cMetxEGFR bispecific antibodies were tested for ADCC activity against tumor cell lines N87 (high EGFR, low cMET) and MKN-45 (low EGFR, amplified cMET). ADCC assays were performed in 384-well plate format using the Promega ADCC bioassay kit. Antibodies were tested in duplicate at 9 different concentrations in half-log serial dilutions (10 μg / ml to 1 ng / ml).

[0279] A reference cetuximab antibody was included as a positive control for the assay, and PG2708 was used as a negative control antibody. Antibodies or assay medium controls (no IgG) were incubated with ADCC effector cells and target cells (N87 or MKN-45) for 6 hours at 37°C. Luciferase activity was quantified using the Bio-Glo luciferase reagent.

[0280] An example of ADCC assay is shown in Figure 15. The cMETxEGFR bispecific antibody did not show significant ADCC activity in both cell lines. The positive control reference cetuximab antibody showed dose-dependent ADCC activity on both cell lines.

[0281] Five bispecific antibodies consisting of EGFR and cMet arms that did show high potency in the N87 HGF / EGF assay and showed high sequence diversity were selected for further analysis (Table 5). Two of the five bispecific antibodies included MF4356, which competed with 5D5 for binding to cMET (Table 2). Table 5 summarizes the characteristics of the selected candidates.

[0282] Wound healing cell migration assay

[0283] Two NSCLC cell lines were tested in the wound healing assay: EBC-1 and H358. These cell lines were selected because they express high levels of EGFR and c-Met (Zhang et al, 2010; Fong et al, 2013). The assay was performed using the CytoSelectTM 24-well plate wound healing assay (Cell Biolabs, CBA-120) according to the manufacturer's instructions. Briefly, 2.5×10 5 Up to 4×10 5 Cancer cells were seeded in each well and incubated overnight at 37°C to form a monolayer. The well insert was then removed to produce a 0.9 mm wound area. After washing with PBS to remove dead cells and debris in the wound area, the cells were incubated with complete medium (0.5% FBS) containing bispecific antibodies (100 nM) or cetuximab: Fab2994 control antibody mixture (100 nM, 1: 1 molar ratio) at 37°C for 15 minutes. Each well was then supplemented with growth factors: HGF (15 ng / ml), EGF (12.5 ng / ml) or a combination of the two (15 and 12.5 ng / ml). Wound closure was monitored over time at 37°C using a confocal microscope (Zeiss LSM780) for 14 h. The degree of wound closure (%) is shown relative to the untreated control.

[0284] H358 cells showed increased migration (percent wound closure) after exposure to HGF or EGF alone, which was most effective with the combination of HGF and EGF ( Figure 3 ). This increased migration was abolished by addition of most bispecific antibodies and was most pronounced with PB8532. Except for the inhibition of migration in the presence of EGF / HGF, this inhibition was comparable to the combination of cetuximab and 5D5 Fab.

[0285] EBC-1 cells showed a slight increase in migration by the addition of HGF, whereas migration was not increased in the presence of EGF or the EGF / HGF combination. However, wound closure could be inhibited by the bispecific antibodies, particularly by PB8532, under all assay conditions. PB8532 was as effective or more effective than the combination of cetuximab / 5D5 Fab (EGF / HGF).

[0286] Analysis of EGFR and cMET expression on PC-9 and HCC827 cells by flow cytometry

[0287] Acquired resistance to erlotinib may be caused by abnormal activation of HGF-mediated c-MET activation. NSCLC cell lines PC-9 and HCC827 were selected to study the ability of cMET xEGFR bispecific antibodies to inhibit ligand-mediated proliferation in a tyrosine kinase inhibitor (TKI)-resistant environment. Both cell lines do not have EGFR mutations and are resistant to both erlotinib and gefitinib or the combination of gefitinib and erlotinib in the presence of HGF (PC-9 only). In PC-9 cells, it has been reported that HGF-induced erlotinib resistance can be eliminated by cMET inhibitors (Nakade et al., 2014), and gefitinib resistance can be eliminated by anti-HGF antibodies (Yano, 2008). The expression of cMET and EGFR of PC-9 and HCC827 was characterized by FACS analysis using fluorescently labeled antibodies. Cells were harvested with PBS 2mM EDTA. Single cell suspensions (10e6 cells in 50 μl) were incubated with fluorescently labeled antibodies in staining buffer (PBS 2% FBS 2mM EDTA) on ice for 20 minutes. The following antibodies were used alone or in combination: Met Alexa Fluor 488 conjugate (clone D1C2, Cell Signaling, 1:50 dilution); EGF receptor Alexa Fluor 647 conjugate (clone D38B1, Cell Signaling, 1:50 dilution). After incubation, cells were washed with staining buffer and FACS analysis was performed on a BD FACSVerse flow cytometer.

[0288] All HCC827 cells showed EGFR expression and could be subdivided into EGFR high cMET pos EGFR pos cMET neg group( Figure 4 ). PC-9 cells contain a small population of EGFR high and cMET pos cells, and the smallest group of EGFR pos and cMET neg cell.

[0289] PC-9 and HCC827 proliferation assays

[0290] Initial experiments were performed to determine the HGF concentrations that established resistance to erlotinib and gefitinib in PC-9 and HCC827 cells. After overnight starvation in a culture medium containing 0.5% FBS, cells were incubated with an increased concentration of HGF (0 to 120 ng / mL) supplemented with 300 nM erlotinib or gefitinib in 10% FBS. After incubation for 72 hours, WST-1 reagent was used to assess cell proliferation according to the manufacturer's instructions. The absorbance was measured at 450 and 630 nm, respectively, at the test wavelength and reference wavelength, using a microplate reader. In PC-9 ( Fig.16A ) and HCC287( Fig. 16B ) In both studies, the addition of HGF induced resistance to TKI in a dose-dependent manner.

[0291] PB8532 and PB8388 were tested for their efficacy in the setting of TKI resistance. 3 Cancer cells were seeded in 96-well plates containing 100 μL complete RPMI 1640 (10% FBS). After overnight starvation in medium containing 0.5% FBS, the cells were incubated with (100nM) or cetuximab:Fab2994 control monospecific antibody mixture (100nM, 1: 1 molar ratio) was pre-incubated at 37°C for 15 minutes. Then, complete medium (10% FBS) containing gefitinib or erlotinib (300nM) or w / o HGF (30ng / ml), EGF (30ng / ml) or a combination of the two (at 30ng / ml) was supplemented in each well. After incubation for 72 hours, cell proliferation was assessed using the WST-1 method.

[0292] Figure 5 It was shown that PB8532 can inhibit HGF-mediated and EGF-mediated gefitinib resistance in PC-9 cells. In HCC827 cells, PB8532 inhibited HGF-mediated TKI resistance and was more potent than the combination of dual monospecific cetuximab 5D5Fab. Comparable results were obtained using the TKI erlotinib.

[0293] Example 4 Inhibition of EGF and cMET phosphorylation by PB8532

[0294] After overnight starvation in FBS-free medium, cells were incubated with medium (0.5% FBS) containing PB8532 (100 nM) or cetuximab:Fab2994 control monospecific antibody mixture (100 nM, 1: 1 molar ratio) at 37 ° C for 15 minutes. The cells were then stimulated with the following growth factors at 37 ° C for 15 minutes: HGF (30 ng / ml) or EGF (50 ng / ml). After stimulation, cells were washed with PBS in the presence of 1 mM orthovanadate (Sigma-Aldrich). Protein extraction was performed using RIPA lysis buffer (50 mM Tris HCl pH 8, 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS) supplemented with a complete protease inhibitor cocktail (Roche), PhosSTOP phosphatase inhibitor (Roche) and orthovanadate (1 mM, Sigma-Aldrich). The lysate was incubated on ice for 30 minutes and then centrifuged at 4°C for 15 minutes to remove cell debris. After centrifugation, the supernatant was collected and the protein concentration was determined using bicinchoninic acid (BCA) reagent (Pierce) according to the manufacturer's instructions. The protein samples were denatured by adding loading buffer 6X (β-mercaptoethanol 0.6M; SDS 8%; Tris-HCl 0.25M pH 6,8; glycerol 40%; bromophenol blue 0.2%) and incubated at 95°C for 5 minutes. After electrophoresis, the proteins were washed with Turbo TMThe membrane was transferred to a nitrocellulose membrane using a blotting system (Bio-Rad). The membrane was blocked for nonspecific binding in 5% non-fat dry milk in Tris-buffered saline-Tween 0.1% (50 mM Tris HCl ph 7.6, 150 mM NaCl, 0.1% Tween; TBS-T) for 1 h at room temperature (RT) and incubated with the primary antibody overnight at 4°C (ON). The following primary antibodies were used: Phospho-Met (Tyr1234 / 1235, clone D26, Cell Signaling) at 1:500 in TBS-T 5% BSA; Met (clone D1C2, Cell Signaling) at 1:1000 in TBS-T 5% BSA; Phospho-EGF receptor (Tyr1068, clone D7A5, Cell Signaling) at 1:1000 in TBS-T 5% non-fat dry milk; EGF receptor (clone D38B1, Cell Signaling) at 1:1000 in TBS-T 5% BSA; Vinculin (monoclonal anti-vinculin, V9131, SIGMA Aldrich) at 1:4000 in TBS-T 5% non-fat dry milk. After incubation with the designated primary antibodies, the membranes were washed in TBS-T for 15 minutes and then incubated with secondary antibodies (1:5000 in TBS-T 5% non-fat dry milk) at RT for 1H. The following secondary antibodies were used: goat anti-rabbit IgG-HRP (sc-2004, Santa Cruz biotechnology); goat anti-mouse IgG-HRP (sc-2005, Santa Cruz biotechnology). Signals were visualized using enhanced chemiluminescence reagent (ECL; Invitrogen) or SuperSignal West Femto chemiluminescent substrate (Thermo Scientific) with a digital imager (ImageQuant LAS 4000, GE Health Care Life Science Technologies).

[0295] Figure 6 Western blot analysis of the performed experiments is shown.

[0296] In PC-9 cells, PB8532 and 5D5 / cetuximab were able to reduce HGF-induced phosphorylation. In addition, both antibodies slightly reduced EGF phosphorylation in the absence and presence of EGF.

[0297] In HCC827 cells, PB8532 reduced phosphorylation of cMET in the presence and absence of HGF. No effect of the combination of 5D5 / cetuximab was observed. In addition, in contrast to the combination of 5D5 and cetuximab, PB8532 reduced phosphorylation of EGFR induced by EGF in this cell line.

[0298] Example 5

[0299] Figure 7 Various sequences of alternative variable regions of the heavy chain of the EGFR binding variable domain as disclosed herein are depicted. Figure 8 Various sequences of alternative variable regions of the heavy chain of the cMET binding variable domain as disclosed herein are depicted. The heavy chain variable region was used to generate a number of different cMET x EGFR bispecific antibodies. The light chains in these antibodies have Fig. 9B The bispecific antibodies were generated as described in Example 1. The antibodies were also generated as ADCC-enhanced forms. The ADCC-enhanced form was generated by co-transfection of an antibody construct containing DNA encoding a reductase that removes fucose residues from the Fc region of IgG1.

[0300] FIG. 17 depicts the titration of various bispecific antibodies produced on CHO-K1 EGFR cells (Panel A) and on MKN-45 cells endogenously expressing c-MET as described in Example 2 (Panel B). Cells were cultured at 2×10 5 Cells / well were incubated with increasing concentrations of antibodies as specified. After washing, binding was detected with anti-human IgG-PE (3 μg / ml). The stained cells were analyzed on the iQue system and the mean fluorescence intensity (MFI) and area under the curve (AUC) were calculated. Control antibodies were MF1337xMF1337 (PG1337p218; TTxTT negative control; black triangles at the bottom) and MF4356xMF3770 (PB8532p04; c-METxEGFR positive control; black triangles). TT stands for tetanus toxoid and contains the VH of MF1337 as described herein (see Figure 1 ) and the variable domain of common light chain, bind to tetanus toxoid, and therefore it is not expected to bind to CHO-K1 EGFR cells and MKN-45 cells. Note (ADCC) represents the antibody with enhanced ADCC function by co-transfection with the DNA encoding the RMD enzyme that removes fucose residues from the Fc region of IgG1. For the list of bispecific antibodies used and their PB numbers, see Table 6.

[0301] ADCC reporter assay

[0302] ADCC reporter assay is carried out to determine whether the cotransfection of the DNA encoding RMD has successfully enhanced ADCC effector function. Both high-affinity and low-affinity assays are used, and all samples are tested in duplicate for BxPC3 cells (which express EGFR) and MKN-45 cells (which express c-MET). The high-affinity effector cells of the assay express the V variant of human FcγRIIIa, while the low-affinity effector cells express the F variant.

[0303] In brief, harvest BxPC3 and MKN-45 target cells, and 1000 cells / well in 30 μ L are seeded and incubated overnight at 95% relative humidity, 5% CO2, 37 ° C. The next day, the culture medium is removed, and 10 μ L of antibody diluent is added to each well (antibody dilution 1.5x; 9-step serial titration with half-logarithmic dilution step produces assay concentration of 1 ng / ml to 10 μ g / ml). On the same day, the effector cells are thawed at 37 ° C, and then 630 μ L is added to 3.6 ml of assay buffer in a 15 ml tube, and mixed by inversion; then 5 μ L of this solution (15,000 cells) is added to the wells of the assay plate. The plate is incubated at 95% relative humidity, 5% CO2, 37 ° C for 6 hours, and then 15 μ L of Bio-Glo reagent is added to the assay wells. Luminescence is measured using an EnVision plate reader.

[0304] A list of the samples tested is provided in Table 6, and Fig.18 The results of the assays are provided. A non-ADCC enhanced anti-HER3 x EGFR control antibody (lot PB4522p25; MF4280 x MF3178 described in WO2015 / 130172) was negative in all four assays ( Fig.18 In contrast, the ADCC-enhanced version of the antibody (PB4522p34) was positive in all four assays (indicated by the solid orange circles). Similarly, the non-ADCC-enhanced anti-c-MET x EGFR control antibody (PB8532p04) was negative in all four assays (indicated by the solid orange circles). Fig.18 ) and the PB8532p05 batch is also non-ADCC-enhanced (indicated by a green asterisk). The three lines with asterisks are all at the bottom of the four figures. However, the ADCC-enhanced p06 variant (PB8532p06) is positive in all assays (indicated by a hollow orange circle). For the five bispecific antibodies (PB19474 to PB19478), enhanced ADCC effector functions similar to those of PB8532p06 were also found. This means that the co-transfection of the DNA encoding the RMD successfully enhanced the ADCC effector function.

[0305] Example 6

[0306] The heavy chain variable region (VH) of the cMET variable domain of PB8532 comprises, for example Figure 8 The amino acids of MF4356 are shown in . The VH of the cMET variable domain of PB19748 comprises the amino acid sequence of MF8230 (see Figure 8 ). The VH of the EGFR variable domain of PB8532 comprises, for example Figure 7 The amino acids of MF3370 are shown in FIG. The VH of the EGFR variable domain of PB19748 comprises Figure 7 The amino acid sequence of MF8233 of PB8532 and PB19748 is the same and is shown in Fig. 9B . The cMET antibody LY2875358 antibody was described in particular in Kim and Kim 2017. The ability of the cMETxEGFR bispecific antibodies PB8532 or PB19748 to inhibit tumor growth in vivo, alone or in combination with the receptor tyrosine kinase inhibitor erlotinib, was tested in a xenograft mouse model. In the chosen model, HCC827 tumor cells were transplanted into immunodeficient NOD SCID gamma (NOD SCID gamma, NSG) human hepatocyte growth factor knock-in (hHGFki) mice, which express human HGF (the ligand of cMET) instead of endogenous mouse HGF. The full name of NSG-hHGFki mice is NOD.Cg-Hgf tm1.1(HGF)Aveo Prk dcscid Il2rgtm1Wjl / J (stk#014553)(NOD.Cg-Hgftm1.1(HGF)Aveo Prkdcscid Il2rgtm1Wjl / J). It has no T or B cells, lacks functional NK cells, and has deficient cytokine signaling, which allows for better tumor engraftment. HCC827 is an established human non-small-cell lung carcinoma (NSCLC) cell line expressing EGFR and cMET and is known to be resistant to erlotinib in the presence of HGF.

[0307] In order to establish the effect of erlotinib on tumor growth in this model, the first experiment was conducted on two groups of mice. Before tumor cell transplantation, the cell cycle of HCC827 cells was enhanced by culturing cells overnight in a medium supplemented with 20% fetal bovine serum (FBS) when the confluence did not exceed 80%. The next day, NSG-hHGFki mice (The Jackson Laboratorian) were subcutaneously inoculated with 17×10e6 HCC827 tumor cells suspended in 300 μl PBS plus 30% matrix gel containing a high concentration of basement membrane matrix. The tumors produced were measured twice a week using a caliper. When the average tumor volume reached about 200mm 3 At 4 hr, mice were randomly divided into two groups (4 to 7 mice per group, depending on the growth and volume of the tumor) and drug treatment was started.

[0308] A fine suspension of erlotinib was prepared freshly weekly in 0.05% hydroxypropyl methylcellulose (HPMC) and 0.2% Tween 80 in water by sonication.

[0309] From day 19, 5 mice (n=5) were treated with erlotinib solution at a dose of 6 mg / kg by gavage once a day (QD), and groups of 4 mice were gavaged once a day with 200 μl of vehicle (0.05% HPMC and 0.1% Tween 80 in water). Tumor volume was measured twice a week using a caliper, and the mean tumor volume (and SEM) was calculated for each group. When the tumor size reached 1500 mm 3 The mice were euthanized at 4:40 p.m. after day 48, and the tumor volumes in the surviving mice were measured until day 62.

[0310] In a second experiment to test the ability of the PB8532 bispecific antibody (alone and in combination with erlotinib) to inhibit tumor growth in this model, NSG-hHGFki mice with tumors (generated as described above) were given one of the six treatments or a combination of treatments, whereby the antibody was given weekly by intraperitoneal (ip) injection, and erlotinib or vehicle was given once a day (QD) by gavage. As a negative control, mice were also treated with PB17160 (a bispecific antibody consisting of the same anti-cMET Fab arms as in PB8532 combined with Fab arms specific for an irrelevant target). An irrelevant target is, for example, a target that is not present in mice and tumors. Variable domains specific to tetanus toxoid are typically used. Suitable tetanus toxoid binding variable domains have the VH of MF1337 as disclosed herein (see Figure 1) and a common light chain, preferably the sequence of Figure 9. Referring to MF1337, bispecific antibodies with a targeting arm and a non-targeting (TT) arm are described in particular in WO2017 / 069628, which is incorporated herein by reference. Another unrelated target is RSV-G. Suitable RSV-G variable domains have Figure 1 The VH of MF2708 and preferably one of FIG. 9 (preferably Fig. 9B )'s common light chain.

[0311] On day 21, when the average tumor volume reached approximately 200 mm 3 At 14:00, mice were randomly divided into 6 groups (5 to 7 mice per group, depending on the growth and volume of the tumor) and drug treatment was started: daily gavage with vehicle alone (n=5); weekly ip injection of 25 mg / kg PB8532 antibody plus daily gavage with vehicle (n=7); weekly ip injection of 25 mg / kg PB17160 antibody plus daily oral gavage with vehicle (n=6); daily oral gavage with 6 mg / kg erlotinib (n=7); weekly ip injection of 25 mg / kg PB8532 antibody plus daily oral gavage with 6 mg / kg erlotinib (n=7); or weekly ip injection of 25 mg / kg PB17160 antibody plus daily oral gavage with 6 mg / kg erlotinib (n=7). As described above, tumor volume was monitored and the average tumor volume (and SEM) in each group was calculated. All treatments were stopped after day 60 until the tumor volume in surviving mice was measured on day 82.

[0312] In the third experiment, the ability of PB19478 bispecific antibody (alone and in combination with erlotinib) to inhibit tumor growth in this model was tested. NSG-hHGFki mice bearing tumors (generated as described above) were given one of the six treatments or a combination of treatments, whereby the antibody was given weekly by intraperitoneal (ip) injection, and erlotinib or vehicle was given once a day (QD) by gavage.

[0313] On day 23, when the average tumor volume reached approximately 200 mm 3At 14:00, mice were randomly divided into 6 groups (5 to 6 mice per group, depending on tumor growth and volume) and drug treatment was started: daily gavage with vehicle alone (n=5); weekly ip injection of 25 mg / kg PB19478 antibody plus daily gavage with vehicle (n=5); daily oral gavage with 6 mg / kg erlotinib (n=6); weekly ip injection of 25 mg / kg PB19478 antibody plus daily oral gavage with 6 mg / kg erlotinib (n=4, one mouse died during the experiment); weekly ip injection of 25 mg / kg LY2875358 antibody plus daily gavage with vehicle (n=5); weekly ip injection of 25 mg / kg LY2875358 antibody plus daily oral gavage with 6 mg / kg erlotinib (n=3, two mice died during the experiment). Tumor volume was monitored and the mean tumor volume (and SEM) in each group was calculated as described above. All treatments were stopped after day 93, and tumor volumes were measured in surviving mice until day 93.

[0314] In the fourth experiment, the effect of PB19478 administered later was tested. NSG-hHGFki mice bearing tumors (generated as described above) were given one of two treatments, whereby the antibody was given weekly by intraperitoneal (ip) injection and erlotinib or vehicle was given once a day (QD) by gavage.

[0315] On day 21, when the average tumor volume reached approximately 200 mm 3 All 14 mice were started on day 51 with 6 mg / kg erlotinib by oral gavage. 3 At the time of labeling, mice were randomly divided into two groups. One of the six groups was treated with 6 mg / kg erlotinib per oral gavage every day, while the other group of 8 received weekly ip injections of 25 mg / kg PB19478 antibody plus gavage with 6 mg / kg erlotinib every day. As previously described, tumor volume was monitored and the average tumor volume (and SEM) in each group was calculated. All treatments were stopped after the 72nd day.

[0316] The results of the first experiment showed that erlotinib was able to induce an antitumor response in NGS-hHGFki mice transplanted with HCC827 cells, but only for as long as the mice were treated ( Fig.19 ). During the drug-free period that began after approximately 4 weeks of treatment, tumor volume increased significantly in mice treated with erlotinib.

[0317] The anti-cMETxEGFR bispecific antibody PB8532 was also able to induce anti-tumor responses in NGS-hHGFki mice transplanted with HCC827 cells ( Fig. 20). The effect was greater when the antibody was given in combination with a daily dose of erlotinib. Within 2.5 weeks, all tumors disappeared from the combination of PB8532 and erlotinib. The control antibody PB17160, which has one Fab arm targeting cMet, did not induce an anti-tumor response, whether combined with erlotinib or not. Therefore, HGF-mediated erlotinib resistance can be overcome by specifically targeting cMet Fab arm in combination with EGFR targeting Fab arm in bispecific antibody PB8532.

[0318] At about 5 1 During the drug-free period starting after 2 weeks of treatment, PB8532 was more effective than erlotinib in reducing tumor volume ( Fig.21 ), and no tumor regrowth was observed in the PB8532+erlotinib combination group.

[0319] The anti-cMETxEGFR bispecific antibody PB19478 was also able to induce anti-tumor responses in NGS-hHGFki mice transplanted with HCC827 cells ( Fig. 22 ). The effect was greater when the antibody was given in combination with a daily dose of erlotinib. Within 2 weeks, all tumors disappeared from PB19478, whether combined with erlotinib or not. In combination with erlotinib, tumors did not reappear during the test period. Without erlotinib, tumors reappeared quickly at about day 50 and remained at detectable levels until they finally began to grow further at day 80. The humanized monoclonal antibody emibetuzumab (LY3875358) was also less effective when combined with the EGFR inhibitor erlotinib. Therefore, HGF-mediated erlotinib resistance can be overcome by specifically targeting cMet Fab arms in combination with the EGFR targeting Fab arms in the bispecific antibodies PB8532 or PB19478.

[0320] Fig.23 It shows that when you treat the tumor at a time point where erlotinib resistance starts to develop, the bispecific antibody PB19478 works immediately.

[0321] In summary, data from the HCC827 tumor cell xenograft model transplanted into immunodeficient NSG-hHGFki mice indicate that PB8532, PB19478, and Figure 7 and Figure 8 Antibodies with similar VH sequences given in have the ability to overcome HGF-mediated erlotinib resistance in vivo. Combination therapy continues to be effective after cessation of treatment.

[0322] References

[0323] Ahmed,M et al.Lack of in Vivo Antibody Dependent CellularCytotoxicity with Antibody containing gold particles / Bioconjuaate chemistry(2015):26 812-816.DOI:10.1021 / acs.bioconjchem.5b00139.

[0324] Buday L,Downward J.Epidermal growth factor regulates p21ras throughthe formation of a complex of receptor,Grb2 adapter protein,and Sosnucleotide exchange factor,Cell 1993;73:611-20.

[0325] Castoldi,R.,et al.″A novel bispecific EGFR / Met antibody blocks tumor-promoting phenotypic effects induced by resistance to EGFR inhibition and haspotent antitumor activity,”Oncogene 32,50(2013):5593-5601.

[0326] Chen et al.,MET activation mediates resistance to lapatinibinhibition of HER2-amplified gastric cancer cells.Mol Cancer Ther 2012;11(3):650-669

[0327] Ferguson KM.Structure-based view of epidermal growth factor receptorregulation.Annu Rev Biophys 2008;37:353-73.

[0328] Fong JT,etal.,Alternative signaling pathways as potential therapeutictargets for overcoming EGFR and c-Met inhibitor resistance in non-small celllung cancer.PLoS One.2013 Nov 4;8(11):e78398.

[0329] Gale NW,Kaplan S,Lowenstein EJ,Schlessinger J,Bar-Sagi D,Grb2mediates the EGF-dependent activation of guanine nueleotide exchange onRas.Nature 1993;363:88-92.

[0330] Gulli LF1,Palmer KC,Chen YQ,Reddy KB,1996.Epidermal growth factor-induced apoptosis in A431 cells can be reversed by reducing the tyrosinekinase activity.Cell Growth Differ.Feb;7(2):173-8.

[0331] Jorissen RN,Walker F,Pouliot N,Garrett TP,Ward CW,BurgessAW.Epidermal growth factor receptor:mechanisms of activation andsignalling.Exp Cell Res 2003;284:31-53.

[0332] Kim,George P.,and Axel Grothey。″Targeting coloreetal caneer withhuman anti-EGFR monoclonal antibodies:focus on panitumumab.“Biologics 2.2(2008):223-228.

[0333] Kim,Ki-Hyan and Kim,Hyori.Progress of antibody-based inhibitors ofthe HGF-cMET axis in cancer therapy.Experimental &Molecular medicine(2017),e307;doi:10.1038 / emm.2017.17)

[0334] Morgillo F,et al.ESMO Open 201.6;1:e000060.doi:10.1136 / esmoopen-2016-000060

[0335] Moores,Sheri L.,et al″A novel bispecific antibody targeting EGFR andcMet is effective against EGFR inhibitor-resistant lung tumors.″Cancerresearch 76.13(2016):3942-3953.

[0336] Nakade et al.Triple inhibition of EGFR,Met,and VEGF Suppressesregrowth of HGF-triggered,erlotinib-resistant lung cancer harboring an EGFRmutation.J Thorac Oncol 2014,9(6):775-83.

[0337] Prigent SA,Gullick WJ.Identification of eErbB-3 binding sites forphosphatidylinositol 30-kinase and SHC using an EGF reeeptor / c-ErbB-3chimera.EMBO J 1994;13:2831-41.

[0338] Robertson SC,Tynan J,Donoghue DJ.RTK mutations and human syndromes:when good receptors turn bad.Trends Genet 2000;16:368.

[0339] Soltoff SP,Carraway KL,III,Prigent SA,Gullick WG, Cantley LC.ErbB3 ismvolved in activation of phosphatidylinositol 3-kinase by epidermal growthfactor.Mol Cell Biol 1994;14:3550-8.

[0340] Yano et al.Hepatocyte growth factor induces gefitinib resistance oflung adenocarcinoma with epidermal growth factor receptor-activatingmutations.Cancrr Res.2008,15;68(22):9479-87.

[0341] Yarden Y.The EGFR family and its ligands in human cancer.Signallingmechanisms and therapeutic opportunities.Eur J Cancer 2001;37(Suppl 4):S3-S8.

[0342] Zhang et al,MET kinase inhibitor SGX523 synergizes with epidermalgrowth factor receptor inhibitor erlotinib in a hepatocyte growth factor-dependent fashion to suppress carcinoma growth.Cancer Res 2010,70:6880-6890.

[0343] US20120107234A1_Symphogen patent “PAN HER antibody composition”.

[0344] 125084 Erbitux Pharmacology Review Part 2 - FDA.

[0345]

[0346] Table 1. Reference antibodies with reported specificities against the cMET extracellular domain.

[0347]

[0348] Table 2. Competition of cMet reference antibodies with cMET cLC antibodies. Shown are OD450 values. OD450 values ​​indicate the presence or lack of competition with the antibody. MF4506 was not tested.

[0349]

[0350] Table 3: List of 24 cMETxEGFR bispecific antibodies selected after dose-dependent titration experiments in N87 HGF / EGF proliferation assay. The MF numbers of the EGFR and cMET arms in each individual PB and their HCDR3 sequences are indicated.

[0351]

[0352] Table 4: Overview of antibody titration experiments performed using 24 cMETxEGFR bispecific antibodies for N87 HGF / EGF, HGF and EGF proliferation assays. Bispecific antibodies are denoted as PBXXXX and the different Fab arms are denoted as MGXXXX. The activity of the bispecific antibodies in the individual assays is denoted as: - no effect; + less inhibition of proliferation than the positive control; ++ = comparable inhibition of proliferation to the positive control antibody 5D5 Fab; +++ = greater inhibition of proliferation than the positive control antibody 5D5 Fab.

[0353]

[0354] Table 5: Composition of the most potent EGFRxcMET bispecific antibodies and their competition with reference antibodies.

[0355] Bispecific Antibodies cMET Arm EGFR arm ADCC enhancement Cetuximab - - - PB8532p05 MF4356 MF3370 no PB19474p01 MF4356 MF8232 yes PB19475p01 MF4356 MF8233 yes PB19476p01 MF8230 MF3370 yes PB19477p01 MF8230 MF8232 yes PB19478p01 MF8230 MF8233 yes PB8532p06 MF4356 MF3370 yes PB8532p04 MF4356 MF3370 no HER-3 arm EGFR arm PB4522p34 MF3178 MF4280 yes PB4522p25 MF3178 MF4280 no TT arm TT arm PG1337p218 MF1337 MF1337 no

[0356] Table 6: Composition of bispecific antibodies. pXX numbers indicate the number of the production run and can be used to identify whether the antibody was produced in an ADCC format.

Claims

1. A bispecific antibody comprising a first variable domain that can bind to the extracellular portion of human epidermal growth factor receptor (EGFR) and a second variable domain that can bind to the extracellular portion of human MET proto-oncogene receptor tyrosine kinase (cMET). The bispecific antibody of claim 1 , comprising a common light chain.

3. The bispecific antibody of claim 1 or claim 2, which is a human antibody. The bispecific antibody according to any one of claims 1 to 3, which is a full-length antibody. The bispecific antibody according to any one of claims 1 to 4, which is an IgG1 antibody with an anti-EGFR and anti-cMET stoichiometry of 1:

1. 6 . The bispecific antibody of claim 1 , which has one variable domain that can bind to EGFR and one variable domain that can bind to cMET.

7. The bispecific antibody of any one of claims 1 to 6, wherein the variable domain that can bind to human EGFR can also bind to cynomolgus monkey and mouse EGFR.

8. The bispecific antibody of any one of claims 1 to 7, wherein the variable domain that can bind to human EGFR binds to domain III of human EGFR. 9 . The bispecific antibody of any one of claims 1 to 8 , wherein the variable domain that can bind to cMET blocks the binding of antibody 5D5 to cMET.

10. The bispecific antibody of any one of claims 1 to 9, wherein the variable domain that can bind to cMET blocks the binding of HGF to cMET.

11. The bispecific antibody of any one of claims 1 to 10, wherein the amino acids at positions 405 and 409 in one CH3 domain are identical to the amino acids at the corresponding positions in the other CH3 domain (EU numbering).

12. The bispecific antibody of any one of claims 1 to 11, wherein the first variable domain comprises a heavy chain variable region having a CDR1 sequence of SYGIS, a CDR2 sequence of WISAYX1X2NTNYAQKLQG, and a CDR3 comprising the sequence X3X4X5X6HWWLX7A, wherein X1=N or S; X2=A or G; X3=D or G; X4=R, S or Y; X5=H, L or Y; X6=D or W and X7=D or G; there are 0 to 5 amino acid insertions, deletions, substitutions, additions or a combination thereof at positions other than X1 to X7, and wherein the second variable domain comprises a heavy chain variable region having an amino acid sequence of one of the sequences of SEQ ID NOs: 1 to 23 with 0 to 10, preferably 0 to 5, amino acid insertions, deletions, substitutions, additions or a combination thereof.

13. The bispecific antibody of claim 12, wherein X1=N; X2=G; X3=D; X4=S; X5=Y; X6=W and X7=G; X1=N; X2=A; X3=D; X4=S; X5=Y; X6=W and X7=G; X1=S; X2=G; X3=D; X4=S; X5=Y; X6=W and X7=G; X1=N; X2=G; X3=D; X4=R; X5=H; X6=W and X7=D; X1=N; X2=A; X3=D; X4=R; X5=H; X6=W and X7=D; X1=S; X2=G; X3=D; X4=R; X5=H; X6=W and X7=D; X1=N; X2=G; X3=G; X4=Y; X5=L; X6=D and X7=G; X1=N; X2=A; X3=G; X4=Y; X5=L; X6=D and X7=G; or X1=S; X2=G; X3=G; X4=Y; X5=L; X6=D and X7=G.

14. The bispecific antibody of claim 12 or claim 13, wherein X1=N; X2=G; X3=D; X4=R; X5=H; X6=W and X7=D; X1=N; X2=A; X3=D; X4=R; X5=H; X6=W and X7=D; or X1=S; X2=G; X3=D; X4=R; X5=H; X6=W and X7=D.

15. The bispecific antibody of any one of claims 12 to 14, wherein X1=N; X2=G; X3=D; X4=R; X5=H; X6=W and X7=D; or X1=N; X2=A; X3=D; X4=R; X5=H; X6=W and X7=D.

16. The bispecific antibody of any one of claims 1 to 15, wherein the heavy chain variable region of the second variable domain comprises an amino acid sequence of one of the sequences of SEQ ID NO: 1 to 3, 7, 8, 10, 13, 15, 16, 17, 21, 22 or 23 having 0 to 10, preferably 0 to 5 amino acid insertions, deletions, substitutions, additions or a combination thereof.

17. The bispecific antibody of any one of claims 1 to 16, wherein the heavy chain variable region of the second variable domain comprises an amino acid sequence of one of the sequences of SEQ ID NO: 2, 7, 8, 10, 13 or 23 having 0 to 10, preferably 0 to 5, amino acid insertions, deletions, substitutions, additions or a combination thereof.

18. The bispecific antibody of any one of claims 1 to 17, wherein the first variable domain comprises a heavy chain variable region having a CDR1 sequence of SYGIS, a CDR2 sequence of WISAYNGNTNYAQKLQG, and a CDR3 comprising the sequence DRHWHWWLDA, and wherein the second variable domain comprises a heavy chain variable region having a CDR1 sequence of SYSMN, a CDR2 sequence of WINTYTGDPTYAQGFTG, and a CDR3 sequence of ETYYYDRGGYPFDP.

19. The bispecific antibody of any one of claims 1 to 17, wherein the first variable domain comprises a heavy chain variable region having a CDR1 sequence of SYGIS, a CDR2 sequence of WISAYNANTNYAQKLQG, and a CDR3 comprising the sequence DRHWHWWLDA, and wherein the second variable domain comprises a heavy chain variable region having a CDR1 sequence of TYSMN, a CDR2 sequence of WINTYTGDPTYAQGFTG, and a CDR3 comprising the sequence ETYFYDRGGYPFDP.

20. The bispecific antibody of any one of claims 1 to 19, wherein the first and second variable domains comprise a common light chain, preferably the light chain of Figure 9B.

21. The antibody of any one of claims 1 to 20, which inhibits HGF-induced growth of HGF growth-responsive cells.

22. The antibody of any one of claims 1 to 21, which inhibits EGF-induced growth of EGF growth-responsive cells.

23. The antibody of any one of claims 1 to 22, for use in treating a disease involving abnormal cells.

24. The bispecific antibody of any one of claims 1 to 23, for use in treating a subject suffering from a tumor.

25. The bispecific antibody for use according to claim 24, wherein the tumor is an EGFR-positive tumor, a cMET-positive tumor, or an EGFR- and cMET-positive tumor.

26. The bispecific antibody for use according to claim 24 or claim 25, wherein the tumor is breast cancer; colon cancer; pancreatic cancer; gastric cancer; ovarian cancer; colorectal cancer; head and neck cancer; lung cancer, including non-small cell lung cancer; or bladder cancer.

27. The bispecific antibody for use according to any one of claims 24 to 26, wherein the tumor is resistant to treatment with an EGFR tyrosine kinase inhibitor.

28. The bispecific antibody of claim 27, wherein the EGFR tyrosine kinase inhibitor is erlotinib, gefitinib or afatinib; an analog of erlotinib, gefitinib or afatinib; or a combination of one or more of their corresponding compounds and / or analogs.

29. The bispecific antibody for use according to claim 28, wherein the EGFR tyrosine kinase inhibitor is erlotinib.

30. The bispecific antibody for use according to any one of claims 23 to 29, wherein the treatment further comprises treatment with an EGFR tyrosine kinase inhibitor.

31. The bispecific antibody for use according to claim 30, wherein the EGFR tyrosine kinase inhibitor is erlotinib.

32. The bispecific antibody for use according to claim 30 or 31, wherein the bispecific antibody and the EGFR tyrosine kinase inhibitor are administered simultaneously, sequentially or separately.

33. The bispecific antibody of any one of claims 1 to 22, for use in treating a disease involving abnormal cells, wherein the bispecific antibody and an EGFR tyrosine kinase inhibitor are administered simultaneously, sequentially or separately.

34. A method of treating a subject having a tumor, the method comprising administering to an individual in need thereof the bispecific antibody of any one of claims 1 to 22.

35. The method of claim 34, wherein the individual has a disease involving abnormal cells.

36. The method of claim 34 or claim 35, wherein the tumor is an EGFR-positive tumor, a cMET-positive tumor, or an EGFR- and cMET-positive tumor.

37. The method of any one of claims 34 to 36, wherein the tumor is breast cancer; colon cancer; pancreatic cancer; gastric cancer; ovarian cancer; colorectal cancer; head and neck cancer; lung cancer, including non-small cell lung cancer; or bladder cancer.

38. The method of any one of claims 34 to 37, wherein the tumor is resistant to treatment with an EGFR tyrosine kinase inhibitor.

39. The method of claim 38, wherein the EGFR tyrosine kinase inhibitor is erlotinib, gefitinib or afatinib; an analog of erlotinib, gefitinib or afatinib; or a combination of one or more of their corresponding compounds and / or analogs.

40. The method of claim 39, wherein the EGFR tyrosine kinase inhibitor is erlotinib.

41. The method of any one of claims 34 to 40, wherein the treatment further comprises administering to the individual in need thereof an EGFR tyrosine kinase inhibitor.

42. The method of claim 41, wherein the EGFR tyrosine kinase inhibitor is erlotinib.

43. The method of claim 41 or claim 42, wherein the bispecific antibody and the EGFR tyrosine kinase inhibitor are administered simultaneously, sequentially or separately.

44. The bispecific antibody of any one of claims 1 to 22, for use in the preparation of a medicament for treating a disease involving abnormal cells.

45. An antibody comprising a variable domain capable of binding to the extracellular portion of human EGFR, mouse EGFR and cynomolgus monkey EGFR.

Images