Proteins that bind NKG2D, CD16 and CEACAM5

Abstract

Multispecific binding proteins that bind to NKG2D, CD16, and CEACAM5 are described, as well as pharmaceutical compositions and methods of treatment useful for the treatment of cancer.

Description

Technical Field

[0001] The present invention relates to multispecific binding proteins that bind NKG2D, CD16 and CEACAM5. Background Art

[0002] Despite extensive research efforts, cancer remains a significant clinical and financial burden for countries around the world. According to the World Health Organization (WHO), it is the second leading cause of death. Surgery, radiotherapy, chemotherapy, biological therapy, immunotherapy, hormone therapy, stem cell transplantation and precision medicine are among the existing treatment modalities. Despite extensive research in these areas, highly effective curative regimens have not yet been identified, especially for the most aggressive cancers. In addition, many existing anti-cancer treatments have significant side effects.

[0003] Cancer immunotherapy is desirable because they are highly specific and can use the patient's own immune system to promote the destruction of cancer cells. Fusion proteins (such as bispecific T cell engagers) are cancer immunotherapies described in the literature that bind to tumor cells and T cells to promote the destruction of tumor cells. Antibodies that bind to certain tumor-associated antigens have been described in the literature. See, for example, WO 2016 / 134371 and WO 2015 / 095412.

[0004] Natural killer (NK) cells are components of the innate immune system and account for approximately 15% of circulating lymphocytes. NK cells infiltrate almost all tissues and are initially characterized by their ability to effectively kill tumor cells without prior sensitization. Activated NK cells kill target cells in a manner similar to cytotoxic T cells, i.e., by cell lytic granules containing perforins and granzymes and by death receptor pathways. Activated NK cells also secrete inflammatory cytokines, such as IFN-γ and chemokines, which promote the recruitment of other leukocytes to target tissues.

[0005] NK cells respond to signals through a variety of activating receptors and inhibitory receptors on their surface. For example, when NK cells encounter healthy self cells, their activity is inhibited by the activation of killer cell immunoglobulin-like receptors (KIR). Alternatively, when NK cells encounter foreign cells or cancer cells, they are activated through their activating receptors (e.g., NKG2D, NCR, DNAM1). NK cells are also activated by the constant regions of some immunoglobulins through CD16, an Fc receptor (Fcγ receptor III) present on the surface of NK cells. The overall sensitivity of NK cells to activation depends on the sum of stimulatory and inhibitory signals. NKG2D is a type II transmembrane protein expressed by essentially all natural killer cells, and NKG2D acts as an activating receptor on these natural killer cells. NKG2D is also present on T cells as a co-stimulatory receptor. The ability to regulate NK cell function by NKG2D can be used in various therapeutic environments, including malignant tumors.

[0006] Carcinoembryonic antigen-related cell adhesion molecule 5 (CEACAM5), also known as meconium antigen 100, CEA, carcinoembryonic antigen, CD66e or CD66e antigen, is a member of the immunoglobulin superfamily. It is a large cell surface glycoprotein and acts primarily as a cell adhesion molecule that mediates cell-cell contacts. In addition to its functions in cell adhesion and migration, CEACAM5 has been found to be overexpressed in a high percentage of human cancers, including 90% of gastrointestinal, colorectal and pancreatic cancers, 70% of non-small cell lung cancer cells and 50% of breast cancers. Overexpression of CEACAM5 has been shown to be positively correlated with tumorigenicity and enhanced tumor invasiveness.

[0007] Although proteins (e.g., antibodies) that bind to CEACAM5 are under development as potential anti-cancer therapeutics, they face significant challenges. Some of these challenges, such as the high level of homology with other CEACAM family members, the low percentage homology with cynomolgus monkey (cyno) CEACAM5, and the highly glycosylated nature of the protein, make it difficult to achieve monospecificity for human CEACAM5 and cross-reactivity with cynomolgus monkey CEACAM5. These challenges highlight the need in the art for new and available antibodies for the treatment of CEACAM5-related cancers. Summary of the invention

[0008] The present invention provides multispecific binding proteins that bind to the NKG2D receptor and CD16 on natural killer cells and the tumor-associated antigen CEACAM5 (carcinoembryonic antigen-related cell adhesion molecule 5). Such proteins can engage more than one NK activating receptor and can block the binding of natural ligands to NKG2D. In certain embodiments, these proteins can agonize NK cells in humans. In some embodiments, these proteins can agonize NK cells in humans and other species (such as rodents and cynomolgus monkeys). The present invention also provides preparations containing any of the proteins described herein; cells containing one or more nucleic acids expressing these proteins, and methods of using these proteins to increase tumor cell death.

[0009] Thus, in one aspect, the present disclosure provides a protein comprising a first antigen binding site that binds NKG2D, a second antigen binding site that binds CEACAM5, and each (i.e., the third antigen binding site, antibody Fc domain, or portion thereof) a third antigen binding site, or antibody Fc domain or portion thereof that binds CD16.

[0010] In some embodiments, the second antigen binding site that binds CEACAM5 comprises a heavy chain variable domain (VH) comprising a complementarity determining region (CDR) 1 (CDRH1), CDRH2 and CDRH3, wherein CDRH1 comprises an amino acid sequence selected from the group consisting of: SEQ ID NOs: 3 and 102, CDRH2 comprises an amino acid sequence selected from the group consisting of: SEQ ID NOs: 37, 104 and 718, and CDRH3 comprises an amino acid sequence selected from the group consisting of: SEQ ID NOs: 6, 38 and 105.

[0011] In some embodiments, the second antigen binding site that binds CEACAM5 comprises a light chain variable domain (VL) comprising CDR 1 (CDL1), CDRL2 and CDRL3, wherein the CDRL1 comprises an amino acid sequence selected from the group consisting of: SEQ ID NOs: 7, 40 and 107, CDRL2 comprises an amino acid sequence selected from the group consisting of: SEQ ID NOs: 8, 41 and 108, and CDRL3 comprises an amino acid sequence selected from the group consisting of: SEQ ID NOs: 9, 42 and 109.

[0012] In some embodiments, the CDRH1, CDRH2 and CDRH3 of the second antigen binding site are: (i) SEQ ID NOs: 3, 37 and 38; (ii) SEQ ID NOs: 3, 718 and 6; or (iii) SEQ ID NOs: 102, 104 and 105, respectively.

[0013] In some embodiments, the CDRL1, CDRL2 and CDRL3 of the second antigen binding site are: (i) SEQ ID NOs: 7, 8 and 9; (ii) SEQ ID NOs: 40, 41 and 42; or (iii) SEQ ID NOs: 107, 108 and 109, respectively.

[0014] In some embodiments, the CDRH1, CDRH2, CDRH3, CDRL1, CDRL2 and CDRL3 are: (i) SEQ ID NO: 3, 37, 38, 40, 41 and 42; (ii) SEQ ID NO: 3, 718, 6, 7, 8 and 9; or (iii) SEQ ID NO: 102, 104, 105, 107, 108 and 109, respectively.

[0015] In some embodiments, the VH comprises an amino acid sequence selected from the group consisting of: SEQ ID NOs: 704, 708, 711, and 715; and the VL comprises an amino acid sequence selected from the group consisting of: SEQ ID NOs: 591, 705, 712, and 716.

[0016] In some embodiments, the VH and VL are: (i) SEQ ID NOs: 704 and 705; (ii) SEQ ID NOs: 708 and 591; (iii) SEQ ID NOs: 711 and 712; or (iv) SEQ ID NOs: 715 and 716, respectively.

[0017] In some embodiments, the second antigen binding site is a single chain variable fragment (scFv), wherein the scFv comprises an amino acid sequence selected from the group consisting of: SEQ ID NO: 703, 707, 710 and 714.

[0018] In some embodiments, the second antigen binding site binds to a human CEACAM5 variant comprising the amino acid sequence of SEQ ID NO:391.

[0019] In some embodiments, the protein comprises an antibody Fc domain or a portion thereof that binds CD16.

[0020] In some embodiments, the first antigen binding site that binds NKG2D is a Fab fragment, and the second antigen binding site that binds CEACAM5 is a scFv. In some embodiments, the first antigen binding site that binds NKG2D is a scFv, and the second antigen binding site that binds CEACAM5 is a Fab fragment.

[0021] In some embodiments, the protein comprises: (i) a first antigen binding site that binds NKG2D, (ii) a second antigen binding site that binds CEACAM5, and (iii) a third antigen binding site that each binds CD16, or an antibody Fc domain or portion thereof, which further comprises an additional antigen binding site that binds CEACAM5. In some embodiments, the first antigen binding site that binds NKG2D is a scFv, and the second antigen binding site and the additional antigen binding site that bind CEACAM5 are each Fab fragments. In some embodiments, the first antigen binding site that binds NKG2D is a scFv, and the second antigen binding site and the additional antigen binding site that binds CEACAM5 are each scFv.

[0022] In some embodiments, the scFv that binds CEACAM5 and / or the scFv that binds NKG2D comprises a heavy chain variable domain and a light chain variable domain.

[0023] In some embodiments, the scFv is connected to an antibody Fc domain or portion thereof that binds CD16 via a hinge comprising Ala-Ser or Gly-Ser. In some embodiments, the hinge further comprises the amino acid sequence Thr-Lys-Gly. In a specific embodiment, the Thr-Lys-Gly is at the N-terminus or C-terminus of the Ala-Ser or Gly-Ser.

[0024] In some embodiments, the heavy chain variable domain of the scFv forms a disulfide bridge with the light chain variable domain of the scFv. In some embodiments, the disulfide bridge is formed between C44 of the heavy chain variable domain and C100 of the light chain variable domain according to the Kabat numbering scheme.

[0025] In some embodiments, the heavy chain variable domain of the scFv is connected to the light chain variable domain of the scFv via a flexible linker. In some embodiments, the flexible linker comprises (G4S)4 (SEQ ID NO: 532).

[0026] In some embodiments, the heavy chain variable domain of the scFv is located at the C-terminus of the light chain variable domain. In some embodiments, the heavy chain variable domain of the scFv is located at the N-terminus of the light chain variable domain. In some embodiments, the Fab is not located between the antigen binding site and the antibody Fc domain or a portion thereof.

[0027] In some embodiments, the first antigen binding site that binds to NKG2D comprises a VH comprising an amino acid sequence at least 90% identical to a VH sequence selected from Table 1, and a VL comprising an amino acid sequence at least 90% identical to a VL sequence selected from Table 1, wherein the VH sequence and the VL sequence selected from Table 1 are from the same clone. In some embodiments, the first antigen binding site that binds to NKG2D comprises: a VH comprising CDR1, CDR2, and CDR3 comprising the amino acid sequences of SEQ ID NOs: 494 or 495, 496, and 524 or 525, respectively; and a VL comprising CDR1, CDR2, and CDR3 comprising the amino acid sequences of SEQ ID NOs: 530, 224, and 499, respectively.

[0028] In some embodiments, the first antigen binding site that binds to NKG2D comprises (i) a VH comprising CDR1, CDR2, and CDR3 comprising the amino acid sequences of SEQ ID NOs: 494 or 495, 496, and 509 or 510, respectively; and a VL comprising CDR1, CDR2, and CDR3 comprising the amino acid sequences of SEQ ID NOs: 530, 224, and 499, respectively. In some embodiments, the first antigen binding site that binds to NKG2D comprises: a VH comprising CDRH1, CDRH2, and CDRH3 comprising the amino acid sequences of SEQ ID NOs: 495, 496, and 510, respectively; and a VL comprising CDRL1, CDRL2, and CDRL3 comprising the amino acid sequences of SEQ ID NOs: 530, 224, and 499, respectively. In some embodiments, the VH of the first antigen binding site comprises an amino acid sequence that is at least 90% identical to SEQ ID NO:508, and the VL of the first antigen binding site comprises an amino acid sequence that is at least 90% identical to SEQ ID NO:493.

[0029] In some embodiments, the VH of the first antigen binding site comprises the amino acid sequence of SEQ ID NO:508, and the VL of the first antigen binding site comprises the amino acid sequence of SEQ ID NO:493.

[0030] In some embodiments, the antibody Fc domain is a human IgG1 antibody Fc domain.In some embodiments, the antibody Fc domain or a portion thereof comprises an amino acid sequence that is at least 90% identical to SEQ ID NO:531.

[0031] In some embodiments, at least one polypeptide chain of the antibody Fc domain or portion thereof comprises one or more mutations at one or more positions selected from the group consisting of: Q347, Y349, L351, S354, E356, E357, K360, Q362, S364, T366, L368, K370, N390, K392, T394, D399, S400, D401, F405, Y407, K409, T411, and K439 relative to SEQ ID NO: 531, numbered by the EU numbering system.

[0032] In some embodiments, at least one polypeptide chain of the antibody Fc domain or portion thereof comprises one or more mutations relative to SEQ ID NO: 531 selected from the group consisting of Q347E, Q347R, Y349S, Y349K, Y349T, Y349D, Y349E, Y349C, L351K, L351D, L351Y, S354C, E356K, E357Q, E357L, E357W, K360E, K360W, Q362E, S364K, S364E, S364H, S364D, T366V, T366I, T366L, T366M, T366K, T366V, 90S, N390D, N390E, K392L, K392M, K392V, K392F, K392D, K392E, T394F, D399R, D399K, D399V, S400K, S400R, D401K, F405A, F405T, Y407A, Y407I, Y407V, K409F, K409W, K409D, T411D, T411E, K439D, and K439E.

[0033] In some embodiments, one polypeptide chain of the antibody Fc domain or portion thereof comprises one or more mutations at one or more positions selected from the group consisting of: Q347, Y349, L351, S354, E356, E357, K360, Q362, S364, T366, L368, K370, K392, T394, D399, S400, D401, F405, Y407, K409, T411, and K439 relative to SEQ ID NO: 531, according to the EU numbering system, and the other polypeptide chain of the antibody Fc domain or portion thereof comprises one or more mutations at one or more positions selected from the group consisting of: NO:531 comprises one or more mutations at one or more positions selected from the group consisting of: Q347, Y349, L351, S354, E356, E357, S364, T366, L368, K370, N390, K392, T394, D399, D401, F405, Y407, K409, T411 and K439.

[0034] In some embodiments, one polypeptide chain of the antibody Fc domain or portion thereof comprises K360E and K409W substitutions relative to SEQ ID NO: 531, and the other polypeptide chain of the antibody Fc domain or portion thereof comprises Q347R, D399V and F405T substitutions relative to SEQ ID NO: 531, according to the EU numbering system. In some embodiments, one polypeptide chain of the antibody Fc domain or portion thereof comprises Y349C substitution relative to SEQ ID NO: 531, and the other polypeptide chain of the antibody Fc domain or portion thereof comprises S354C substitution relative to SEQ ID NO: 531, according to the EU numbering system.

[0035] Another aspect of the present disclosure provides a protein comprising: (a) a first polypeptide comprising the amino acid sequence of SEQ ID NO:549; (b) a second polypeptide comprising the amino acid sequence of SEQ ID NO:550; and (c) a third polypeptide comprising an amino acid sequence selected from the group consisting of: SEQ ID NO:702, 706, 709 and 713.

[0036] Another aspect of the present disclosure provides an isolated nucleic acid molecule or a plurality of isolated nucleic acid molecules encoding any of the disclosed proteins.

[0037] Another aspect of the present disclosure provides an expression vector comprising an isolated nucleic acid molecule or a plurality of isolated nucleic acid molecules encoding any of the disclosed proteins.

[0038] Another aspect of the present disclosure provides a plurality of expression vectors comprising the plurality of isolated nucleic acid molecules.

[0039] Another aspect of the present disclosure provides a host cell comprising one or more expression vectors. In some embodiments, the host cell comprises multiple expression vectors. In some embodiments, the host cell is a Chinese hamster ovary (CHO) cell.

[0040] Another aspect of the present disclosure provides a method for producing a protein comprising: (a) a first antigen binding site that binds to NKG2D; (b) a second antigen binding site that binds to CEACAM5; and (c) a third antigen binding site that binds to CD16, or an antibody Fc domain or a portion thereof; wherein the method comprises: (i) providing a host cell of the present disclosure; (ii) culturing the host cell in a culture medium under conditions suitable for expressing the protein; and (iii) isolating the protein from the culture medium.

[0041] Another aspect of the present disclosure provides a method for producing a protein comprising a first polypeptide, a second polypeptide, and a third polypeptide, wherein the method comprises: (a) providing one or more host cells, wherein the one or more host cells comprise an expression vector or multiple expression vectors comprising: (i) a first isolated nucleic acid molecule encoding a first polypeptide comprising the amino acid sequence of SEQ ID NO: 549; (ii) a second isolated nucleic acid molecule encoding a second polypeptide comprising the amino acid sequence of SEQ ID NO: 550; and (iii) a third nucleic acid molecule encoding a third polypeptide comprising an amino acid sequence selected from the group consisting of: SEQ ID NOs: 702, 706, 709, and 713; (b) culturing the one or more cells in a culture medium under conditions suitable for expressing the first polypeptide, the second polypeptide, and the third polypeptide; (c) recovering the polypeptides from the host cells and / or the culture medium; and (d) purifying the recovered polypeptides under conditions whereby the protein is produced.

[0042] Another aspect of the present disclosure provides a pharmaceutical composition comprising a protein as described herein and a pharmaceutically acceptable carrier.

[0043] Another aspect of the present disclosure provides a method of increasing tumor cell death, the method comprising exposing the tumor cells and natural killer cells to an effective amount of a protein as described herein or a pharmaceutical composition as described herein.

[0044] Another aspect of the present disclosure provides a method of treating cancer, comprising administering an effective amount of a protein as described herein or a pharmaceutical composition as described herein to a patient in need thereof.

[0045] Another aspect of the present disclosure provides the use of a protein in the preparation of a medicament for treating cancer in a human subject, wherein the protein comprises: (a) a first polypeptide comprising the amino acid sequence of SEQ ID NO: 549; (b) a second polypeptide comprising the amino acid sequence of SEQ ID NO: 550; and (c) a third polypeptide comprising an amino acid sequence selected from the group consisting of: SEQ ID NO: 702, 706, 709 and 713.

[0046] In some embodiments, the cancer is selected from the group consisting of gastrointestinal cancer, colorectal cancer, pancreatic cancer, non-small cell lung cancer, and esophageal cancer. In some embodiments, the cancer expresses CEACAM5.

[0047] The present invention also provides binding proteins that bind to CEACAM5. These binding proteins contain antigen binding sites as disclosed herein. In some aspects, these binding proteins are antibodies or antigen binding fragments thereof having antigen binding sites as disclosed herein. In other aspects, the antigen binding site is an antigen binding fragment of an antibody. The present invention also relates to nucleic acids encoding these binding proteins, methods for preparing these binding proteins, and uses of these binding proteins in treating diseases.

[0048] In some embodiments, the antigen binding site comprises a VH comprising CDRH1, CDRH2 and CDRH3, and a VL comprising CDRL1, CDRL2 and CDRL3, wherein: (i) CDRH1 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 3 and 102; (ii) CDRH2 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 37, 104 and 718; (iii) CDRH3 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 6, 38 and 105; (iv) CDRL1 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 7, 40 and 107; (v) CDRL2 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 8, 41 and 108; and (vi) CDRL3 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 9, 42 and 109.

[0049] In some embodiments, the CDRH1, CDRH2 and CDRH3 are: (i) SEQ ID NO: 3, 37 and 38; (ii) SEQ ID NO: 3, 718 and 6; or (iii) SEQ ID NO: 102, 104 and 105; and the CDRL1, CDRL2 and CDRL3 of the second antigen binding site are: (iv) SEQ ID NO: 7, 8 and 9; (v) SEQ ID NO: 40, 41 and 42; or (vi) SEQ ID NO: 107, 108 and 109.

[0050] In some embodiments, the CDRH1, CDRH2, CDRH3, CDRL1, CDRL2 and CDRL3 are: (i) SEQ ID NO: 3, 37, 38, 40, 41 and 42; (ii) SEQ ID NO: 3, 718, 6, 7, 8 and 9; or (iii) SEQ ID NO: 102, 104, 105, 107, 108 and 109, respectively.

[0051] In some embodiments, the VH comprises an amino acid sequence selected from the group consisting of: SEQ ID NOs: 704, 708, 711, and 715; and the VL comprises an amino acid sequence selected from the group consisting of: SEQ ID NOs: 591, 705, 712, and 716.

[0052] In some embodiments, the VH and VL are: (i) SEQ ID NOs: 704 and 705; (ii) SEQ ID NOs: 708 and 591; (iii) SEQ ID NOs: 711 and 712; or (iv) SEQ ID NOs: 715 and 716, respectively.

[0053] In some embodiments, the antigen binding site is a Fab fragment or a scFv. In some embodiments, the antigen binding site is a scFv comprising an amino acid sequence selected from the group consisting of: SEQ ID NO: 703, 707, 710 and 714.

[0054] In some embodiments, the antigen binding site binds to a human CEACAM5 variant comprising the amino acid sequence of SEQ ID NO:391.

[0055] In some embodiments, the protein comprises an antibody Fc domain or a portion thereof that binds to CD16. In some embodiments, the antibody Fc domain or a portion thereof that binds to CD16 is connected to the antigen binding site via a hinge comprising Ala-Ser or Gly-Ser. In some embodiments, the hinge further comprises the amino acid sequence Thr-Lys-Gly.

[0056] In some embodiments, the VH domain of the scFv forms a disulfide bridge with the VL domain of the scFv. In some embodiments, the disulfide bridge is formed between C44 of the VH and C100 of the VL according to the Kabat numbering scheme.

[0057] In some embodiments, the VH of the scFv is connected to the VL of the scFv via a flexible linker. In some embodiments, the flexible linker comprises (G4S)4 (SEQ ID NO: 532).

[0058] In some embodiments, the VH of the scFv is located at the C-terminus of the VL. In some embodiments, the VH of the scFv is located at the N-terminus of the VL.

[0059] In some embodiments, the antibody Fc domain is a human IgG1 antibody Fc domain.In some embodiments, the antibody Fc domain or a portion thereof comprises an amino acid sequence that is at least 90% identical to SEQ ID NO:531.

[0060] In some embodiments, at least one polypeptide chain of the antibody Fc domain or portion thereof comprises one or more mutations at one or more positions selected from the group consisting of: Q347, Y349, L351, S354, E356, E357, K360, Q362, S364, T366, L368, K370, N390, K392, T394, D399, S400, D401, F405, Y407, K409, T411, and K439 relative to SEQ ID NO: 531, numbered by the EU numbering system.

[0061] In some embodiments, at least one polypeptide chain of the antibody Fc domain or portion thereof comprises one or more mutations relative to SEQ ID NO: 531 selected from the group consisting of Q347E, Q347R, Y349S, Y349K, Y349T, Y349D, Y349E, Y349C, L351K, L351D, L351Y, S354C, E356K, E357Q, E357L, E357W, K360E, K360W, Q362E, S364K, S364E, S364H, S364D, T366V, T366I, T366L, T366M, T366K, T366V, 90S, N390D, N390E, K392L, K392M, K392V, K392F, K392D, K392E, T394F, D399R, D399K, D399V, S400K, S400R, D401K, F405A, F405T, Y407A, Y407I, Y407V, K409F, K409W, K409D, T411D, T411E, K439D, and K439E.

[0062] In some embodiments, one polypeptide chain of the antibody Fc domain or portion thereof comprises one or more mutations at one or more positions selected from the group consisting of: Q347, Y349, L351, S354, E356, E357, K360, Q362, S364, T366, L368, K370, K392, T394, D399, S400, D401, F405, Y407, K409, T411, and K439 relative to SEQ ID NO: 531, according to the EU numbering system, and the other polypeptide chain of the antibody Fc domain or portion thereof comprises one or more mutations at one or more positions selected from the group consisting of: NO:531 comprises one or more mutations at one or more positions selected from the group consisting of: Q347, Y349, L351, S354, E356, E357, S364, T366, L368, K370, N390, K392, T394, D399, D401, F405, Y407, K409, T411 and K439.

[0063] In some embodiments, one polypeptide chain of the antibody Fc domain or a portion thereof comprises K360E and K409W substitutions relative to SEQ ID NO:531 according to the EU numbering system; and the other polypeptide chain of the antibody Fc domain or a portion thereof comprises Q347R, D399V and F405T substitutions relative to SEQ ID NO:531.

[0064] In some embodiments, one polypeptide chain of the antibody heavy chain constant region comprises a Y349C substitution relative to SEQ ID NO: 531 according to the EU numbering system; and the other polypeptide chain of the antibody heavy chain constant region comprises a S354C substitution relative to SEQ ID NO: 531.

[0065] In some embodiments, the protein further comprises a second antigen binding site that binds CEACAM5.

[0066] Another aspect of the present disclosure provides an isolated nucleic acid molecule encoding any of the disclosed proteins, including but not limited to these binding proteins, antibodies, antigen-binding fragments or antigen-binding sites.

[0067] Another aspect of the present disclosure provides an expression vector comprising an isolated nucleic acid molecule encoding any of the disclosed proteins.

[0068] Another aspect of the present disclosure provides a host cell comprising an expression vector containing an isolated nucleic acid molecule encoding any of the disclosed proteins. In some embodiments, the host cell is a Chinese Hamster Ovary (CHO) cell.

[0069] Another aspect of the present disclosure provides a method for producing a protein, comprising: (a) providing a host cell comprising an expression vector containing an isolated nucleic acid molecule encoding any of the disclosed proteins; (b) culturing the host cell in a culture medium under conditions suitable for expression of the protein; and (c) isolating the protein from the culture medium.

[0070] Another aspect of the present disclosure provides a method of increasing tumor cell death, the method comprising exposing the tumor cells and natural killer cells to an effective amount of a protein as described herein or a pharmaceutical composition as described herein.

[0071] Another aspect of the present disclosure provides a method of treating cancer, comprising administering an effective amount of a protein as described herein or a pharmaceutical composition as described herein to a patient in need thereof.

[0072] Another aspect of the present disclosure provides use of a protein as described herein in the preparation of a medicament for treating cancer in a human subject, wherein the protein comprises: (a) a first polypeptide comprising the amino acid sequence of SEQ ID NO: 549; (b) a second polypeptide comprising the amino acid sequence of SEQ ID NO: 550; and (c) a third polypeptide comprising an amino acid sequence selected from the group consisting of: SEQ ID NOs: 702, 706, 709 and 713.

[0073] In some embodiments, the cancer is selected from the group consisting of gastrointestinal cancer, colorectal cancer, pancreatic cancer, non-small cell lung cancer, and esophageal cancer. In some embodiments, the cancer expresses CEACAM5.

[0074] Various aspects and embodiments of the invention are described in further detail below. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 is a representative diagram of a heterodimeric multispecific antibody (e.g., a trispecific binding protein (TriNKET)). Each arm can represent a NKG2D binding domain or a CEACAM5 binding domain. In some embodiments, the NKG2D binding domain and the CEACAM5 binding domain can share a common light chain.

[0076] Figure 2A-2E Five exemplary formats of multispecific binding proteins (e.g., TriNKET) are described. Figure 2A As shown, the NKG2D binding domain or the CEACAM5 binding domain can be in scFv format (left arm). The antibody containing NKG2D targeting scFv, CEACAM5 targeting Fab fragment and CD16 targeting heterodimeric antibody Fc domain or its part is referred to herein as F3-TriNKET. The antibody containing CEACAM5 targeting scFv, NKG2D targeting Fab fragment and CD16 binding heterodimeric antibody Fc domain or its part is referred to herein as F3'-TriNKET ( Figure 2E ).like Figure 2B As shown, both the NKG2D binding domain and the CEACAM5 binding domain can be in scFv format. Figure 2C to Figure 2D is a diagrammatic representation of an antibody with three antigen binding sites comprising two antigen binding sites that bind CEACAM5 and a NKG2D binding site fused to the heterodimeric antibody Fc domain or portion thereof that binds CD 16. These antibody formats are referred to herein as F4-TriNKET. Figure 2C The two CEACAM5 binding sites are illustrated in Fab fragment format and the NKG2D binding site is in scFv format. Figure 2D The CEACAM5 binding sites are described in scFv format and the NKG2D binding sites are described in scFv format. Figure 2ERepresents a TriNKET containing a tumor targeting scFv, a NKG2D targeting Fab fragment, and a heterodimeric antibody Fc domain or portion thereof (otherwise referred to herein as a constant region / domain ("CD domain")) that binds CD16. The antibody format is referred to herein as F3'-TriNKET. In certain exemplary multispecific binding proteins, the heterodimerization mutations on the antibody Fc domain or portion thereof include K360E and K409W on one polypeptide chain of the Fc domain or portion thereof; and Q347R, D399V, and F405T (shown as a triangular lock and key shape in these CD domains) on the other polypeptide chain of the Fc domain or portion thereof. The bold bars between the heavy chain variable domain and the light chain variable domain of these Fab fragments represent disulfide bonds.

[0077] Figure 3 Figure 1 is a representative diagram of TriNKET in the form of a triomab, a trifunctional bispecific antibody that retains an IgG-like shape. The chimera consists of two half antibodies derived from two corresponding antibodies, each with one light chain and one heavy chain. The triomab format can be a heterodimeric construct containing 1 / 2 rat antibody and 1 / 2 mouse antibody.

[0078] Figure 4 is a representation of a KiH common light chain format of TriNKET involving knobs-into-holes (KiH) technology. KiH is a heterodimer containing two Fab fragments that bind to targets 1 and 2 and an Fc stabilized by heterodimerization mutations. TriNKET in KiH format can be a heterodimeric construct with two Fab fragments that bind to targets 1 and 2, containing two different heavy chains and a common light chain paired with the two heavy chains.

[0079] Figure 5 DVD-Ig TM ) represents a TriNKET in a form that combines the target binding domains of two monoclonal antibodies through a flexible naturally occurring linker and produces a tetravalent IgG-like molecule. DVD-Ig TM It is a homodimeric construct in which the variable domain targeting antigen 2 is fused to the N-terminus of the variable domain of the Fab fragment targeting antigen 1. DVD-Ig TM The format contains normal Fc.

[0080] Figure 6Is a representative diagram of TriNKET in the form of an orthogonal Fab fragment interface (Ortho-Fab), a heterodimeric construct containing two Fab fragments binding to target 1 and target 2 fused to an Fc. Light chain (LC)-heavy chain (HC) pairing is ensured by the orthogonal interface. Heterodimerization is ensured by mutations in the Fc.

[0081] Figure 7 This is a representative image of TriNKET in 2-in-1 Ig format.

[0082] Figure 8 is a representative diagram of the ES format of TriNKET, a heterodimeric construct containing two different Fab fragments binding to target 1 and target 2 fused to the Fc. Heterodimerization is ensured by electrostatic steering mutations in the Fc.

[0083] Fig. 9 is a representative diagram of TriNKET in Fab arm exchange format: an antibody in which the Fab fragment arms are exchanged (thus generating a bispecific antibody) by swapping the heavy chain and attached light chain (half-molecule) with the heavy-light chain pairing from another molecule. The Fab arm exchange format (cFae) is a heterodimer containing 2 Fab fragments binding to targets 1 and 2 and an Fc stabilized by heterodimerization mutations.

[0084] Fig.10 is a representative image of TriNKET in SEED Body format, a heterodimer containing two Fab fragments binding to targets 1 and 2 and an Fc stabilized by heterodimerization mutations.

[0085] Fig.11 is a representative diagram of the LuZ-Y format of TriNKET in which a leucine zipper is used to induce heterodimerization of two different HCs. The LuZ-Y format is a heterodimer containing two different scFabs binding to targets 1 and 2 fused to Fc. Heterodimerization is ensured by a leucine zipper motif fused to the C-terminus of the Fc.

[0086] Fig.12 This is a representative image of TriNKET in Cov-X-Body format.

[0087] Figure 13A-13Bare representative images of TriNKETs in κλ-Body format, which are heterodimeric constructs in which two different Fab fragments are fused to an Fc stabilized by heterodimerization mutations: one Fab fragment targeting antigen 1 contains κLC, while the second Fab fragment targeting antigen 2 contains λLC. Fig.13A is an exemplary representation of a form of κλ-Body; Fig. 13B is another exemplary representation of a κλ-Body.

[0088] Fig.14 is an Oasc-Fab heterodimeric construct comprising a Fab fragment binding to target 1 and a scFab binding to target 2, both fused to an Fc domain. Heterodimerization is ensured by mutations in the Fc domain.

[0089] Fig.15 DuetMab is a heterodimeric construct containing two different Fab fragments binding to antigens 1 and 2 and an Fc stabilized by heterodimerization mutations. Fab fragment 1 and Fab fragment 2 contain differential SS bridges that ensure correct light and heavy chain pairing.

[0090] Fig.16 is a CrossmAb, which is a heterodimeric construct with two different Fab fragments binding to target 1 and target 2 and an Fc stabilized by heterodimerization mutations. The CL domain and CH1 domain, as well as the VH domain and VL domain are swapped, for example, CH1 is fused in-line with VL, and CL is fused in-line with VH.

[0091] Fig.17 is Fit-Ig, which is a homodimeric construct in which the Fab fragment that binds to Antigen 2 is fused to the N-terminus of the HC of the Fab fragment that binds to Antigen 1. This construct contains a wild-type Fc.

[0092] Fig.18 is a line graph showing the binding affinity of NKG2D binding domains (listed as clones) to human recombinant NKG2D in an ELISA assay.

[0093] Fig.19 is a line graph showing the binding affinity of NKG2D binding domains (listed as clones) to cynomolgus monkey recombinant NKG2D in an ELISA assay.

[0094] Fig. 20 is a line graph showing the binding affinity of NKG2D binding domains (listed as clones) to mouse recombinant NKG2D in an ELISA assay.

[0095] Fig.21Is a bar graph showing binding of NKG2D binding domains (listed as clones) to EL4 cells expressing human NKG2D by flow cytometry, showing mean fluorescence intensity (MFI) over multiple of background (FOB).

[0096] Fig. 22 Is a bar graph showing binding of NKG2D binding domains (listed as clones) to EL4 cells expressing mouse NKG2D by flow cytometry, showing mean fluorescence intensity (MFI) over multiple of background (FOB).

[0097] Fig.23 is a line graph showing the specific binding affinity of the NKG2D binding domain (listed as clones) to recombinant human NKG2D-Fc by competition with the natural ligand ULBP-6.

[0098] Fig.24 is a line graph showing the specific binding affinity of the NKG2D binding domain (listed as clones) to recombinant human NKG2D-Fc by competition with the natural ligand MICA.

[0099] Fig.25 is a line graph showing the specific binding affinity of the NKG2D binding domain (listed as clones) to recombinant mouse NKG2D-Fc by competition with the natural ligand Rae-1δ.

[0100] Fig.26 Is a bar graph showing activation of human NKG2D by the NKG2D binding domain (listed as clones) by quantifying the percentage of TNF-α positive cells expressing human NKG2D-CD3ζ fusion protein.

[0101] Fig. 27 Is a bar graph showing activation of mouse NKG2D by the NKG2D binding domain (listed as clones) by quantifying the percentage of TNF-α positive cells expressing the mouse NKG2D-CD3ζ fusion protein.

[0102] Fig.28 is a bar graph showing activation of human NK cells by NKG2D binding domains (listed as clones).

[0103] Fig.29 is a bar graph showing activation of human NK cells by NKG2D binding domains (listed as clones).

[0104] Fig.30 is a bar graph showing activation of mouse NK cells by NKG2D binding domains (listed as clones).

[0105] Fig.31is a bar graph showing activation of mouse NK cells by NKG2D binding domains (listed as clones).

[0106] Fig.32 is a bar graph showing the cytotoxic effects of NKG2D binding domains (listed as clones) on tumor cells.

[0107] Fig.33 is a bar graph showing the melting temperatures of NKG2D binding domains (listed as clones) measured by differential scanning fluorimetry.

[0108] Figure 34A-Figure 34C It is a bar graph showing the synergistic activation of NK cells using CD16 and NKG2D combination. Fig.34A The levels of CD107a are shown; Fig.34B The levels of IFN-γ are shown; Fig.34C Levels of CD107a and IFN-γ are shown. Graphs show mean (n=2) ± SD. Data are representative of five independent experiments using five different healthy donors.

[0109] Figure 35A-Figure 35B is a Biacore sensorgram showing simultaneous binding of CEACAM5, NKG2D and CD16 target proteins to CEACAM5 TriNKET. Fig.35B yes Fig.35A , both showing the formation of a heterotetrameric complex.

[0110] Figure 36A-Figure 36E is a line graph showing flow cytometry experiments binding of CEACAM5 TriNKET to various human and cynomolgus monkey CEACAM family member proteins. Fig.36A AB0411 and AB0466 were shown to bind to human CEACAM1; Fig.36B AB0411 binding to human CEACAM6 was demonstrated; Fig.36C demonstrated that CEACAM5 TriNKET does not bind to human CEACAM8; Fig.36D AB0264 and AB0621 were shown to bind to cynomolgus monkey CEACAM5, but AB0466 and AB0411 did not; Fig.36E Demonstrates that CEACAM5 TriNKET does not bind to cells lacking CEACAM protein expression. Data represent the mean of duplicate wells and error bars represent SD.

[0111] Figure 37A-Figure 37D is a line graph from a DELFIA assay showing that CEACAM5TriNKET promotes the target cancer line SK-CO-1 ( Fig.37A )、LS-147T( Fig.37B )、ZR-75-30( Fig.37C ) and HPAF-II( Fig.37D ) cracking.

[0112] Figure 38A-Figure 38B is a line graph from a DELFIA assay showing that AB0264 promotes lysis of the target cancer line ZR-75-30 better than AB0755 (its corresponding mAb). Fig.38A ) compared to IL-2 activated NK cells ( Fig.38B ) showed more potent killing of ZR-75-30 cancer cells. Data points represent mean ± SD.

[0113] Figure 39A-Figure 39E is a line graph from a DELFIA assay showing that AB0264 promotes the growth of the target cancer line MKN-45 better than its corresponding mAb ( Fig.39A )、SK-CO-1( Fig.39B )、LS-147T( Fig.39C )、ZR-75-30( Fig.39D ) and HPAF-II( Fig.39E ). Data points represent mean ± SD.

[0114] Figure 40A-40E is a line graph from a DELFIA assay showing that AB0411 promotes the growth of the target cancer line MKN-45 ( Fig.40A )、SK-CO-1( Fig.40B )、LS-147T( Fig.40C )、ZR-75-30( Fig.40D ) and HPAF-II( Fig.40E ). Data points represent mean ± SD.

[0115] Fig.41 is a line graph from a DELFIA assay showing that NK-mediated killing of target cells is dependent on the binding of TriNKET to CD16, NKG2D, and CEACAM5. Data points represent mean ± SD.

[0116] Figure 42A-42D are line graphs from IFNγ and CD107a activation assays. Fig.42A We show that primary NK cells induce IFNγ secretion following co-engagement with CEACAM5TriNKET and SK-CO-1 target cells; Fig.42BWe show that primary NK cells induce IFNγ production and CD107a degranulation after co-engagement with CEACAM5 TriNKET and MKN-45 target cells; Fig.42C ) or SK-CO-1( Fig.42D ) cells, TriNKET enhanced the expression of CD8 + Degranulation of NK cells. Data represent the results of an experiment using PBMCs from three animals. Data represent mean ± SD.

[0117] Fig.43A is a graph of a flow cytometry experiment showing the expression of CEACAM5 protein on the cancer cell line SK-CO-1 and patient-derived primary non-small cell lung cancer tumor organoid lines 10910 and 3222. A DELFIA assay shows that CEACAM5 TriNKET promotes the expression of CEACAM5 protein on the patient-derived primary non-small cell lung cancer tumor organoid line 3222 ( Fig.43B ) and 10910( Fig.43C ) cracking.

[0118] Fig.44A Describes the generation of activated CD8 + Schematic representation of the T cell approach. Fig.44B is a line graph from a DELFIA assay showing pre-activated CD8 + Enhancement of T cell-mediated target cell lysis is dependent on the binding of TriNKET to NKG2D and CEACAM5. Maximum lysis values ​​represent the mean ± SD of three donors.

[0119] Fig.45 is a Kaplan-Meier curve showing the remaining percentage of hCEACAM5 transgenic mice bearing subcutaneous B16F10-hCEACAM5 tumors in each AB0621 mouse IgG2a surrogate (mAB0621) treatment group over time.

[0120] Figure 46A-Figure 46E In the case of administration of 15 mg / kg of isotype control ( Fig.46A ) or 15 mg / kg( Fig.46B )、5mg / kg( Fig.46C )、1.5mg / kg( Fig.46D ) or 0.5mg / kg( Fig.46E Figure 2. Single curve of tumor volume in B16F10-hCEACAM5 tumor-bearing mice in the hCEACAM5 transgenic model up to day 26 after mAB0621 administration.

[0121] Fig.47 is a Kaplan-Meier curve showing the change in the percentage of remaining animals over time in each group treated with mAB0621 and / or anti-PD1 antibody alone or dually.

[0122] Figure 48A-48D In the isotype control treatment group ( Fig.48A ), anti-PD1 treatment group ( Fig.48B ), mAB0621 treatment group ( Fig.48C ) or mAB0621+anti-PD-1 treatment group ( Fig.48D ) The volume of a single B16F10-hCEACAM5 tumor was measured for each animal.

[0123] Figure 49A-49B Figure 2 shows free and total AB0264 in cynomolgus monkeys ( Fig.49A ) and AB0411( Fig.49B ) over time.

[0124] Figure 50A-Figure 50C Figure 2 shows free and total AB0621 in B6.Cg-Tg(hCEACAM5)2682Wzm / Ieg transgenic mice ( Fig.50A )、AB0411( Fig.50B ) and AB0466( Fig.50C ) over time. DETAILED DESCRIPTION

[0125] The present invention provides multispecific binding proteins that bind to NKG2D receptors and CD16 on natural killer cells and the tumor-associated antigen CEACAM5. In some embodiments, these multispecific proteins further comprise an additional antigen binding site that binds CEACAM5. The present invention also provides pharmaceutical compositions comprising such multispecific binding proteins, and methods of treatment using such multispecific proteins and pharmaceutical compositions, for purposes such as treating autoimmune diseases and cancer. Various aspects of the present invention are described below in sections; however, aspects of the present invention described in a particular section are not limited to any particular section.

[0126] To facilitate understanding of the present invention, certain terms and phrases are defined below.

[0127] As used herein, the terms "a" and "an" mean "one or more" and "one or more" and include the plural forms unless the context is inappropriate.

[0128] As used herein, the terms "plurality" or "plurality" mean more than one or more than one.

[0129] As used herein, the term "antigen binding site" refers to the portion of an immunoglobulin molecule that participates in antigen binding. In human antibodies, the antigen binding site is formed by amino acid residues in the N-terminal variable ("V") region of the heavy ("H") and light ("L") chains. The three highly distinct segments within these heavy and light chain V regions are called "hypervariable regions," which are inserted between more conservative flanking segments called "framework regions" or "FRs." Therefore, the term "FR" refers to the amino acid sequence naturally present between and near the hypervariable regions in immunoglobulins. In human antibody molecules, the three hypervariable regions of the light chain and the three hypervariable regions of the heavy chain are arranged relative to each other in three-dimensional space to form an antigen binding surface. This antigen binding surface is complementary to the three-dimensional surface of the bound antigen, and the three hypervariable regions of each heavy chain and light chain are called "complementarity determining regions" or "CDRs." In some animals (such as camels and cartilaginous fish), the antigen binding site is formed by a single antibody chain, providing a "single domain antibody." The antigen binding site may exist in an intact antibody, in an antigen binding fragment of an antibody that retains an antigen binding surface, or in a recombinant polypeptide (such as a scFv) using a peptide linker to connect the heavy chain variable domain to the light chain variable domain in a single polypeptide.

[0130] As used herein, the term "Fc domain" or "Fc region" refers to the C-terminal region of the immunoglobulin heavy chain derived from the second constant domain and the third constant domain. The term includes native sequence Fc regions and variant Fc regions. In some embodiments, the variant Fc region comprises an amino acid sequence that is at least 90% identical (i.e., at least 90%, at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98% or at least 99%) to a native sequence Fc region of a human (e.g., human IgG1, IgG2, IgG3 or IgG4 Fc region). Although the boundaries of the Fc region of an IgG heavy chain may be slightly different, the human IgG heavy chain Fc region is generally defined as extending from Cys226 or from Pro230 to the carboxyl terminus of the heavy chain. However, the C-terminal lysine (Lys447) in the Fc region may or may not be present. Unless otherwise specified herein, the numbering of amino acid residues in the Fc region or constant region is according to the EU numbering system (also known as the EU index), as described in Kabat et al., Sequences of Proteins of Immunological Interest, 5th Ed. Public Health Service, National Institutes of Health, Bethesda, Md., 1991. As used herein, an Fc domain comprises two polypeptide chains that together form a dimeric Fc domain, i.e., each polypeptide comprises a C-terminal constant region of an immunoglobulin heavy chain and is capable of self-association. In a specific embodiment, the subunit of an IgG Fc domain comprises an IgG CH2 and an IgG CH3 constant domain.

[0131] As used herein, the term "tumor-associated antigen" refers to any antigen, including but not limited to proteins, glycoproteins, gangliosides, carbohydrates or lipids associated with cancer. Such antigens can be expressed on malignant cells or in the tumor microenvironment (such as on tumor-associated blood vessels, extracellular matrix, mesenchymal matrix or immune infiltrates).

[0132] As used herein, the terms "subject" and "patient" refer to an organism to be treated by the methods and compositions described herein. Such organisms preferably include, but are not limited to, mammals (e.g., mice, apes, horses, cows, pigs, dogs, cats, etc.), and more preferably include humans.

[0133] As used herein, the term "effective amount" refers to an amount of a compound (e.g., a protein, binding protein, antibody, or antigen-binding fragment of the invention) sufficient to achieve a beneficial or desired result. An effective amount can be administered by one or more administrations, applications, or dosages, and is not intended to be limited to a particular formulation or route of administration. As used herein, the term "treatment" includes any effect that results in improvement or alleviation of symptoms of a condition, disease, illness, or the like, such as, for example, alleviation, reduction, regulation, relief, or elimination.

[0134] As used herein, the term "pharmaceutical composition" refers to the combination of an active agent with an inert or active pharmaceutically acceptable carrier, rendering the composition particularly suitable for diagnostic or therapeutic use, either in vivo or ex vivo.

[0135] As used herein, the term "pharmaceutically acceptable carrier" refers to any standard pharmaceutical carrier, such as phosphate buffered saline, water, emulsions (e.g., oil / water or water / oil emulsions), and various types of wetting agents. The composition may also contain stabilizers and preservatives. For examples of carriers, stabilizers, and adjuvants, see, e.g., Martin, Remington's Pharmaceutical Sciences, 15th Edition, Mack Publ. Co., Easton, PA

[1975] .

[0136] As used herein, the term "pharmaceutically acceptable salt" refers to any pharmaceutically acceptable salt (e.g., acid or base) of a compound of the present invention, which can provide a compound of the present invention or its active metabolite or residue after being administered to a subject. As known to those skilled in the art, the "salt" of a compound of the present invention can be derived from an inorganic or organic acid and base. Exemplary acids include, but are not limited to, hydrochloric acid, hydrobromic acid, sulfuric acid, nitric acid, perchloric acid, fumaric acid, maleic acid, phosphoric acid, glycolic acid, lactic acid, salicylic acid, succinic acid, p-toluenesulfonic acid, tartaric acid, acetic acid, citric acid, methanesulfonic acid, ethanesulfonic acid, formic acid, benzoic acid, malonic acid, naphthalene-2-sulfonic acid, benzenesulfonic acid, etc. Other acids (such as oxalic acid), although not pharmaceutically acceptable in themselves, can be used to prepare salts that can be used as intermediates for obtaining compounds of the present invention and pharmaceutically acceptable acid addition salts thereof.

[0137] Exemplary bases include, but are not limited to, alkali metal (e.g., sodium) hydroxides, alkaline earth metal (e.g., magnesium) hydroxides, ammonia, and compounds of formula NW4 + Compounds (wherein W is C 1-4 Alkyl) and so on.

[0138] Exemplary salts include, but are not limited to, acetate, adipate, alginate, aspartate, benzoate, benzenesulfonate, bisulfate, butyrate, citrate, camphorate, camphorsulfonate, cyclopentanepropionate, digluconate, dodecylsulfate, ethanesulfonate, fumarate, fluoroheptanoate, glycerophosphate, hemisulfate, heptanoate, hexanoate, hydrochloride, hydrobromide, hydroiodide, 2-hydroxyethanesulfonate, lactate, maleate, methanesulfonate, 2-naphthalenesulfonate, nicotinate, oxalate, palmitate, pectinate, persulfate, phenylpropionate, picrate, pivalate, propionate, succinate, tartrate, thiocyanate, toluenesulfonate, undecanoate, and the like. Other examples of salts include those with a suitable cation (e.g., Na + NH4 + and NW4 + (W is C 1-4 alkyl)) to form anions of the compounds of the present invention and the like.

[0139] For therapeutic use, salts of the compounds of the invention are considered pharmaceutically acceptable. However, salts of acids and bases that are non-pharmaceutically acceptable may also be used, for example, in the preparation or purification of pharmaceutically acceptable compounds.

[0140] As used herein, "CEACAM5" (also known as carcinoembryonic antigen-related cell adhesion molecule 5, meconium antigen 100, CEA, carcinoembryonic antigen, CD66e, or CD66e antigen) refers to the protein of Uniprot Accession No. P06731 and related isoforms.

[0141] Throughout the specification, when compositions are described as having, including, or comprising specific ingredients, or when processes and methods are described as having, including, or comprising specific steps, it is considered that, in addition, there are compositions of the invention consisting essentially of or consisting of said ingredients, and there are processes and methods according to the invention consisting essentially of or consisting of said processing steps.

[0142] In general, compositions specifying percentages are by weight unless otherwise indicated. In addition, if a variable is not accompanied by a definition, the preceding definition for that variable controls.

[0143] I. Protein

[0144] The present invention provides multispecific binding proteins that bind to NKG2D receptors and CD16 on natural killer cells and the tumor-associated antigen CEACAM5. These multispecific binding proteins can be used in the pharmaceutical compositions and treatment methods described herein. The binding of these multispecific binding proteins to NKG2D receptors and CD16 on natural killer cells enhances the activity of the natural killer cells in destroying tumor cells expressing CEACAM5. The binding of these multispecific binding proteins to tumor cells expressing CEACAM5 brings these cells close to the natural killer cells, which helps the natural killer cells to directly and indirectly destroy these tumor cells. Multispecific binding proteins that bind to NKG2D, CD16 and another target are disclosed in International Application Publication Nos. WO2018148445 and WO2019157366, which are not incorporated herein by reference. A further description of some exemplary multispecific binding proteins is provided below.

[0145] The first component of the multispecific binding protein binds to cells expressing the NKG2D receptor (which may include but are not limited to NK cells, γδ T cells and CD8 + After NKG2D binding, these multispecific binding proteins can block natural ligands (such as ULBP6 and MICA) from binding to NKG2D and activating NK cells.

[0146] The second component of these multispecific binding proteins is an antigen binding site that binds CEACAM5. Cells expressing CEACAM5 can be found in, for example, gastrointestinal cancer, colorectal cancer, pancreatic cancer, non-small cell lung cancer, and esophageal cancer.

[0147] The third component of these multispecific binding proteins is an antibody Fc domain or portion thereof or antigen binding site, each of which binds to Fc receptors on the surface of cells expressing CD16, including, as specific examples, leukocytes (including natural killer cells, macrophages, neutrophils, eosinophils, mast cells and follicular dendritic cells).

[0148] The additional antigen binding site of these multispecific binding proteins can bind to CEACAM5. In certain embodiments, the first antigen binding site that binds NKG2D is a scFv, and the second antigen binding site and the additional antigen binding site bind to CEACAM5, each of which is a Fab fragment. In certain embodiments, the first antigen binding site that binds NKG2D is a scFv, and the second antigen binding site and the additional antigen binding site bind to CEACAM5, each of which is a scFv. In certain embodiments, the first antigen binding site that binds NKG2D is a Fab fragment, and the second antigen binding site and the additional antigen binding site bind to CEACAM5, each of which is a scFv. In certain embodiments, the first antigen binding site that binds NKG2D is a Fab, and the second antigen binding site and the additional antigen binding site bind to CEACAM5, each of which is a Fab fragment.

[0149] These antigen binding sites can each be incorporated into an antibody heavy chain variable domain and an antibody light chain variable domain (e.g., arranged as in an antibody, or fused together to form an scFv), or one or more of these antigen binding sites can be a single domain antibody, such as a V H H antibody (such as camelid antibodies) or V NAR Antibodies (such as those found in cartilaginous fish).

[0150] The multispecific binding proteins described herein can take various formats. For example, one format is a heterodimeric multispecific antibody ( Figure 1 ). The first immunoglobulin heavy chain comprises a first Fc (hinge-CH2-CH3) domain, a first heavy chain variable domain, and an optional first CH1 heavy chain domain. The first immunoglobulin light chain comprises a first light chain variable domain and an optional first light chain antibody Fc domain. The first immunoglobulin light chain forms an antigen binding site that binds NKG2D together with the first immunoglobulin heavy chain. The second immunoglobulin heavy chain comprises a second Fc (hinge-CH2-CH3) domain, a second heavy chain variable domain, and an optional second CH1 heavy chain domain. The second immunoglobulin light chain comprises a second light chain variable domain and an optional second light chain constant domain. The second immunoglobulin light chain forms an antigen binding site that binds CEACAM5 together with the second immunoglobulin heavy chain. The first Fc domain and the second Fc domain together are capable of binding to CD16 ( Figure 1 ).

[0151] Another exemplary format involves a heterodimeric multispecific antibody comprising a first immunoglobulin heavy chain, a second immunoglobulin heavy chain, and an immunoglobulin light chain ( Figure 2A ). The first immunoglobulin heavy chain comprises a first Fc (hinge-CH2-CH3) domain, which is fused to a scFv consisting of a heavy chain variable domain and a light chain variable domain (which pairs and binds to NKG2D, or binds to CEACAM5) via a linker or antibody hinge. The second immunoglobulin heavy chain comprises a second Fc (hinge-CH2-CH3) domain, a second heavy chain variable domain, and a CH1 heavy chain domain. The immunoglobulin light chain comprises a light chain variable domain and a light chain constant domain. The second immunoglobulin heavy chain pairs with the immunoglobulin light chain and binds to NKG2D or binds to CEACAM5. The first Fc domain and the second Fc domain together are capable of binding to CD16 ( Figure 2A ).

[0152] Another exemplary format involves a heterodimeric multispecific antibody comprising a first immunoglobulin heavy chain and a second immunoglobulin heavy chain ( Figure 2B ). The first immunoglobulin heavy chain comprises a first Fc (hinge-CH2-CH3) domain, which is fused to a scFv consisting of a heavy chain variable domain and a light chain variable domain (which pair and bind to NKG2D, or bind to CEACAM5) via a linker or an antibody hinge. The second immunoglobulin heavy chain comprises a second Fc (hinge-CH2-CH3) domain, which is fused to a scFv consisting of a heavy chain variable domain and a light chain variable domain (which pair and bind to NKG2D, or bind to CEACAM5) via a linker or an antibody hinge. The first Fc domain and the second Fc domain together are capable of binding to CD16 ( Figure 2B ).

[0153] In some embodiments, the above-mentioned scFv is connected to the antibody constant domain via a hinge sequence. In some embodiments, the hinge comprises amino acids Ala-Ser or Gly-Ser. In some embodiments, the hinge connects the scFv that binds NKG2D, and the antibody heavy chain constant domain comprises amino acids Ala-Ser. In some embodiments, the hinge connects the scFv that binds CEACAM5, and the antibody heavy chain constant domain comprises amino acids Gly-Ser. In some other embodiments, the hinge comprises amino acids Ala-Ser and Thr-Lys-Gly. The hinge sequence can provide flexibility in binding to the target antigen, as well as a balance between flexibility and optimal geometry.

[0154] In some embodiments, the above-mentioned scFv comprises a heavy chain variable domain and a light chain variable domain. In some embodiments, the heavy chain variable domain forms a disulfide bridge with the light chain variable domain to enhance the stability of the scFv. For example, the amino acid position is numbered according to Kabat, and a disulfide bridge can be formed between the C44 residue of the heavy chain variable domain and the C100 residue of the light chain variable domain. In some embodiments, the heavy chain variable domain is connected to the light chain variable domain by a flexible joint. Any suitable joint can be used, for example, (G4S)4 joint (SEQ ID NO:532). In some embodiments of the scFv, the heavy chain variable domain is located at the N-terminal of the light chain variable domain. In some embodiments of the scFv, the heavy chain variable domain is located at the C-terminal of the light chain variable domain.

[0155] The multispecific binding proteins described herein may further comprise one or more additional antigen binding sites. The (these) additional antigen binding sites may be fused to the N-terminus of the constant region CH2 domain or the C-terminus of the constant region CH3 domain, optionally fused via a linker sequence. In certain embodiments, the (these) additional antigen binding sites take the form of a single-chain variable region (scFv), which is optionally disulfide stabilized to produce a tetravalent or trivalent multispecific binding protein. For example, the multispecific binding protein comprises a first antigen binding site that binds NKG2D, a second antigen binding site that binds CEACAM5, an additional antigen binding site that binds CEACAM5, and an antibody constant region or portion thereof sufficient to bind CD16 or a fourth antigen binding site that binds CD16. Any of these antigen binding sites may take the form of a Fab fragment or scFv, such as the above-mentioned scFv.

[0156] In some embodiments, the additional antigen binding site binds to a different epitope of CEACAM5 than the second antigen binding site. In some embodiments, the additional antigen binding site binds to the same epitope as the second antigen binding site. In some embodiments, the additional antigen binding site comprises the same heavy chain CDR sequence and light chain CDR sequence as the second antigen binding site. In some embodiments, the additional antigen binding site comprises the same heavy chain variable domain sequence and light chain variable domain sequence as the second antigen binding site. In some embodiments, the additional antigen binding site has the same amino acid sequence as the second antigen binding site. An exemplary format is shown in Figure 2C and Figure 2DThus, these multispecific binding proteins can provide bivalent engagement of CEACAM5. The bivalent engagement of CEACAM5 by these multispecific proteins can stabilize CEACAM5 on the surface of tumor cells and enhance the cytotoxicity of NK cells to these tumor cells. The bivalent engagement of CEACAM5 by these multispecific proteins can confer stronger binding of these multispecific proteins to these tumor cells, thereby promoting a stronger cytotoxic response of NK cells to these tumor cells (especially to tumor cells expressing low levels of CEACAM5).

[0157] These multispecific binding proteins can take other formats. In some embodiments, the multispecific binding protein is in the form of a Triomab, which is a trifunctional bispecific antibody that retains an IgG-like shape. The chimera consists of two half antibodies derived from two corresponding antibodies, each half antibody having a light chain and a heavy chain.

[0158] In some embodiments, the multispecific binding protein is in KiH format, which involves knob-in-hole (KiH) technology. KiH involves engineering C H The concept behind the "knob-hole" Fc technology is to replace a small residue with a large residue (e.g., T366W in EU numbering). CH3A ) introduced a "knob" in one CH3 domain (CH3A). To accommodate the "hole", the nearest neighboring residues to the hole were replaced with smaller residues (e.g., T366S / L368A / Y407V CH3B) creates a complementary "hole" surface on the other CH3 domain (CH3B). The "hole" mutation was optimized by structure-guided phage library screening (Atwell S, Ridgway JB, Wells JA, Carter P., Stable heterodimers from remodeling the domain interface of a homodimer using a phage display library, J. Mol. Biol. (1997) 270 (1): 26-35). X-ray crystal structure of a KiH Fc variant (Elliott JM, Ultsch M, Lee J, Tong R, Takeda K, Spiess C, et al., Antiparallel conformation of knob and hole aglycosylated half-antibody homomodimers is mediated by a CH2-CH3 hydrophobic interaction. J. Mol. Biol. (2014) 426(9): 1947–57; Mimoto F, Kadono S, Katada H, Igawa T, Kamikawa T, Hattori K. Crystal structure of a novel asymmetrically engineered Fc variant with improved affinity for FcγRs. Mol. Immunol. (2014) 58(1):132–8) demonstrated that hydrophobic interactions driven by steric complementarity at the core interface between CH3 domains thermodynamically favor heterodimerization, whereas the knob-knob and hole-hole interfaces are not favored towards homodimerization due to steric hindrance and disruption of favorable interactions, respectively.

[0159] In some embodiments, the multispecific binding protein is a dual variable domain immunoglobulin (DVD-Ig TM ) format, which combines the target binding domains of two monoclonal antibodies through a flexible, naturally occurring linker and produces a tetravalent IgG-like molecule.

[0160] In some embodiments, the multispecific binding protein is in the form of an orthogonal Fab interface (Ortho-Fab). In the ortho-Fab IgG method (Lewis SM, Wu X, Pustilnik A, Sereno A, Huang F, Rick HL et al., Generation of bispecific IgG antibodies by structure-based design of anorthogonal Fab interface. Nat. Biotechnol. (2014) 32(2): 191–8), the structure-based region design is performed on only the LC and HC of one Fab fragment. VH-CH1 Complementary mutations were introduced at the interface without making any changes to the other Fab fragment.

[0161] In some embodiments, the multispecific binding protein is in a 2-in-1 Ig format. In some embodiments, the multispecific binding protein is in ES format, which is a heterodimeric construct containing two different Fab fragments binding to target 1 and target 2 fused to Fc. Heterodimerization is ensured by electrostatic steering mutations in the Fc.

[0162] In some embodiments, the multispecific binding protein is in the form of a κλ-Body, which is a heterodimeric construct in which two different Fab fragments are fused to an Fc stabilized by heterodimerization mutations: Fab fragment 1 targeting antigen 1 contains κLC, while Fab fragment 2 targeting antigen 2 contains λLC. Fig.13A is an exemplary representation of a form of κλ-Body; Fig. 13B is another exemplary representation of a κλ-Body.

[0163] In some embodiments, the multispecific binding protein is in a Fab arm exchange format (an antibody in which the Fab fragment arms are exchanged (thereby generating a bispecific antibody) by swapping the heavy chain and attached light chain (half-molecule) with a heavy-light chain pairing from another molecule).

[0164] In some embodiments, the multispecific binding protein is in the form of a SEED Body. The chain exchange engineered domain (SEED) platform is designed to generate asymmetric and bispecific antibody-like molecules, which expands the therapeutic application of natural antibodies. The protein engineering platform is based on exchanging structurally related sequences of immunoglobulins within a conserved CH3 domain. The SEED design allows for the efficient generation of AG / GA heterodimers without biasing towards homodimerization of AG and GA SEED CH3 domains. (Muda M. et al., Protein Eng. Des. Sel. (2011, 24 (5): 447-54)).

[0165] In some embodiments, the multispecific binding protein is in the form of LuZ-Y, in which a leucine zipper is used to induce heterodimerization of two different HCs (Wranik, BJ. et al., J. Biol. Chem. (2012), 287:43331-9).

[0166] In some embodiments, the multispecific binding protein is in the form of Cov-X-Body. In the bispecific CovX-Body, two different peptides are linked together using a branched azetidinone linker and fused to the scaffold antibody in a site-specific manner under mild conditions. Although the pharmacophore is responsible for functional activity, the antibody scaffold confers a long half-life and Ig-like distribution. These pharmacophores can be chemically optimized or replaced with other pharmacophores to generate optimized or unique bispecific antibodies. (Doppalapudi VR et al., PNAS (2010), 107 (52); 22611-22616).

[0167] In some embodiments, the multispecific binding protein is in the form of an Oasc-Fab heterodimer, which comprises a Fab fragment that binds to target 1 and a scFab that binds to target 2 fused to an Fc. Heterodimerization is ensured by mutations in the Fc.

[0168] In some embodiments, the multispecific binding protein is in the form of DuetMab, which is a heterodimeric construct containing two different Fab fragments that bind to antigen 1 and antigen 2 and an Fc stabilized by heterodimerization mutations. Fab fragment 1 and Fab fragment 2 contain an SS bridge that ensures correct LC and HC pairing.

[0169] In some embodiments, the multispecific binding protein is in the form of a CrossmAb, which is a heterodimeric construct with two different Fab fragments (fused to an Fc stabilized by heterodimerization) that bind to target 1 and target 2. The CL domain and the CH1 domain, as well as the VH domain and the VL domain are interchanged, e.g., CH1 is fused in-frame with VL, and CL is fused in-frame with VH.

[0170] In some embodiments, the multispecific binding protein is in the form of Fit-Ig, which is a homodimeric construct in which the Fab fragment that binds to Antigen 2 is fused to the N-terminus of the HC of the Fab fragment that binds to Antigen 1. The construct contains a wild-type Fc.

[0171] The individual components of these multispecific binding proteins are described in more detail below.

[0172] NKG2D binding site

[0173] After binding to the NKG2D receptor and CD16 on natural killer cells and tumor-associated antigens on cancer cells, these multispecific binding proteins can engage more than one NK activating receptor and can block the binding of natural ligands to NKG2D. In certain embodiments, these proteins can excite NK cells in humans. In some embodiments, these proteins can excite NK cells in humans and other species (such as rodents and cynomolgus monkeys). In some embodiments, these proteins can excite NK cells in humans and other species (such as cynomolgus monkeys).

[0174] Table 1 lists the peptide sequences of the heavy chain variable domain and the light chain variable domain (which can bind to NKG2D after combination). In some embodiments, the heavy chain variable domain and the light chain variable domain are arranged in a Fab format. In some embodiments, the heavy chain variable domain and the light chain variable domain are fused together to form a scFv.

[0175] The NKG2D binding sites listed in Table 1 may have different binding affinities to NKG2D, nevertheless, they all activate human NK cells.

[0176] Unless otherwise indicated, the CDR sequences provided in Table 1 are identified according to Kabat numbering.

[0177]

[0178]

[0179]

[0180]

[0181]

[0182]

[0183]

[0184]

[0185]

[0186]

[0187]

[0188]

[0189]

[0190] In certain embodiments, the first antigen binding site that binds NKG2D (e.g., human NKG2D) comprises an antibody heavy chain variable domain (VH) comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to the VH of an antibody disclosed in Table 1, and an antibody light chain variable domain (VL) comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to the VL of the same antibody disclosed in Table 1. In certain embodiments, the first antigen binding site comprises the heavy chain CDR1, CDR2 and CDR3 and the light chain CDR1, CDR2 and CDR3 of the VH and VL sequences of the antibodies disclosed in Table 1, as determined by Kabat (see Kabat et al., (1991) Sequences of Proteins of Immunological Interest, NIH Publication No. 91-3242, Bethesda), Chothia (see, e.g., Chothia C & Lesk AM, (1987), J. Mol. Biol. 196: 901-917), MacCallum (see MacCallum RM et al., (1996) J. Mol. Biol. 262: 732-745) or any other CDR determination method known in the art. In certain embodiments, the first antigen binding site comprises the heavy chain CDR1, CDR2 and CDR3 and the light chain CDR1, CDR2 and CDR3 of the antibodies disclosed in Table 1.

[0191] In certain embodiments, the first antigen binding site that binds NKG2D comprises a heavy chain variable domain related to SEQ ID NO: 392, such as by having an amino acid sequence that is at least 90% (e.g., at least 90%, at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to SEQ ID NO: 392, and / or incorporating an amino acid sequence that is identical to the CDR1 (SEQ ID NO: 394), CDR2 (SEQ ID NO: 395), and CDR3 (SEQ ID NO: 396) sequences of SEQ ID NO: 392. A heavy chain variable domain having at least 90% sequence identity to SEQ ID NO: 392 can be coupled to a variety of light chain variable domains to form a NKG2D binding site. For example, a first antigen binding site incorporating a heavy chain variable domain having at least 90% sequence identity to SEQ ID NO:392 can further incorporate a light chain variable domain having at least 90% sequence identity to any one of the sequences selected from the group consisting of SEQ ID NO:393, 398, 400, 402, 404, 408, 410, 412, 414, 416, 418, 420, 422, 424, 426, 428, 430, 432, and 434. For example, the first antigen binding site incorporates a heavy chain variable domain having an amino acid sequence that is at least 90% (e.g., at least 90%, at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to SEQ ID NO: 392 and a light chain variable domain having an amino acid sequence that is at least 90% (e.g., at least 90%, at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to any one sequence selected from the group consisting of: SEQ ID NO: 392. NO: 393, 398, 400, 402, 404, 408, 410, 412, 414, 416, 418, 420, 422, 424, 426, 428, 430, 432 and 434.

[0192] In certain embodiments, the first antigen binding site that binds to NKG2D comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to the amino acid sequence of SEQ ID NO: 435, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to SEQ ID NO: 436. In certain embodiments, the VH comprises CDR1, CDR2, and CDR3 comprising the amino acid sequences of SEQ ID NO: 437 or 438, 439, and 442 or 443, respectively. In certain embodiments, the VL comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 440, 441 and 444, respectively. In certain embodiments, the first antigen binding site comprises (a) a VH comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 437 or 438, 439 and 442 or 443, respectively; and (b) a VL comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 440, 441 and 444, respectively.

[0193] In certain embodiments, the first antigen binding site that binds to NKG2D comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to the amino acid sequence of SEQ ID NO: 445, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to SEQ ID NO: 454. In certain embodiments, the VH comprises CDR1, CDR2, and CDR3 comprising the amino acid sequences of SEQ ID NO: 446 or 447, 448, and 449 or 450, respectively. In certain embodiments, the VL comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 451, 452 and 453, respectively. In certain embodiments, the first antigen binding site comprises (a) a VH comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 446 or 447, 448 and 449 or 450, respectively; and (b) a VL comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 451, 452 and 453, respectively.

[0194] In certain embodiments, the first antigen binding site that binds NKG2D comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to the amino acid sequence of SEQ ID NO:455, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to SEQ ID NO:456.

[0195] In certain embodiments, the first antigen binding site that binds to NKG2D comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to the amino acid sequence of SEQ ID NO: 457, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to SEQ ID NO: 458. In certain embodiments, the VH comprises CDR1, CDR2, and CDR3 comprising the amino acid sequences of SEQ ID NOs: 437, 459, and 460, respectively. In certain embodiments, the VL comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 461, 441 and 462, respectively. In certain embodiments, the first antigen binding site comprises (a) a VH comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 437, 459 and 460, respectively; and (b) a VL comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 461, 441 and 462, respectively.

[0196] In certain embodiments, the first antigen binding site that binds to NKG2D comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to the amino acid sequence of SEQ ID NO: 463, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to SEQ ID NO: 464. In certain embodiments, the VH comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NO: 465 or 466, 467 and 468 or 469, respectively. In certain embodiments, the VL comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 470, 63 and 472, respectively. In certain embodiments, the first antigen binding site comprises (a) a VH comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 465 or 466, 467 and 468 or 469, respectively; and (b) a VL comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 470, 63 and 472, respectively.

[0197] In certain embodiments, the first antigen binding site that binds to NKG2D comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to the amino acid sequence of SEQ ID NO: 473, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to SEQ ID NO: 474. In certain embodiments, the VH comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NO: 475 or 476, 477 and 478 or 479, respectively. In certain embodiments, the VL comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 480, 63 and 481, respectively. In certain embodiments, the first antigen binding site comprises (a) a VH comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 475 or 476, 477 and 478 or 479, respectively; and (b) a VL comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 480, 63 and 481, respectively.

[0198] In certain embodiments, the first antigen binding site that binds to NKG2D comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to the amino acid sequence of SEQ ID NO: 501, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to SEQ ID NO: 502. In certain embodiments, the VH comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NO: 465 or 466, 503 and 504 or 505, respectively. In certain embodiments, the VL comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 506, 452 and 507, respectively. In certain embodiments, the first antigen binding site comprises (a) a VH comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 465 or 466, 503 and 504 or 505, respectively; and (b) a VL comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 506, 452 and 507, respectively.

[0199] In certain embodiments, the first antigen binding site that binds to NKG2D comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to the amino acid sequence of SEQ ID NO: 482, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to SEQ ID NO: 483. In certain embodiments, the VH comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NO: 484 or 3, 486 and 487 or 488, respectively. In certain embodiments, the VL comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 489, 490 and 491, respectively. In certain embodiments, the first antigen binding site comprises (a) a VH comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 484 or 3, 486 and 487 or 488, respectively; and (b) a VL comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 489, 490 and 491, respectively.

[0200] In certain embodiments, the first antigen binding site that binds to NKG2D comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to the amino acid sequence of SEQ ID NO: 492, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to SEQ ID NO: 493. In certain embodiments, the VH comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NO: 494 or 495, 496 and 497 or 498, respectively. In certain embodiments, the VL comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 530, 224 and 499, respectively. In certain embodiments, the first antigen binding site comprises (a) a VH comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 494 or 495, 496 and 497 or 498, respectively; and (b) a VL comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 530, 224 and 499, respectively.

[0201] In certain embodiments, the first antigen binding site that binds to NKG2D comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to the amino acid sequence of SEQ ID NO: 508, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to SEQ ID NO: 493. In certain embodiments, the VH comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NO: 494 or 495, 496 and 509 or 510, respectively. In certain embodiments, the VL comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 530, 224 and 499, respectively. In certain embodiments, the first antigen binding site comprises (a) a VH comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 494 or 495, 496 and 509 or 510, respectively; and (b) a VL comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 530, 224 and 499, respectively.

[0202] In certain embodiments, the first antigen binding site that binds to NKG2D comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to the amino acid sequence of SEQ ID NO: 511, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to SEQ ID NO: 493. In certain embodiments, the VH comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NO: 494 or 495, 496 and 512 or 513, respectively. In certain embodiments, the VL comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 530, 224 and 499, respectively. In certain embodiments, the first antigen binding site comprises (a) a VH comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 494 or 495, 496 and 512 or 513, respectively; and (b) a VL comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 530, 224 and 499, respectively.

[0203] In certain embodiments, the first antigen binding site that binds to NKG2D comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to the amino acid sequence of SEQ ID NO: 514, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to SEQ ID NO: 493. In certain embodiments, the VH comprises CDR1, CDR2, and CDR3 comprising the amino acid sequences of SEQ ID NO: 494 or 495, 496, and 515 or 516, respectively. In certain embodiments, the VL comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 530, 224 and 499, respectively. In certain embodiments, the first antigen binding site comprises (a) a VH comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 494 or 495, 496 and 515 or 516, respectively; and (b) a VL comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 530, 224 and 499, respectively.

[0204] In certain embodiments, the first antigen binding site that binds to NKG2D comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to the amino acid sequence of SEQ ID NO: 517, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to SEQ ID NO: 493. In certain embodiments, the VH comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NO: 494 or 495, 496 and 518 or 519, respectively. In certain embodiments, the VL comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 530, 224 and 499, respectively. In certain embodiments, the first antigen binding site comprises (a) a VH comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 494 or 495, 496 and 518 or 519, respectively; and (b) a VL comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 530, 224 and 499, respectively.

[0205] In certain embodiments, the first antigen binding site that binds to NKG2D comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to the amino acid sequence of SEQ ID NO: 520, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to SEQ ID NO: 493. In certain embodiments, the VH comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NO: 494 or 495, 496 and 521 or 522, respectively. In certain embodiments, the VL comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 530, 224 and 499, respectively. In certain embodiments, the first antigen binding site comprises (a) a VH comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 494 or 495, 496 and 521 or 522, respectively; and (b) a VL comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 530, 224 and 499, respectively.

[0206] In certain embodiments, the first antigen binding site that binds to NKG2D comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to the amino acid sequence of SEQ ID NO: 523, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to SEQ ID NO: 493. In certain embodiments, the VH comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NO: 494 or 495, 496 and 524 or 525, respectively. In certain embodiments, the VL comprises CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 530, 224 and 499, respectively. In certain embodiments, the first antigen binding site comprises (a) a VH comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 494 or 495, 496 and 524 or 525, respectively; and (b) a VL comprising CDR1, CDR2 and CDR3 comprising the amino acid sequences of SEQ ID NOs: 530, 224 and 499, respectively.

[0207] In certain embodiments, the first antigen binding site that binds NKG2D comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to the amino acid sequence of SEQ ID NO:526, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to SEQ ID NO:527.

[0208] In certain embodiments, the first antigen binding site that binds NKG2D comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to the amino acid sequence of SEQ ID NO:528, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to SEQ ID NO:529.

[0209] These multispecific binding proteins can bind to cells expressing NKG2D, including but not limited to NK cells, γδ T cells, and CD8 + αβ T cells. After NKG2D binding, these multispecific binding proteins can block natural ligands (such as ULBP6 and MICA) from binding to NKG2D and activating NK cells.

[0210] These multispecific binding proteins bind to cells expressing CD16, an Fc receptor on the surface of leukocytes including natural killer cells, macrophages, neutrophils, eosinophils, mast cells and follicular dendritic cells. D binds to NKG2D with an affinity of 2 nM to 120 nM, e.g., 2 nM to 110 nM, 2 nM to 100 nM, 2 nM to 90 nM, 2 nM to 80 nM, 2 nM to 70 nM, 2 nM to 60 nM, 2 nM to 50 nM, 2 nM to 40 nM, 2 nM to 30 nM, 2 nM to 20 nM, 2 nM to 10 nM, about 15 nM, about 14 nM, about 13 nM, about 12 nM, about 11 nM, about 10 nM, about 9 nM, about 8 nM, about 7 nM, about 6 nM, about 5 nM, about 4.5 nM, about 4 nM, about 3.5 nM, about 3 nM, In some embodiments, the NKG2D binding site has a K of 10 to 62 nM. D Binds to NKG2D.

[0211] CEACAM5-binding site

[0212] The CEACAM5 binding site of the multispecific binding protein disclosed herein comprises a heavy chain variable domain and a light chain variable domain. Table 2 lists some exemplary sequences of heavy chain variable domains and light chain variable domains, which, when combined, can bind to CEACAM5. As shown, CDR sequences are identified according to Chothia and Kabat numbering. Cysteine ​​mutations for disulfide bond formation are underlined. The scFv sequence comprises a (G4S)4 linker (SEQ ID NO: 532) (italics) between VH and VL.

[0213] Table 2. Sequences of exemplary antigen binding sites that bind to CEACAM5

[0214]

[0215]

[0216]

[0217]

[0218]

[0219]

[0220]

[0221]

[0222]

[0223]

[0224]

[0225]

[0226]

[0227]

[0228] Alternatively, novel antigen binding sites that can bind to CEACAM5 can be identified by screening for binding to the amino acid sequence defined by SEQ ID NO: 391, its mature extracellular fragment or a fragment containing the CEACAM5 domain (see, e.g., U.S. Pat. Nos. 9,771,431, 9,617,345 and 8,470,994, and U.S. Application Nos. 15 / 683,087 and 14 / 515,765).

[0229] An exemplary sequence of a human CEACAM5 isoform is provided below and is available from the GenBank database under Accession No. NP_004354.

[0230]

[0231]

[0232] In certain embodiments, the second antigen binding site that binds CEACAM5 (e.g., human CEACAM5, e.g., cynomolgus monkey CEACAM5) comprises a heavy chain variable domain (VH) comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to the VH of the antigen binding site disclosed in Table 2, and a light chain variable domain (VL) comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to the VL of the same antigen binding site disclosed in Table 2. In certain embodiments, the second antigen binding site comprises the heavy chain CDR1, CDR2 and CDR3 and the light chain CDR1, CDR2 and CDR3 of the VH and VL sequences of the antigen binding sites disclosed in Table 2, as determined by Kabat (see Kabat et al., (1991) Sequences of Proteins of Immunological Interest, NIH Publication No. 91-3242, Bethesda), Chothia (see, e.g., Chothia C & Lesk AM, (1987), J Mol Biol 196: 901-917), MacCallum (see MacCallum RM et al., (1996) J Mol Biol 262: 732-745) or any other CDR determination method known in the art. In certain embodiments, the second antigen binding site comprises the heavy chain CDR1, CDR2 and CDR3 sequences and the light chain CDR1, CDR2 and CDR3 sequences of the antigen binding sites disclosed in Table 2.

[0233] In certain embodiments, the second antigen binding site is associated with an scFv in Table 2. For example, in certain embodiments, the second antigen binding site comprises a VH sequence comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to an amino acid sequence of a VH in Table 2. In certain embodiments, the second antigen binding site comprises a VL sequence comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to an amino acid sequence of a VL in Table 2. In certain embodiments, the second antigen-binding site comprises a VH sequence having an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to the amino acid sequence of the VH in Table 2, and a VL sequence having an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to the amino acid sequence of the VL in Table 2, wherein the VH and VL in Table 2 are selected from sequence homologous pairs.

[0234] In certain embodiments, the VH comprises CDR1, CDR2 and CDR3 selected from the sequence homology pairs listed in Table 2. In certain embodiments, the VL comprises CDR1, CDR2 and CDR3 selected from the sequence homology pairs listed in Table 2. In certain embodiments, the second antigen binding site comprises (a) a VH comprising CDR1, CDR2 and CDR3 of a VH sequence listed in Table 2; and (b) a VL comprising CDR1, CDR2 and CDR3 of a VL sequence listed in Table 2, wherein the VH sequence and the VL sequence are selected from the sequence homology pairs listed in Table 2.

[0235] As used herein, the term "cognate pairing" refers to VH and VL that form an antigen binding site. In some embodiments, "cognate pairing" refers to VH and VL pairing as shown in Table 2. In some embodiments, "cognate pairing" refers to VH and VL pairing as shown in Table 3.

[0236] As used in Table 2, the term "derived from" when applied to VH, VL or CDR refers to an amino acid sequence with additional mutations (e.g., substitutions, deletions, etc.) relative to the reference sequence. For example, the VH of the homologous pairing A1 in Table 2 (SEQ ID NO: 567) is derived from the VH of PH_420-CEACAM5 (shown in Table 3). SEQ ID NO: 567 has a cysteine ​​mutation relative to the VH of PH_420-CEACAM5. The VL of the homologous pairing A1 in Table 2 (SEQ ID NO: 568) is derived from the VL of PH_420-CEACAM5 (shown in Table 3). SEQ ID NO: 568 has a cysteine ​​mutation relative to the VL of PH_420-CEACAM5.

[0237] As used in Table 2, the term "derived from" when applied to scFv refers to an amino acid sequence with additional mutations and / or linker sequences. For example, the scFv of GB1 is derived from the homologous pairing A1 in Table 2 and comprises the VH sequence and VL sequence of the homologous pairing A1 in Table 2 and a linker sequence (e.g., (G4S)4 linker sequence (SEQ ID NO: 532)).

[0238] Table 3 lists exemplary sequences of heavy chain variable domains and light chain variable domains that, when combined (eg, as a cognate pair), can bind to CEACAM5. As indicated, CDR sequences are identified according to Chothia and Kabat numbering.

[0239] Table 3. Sequences of exemplary antigen binding sites that bind to CEACAM5

[0240]

[0241]

[0242]

[0243]

[0244]

[0245]

[0246]

[0247]

[0248]

[0249]

[0250]

[0251]

[0252]

[0253]

[0254]

[0255]

[0256]

[0257]

[0258]

[0259]

[0260]

[0261]

[0262]

[0263]

[0264]

[0265]

[0266]

[0267]

[0268]

[0269]

[0270]

[0271]

[0272]

[0273]

[0274]

[0275]

[0276]

[0277]

[0278]

[0279]

[0280]

[0281]

[0282]

[0283]

[0284]

[0285]

[0286]

[0287]

[0288]

[0289]

[0290]

[0291]

[0292]

[0293]

[0294]

[0295]

[0296]

[0297]

[0298]

[0299]

[0300]

[0301]

[0302]

[0303]

[0304]

[0305]

[0306]

[0307]

[0308]

[0309]

[0310]

[0311]

[0312]

[0313]

[0314]

[0315]

[0316]

[0317]

[0318]

[0319]

[0320]

[0321]

[0322] Table 4. Sequences of exemplary antigen binding sites that bind to CEACAM5

[0323]

[0324]

[0325]

[0326]

[0327]

[0328]

[0329]

[0330]

[0331]

[0332]

[0333]

[0334]

[0335]

[0336]

[0337]

[0338]

[0339]

[0340]

[0341]

[0342]

[0343]

[0344]

[0345]

[0346]

[0347]

[0348]

[0349]

[0350]

[0351]

[0352]

[0353]

[0354]

[0355]

[0356]

[0357]

[0358]

[0359]

[0360]

[0361]

[0362]

[0363]

[0364]

[0365]

[0366]

[0367]

[0368]

[0369]

[0370]

[0371]

[0372]

[0373]

[0374]

[0375]

[0376]

[0377]

[0378]

[0379]

[0380]

[0381]

[0382]

[0383]

[0384]

[0385]

[0386]

[0387]

[0388]

[0389]

[0390]

[0391]

[0392]

[0393] In certain embodiments, the second antigen binding site that binds CEACAM5 (e.g., human CEACAM5, e.g., cynomolgus monkey CEACAM5) comprises an antibody heavy chain variable domain (VH) comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to the VH of an antibody disclosed in Table 3, and an antibody light chain variable domain (VL) comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to the VL of the same antibody disclosed in Table 3 or Table 4. In certain embodiments, the second antigen binding site comprises the heavy chain CDR1, CDR2 and CDR3 and the light chain CDR1, CDR2 and CDR3 of the VH and VL sequences of the antigen binding sites disclosed in Table 3 or Table 4, as determined by Kabat (see Kabat et al., (1991) Sequences of Proteins of Immunological Interest, NIH Publication No. 91-3242, Bethesda), Chothia (see, e.g., Chothia C & Lesk AM, (1987), J Mol Biol 196: 901-917), MacCallum (see MacCallum RM et al., (1996) J Mol Biol 262: 732-745) or any other CDR determination method known in the art. In certain embodiments, the second antigen binding site comprises the heavy chain CDR1, CDR2 and CDR3 and the light chain CDR1, CDR2 and CDR3 of the antigen binding sites disclosed in Table 3 or Table 4.

[0394] In certain embodiments, the second antigen binding site is associated with a scFv having a VH and a VL in Table 3. For example, in certain embodiments, the second antigen binding site comprises a VH having an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to an amino acid sequence of a VH in Table 3. In certain embodiments, the second antigen binding site comprises a VL having an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to an amino acid sequence of a VL in Table 3. In certain embodiments, the second antigen-binding site comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to the amino acid sequence of the VL in Table 3, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to the amino acid sequence of the VL in Table 3, wherein the VH sequence and VL sequence are selected from the sequence homologous pairs listed in Table 3.

[0395] In certain embodiments, the VH comprises a CDR1, CDR2, and CDR3 selected from the VH sequences listed in Table 3. In certain embodiments, the VL comprises a CDR1, CDR2, and CDR3 selected from the VL sequences listed in Table 3. In certain embodiments, the second antigen binding site comprises (a) a VH comprising a CDR1, CDR2, and CDR3 selected from the VH sequences listed in Table 3; and (b) a VL comprising a CDR1, CDR2, and CDR3 selected from the VL sequences listed in Table 3, wherein the VH sequence and the VL sequence are selected from the sequence cognate pairs listed in Table 3.

[0396] In certain embodiments, the second antigen binding site is associated with a scFv having a VH and a VL in Table 4. For example, in certain embodiments, the second antigen binding site comprises a VH comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to an amino acid sequence of a VH in Table 4. In certain embodiments, the second antigen binding site comprises a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to an amino acid sequence of a VL in Table 4. In certain embodiments, the second antigen-binding site comprises a VH having an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to the amino acid sequence of the VH in Table 4, and a VL comprising an amino acid sequence that is at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to the amino acid sequence of the VL in Table 4, wherein the VH and VL in Table 4 are selected from the same clone listed in Table 4.

[0397] In certain embodiments, the VH comprises CDR1, CDR2, and CDR3 selected from the VH sequences of Table 4. In certain embodiments, the VL comprises CDR1, CDR2, and CDR3 selected from the VL sequences of Table 4. In certain embodiments, the second antigen binding site comprises (a) a VH comprising CDR1, CDR2, and CDR3 selected from the VH sequences of Table 4; and (b) a VL comprising CDR1, CDR2, and CDR3 selected from the VL sequences of Table 4, wherein the VH and VL in Table 4 are selected from the same clone listed in Table 4.

[0398] In each of the foregoing embodiments, it is contemplated herein that the VH and / or VL sequences that together bind CEACAM5 may contain amino acid alterations (e.g., at least 1, 2, 3, 4, 5 or 10 amino acid substitutions, deletions or additions) in the framework regions of the VH and / or VL without significantly affecting their ability to bind to CEACAM5.

[0399] In certain embodiments, an antigen binding site of the invention derived from a cognate pair of Tier 1 in Table 3 or from a clone of Tier 1 in Table 4 has a K of less than or equal to 25 nM (e.g., less than or equal to 24 nM, 23 nM, 22 nM, 21 nM, 20 nM, 19 nM, 18 nM, 17 nM, 16 nM, 15 nM, 14 nM, 13 nM, 12 nM, 11 nM, 10 nM, 9 nM, 8 nM, 7 nM, 6 nM, 5 nM, 4 nM, 3 nM, 2 nM, or 1 nM). D The value of binding to human CEACAM5 or CEACAM5 variants or their extracellular regions is understood to be smaller than D In certain embodiments, an antigen binding site of the invention derived from a homologous pairing of Tier 2 in Table 3 or from a clone of Tier 2 in Table 4 has an affinity of greater than or equal to 15 nM (e.g., greater than or equal to 20 nM, 25 nM, 30 nM, 40 nM, 50 nM, 60 nM, 70 nM, 80 nM, 90 nM, 100 nM, 110 nM, 120 nM, 130 nM, 140 nM, 150 nM, 500nM, 520nM, 540nM or 560nM) D In certain embodiments, the antigen binding sites of the invention derived from the homologous pairs of level 2 in Table 3 or from the clones of level 2 in Table 4 are expressed at 15-560 nM, 15-400 nM, 15-300 nM, 15-200 nM, 15-100 nM, 15-80 nM, 20-560 nM, 20-400 nM, 20-300 nM, 20-200 nM, 20-100 nM, 20-80 nM, 25-100 nM, 25-560 nM, 25-400 nM, 20-300 nM, 20-200 nM, 20-100 nM, 20-80 nM, 25-100 nM, 25-560 nM, 25-400 nM, 0nM, 25-300nM, 25-200nM, 25-100nM, 25-80nM, 50-560nM, 50-400nM, 50-300nM, 50-200nM, 50-100nM, 50-80nM, 100-560nM, 100-400nM, 100-300nM, 100-200nM, 120-560nM, 120-400nM, 120-300nM, or a K in the range of 120-200nM. D The value binds to human CEACAM5 or CEACAM5 variants or their extracellular regions.

[0400] In certain embodiments, as disclosed herein, the antigen binding site comprises a VH sequence and a VL sequence that are at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to a cognate pair of Tier 1 in Table 3 or a clone of Tier 1 in Table 4, respectively, and has a K of less than or equal to 25 nM (e.g., less than or equal to 24 nM, 23 nM, 22 nM, 21 nM, 20 nM, 19 nM, 18 nM, 17 nM, 16 nM, 15 nM, 14 nM, 13 nM, 12 nM, 11 nM, 10 nM, 9 nM, 8 nM, 7 nM, 6 nM, 5 nM, 4 nM, 3 nM, 2 nM, or 1 nM). D In some embodiments, the antigen binding site comprises a VH sequence and a VL sequence that are at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to a homologous pair of Tier 2 in Table 3 or a clone of Tier 2 in Table 4, respectively, as disclosed herein, and has a binding affinity of greater than or equal to 15 nM (e.g., greater than or equal to 20 nM, 25 nM, 30 nM, 40 nM, 50 nM, 60 nM, 70 nM, 80 nM, 90 nM, 100 nM, 150 nM, 160 nM, 170 nM, 180 nM, 190 nM, 200 nM, 210 nM, 220 nM, 230 nM, 240 nM, 250 nM, 260 nM, 270 nM, 280 nM, 290 nM, 300 nM, 310 nM, 320 nM, 330 nM, 340 nM, 350 nM, 360 nM, 370 nM, 380 nM, 390 nM, 400 nM, 410 nM, 420 nM, 430 nM, 440 nM, 450 nM, 460 nM, 470 nM, 480 nM, 490 nM, 500 nM, 510 nM, 520 nM, 530 nM, 540 nM, 560 nM, 570 nM, 5 , 70nM, 80nM, 90nM, 100nM, 110nM, 120nM, 130nM, 140nM, 150nM, 160nM, 170nM, 180nM, 190nM, 200nM, 220nM, 240nM, 260nM, 280nM, 300nM, 320nM, 340nM, 360nM, 380nM, 400nM, 420nM, 440nM, 460nM, 480nM, 500nM, 520nM, 540nM or 560nM) DIn certain embodiments, the antigen binding site comprises a VH sequence and a VL sequence that are at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100%) identical to a homologous pair of Tier 2 in Table 3 or a clone of Tier 2 in Table 4, respectively, and is expressed at 15-560 nM, 15-400 nM, 15-300 nM, 15-200 nM, 15-100 nM, 15-80 nM, 20-560 nM, 20-400 nM, 20-300 nM, 20- K in the range of 20-200 nM, 20-100 nM, 20-80 nM, 25-100 nM, 25-560 nM, 25-400 nM, 25-300 nM, 25-200 nM, 25-100 nM, 25-80 nM, 50-560 nM, 50-400 nM, 50-300 nM, 50-200 nM, 50-100 nM, 50-80 nM, 100-560 nM, 100-400 nM, 100-300 nM, 100-200 nM, 120-560 nM, 120-400 nM, 120-300 nM or 120-200 nM D The value binds to human CEACAM5 or CEACAM5 variants or their extracellular regions.

[0401] In certain embodiments, an antigen binding site of the invention derived from a cognate pair of Tier 1 or Tier 2 in Table 3 or a clone of Tier 1 or Tier 2 in Table 4 does not bind to CEACAM1, CEACAM6 or CEACAM8 at a detectable level (e.g., for in vitro binding as detected by surface plasmon resonance (SPR) or enzyme-linked immunosorbent assay (ELISA), or for binding to cells expressing the corresponding antigen as detected by flow cytometry). In certain embodiments, an antigen binding site of the invention derived from a cognate pair of Tier 3 in Table 3 or a clone of Tier 3 in Table 4 binds to CEACAM1, CEACAM6 or CEACAM8 at a detectable level (e.g., with a K of greater than or equal to 100 nM, 200 nM, 500 nM, 1 μM, 2 μM, 5 μM or 10 μM). D value) in combination with CEACAM1, CEACAM6 or CEACAM8.

[0402] In certain embodiments, as disclosed herein, the antigen binding site comprises a VH sequence and a VL sequence that are at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to a cognate pair of Tier 1 or Tier 2 in Table 3 or a clone of Tier 1 or Tier 2 in Table 4, respectively, and does not bind to CEACAM1, CEACAM6, or CEACAM8 at a detectable level (e.g., for in vitro binding, as detected by surface plasmon resonance (SPR) or enzyme-linked immunosorbent assay (ELISA), or for binding to cells expressing the corresponding antigen, as detected by flow cytometry). In certain embodiments, as disclosed herein, the antigen binding site comprises a VH sequence and a VL sequence that are at least 90% (e.g., at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100%) identical to a cognate pair of Tier 3 in Table 3 or a clone of Tier 3 in Table 4, respectively, and binds to the antigen binding site at a detectable level (e.g., with a K of greater than or equal to 100 nM, 200 nM, 500 nM, 1 μM, 2 μM, 5 μM, or 10 μM). D value) in combination with CEACAM1, CEACAM6 or CEACAM8.

[0403] In certain embodiments of any of the antigen binding sites disclosed herein, when the N-terminal amino acid of the heavy chain variable region is Gln (Q), the Q can be replaced with Glu (E), thereby generating a variant called "QE". In certain embodiments of any of the antigen binding sites disclosed herein, when the N-terminal amino acid of the heavy chain variable region is Gln (Q) or Glu (E), the Q or E can be replaced with pyroglutamic acid (pE). The antigen binding site generated by this replacement falls into the same level as the parent antigen binding site.

[0404] In certain embodiments, the second antigen binding site competes with the above antigen binding site for binding to CEACAM5 (eg, human CEACAM5, eg, cynomolgus monkey CEACAM5). In certain embodiments, the antigen binding sites of the invention compete with antigen binding sites associated with clones selected from Table 2, wherein the clones are selected from the group consisting of clone PH_420-CEACAM5, clone 1078_C04-CEACAM5, clone 1079_H05-CEACAM5, 7A10.A7-CEACAM5-B.01, 8H2.B10-CEACAM5-B.01, murine 16F6.A2-CEACAM5-B.02, humanized 16F6.A2-CEACAM5-B.02-BM, PH_415-CEACAM5, PH_416-CEACAM5, PH_418-CEACAM5, PH_419-CEACAM5, PH_417-CEACAM5, PH_421-CEACAM5 AM5, 1078_G03-CEACAM5, mouse 1A1.A3-CEACAM5-B.02, humanized 1A1.A3-CEACAM5-B.02-BM, 1080 _G01-CEACAM5, 1078_C12-CEACAM5, 1078_F02-CEACAM5, 1079_B08-CEACAM5, 1078_G03 -CEACAM5, 1079_A10-CEACAM5, 1079_A12-CEACAM5, 1078_C04-CEACAM5, 1080_F11-CEA CAM5, 1081_E01-CEACAM5, 1083_A05-CEACAM5, 1085_D12-CEACAM5, 1079_G12-CEACAM5 , 1080_A01-CEACAM5 , 12C7.A2-CEACAM5-B.01 , 12A6.H2-CEACAM5-B.01 , 4G3.C3-CEACAM5-B.01 , 4B10.B3-CEACAM5-A.02 , 13C7.A6-CEACAM5-B.02 , 7E11.B2-CEACAM5-B.02, 10D6.E3-CEACAM5-B.02, 13C7.F2-CEACAM5-B.02, 16B11.G2-CEACAM5-B.01 and 6D10.C8-CEACAM5-B.02. In some embodiments, an antigen binding site of the invention competes with an antigen binding site comprising a VL sequence and a VH sequence selected from Table 1, wherein the VH sequence and the VL sequence are from a cognate pair in Table 1 or a clone in Table 2.In some embodiments, an antigen binding site of the invention competes with an antigen binding site comprising CDRH1, CDRH2, CDRH3, CDRL1, CDRL2, and CDRL3, wherein these CDRs are selected from the cognate pairs listed in Table 1 or the clones listed in Table 2.

[0405] Fc domain

[0406] In the Fc domain, CD16 binding is mediated by the hinge region and the CH2 domain. For example, in human IgG1, the interaction with CD16 is mainly concentrated on the carbohydrate residues N-acetyl-D-glucosamine in the amino acid residues Asp 265-Glu269, Asn 297-Thr 299, Ala 327-Ile 332, Leu 234-Ser 239 and CH2 domain (see, Sondermann et al., Nature, 406 (6793): 267-273). Therefore, in some embodiments, the antibody Fc domain or a portion thereof comprises a hinge and a CH2 domain. Based on a known domain, mutations can be selected to enhance or weaken the binding affinity to CD16, such as by using a phage display library or a yeast surface display cDNA library to select, or can be designed based on a known three-dimensional structure of the interaction. Binding of CD16 to an Fc domain can be determined by conventional methods, for example, by surface plasmon resonance (SPR) or enzyme-linked immunosorbent assay (ELISA) for in vitro binding, or by flow cytometry for binding to cells expressing CD16 on their cell surface.

[0407] The assembly of heterodimeric antibody heavy chain can be completed by expressing two different antibody heavy chain sequences in the same cell, which can result in the assembly of homodimers of every kind of antibody heavy chain and the assembly of heterodimers. The preferred assembly of heterodimers can be promoted by incorporating different mutations into the CH3 domain of each antibody heavy chain constant region to achieve, as shown in US13 / 494870, US16 / 028850, US11 / 533709, US12 / 875015, US13 / 289934, US14 / 773418, US12 / 811207, US13 / 866756, US14 / 647480 and US14 / 830336. For example, mutations can be carried out in the CH3 domain based on human IgG1, and different amino acid replacement pairs are incorporated into the first polypeptide and the second polypeptide, thereby allowing the two chains to selectively heterodimerize each other. As in Kabat, the amino acid replacement positions described below are all numbered according to the EU index.

[0408] In one case, the amino acid substitution in the first polypeptide replaces the original amino acid with a larger amino acid selected from arginine (R), phenylalanine (F), tyrosine (Y) or tryptophan (W), and at least one amino acid substitution in the second polypeptide replaces the original amino acid with a smaller amino acid selected from alanine (a), serine (S), threonine (T) or valine (V), such that the larger amino acid substitution (protrusion) is incorporated into the surface of the (these) smaller amino acid substitutions (cavities). For example, one polypeptide can incorporate a T366W substitution, while another polypeptide can incorporate three substitutions, including T366S, L368A and Y407V.

[0409] The antibody heavy chain variable domain of the present invention can be optionally coupled to an amino acid sequence that is at least 90% identical to an antibody constant region, such as an IgG constant region comprising a hinge, a CH2 domain, and a CH3 domain (with or without a CH1 domain). In some embodiments, the amino acid sequence of the constant region is at least 90% identical to a human antibody constant region (such as a human IgG1 constant region, an IgG2 constant region, an IgG3 constant region, or an IgG4 constant region). In one embodiment, an antibody Fc domain or a portion thereof sufficient to bind CD16 comprises a sequence identical to that of a wild-type human IgG1 Fc region.

[0410] In some other embodiments, the amino acid sequence of the constant region is at least 90% identical to the constant region of an antibody from another mammal, such as a rabbit, dog, cat, mouse, or horse.

[0411] In some embodiments, the antibody constant domain connected to the scFv or Fab fragment is capable of binding to CD 16. In some embodiments, the protein incorporates a portion of an antibody Fc domain (e.g., a portion of an antibody Fc domain sufficient to bind to CD 16), wherein the antibody Fc domain comprises a hinge and a CH2 domain (e.g., a hinge and a CH2 domain of a human IgG1 antibody) and / or an amino acid sequence that is at least 90% identical to amino acid sequence 234-332 of a human IgG antibody.

[0412] One or more mutations can be incorporated into the constant region compared to a human IgG1 constant region, for example at Q347, Y349, L351, S354, E356, E357, K360, Q362, S364, T366, L368, K370, N390, K392, T394, D399, S400, D401, F405, Y407, K409, T411 and / or K439. Exemplary substitutions include, e.g., Q347E, Q347R, Y349S, Y349K, Y349T, Y349D, Y349E, Y349C, T350V, L351K, L351D, L351Y, S354C, E356K, E357Q, E357L, E357W, K360E, K360W, Q362E, S364K, S364E, S364H, S364D, T366V, T366I, T366L, T366M, T366K, T366W, T366S, L368E, L368A, L368D, K370S, N390D, N390E, K392L, K392M, K392V, K392F, K392D, K392E, T394F, T394W, D399R, D399K, D399V, S400K, S400R, D401K, F405A, F405T, Y407A, Y407I, Y407V, K409F, K409W, K409D, K409R, T411D, T411E, K439D, and K439E.

[0413] In certain embodiments, the mutations of CH1 that can be incorporated into the human IgG1 constant region can be located at amino acids V125, F126, P127, T135, T139, A140, F170, P171 and / or V173. In certain embodiments, the mutations of Cκ that can be incorporated into the human IgG1 constant region can be located at amino acids E123, F116, S176, V163, S174 and / or T164.

[0414] Alternatively, the amino acid substitutions may be selected from the following substitution groups shown in Table 5.

[0415]

[0416]

[0417] Alternatively, the amino acid substitutions may be selected from the following substitution groups shown in Table 6.

[0418]

[0419] Alternatively, the amino acid substitutions may be selected from the following substitution groups shown in Table 7.

[0420]

[0421] Alternatively, at least one amino acid substitution in each polypeptide chain may be selected from Table 8.

[0422]

[0423] Alternatively, at least one amino acid substitution can be selected from the following substitution group in Table 9, wherein the position(s) indicated in the first polypeptide column is substituted with any known negatively charged amino acid, and the position(s) indicated in the second polypeptide column is substituted with any known positively charged amino acid.

[0424]

[0425] Alternatively, at least one amino acid substitution can be selected from the following group in Table 10, wherein the position(s) indicated in the first polypeptide column is substituted with any known positively charged amino acid, and the position(s) indicated in the second polypeptide column is substituted with any known negatively charged amino acid.

[0426]

[0427] Alternatively, the amino acid substitutions may be selected from the following groups in Table 11.

[0428]

[0429] Alternatively, or additionally, the structural stability of the heteromultimeric protein can be improved by introducing S354C on the first or second polypeptide chain and Y349C on the opposite polypeptide chain, which forms an artificial disulfide bridge within the interface of the two polypeptides.

[0430] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at position T366, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: T366, L368 and Y407.

[0431] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: T366, L368 and Y407, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at position T366.

[0432] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: E357, K360, Q362, S364, L368, K370, T394, D401, F405 and T411, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: Y349, E357, S364, L368, K370, T394, D401, F405 and T411.

[0433] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: Y349, E357, S364, L368, K370, T394, D401, F405 and T411, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: E357, K360, Q362, S364, L368, K370, T394, D401, F405 and T411.

[0434] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: L351, D399, S400 and Y407, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: T366, N390, K392, K409 and T411.

[0435] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: T366, N390, K392, K409 and T411, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: L351, D399, S400 and Y407.

[0436] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: Q347, Y349, K360 and K409, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: Q347, E357, D399 and F405.

[0437] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: Q347, E357, D399 and F405, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: Y349, K360, Q347 and K409.

[0438] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: K370, K392, K409 and K439, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: D356, E357 and D399.

[0439] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: D356, E357 and D399, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: K370, K392, K409 and K439.

[0440] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: L351, E356, T366 and D399, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: Y349, L351, L368, K392 and K409.

[0441] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: Y349, L351, L368, K392 and K409, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region at one or more positions selected from the group consisting of: L351, E356, T366 and D399.

[0442] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region by an S354C substitution, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region by a Y349C substitution.

[0443] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region by a Y349C substitution, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region by a S354C substitution.

[0444] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region by K360E and K409W substitutions, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region by Q347R, D399V and F405T substitutions.

[0445] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region by Q347R, D399V and F405T substitutions, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region by K360E and K409W substitutions.

[0446] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region by T366W substitution, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region by T366S, T368A and Y407V substitutions.

[0447] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region by T366S, T368A and Y407V substitutions, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region by T366W substitution.

[0448] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region by T350V, L351Y, F405A and Y407V substitutions, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region by T350V, T366L, K392L and T394W substitutions.

[0449] In some embodiments, the amino acid sequence of one polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region by T350V, T366L, K392L and T394W substitutions, and wherein the amino acid sequence of the other polypeptide chain of the antibody constant region differs from the amino acid sequence of the IgG1 constant region by T350V, L351Y, F405A and Y407V substitutions.

[0450] Exemplary Multispecific Binding Proteins

[0451] Listed below are examples of TriNKETs comprising an antigen binding site that binds CEACAM5 and an antigen binding site that binds NKG2D, each linked to an antibody constant region comprising mutations that enable heterodimerization of the two Fc chains. The CDR sequences according to Chothia are in bold, while the CDR sequences according to Kabat are underlined.

[0452] It is contemplated that the TriNKET is in F3 format, i.e., the antigen binding site that binds CEACAM5 is a Fab, and the antigen binding site that binds NKG2D is a scFv. All TriNKETs shown below are in F3' format, i.e., the antigen binding site that binds CEACAM5 is a scFv, and the antigen binding site that binds NKG2D is a Fab. In each TriNKET, the scFv contains a replacement of Cys in the VH and VL regions, promoting the formation of a disulfide bridge between the VH and VL of the scFv.

[0453] The VH and VL of the scFv can be connected by a linker (e.g., a peptide linker). In some embodiments, the peptide linker is a flexible linker. Regarding the amino acid composition of the linker, a peptide with the following properties is selected: flexibility is imparted, the structure and function of other domains of the protein of the present invention are not interfered with, and cleavage from proteases is resisted. For example, glycine residues and serine residues provide protease resistance together. In some embodiments, the VL is connected to the N-terminus or C-terminus of the VH by a (GlyGlyGlyGlySer)4((G4S)4) linker (SEQ ID NO:532).

[0454] The length of the linker (e.g., a flexible linker) can be "short", e.g., 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 or 12 amino acid residues, or "long", e.g., at least 13 amino acid residues. In certain embodiments, the length of the linker is 10-50, 10-40, 10-30, 10-25, 10-20, 15-50, 15-30, 15-25, 15-20, 20-50, 20-40, 20-30 or 20-25 amino acid residues.

[0455] In certain embodiments, the linker comprises or consists of the following sequence: (GS) n (SEQ ID NO:533), (GGS) n (SEQ ID NO:534), (GGGS) n (SEQ ID NO:535), (GGSG) n (SEQ ID NO:536), (GGSGG) n (SEQ ID NO:537) and (GGGS) n (SEQ ID NO:538), wherein n is 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 or 20. In certain embodiments, the linker comprises or consists of an amino acid sequence selected from the group consisting of SEQ ID NO:532, SEQ ID NO:539, SEQ ID NO:540, SEQ ID NO:541, SEQ ID NO:542, SEQ ID NO:543, SEQ ID NO:544, SEQ ID NO:545, SEQ ID NO:546, and SEQ ID NO:547 as listed in Table 12.

[0456]

[0457] In F3'-TriNKET, the scFv binding to CEACAM5 is linked to the N-terminus of Fc via a Gly-Ser linker. An Ala-Ser or Gly-Ser linker is included in the elbow hinge region sequence to strike a balance between flexibility and optimal geometry. In certain embodiments, an additional sequence Thr-Lys-Gly can be added to the N-terminus or C-terminus of the Ala-Ser or Gly-Ser sequence at the hinge.

[0458] As used herein to describe these exemplary TriNKETs, Fc comprises antibody hinge, CH2 and CH3. In each exemplary TriNKET, the Fc domain connected to scFv comprises the mutations of Q347R, D399V and F405T, and the Fc domain connected to Fab comprises the mutations K360E and K409W of matching for forming heterodimers. The Fc domain connected to scFv further comprises S354C displacement in the CH3 domain, which is connected to the Y349C displacement on the Fc of Fab to form a disulfide bond. These displacements are in bold in the sequence described in this subsection.

[0459] For example, the TriNKET of the present disclosure is F3'-GB1. F3'-GB1 comprises (a) a scFv sequence that binds to CEACAM5 derived from a VH sequence and a VL sequence in the homologous pairing A1 of Table 2, which is connected to an Fc domain, and (b) a Fab fragment that binds to NKG2D derived from A49MI, which comprises a heavy chain portion comprising a heavy chain variable domain and a CH1 domain, and a light chain portion comprising a light chain variable domain and a light chain constant domain, wherein the CH1 domain is connected to the Fc domain. F3'-GB1 comprises three polypeptides: GB1-VL-VH-Fc, A49MI-VH-CH1-Fc and A49MI-VL-CL.

[0460] GBI-VL-VH-Fc (SEQ ID NO:548)

[0461]

[0462] A49MI-VH-CHI-Fc (SEQ ID NO:549)

[0463]

[0464] A49MI-VL-CL (SEO ID NO: 550)

[0465]

[0466] GB1-VL-VH-Fc represents the complete sequence of a scFv that binds CEACAM5 connected to an Fc domain via a hinge comprising Gly-Ser. As described below, the Fc domain connected to the scFv comprises Q347R, D399V and F405T substitutions for heterodimerization, and an S354C substitution for forming a disulfide bond with the Y349C substitution in A49MI-VH-CH1-Fc. The scFv comprises the heavy chain variable domain of GB1, which is connected to the C-terminus of the light chain variable domain of GB1 via a (G4S)4 linker (SEQ ID NO: 532). The heavy and light variable domains of the scFv are also connected by a disulfide bridge formed by a cysteine ​​heterodimerization mutation, indicated in bold underline in the above sequence.

[0467] A49MI-VH-CH1-Fc represents the heavy chain portion of the Fab fragment, which comprises the heavy chain variable domain of A49MI that binds NKG2D (SEQ ID NO: 508) and the CH1 domain connected to the Fc domain. The Fc domain in A49MI-VH-CH1-Fc comprises a Y349C substitution in the CH3 domain, which forms a disulfide bond with the S354C substitution on the Fc in GB1-VL-VH-Fc. In A49MI-VH-CH1-Fc, the Fc domain also comprises K360E and K409W substitutions for heterodimerization with the Fc in GB1-VL-VH-Fc.

[0468] A49MI-VL-CL represents the light chain portion of the Fab fragment, which comprises the light chain variable domain of A49MI that binds NKG2D (SEQ ID NO: 493) and the light chain constant domain.

[0469] Another TriNKET disclosed herein is F3'-GB3. F3'-GB3 comprises (a) a CEACAM5-binding scFv sequence derived from a VH sequence and a VL sequence in a corresponding cognate pair in Table 2, connected to an Fc domain, and (b) an NKG2D-binding Fab fragment derived from A49MI, comprising a heavy chain portion comprising a heavy chain variable domain and a CH1 domain, and a light chain portion comprising a light chain variable domain and a light chain constant domain, wherein the CH1 domain is connected to the Fc domain. F3'-GB3 comprises three polypeptides: GB3-VH-VL-Fc, A49MI-VH-CH1-Fc, and A49MI-VL-CL. A49MI-VH-CH1-Fc and A49MI-VL-CL are described above in the context of F3'GB1. The polypeptides of GB3-VH-VL-Fc are set forth below.

[0470] GB3-VH-VL-Fc (SEQ ID NO: 551)

[0471]

[0472] GB3-VH-VL-Fc represents the complete sequence of a scFv that binds CEACAM5 connected to an Fc domain via a hinge comprising Gly-Ser. As described above, the Fc domain connected to the scFv comprises Q347R, D399V and F405T substitutions for heterodimerization, and an S354C substitution for forming a disulfide bond with the Y349C substitution in A49MI-VH-CH1-Fc. The scFv comprises the light chain variable domain of GB3, which is connected to the C-terminus of the heavy chain variable domain of GB3 via a (G4S)4 linker (SEQ ID NO: 532). The heavy and light variable domains of the scFv are also connected by a disulfide bridge formed by a cysteine ​​heterodimerization mutation, indicated in bold underline in the above sequence.

[0473] Another TriNKET disclosed herein is F3'-GB5. F3'-GB5 comprises (a) a CEACAM5-binding scFv sequence derived from a VH sequence and a VL sequence in a corresponding cognate pair in Table 2, connected to an Fc domain, and (b) an NKG2D-binding Fab fragment derived from A49MI, comprising a heavy chain portion comprising a heavy chain variable domain and a CH1 domain, and a light chain portion comprising a light chain variable domain and a light chain constant domain, wherein the CH1 domain is connected to the Fc domain. F3'-GB5 comprises three polypeptides: GB5-VH-VL-Fc, A49MI-VH-CH1-Fc, and A49MI-VL-CL. A49MI-VH-CH1-Fc and A49MI-VL-CL are described above in the context of F3'GB1. The polypeptides of GB5-VH-VL-Fc are set forth below.

[0474] GB5-VH-VL-Fc (SEQ ID NO: 552)

[0475]

[0476] GB5-VH-VL-Fc represents the complete sequence of a scFv that binds CEACAM5 connected to an Fc domain via a hinge comprising Gly-Ser. As described above, the Fc domain connected to the scFv comprises Q347R, D399V and F405T substitutions for heterodimerization, and an S354C substitution for forming a disulfide bond with the Y349C substitution in A49MI-VH-CH1-Fc. The scFv comprises the light chain variable domain of GB5, which is connected to the C-terminus of the heavy chain variable domain of GB5 via a (G4S)4 linker (SEQ ID NO: 532). The heavy and light variable domains of the scFv are also connected by a disulfide bridge formed by a cysteine ​​heterodimerization mutation, indicated by bold underline in the above sequence.

[0477] Another TriNKET disclosed herein is F3'-GB7. F3'-GB7 comprises (a) a CEACAM5-binding scFv sequence derived from a VH sequence and a VL sequence in a corresponding cognate pair in Table 2, connected to an Fc domain, and (b) an NKG2D-binding Fab fragment derived from A49MI, comprising a heavy chain portion comprising a heavy chain variable domain and a CH1 domain, and a light chain portion comprising a light chain variable domain and a light chain constant domain, wherein the CH1 domain is connected to the Fc domain. F3'-GB7 comprises three polypeptides: GB7-VH-VL-Fc, A49MI-VH-CH1-Fc, and A49MI-VL-CL. A49MI-VH-CH1-Fc and A49MI-VL-CL are described above in the context of F3'GB1. The polypeptides of GB7-VH-VL-Fc are set forth below.

[0478] GB7-VH-VL-Fc (SEQ ID NO:553)

[0479]

[0480] GB7-VH-VL-Fc represents the complete sequence of a scFv that binds CEACAM5 connected to an Fc domain via a hinge comprising Gly-Ser. As described above, the Fc domain connected to the scFv comprises Q347R, D399V and F405T substitutions for heterodimerization, and an S354C substitution for forming a disulfide bond with the Y349C substitution in A49MI-VH-CH1-Fc. The scFv comprises the light chain variable domain of GB7, which is connected to the C-terminus of the heavy chain variable domain of GB7 via a (G4S)4 linker (SEQ ID NO: 532). The heavy and light variable domains of the scFv are also connected by a disulfide bridge formed by a cysteine ​​heterodimerization mutation, indicated by bold underline in the above sequence.

[0481] Another TriNKET disclosed herein is F3'-GB9. F3'-GB9 comprises (a) a CEACAM5-binding scFv sequence derived from a VH sequence and a VL sequence in a corresponding cognate pair in Table 2, connected to an Fc domain, and (b) an NKG2D-binding Fab fragment derived from A49MI, comprising a heavy chain portion comprising a heavy chain variable domain and a CH1 domain, and a light chain portion comprising a light chain variable domain and a light chain constant domain, wherein the CH1 domain is connected to the Fc domain. F3'-GB9 comprises three polypeptides: GB9-VH-VL-Fc, A49MI-VH-CH1-Fc, and A49MI-VL-CL. A49MI-VH-CH1-Fc and A49MI-VL-CL are described above in the context of F3'GB1. The polypeptides of GB9-VH-VL-Fc are set forth below.

[0482] GB9-VL-VH-Fc (SEQ ID NO:554)

[0483]

[0484] GB9-VH-VL-Fc represents the complete sequence of a scFv that binds CEACAM5 connected to an Fc domain via a hinge comprising Gly-Ser. As described above, the Fc domain connected to the scFv comprises Q347R, D399V and F405T substitutions for heterodimerization, and an S354C substitution for forming a disulfide bond with the Y349C substitution in A49MI-VH-CH1-Fc. The scFv comprises the light chain variable domain of GB9, which is connected to the C-terminus of the heavy chain variable domain of GB9 via a (G4S)4 linker (SEQ ID NO: 532). The heavy and light variable domains of the scFv are also connected by a disulfide bridge formed by a cysteine ​​heterodimerization mutation, indicated by bold underline in the above sequence.

[0485] Another TriNKET disclosed herein is F3'-GB11. F3'-GB11 comprises (a) a CEACAM5-binding scFv sequence derived from a VH sequence and a VL sequence in a corresponding cognate pair in Table 2, connected to an Fc domain, and (b) an NKG2D-binding Fab fragment derived from A49MI, comprising a heavy chain portion comprising a heavy chain variable domain and a CH1 domain, and a light chain portion comprising a light chain variable domain and a light chain constant domain, wherein the CH1 domain is connected to the Fc domain. F3'-GB11 comprises three polypeptides: GB11-VH-VL-Fc, A49MI-VH-CH1-Fc, and A49MI-VL-CL. A49MI-VH-CH1-Fc and A49MI-VL-CL are described above in the context of F3'GB1. The polypeptides of GB11-VH-VL-Fc are set forth below.

[0486] GBII-VH-VL-Fc (SEQ ID NO:555)

[0487]

[0488] GB11-VH-VL-Fc represents the complete sequence of a scFv that binds CEACAM5 connected to an Fc domain via a hinge comprising Gly-Ser. As described above, the Fc domain connected to the scFv comprises Q347R, D399V and F405T substitutions for heterodimerization, and an S354C substitution for forming a disulfide bond with the Y349C substitution in A49MI-VH-CH1-Fc. The scFv comprises the light chain variable domain of GB11, which is connected to the C-terminus of the heavy chain variable domain of GB11 via a (G4S)4 linker (SEQ ID NO: 532). The heavy and light variable domains of the scFv are also connected by a disulfide bridge formed by a cysteine ​​heterodimerization mutation, indicated by bold underline in the above sequence.

[0489] Another TriNKET disclosed herein is F3'-GB13. F3'-GB13 comprises (a) a CEACAM5-binding scFv sequence derived from a VH sequence and a VL sequence in a corresponding cognate pair in Table 2, connected to an Fc domain, and (b) an NKG2D-binding Fab fragment derived from A49MI, comprising a heavy chain portion comprising a heavy chain variable domain and a CH1 domain, and a light chain portion comprising a light chain variable domain and a light chain constant domain, wherein the CH1 domain is connected to the Fc domain. F3'-GB13 comprises three polypeptides: GB13-VH-VL-Fc, A49MI-VH-CH1-Fc, and A49MI-VL-CL. A49MI-VH-CH1-Fc and A49MI-VL-CL are described above in the context of F3'GB1. The polypeptides of GB13-VH-VL-Fc are set forth below.

[0490] GB13-VH-VL-Fc (SEQ ID NO: 556)

[0491]

[0492] GB13-VH-VL-Fc represents the complete sequence of a scFv that binds CEACAM5 connected to an Fc domain via a hinge comprising Gly-Ser. As described above, the Fc domain connected to the scFv comprises Q347R, D399V and F405T substitutions for heterodimerization, and an S354C substitution for forming a disulfide bond with the Y349C substitution in A49MI-VH-CH1-Fc. The scFv comprises the light chain variable domain of GB13, which is connected to the C-terminus of the heavy chain variable domain of GB13 via a (G4S)4 linker (SEQ ID NO: 532). The heavy and light variable domains of the scFv are also connected by a disulfide bridge formed by a cysteine ​​heterodimerization mutation, indicated by bold underline in the above sequence.

[0493] Another TriNKET disclosed herein is F3'-GB15. F3'-GB15 comprises (a) a CEACAM5-binding scFv sequence derived from a VH sequence and a VL sequence in a corresponding cognate pair in Table 2, connected to an Fc domain, and (b) an NKG2D-binding Fab fragment derived from A49MI, comprising a heavy chain portion comprising a heavy chain variable domain and a CH1 domain, and a light chain portion comprising a light chain variable domain and a light chain constant domain, wherein the CH1 domain is connected to the Fc domain. F3'-GB15 comprises three polypeptides: GB15-VH-VL-Fc, A49MI-VH-CH1-Fc, and A49MI-VL-CL. A49MI-VH-CH1-Fc and A49MI-VL-CL are described above in the context of F3'GB1. The polypeptides of GB15-VH-VL-Fc are set forth below.

[0494] GB15-VL-VH-Fc (SEQ ID NO: 557)

[0495]

[0496]

[0497] GB15-VH-VL-Fc represents the complete sequence of a scFv that binds CEACAM5 connected to an Fc domain via a hinge comprising Gly-Ser. As described above, the Fc domain connected to the scFv comprises Q347R, D399V and F405T substitutions for heterodimerization, and an S354C substitution for forming a disulfide bond with the Y349C substitution in A49MI-VH-CH1-Fc. The scFv comprises the light chain variable domain of GB15, which is connected to the C-terminus of the heavy chain variable domain of GB15 via a (G4S)4 linker (SEQ ID NO: 532). The heavy and light variable domains of the scFv are also connected by a disulfide bridge formed by a cysteine ​​heterodimerization mutation, indicated by bold underline in the above sequence.

[0498] Another TriNKET disclosed herein is F3'-GB17. F3'-GB17 comprises (a) a CEACAM5-binding scFv sequence derived from a VH sequence and a VL sequence in a corresponding cognate pair in Table 2, connected to an Fc domain, and (b) an NKG2D-binding Fab fragment derived from A49MI, comprising a heavy chain portion comprising a heavy chain variable domain and a CH1 domain, and a light chain portion comprising a light chain variable domain and a light chain constant domain, wherein the CH1 domain is connected to the Fc domain. F3'-GB17 comprises three polypeptides: GB17-VH-VL-Fc, A49MI-VH-CH1-Fc, and A49MI-VL-CL. A49MI-VH-CH1-Fc and A49MI-VL-CL are described above in the context of F3'GB1. The polypeptides of GB17-VH-VL-Fc are set forth below.

[0499] GBI7-VH-VL-Fc (SEQ ID NO:558)

[0500]

[0501] GB17-VH-VL-Fc represents the complete sequence of a scFv that binds CEACAM5 connected to an Fc domain via a hinge comprising Gly-Ser. As described above, the Fc domain connected to the scFv comprises Q347R, D399V and F405T substitutions for heterodimerization, and an S354C substitution for forming a disulfide bond with the Y349C substitution in A49MI-VH-CH1-Fc. The scFv comprises the light chain variable domain of GB17, which is connected to the C-terminus of the heavy chain variable domain of GB17 via a (G4S)4 linker (SEQ ID NO: 532). The heavy and light variable domains of the scFv are also connected by a disulfide bridge formed by a cysteine ​​heterodimerization mutation, indicated by bold underline in the above sequence.

[0502] Another TriNKET disclosed herein is F3'-GB19. F3'-GB19 comprises (a) a CEACAM5-binding scFv sequence derived from a VH sequence and a VL sequence in a corresponding cognate pair in Table 2, connected to an Fc domain, and (b) an NKG2D-binding Fab fragment derived from A49MI, comprising a heavy chain portion comprising a heavy chain variable domain and a CH1 domain, and a light chain portion comprising a light chain variable domain and a light chain constant domain, wherein the CH1 domain is connected to the Fc domain. F3'-GB19 comprises three polypeptides: GB19-VH-VL-Fc, A49MI-VH-CH1-Fc, and A49MI-VL-CL. A49MI-VH-CH1-Fc and A49MI-VL-CL are described above in the context of F3'GB1. The polypeptides of GB19-VH-VL-Fc are set forth below.

[0503] GB19-VH-VL-Fc (SEQ ID NO:559)

[0504]

[0505] GB19-VH-VL-Fc represents the complete sequence of a scFv that binds CEACAM5 connected to an Fc domain via a hinge comprising Gly-Ser. As described above, the Fc domain connected to the scFv comprises Q347R, D399V and F405T substitutions for heterodimerization, and an S354C substitution for forming a disulfide bond with the Y349C substitution in A49MI-VH-CH1-Fc. The scFv comprises the light chain variable domain of GB19, which is connected to the C-terminus of the heavy chain variable domain of GB19 via a (G4S)4 linker (SEQ ID NO: 532). The heavy and light variable domains of the scFv are also connected by a disulfide bridge formed by a cysteine ​​heterodimerization mutation, indicated by bold underline in the above sequence.

[0506] Another TriNKET disclosed herein is F3'-GB21. F3'-GB21 comprises (a) a CEACAM5-binding scFv sequence derived from a VH sequence and a VL sequence in a corresponding cognate pair in Table 2, connected to an Fc domain, and (b) an NKG2D-binding Fab fragment derived from A49MI, comprising a heavy chain portion comprising a heavy chain variable domain and a CH1 domain, and a light chain portion comprising a light chain variable domain and a light chain constant domain, wherein the CH1 domain is connected to the Fc domain. F3'-GB21 comprises three polypeptides: GB21-VH-VL-Fc, A49MI-VH-CH1-Fc, and A49MI-VL-CL. A49MI-VH-CH1-Fc and A49MI-VL-CL are described above in the context of F3'GB1. The polypeptides of GB21-VH-VL-Fc are set forth below.

[0507] GB21-VH-VL-Fc (SEQ ID NO:560)

[0508]

[0509] GB21-VH-VL-Fc represents the complete sequence of a scFv that binds CEACAM5 connected to an Fc domain via a hinge comprising Gly-Ser. As described above, the Fc domain connected to the scFv comprises Q347R, D399V and F405T substitutions for heterodimerization, and an S354C substitution for forming a disulfide bond with the Y349C substitution in A49MI-VH-CH1-Fc. The scFv comprises the light chain variable domain of GB21, which is connected to the C-terminus of the heavy chain variable domain of GB21 via a (G4S)4 linker (SEQ ID NO: 532). The heavy and light variable domains of the scFv are also connected by a disulfide bridge formed by a cysteine ​​heterodimerization mutation, indicated by bold underline in the above sequence.

[0510] Another TriNKET disclosed herein is F3'-GB23. F3'-GB23 comprises (a) a CEACAM5-binding scFv sequence derived from a VH sequence and a VL sequence in a corresponding cognate pair in Table 2, connected to an Fc domain, and (b) an NKG2D-binding Fab fragment derived from A49MI, comprising a heavy chain portion comprising a heavy chain variable domain and a CH1 domain, and a light chain portion comprising a light chain variable domain and a light chain constant domain, wherein the CH1 domain is connected to the Fc domain. F3'-GB23 comprises three polypeptides: GB23-VH-VL-Fc, A49MI-VH-CH1-Fc, and A49MI-VL-CL. A49MI-VH-CH1-Fc and A49MI-VL-CL are described above in the context of F3'GB1. The polypeptides of GB23-VH-VL-Fc are set forth below.

[0511] GB23-VH-VL-Fc (SEQ ID NO:561)

[0512]

[0513] GB23-VH-VL-Fc represents the complete sequence of a scFv that binds CEACAM5 connected to an Fc domain via a hinge comprising Gly-Ser. As described above, the Fc domain connected to the scFv comprises Q347R, D399V and F405T substitutions for heterodimerization, and an S354C substitution for forming a disulfide bond with the Y349C substitution in A49MI-VH-CH1-Fc. The scFv comprises the light chain variable domain of GB23, which is connected to the C-terminus of the heavy chain variable domain of GB23 via a (G4S)4 linker (SEQ ID NO: 532). The heavy and light variable domains of the scFv are also connected by a disulfide bridge formed by a cysteine ​​heterodimerization mutation, indicated by bold underline in the above sequence.

[0514] Another TriNKET disclosed herein is F3'-GB25. F3'-GB25 comprises (a) a CEACAM5-binding scFv sequence derived from a VH sequence and a VL sequence in a corresponding cognate pair in Table 2, connected to an Fc domain, and (b) an NKG2D-binding Fab fragment derived from A49MI, comprising a heavy chain portion comprising a heavy chain variable domain and a CH1 domain, and a light chain portion comprising a light chain variable domain and a light chain constant domain, wherein the CH1 domain is connected to the Fc domain. F3'-GB25 comprises three polypeptides: GB25-VH-VL-Fc, A49MI-VH-CH1-Fc, and A49MI-VL-CL. A49MI-VH-CH1-Fc and A49MI-VL-CL are described above in the context of F3'GB1. The polypeptides of GB25-VH-VL-Fc are set forth below.

[0515] GB25-VH-VL-Fc (SEQ ID NQ:562)

[0516]

[0517]

[0518] GB25-VH-VL-Fc represents the complete sequence of a scFv that binds CEACAM5 connected to an Fc domain via a hinge comprising Gly-Ser. As described above, the Fc domain connected to the scFv comprises Q347R, D399V and F405T substitutions for heterodimerization, and an S354C substitution for forming a disulfide bond with the Y349C substitution in A49MI-VH-CH1-Fc. The scFv comprises the light chain variable domain of GB25, which is connected to the C-terminus of the heavy chain variable domain of GB25 via a (G4S)4 linker (SEQ ID NO: 532). The heavy and light variable domains of the scFv are also connected by a disulfide bridge formed by a cysteine ​​heterodimerization mutation, indicated by bold underline in the above sequence.

[0519] Another TriNKET disclosed herein is F3'-GB27. F3'-GB27 comprises (a) a CEACAM5-binding scFv sequence derived from a VH sequence and a VL sequence in a corresponding cognate pair in Table 2, connected to an Fc domain, and (b) an NKG2D-binding Fab fragment derived from A49MI, comprising a heavy chain portion comprising a heavy chain variable domain and a CH1 domain, and a light chain portion comprising a light chain variable domain and a light chain constant domain, wherein the CH1 domain is connected to the Fc domain. F3'-GB27 comprises three polypeptides: GB27-VH-VL-Fc, A49MI-VH-CH1-Fc, and A49MI-VL-CL. A49MI-VH-CH1-Fc and A49MI-VL-CL are described above in the context of F3'GB1. The polypeptides of GB27-VH-VL-Fc are set forth below.

[0520] GB27-VH-VL-Fc (SEQ ID NO:563)

[0521]

[0522] GB27-VH-VL-Fc represents the complete sequence of a scFv that binds CEACAM5 connected to an Fc domain via a hinge comprising Gly-Ser. As described above, the Fc domain connected to the scFv comprises Q347R, D399V and F405T substitutions for heterodimerization, and an S354C substitution for forming a disulfide bond with the Y349C substitution in A49MI-VH-CH1-Fc. The scFv comprises the light chain variable domain of GB27, which is connected to the C-terminus of the heavy chain variable domain of GB27 via a (G4S)4 linker (SEQ ID NO: 532). The heavy and light variable domains of the scFv are also connected by a disulfide bridge formed by a cysteine ​​heterodimerization mutation, indicated by bold underline in the above sequence.

[0523] Another TriNKET disclosed herein is F3'-GB29. F3'-GB29 comprises (a) a CEACAM5-binding scFv sequence derived from a VH sequence and a VL sequence in a corresponding cognate pair in Table 2, connected to an Fc domain, and (b) an NKG2D-binding Fab fragment derived from A49MI, comprising a heavy chain portion comprising a heavy chain variable domain and a CH1 domain, and a light chain portion comprising a light chain variable domain and a light chain constant domain, wherein the CH1 domain is connected to the Fc domain. F3'-GB29 comprises three polypeptides: GB29-VH-VL-Fc, A49MI-VH-CH1-Fc, and A49MI-VL-CL. A49MI-VH-CH1-Fc and A49MI-VL-CL are described above in the context of F3'GB1. The polypeptides of GB29-VH-VL-Fc are set forth below.

[0524] GB29-VH-VL-Fc (SEQ ID NO:564)

[0525]

[0526] GB29-VH-VL-Fc represents the complete sequence of a scFv that binds CEACAM5 connected to an Fc domain via a hinge comprising Gly-Ser. As described above, the Fc domain connected to the scFv comprises Q347R, D399V and F405T substitutions for heterodimerization, and an S354C substitution for forming a disulfide bond with the Y349C substitution in A49MI-VH-CH1-Fc. The scFv comprises the light chain variable domain of GB29, which is connected to the C-terminus of the heavy chain variable domain of GB29 via a (G4S)4 linker (SEQ ID NO: 532). The heavy and light variable domains of the scFv are also connected by a disulfide bridge formed by a cysteine ​​heterodimerization mutation, indicated by bold underline in the above sequence.

[0527] Another TriNKET disclosed herein is F3'-GB31. F3'-GB31 comprises (a) a CEACAM5-binding scFv sequence derived from a VH sequence and a VL sequence in a corresponding cognate pair in Table 2, connected to an Fc domain, and (b) an NKG2D-binding Fab fragment derived from A49MI, comprising a heavy chain portion comprising a heavy chain variable domain and a CH1 domain, and a light chain portion comprising a light chain variable domain and a light chain constant domain, wherein the CH1 domain is connected to the Fc domain. F3'-GB31 comprises three polypeptides: GB31-VH-VL-Fc, A49MI-VH-CH1-Fc, and A49MI-VL-CL. A49MI-VH-CH1-Fc and A49MI-VL-CL are described above in the context of F3'GB1. The polypeptides of GB31-VH-VL-Fc are set forth below.

[0528] GB31-VH-VL-Fc (SEQ ID NO:565)

[0529]

[0530] GB31-VH-VL-Fc represents the complete sequence of a scFv that binds CEACAM5 connected to an Fc domain via a hinge comprising Gly-Ser. As described above, the Fc domain connected to the scFv comprises Q347R, D399V and F405T substitutions for heterodimerization, and an S354C substitution for forming a disulfide bond with the Y349C substitution in A49MI-VH-CH1-Fc. The scFv comprises the light chain variable domain of GB31, which is connected to the C-terminus of the heavy chain variable domain of GB31 via a (G4S)4 linker (SEQ ID NO: 532). The heavy and light variable domains of the scFv are also connected by a disulfide bridge formed by a cysteine ​​heterodimerization mutation, indicated by bold underline in the above sequence.

[0531] In certain embodiments, the TriNKET of the present disclosure is the same as one of the above exemplary TriNKETs comprising the EW-RVT Fc mutation, but differs in that the Fc domain attached to the Fab fragment that binds NKG2D comprises the substitutions Q347R, D399V, and F405T, and the Fc domain attached to the scFv that binds CEACAM5 comprises the matching substitutions K360E and K409W, for forming heterodimers. In certain embodiments, the TriNKET of the present disclosure is the same as one of the above exemplary TriNKETs comprising the KiH Fc mutation, but differs in that the Fc domain attached to the Fab fragment that binds NKG2D comprises the "hole" substitutions T366S, L368A, and Y407V, and the Fc domain attached to the scFv that binds CEACAM5 comprises the "knob" substitution T366W, for forming heterodimers.

[0532] In certain embodiments, a TriNKET of the present disclosure is identical to one of the exemplary TriNKETs described above, except that the Fc domain linked to the Fab fragment that binds NKG2D comprises a S354C substitution in the CH3 domain, and the Fc domain linked to the scFv that binds CEACAM5 comprises a matching Y349C substitution in the CH3 domain for disulfide bond formation.

[0533] It will be appreciated by those skilled in the art that during the production and / or storage of the protein, the N-terminal glutamic acid (E) or glutamine (Q) may be cyclized to form a lactam (e.g., spontaneously or by catalysis of an enzyme present during the production and / or storage process). Therefore, in some embodiments in which the N-terminal residue of the amino acid sequence of the polypeptide is E or Q, the corresponding amino acid sequence in which the E or Q is replaced with pyroglutamic acid is also contemplated herein.

[0534] It will also be appreciated by those skilled in the art that during the production and / or storage of the protein, the C-terminal lysine (K) of the protein may be removed (e.g., spontaneously or catalyzed by an enzyme present during the production and / or storage process). This removal of K is often observed in proteins that contain an Fc domain at the C-terminus. Therefore, in some embodiments in which the C-terminal residue of the amino acid sequence of the polypeptide (e.g., an Fc domain sequence) is K, the corresponding amino acid sequence in which the K is removed is also contemplated herein.

[0535] The above multispecific proteins can be produced using recombinant DNA techniques well known to those skilled in the art. For example, a first nucleic acid sequence encoding a first immunoglobulin heavy chain can be cloned into a first expression vector; a second nucleic acid sequence encoding a second immunoglobulin heavy chain can be cloned into a second expression vector; a third nucleic acid sequence encoding an immunoglobulin light chain can be cloned into a third expression vector; and the first expression vector, the second expression vector, and the third expression vector can be stably transfected together into a host cell to produce a multimeric protein.

[0536] In order to achieve the highest yield of the multispecific protein, different ratios of the first expression vector, the second expression vector, and the third expression vector can be explored to determine the optimal ratio for transfection into host cells. After transfection, individual clones can be isolated using methods known in the art (such as limiting dilution, ELISA, FACS, microscopy, or Clonepix) for cell bank generation.

[0537] The clones can be cultured under conditions suitable for bioreactor scale-up and maintain expression of the multispecific protein. The multispecific proteins can be isolated and purified using methods known in the art, including centrifugation, depth filtration, cell lysis, homogenization, freeze-thaw, affinity purification, gel filtration, ion exchange chromatography, hydrophobic interaction exchange chromatography, and mixed mode chromatography.

[0538] II. Characteristics of Multispecific Proteins

[0539] The multispecific proteins described herein comprise a NKG2D binding site, a CEACAM5 binding site, and an antibody Fc domain or portion thereof sufficient to bind CD 16, or an antigen binding site that binds CD 16. In some embodiments, as exemplified in the F4-TriNKET format, these multispecific proteins contain an additional antigen binding site that binds to CEACAM5.

[0540] In some embodiments, the multispecific proteins exhibit thermal stability similar to that of the corresponding monoclonal antibodies (ie, monoclonal antibodies containing the same CEACAM5 binding site as incorporated into the multispecific proteins).

[0541] In some embodiments, these multispecific proteins simultaneously bind to cells expressing NKG2D and / or CD16 (such as NK cells) and cells expressing CEACAM5 (such as certain tumor cells). The binding of these multispecific proteins to NK cells can enhance the activity of NK cells in destroying tumor cells expressing CEACAM5.

[0542] In some embodiments, the multispecific proteins bind to CEACAM5 with similar affinity as the corresponding anti-CEACAM5 monoclonal antibody (i.e., a monoclonal antibody containing the same CEACAM5 binding site as incorporated into the multispecific proteins). In some embodiments, the multispecific proteins kill tumor cells expressing CEACAM5 more effectively than the corresponding monoclonal antibody.

[0543] In certain embodiments, the multispecific proteins described herein (which contain a binding site for CEACAM5) activate primary human NK cells when co-cultured with cells expressing CEACAM5. The hallmarks of NK cell activation are increased CD107a degranulation and IFN-γ cytokine production. In addition, these multispecific proteins can show better activation of human NK cells in the presence of cells expressing CEACAM5 compared to the corresponding anti-CEACAM5 monoclonal antibodies.

[0544] In some embodiments, the multispecific proteins described herein (which comprise a binding site for CEACAM5) enhance the activity of resting human NK cells and IL-2-activated human NK cells when co-cultured with cells expressing CEACAM5.

[0545] In some embodiments, these multispecific proteins provide advantages in targeting tumor cells expressing intermediate and low levels of CEACAM5 compared to corresponding monoclonal antibodies that bind to CEACAM5.

[0546] In some embodiments, the bivalent F4 format of TriNKET (i.e., TriNKET comprising an additional antigen binding site that binds to CEACAM5) improves the affinity of TriNKET binding to CEACAM5, which in effect stabilizes the expression and maintenance of high levels of CEACAM5 on the surface of tumor cells. In some embodiments, F4-TriNKET mediates more potent tumor cell killing than the corresponding F3-TriNKET or F3'-TriNKET.

[0547] III. Therapeutic Applications

[0548] The present invention provides methods for treating autoimmune diseases or cancer using the multispecific binding proteins described herein and / or the pharmaceutical compositions described herein. These methods can be used to treat a variety of cancers that express CEACAM5.

[0549] The treatment method can be characterized according to the cancer to be treated. For example, in certain embodiments, the cancer is selected from the group consisting of gastrointestinal cancer, colorectal cancer, pancreatic cancer, non-small cell lung cancer and esophageal cancer.

[0550] In certain embodiments, the cancer is a solid tumor. In certain other embodiments, the cancer is brain cancer, bladder cancer, breast cancer, cervical cancer, colon cancer, colorectal cancer, endometrial cancer, esophageal cancer, leukemia, lung cancer, liver cancer, melanoma, ovarian cancer, pancreatic cancer, prostate cancer, rectal cancer, kidney cancer, stomach cancer, testicular cancer, or uterine cancer.In yet other embodiments, the cancer is a vascularized tumor, squamous cell carcinoma, adenocarcinoma, small cell carcinoma, melanoma, glioma, neuroblastoma, sarcoma (e.g., angiosarcoma or chondrosarcoma), laryngeal cancer, parotid cancer, biliary tract cancer, thyroid cancer, acral lentigo melanoma, actinic keratosis, acute lymphocytic leukemia, acute myeloid leukemia, adenoid cystic carcinoma, adenoma, adenosarcoma, adenosquamous cell carcinoma, anal canal cancer, anal cancer, anorectal cancer, astrocytoma, Bartholin's gland cancer, basal cell carcinoma, biliary tract cancer (biliary cancer), bone cancer, bone marrow cancer, bronchial cancer, bronchial adenocarcinoma, carcinoid, bile duct cancer (cholangiocarcinoma), chondrosarcoma, choroid plexus papilloma / carcinoma, chronic lymphocytic leukemia, chronic myeloid leukemia, clear cell carcinoma, connective tissue cancer, cystadenoma, digestive system cancer, duodenal cancer, endocrine system cancer, endodermal sinus tumor, endometrial hyperplasia, endometrial stromal sarcoma, endometrioid adenocarcinoma, endothelial cell cancer, ependymal cancer, epithelial cell cancer, Ewing's sarcoma, eye and orbital cancer, female genital organ cancer, focal nodular hyperplasia, gallbladder cancer, gastric antrum cancer, gastric fundus cancer, gastrinoma, glioblastoma, glucagonoma, heart cancer, hemangioblastoma, hemangioendothelioma, hemangioma, hepatic adenoma, hepatocellular carcinoma, Hodgkin's disease, ileal cancer, insulinoma, intraepithelial neoplasia, intraepithelial squamous cell neoplasia, intrahepatic bile duct cancer, invasive squamous cell carcinoma, jejunal cancer, joint cancer, Kaposi's sarcoma, pelvic cancer, large cell carcinoma, colorectal cancer, leiomyosarcoma, malignant lentigo melanoma, lymphoma, male genital cancer, malignant melanoma, malignant mesothelioma, medulloblastoma, medulloepithelioma, meningeal cancer, mesothelioma, metastatic cancer, oral cancer (mouth cancer), mucoepidermoid carcinoma, multiple myeloma, muscle cancer, nasal cancer, nervous system cancer, neuroepithelial adenocarcinoma nodular melanoma, non-epithelial skin cancer, non-Hodgkin lymphoma, oat cell carcinoma, oligodendroglial carcinoma, oral cavity cancer, osteosarcoma, papillary serous adenocarcinoma, penile cancer, pharyngeal cancer, pituitary tumor, plasmacytoma, pseudosarcoma, pulmonary blastoma, rectal cancer, renal cell cancer, respiratory cancer, retinoblastoma, rhabdomyosarcoma, sarcoma, serous carcinoma, sinus cancer, skin cancer, small cell carcinoma, small intestine cancer, smooth muscle cancer, soft tissue cancer, somatostatin-secreting tumor, spinal cancer, squamous cell carcinoma, rhabdomyosarcoma, submesothelial carcinoma, superficial spreading melanoma, T-cell leukemia, tongue cancer, undifferentiated carcinoma, ureter cancer, urethra cancer, urinary bladder cancer, urinary system cancer, uterine cervix cancer cancer), uterine corpus cancer, uveal melanoma, vaginal cancer, verrucous carcinoma, vipoma, vulvar cancer, well-differentiated carcinoma, or Wilms' tumor.

[0551] The cancer to be treated can be characterized by the presence of specific antigens expressed on the surface of cancer cells. Cancers characterized by the expression of CEACAM5 include, but are not limited to, medullary thyroid carcinoma (MTC), non-medullary thyroid carcinoma (non-MTC), gastric cancer, colorectal cancer, hepatocellular carcinoma, lung cancer, pancreatic cancer, breast cancer and ovarian cancer. In certain embodiments, cancer cells may express one or more of the following: CEACAM1, CEACAM3, CEACAM6 and CEACAM8. In certain embodiments, in addition to CEACAM5, cancer cells may express one or more of the following: CD2, CD19, CD20, CD30, CD38, CD40, CD52, CD70, EGFR / ERBB1, IGF1R, HER3 / ERBB3, HER4 / ERBB4, MUC1, TROP2, cMET, SLAMF7, PSCA, MICA, MICB, TRAILR1, TRAILR2, MAGE-A3, B7.1, B7.2, CTLA4 and PD1.

[0552] It is contemplated that the proteins, conjugates, cells and / or pharmaceutical compositions disclosed herein can be used to treat a variety of cancers, not limited to cancers in which cancer cells express CEACAM5. For example, in certain embodiments, the proteins, conjugates, cells and / or pharmaceutical compositions disclosed herein can be used to treat cancers associated with cells expressing CEACAM5. CEACAM5 is overexpressed in a high percentage of human cancers. Therefore, the methods disclosed herein can be used to treat a variety of cancers in which CEACAM5 is expressed.

[0553] IV. Combination Therapy

[0554] Another aspect of the invention provides combination therapies.The multispecific binding proteins described herein can be used in combination with additional therapeutic agents to treat autoimmune diseases or to treat cancer.

[0555] Exemplary therapeutic agents that can be used as part of a combination therapy for treating autoimmune inflammatory diseases are described in Li et al. (2017) Front. Pharmacol., 8:460, and include, for example, nonsteroidal anti-inflammatory drugs (NSAIDs) (e.g., COX-2 inhibitors), glucocorticoids (e.g., prednisone / prednisolone, methylprednisolone, and fluorinated glucocorticoids such as dexamethasone and betamethasone), disease-modifying antirheumatic drugs (DMARDs) (e.g., methotrexate, leflunomide, gold compounds, sulfasalazine, azathioprine, cyclophosphamide, antimalarials, D-penicillamine, and cyclosporine ), anti-TNF biologics (e.g., infliximab, etanercept, adalimumab, golimumab, certolizumab pegol and their biosimilars), and other biologics targeting CTLA-4 (e.g., abatacept), IL-6 receptor (e.g., tocilizumab), IL-1 (e.g., anakinra), Th1 immune response (IL-12 / IL-23) (e.g., ustekinumab), Th17 immune response (IL-17) (e.g., secukinumab), and CD20 (e.g., rituximab).

[0556] Exemplary therapeutic agents that can be used as part of a combination therapy to treat cancer include, for example, radiation, mitomycin, tretinoin, ribomustin, gemcitabine, vincristine, etoposide, cladribine, dibromomannitol, methotrexate, doxorubicin, carboquinone, pentostatin, diamine nitrazepam, zenastatin, cetrorelix, letrozole, raltitrexed, daunorubicin, fadrozole, fotemustine, thymalfasin, sobuzosane, nedaplatin, cytarabine, bicalutamide, vinorelbine, vesanaline, aminoglutethimide, amsacrine, proglumide, elliptinium acetate, ketanserin, doxifluridine, etretinate, isotretinoin, streptozotocin, nimustine, vindesine, flutamide, flutamide, dr ogenil), butocin, carmofur, razoxane, sizofilan, carboplatin, dibromodulan, tegafur, ifosfamide, prednimustine, picibanil, levamisole, teniposide, improsulfan, enocitabine, lisuride, oxymetholone, tamoxifen, progesterone, melastane, cyclothiodine, formestane, interferon-α, interferon-2-α, interferon-β, interferon-γ (IFN-γ), colony stimulating factor-1, colony stimulating factor-2, denileukin, interleukin-2, luteinizing hormone releasing factor and variants of the above agents, which may exhibit different binding to their cognate receptors, or increased or decreased serum half-life.

[0557] Another class of drugs that can be used as part of a combination therapy for treating cancer is an immune checkpoint inhibitor. Exemplary immune checkpoint inhibitors include agents that inhibit one or more of the following: (i) cytotoxic T lymphocyte-associated antigen 4 (CTLA4), (ii) programmed cell death protein 1 (PD1), (iii) PDL1, (iv) LAG3, (v) B7-H3, (vi) B7-H4, and (vii) TIM3. The CTLA4 inhibitor Ipilimumab has been approved by the U.S. Food and Drug Administration for the treatment of melanoma. In a specific embodiment, the inhibitor can be an antibody, an antigen binding fragment, an immunoadhesin, a fusion protein, or an oligopeptide. In some embodiments, the checkpoint inhibitor is a PD1 inhibitor selected from the group consisting of an anti-PD1 antibody or an anti-PDL1 antibody. In some embodiments, the PD1 inhibitor is selected from nivolumab (OPDIVO, Bristol Myers Squibb, New York, New York), pembrolizumab (KEYTRUDA, Merck Sharp & Dohme Corp, Kenilworth, NJ USA), cemiplimab (Regeneron, Tarrytown, NY) or pidilizumab (CT-011). In some embodiments, the PD1 inhibitor is an immunoadhesin (e.g., an immunoadhesin comprising an extracellular portion of PDL1 or PDL2 fused to a constant region (e.g., an Fc region of an immunoglobulin sequence) or a PD1 binding portion). In some embodiments, the PD1 inhibitor is AMP-224. In some embodiments, the PDL1 inhibitor is an anti-PDL1 antibody, such as durvalumab (IMFINZI, Astrazeneca, Wilmington, DE), atezolizumab (TECENTRIQ, Roche, Zurich, CH) or avelumab (BAVENCIO, EMD Serono, Billerica, MA). In some embodiments, the PDL1 inhibitor is selected from YW243.55.S70, MPDL3280A, MEDI-4736, MSB-0010718C or MDX-1105.

[0558] Still other agents that can be used as part of a combination therapy to treat cancer are monoclonal antibody agents directed against non-checkpoint targets (eg, Herceptin) and non-cytotoxic agents (eg, tyrosine kinase inhibitors).

[0559] Other classes of anticancer agents include, for example: (i) inhibitors selected from the group consisting of ALK inhibitors, ATR inhibitors, A2A antagonists, base excision repair inhibitors, Bcr-Abl tyrosine kinase inhibitors, Bruton's tyrosine kinase inhibitors, CDC7 inhibitors, CHK1 inhibitors, cyclin-dependent kinase inhibitors, DNA-PK inhibitors, inhibitors of DNA-PK and mTOR, DNMT1 inhibitors, DNMT1 inhibitors plus 2-chloro-deoxyadenosine, HDAC inhibitors, Hedgehog signaling pathway inhibitors, IDO inhibitors, JAK inhibitors. preparations, mTOR inhibitors, MEK inhibitors, MELK inhibitors, MTH1 inhibitors, PARP inhibitors, phosphoinositide 3-kinase inhibitors, inhibitors of PARP1 and DHODH, proteasome inhibitors, topoisomerase II inhibitors, tyrosine kinase inhibitors, VEGFR inhibitors and WEE1 inhibitors; (ii) agonists of OX40, CD137, CD40, GITR, CD27, HVEM, TNFRSF25 or ICOS; and (iii) cytokines selected from the group consisting of IL-12, IL-15, GM-CSF and G-CSF.

[0560] The proteins of the invention may also be used as an adjunct to surgical removal of primary lesions.

[0561] The amounts of the multispecific binding protein and the additional therapeutic agent and the relative timing of administration can be selected so as to achieve the desired combined therapeutic effect. For example, when the combination therapy is administered to a patient in need of such administration, the therapeutic agents in the combination or one or more pharmaceutical compositions comprising these therapeutic agents can be administered in any order, e.g., sequentially, simultaneously, together, at the same time, etc. In addition, for example, the multispecific binding protein can be administered during the period when the additional therapeutic agent(s) exert their prophylactic or therapeutic effect, and vice versa.

[0562] V. Pharmaceutical Compositions

[0563] The present disclosure also features pharmaceutical compositions containing a therapeutically effective amount of a protein as described herein. The composition can be formulated for use in a variety of drug delivery systems. One or more physiologically acceptable excipients or carriers can also be included in the composition for appropriate formulation. Formulations suitable for use in the present disclosure are found in Remington's Pharmaceutical Sciences, Mack Publishing Company, Philadelphia, Pa., 17th edition, 1985. For a brief review of drug delivery methods, see, e.g., Langer (Science 249: 1527-1533, 1990).

[0564] Intravenous drug delivery preparations disclosed herein can be contained in a bag, a pen or a syringe. In certain embodiments, the bag can be connected to a passage comprising a pipe and / or a needle. In certain embodiments, the preparation can be a freeze-dried preparation or a liquid preparation. In certain embodiments, the preparation can be freeze-dried (freeze drying). In certain embodiments, the preparation can be a liquid preparation.

[0565] The protein may be present in a liquid aqueous pharmaceutical formulation comprising a therapeutically effective amount of the protein formulated in a buffered solution.

[0566] These compositions can be sterilized by conventional sterilization techniques, or can be sterile filtered. The resulting aqueous solution can be packaged for use as is or freeze-dried, and the freeze-dried preparation is combined with a sterile aqueous carrier before administration. The pH of these preparations will generally be between 3 and 11, more preferably between 5 and 9 or between 6 and 8, and most preferably between 7 and 8, such as 7 to 7.5. The resulting composition in solid form can be packaged in multiple single-dose units, each containing a fixed amount of one or more of the above-mentioned agents. The composition in solid form can also be packaged in a container for flexible amounts.

[0567] In certain embodiments, the present disclosure provides a formulation with an extended shelf life comprising a protein of the present disclosure in combination with mannitol, citric acid monohydrate, sodium citrate, disodium phosphate dihydrate, sodium dihydrogen phosphate dihydrate, sodium chloride, polysorbate 80, water, and sodium hydroxide.

[0568] In certain embodiments, aqueous formulations comprising proteins of the present disclosure in a pH buffered solution are prepared. The pH range of the buffer of the present invention can be from about 4 to about 8, for example, from about 4.5 to about 6.0, or from about 4.8 to about 5.5, or the pH can be from about 5.0 to about 5.2. Ranges in the middle of the above pH are also intended to be part of the present disclosure. For example, numerical ranges using a combination of any of the above values ​​as upper and / or lower limits are intended to be included. Examples of buffers that will control the pH within this range include acetates (e.g., sodium acetate), succinates (such as sodium succinate), gluconates, histidine, citrates, and other organic acid buffers.

[0569] In certain embodiments, the formulation includes a buffer system containing citrate and phosphate to maintain a pH range of about 4 to about 8. In certain embodiments, the pH range can be about 4.5 to about 6.0, or about pH 4.8 to about 5.5, or in a pH range of about 5.0 to about 5.2. In certain embodiments, the buffer system includes citric acid monohydrate, sodium citrate, disodium phosphate dihydrate, and / or sodium dihydrogen phosphate dihydrate. In certain embodiments, the pH of the formulation is adjusted with sodium hydroxide.

[0570] Polyols can also be included in the formulation as tonicifiers and can stabilize antibodies. The amount of the polyol added to the formulation can vary according to the isotonicity of the formulation required. In some embodiments, the aqueous formulation can be isotonic. The amount of the polyol added can also be changed according to the molecular weight of the polyol. For example, compared with disaccharides (such as trehalose), a smaller amount of monosaccharides (e.g., mannitol) can be added. In some embodiments, the polyol that can be used as a tonicity agent in the formulation is mannitol.

[0571] Detergents or surfactants can also be added to the preparation. Exemplary detergents include nonionic detergents, such as polysorbates (e.g., polysorbate 20, polysorbate 80, etc.) or poloxamer (e.g., poloxamer 188). The amount of the detergent added is such that it reduces the aggregation of the antibody formulated and / or minimizes the formation of particles in the preparation and / or reduces adsorption. In certain embodiments, the surfactant that the preparation can include is polysorbate. In certain embodiments, the preparation can contain the detergent polysorbate 80 or Tween 80. Tween 80 is a term used to describe polyoxyethylene (20) sorbitan monooleate (see Fiedler, Lexikon der Hifsstoffe, Editio Cantor Verlag Aulendorf, 4th edition, 1996).

[0572] In some embodiments, the protein product of the present disclosure is formulated as a liquid formulation. The liquid formulation can be presented in a USP / Ph Eur I type 50R bottle, closed with a rubber stopper and sealed with an aluminum crimp seal cap. The stopper can be made of an elastomer that meets the requirements of USP and Ph Eur. In certain embodiments, the bottle can be filled with a protein product solution to allow for extractable volume. In certain embodiments, the liquid formulation can be diluted with a saline solution.

[0573] In certain embodiments, liquid preparations of the present disclosure can be prepared in combination with sugars having a level of stabilizing effect. In certain embodiments, the liquid preparation can be prepared in an aqueous carrier. In certain embodiments, the amount of the stabilizer added does not exceed the amount of viscosity that may cause undesirable or unsuitable intravenous administration. In certain embodiments, the sugar may be a disaccharide, for example, sucrose. In certain embodiments, the liquid preparation may also include one or more of a buffer, a surfactant, and a preservative.

[0574] In certain embodiments, the pH of the liquid formulation can be set by adding a pharmaceutically acceptable acid and / or base. In certain embodiments, the pharmaceutically acceptable acid can be hydrochloric acid. In certain embodiments, the base can be sodium hydroxide.

[0575] In addition to aggregation, deamidation is a common product variant of peptides and proteins that may occur during fermentation, harvesting / cell clarification, purification, storage of drug substance / drug product, and sample analysis. Deamidation is the loss of NH3 from a protein to form a succinimide intermediate that can undergo hydrolysis. The succinimide intermediate results in a mass reduction of 17 Daltons in the parent peptide. Subsequent hydrolysis results in a mass increase of 18 Daltons. Due to its instability under aqueous conditions, it is difficult to separate the succinimide intermediate. Therefore, deamidation is usually detectable as a mass increase of 1 Dalton. Deamidation of asparagine produces aspartic acid or isoaspartic acid. Parameters that affect the rate of deamidation include pH, temperature, solvent dielectric constant, ionic strength, primary sequence, local polypeptide conformation, and tertiary structure. The amino acid residues adjacent to Asn in the peptide chain affect the rate of deamidation. In the protein sequence, Gly and Ser after Asn lead to higher sensitivity to deamidation.

[0576] In certain embodiments, the liquid formulations of the present disclosure may be stored under certain pH and humidity conditions to prevent deamination of the protein product.

[0577] The aqueous carrier of interest herein is an aqueous carrier that is pharmaceutically acceptable (safe and non-toxic to humans) and can be used to prepare liquid formulations. Illustrative carriers include sterile water for injection (SWFI), bacteriostatic water for injection (BWFI), pH buffered solutions (e.g., phosphate buffered saline), sterile saline, Ringer's solution, or dextrose solution.

[0578] Preservatives may optionally be added to the formulations herein to reduce bacterial action. The addition of a preservative may, for example, aid in the production of multi-use (multi-dose) formulations.

[0579] Intravenous (IV) formulations may be the preferred route of administration in certain circumstances, such as when the patient is hospitalized after a transplant and receives all medications by IV. In certain embodiments, the liquid formulation is diluted with 0.9% sodium chloride solution prior to administration. In certain embodiments, the diluted drug product for injection is isotonic and suitable for administration by intravenous infusion.

[0580] In certain embodiments, salt or buffer components can be added in an amount of 10mM-200mM. These salts and / or buffers are pharmaceutically acceptable and are derived from various known acids (inorganic and organic acids) with "alkali-forming" metals or amines. In certain embodiments, the buffer can be a phosphate buffer. In certain embodiments, the buffer can be a glycine salt, carbonate, citrate buffer, in which case sodium, potassium or ammonium ions can be used as counter ions.

[0581] Preservatives may optionally be added to the formulations herein to reduce bacterial action. The addition of a preservative may, for example, aid in the production of multi-use (multi-dose) formulations.

[0582] The aqueous carrier of interest herein is an aqueous carrier that is pharmaceutically acceptable (safe and non-toxic to humans) and can be used to prepare liquid formulations. Illustrative carriers include sterile water for injection (SWFI), bacteriostatic water for injection (BWFI), pH buffered solutions (e.g., phosphate buffered saline), sterile saline, Ringer's solution, or dextrose solution.

[0583] The proteins of the present disclosure may be present in a freeze-dried formulation comprising the proteins and a lyoprotectant. The lyoprotectant may be a sugar, for example, a disaccharide. In certain embodiments, the lyoprotectant may be sucrose or maltose. The freeze-dried formulation may also include one or more of a buffer, a surfactant, a filler and / or a preservative.

[0584] The amount of sucrose or maltose that can be used to stabilize the freeze-dried drug product can be at least a 1:2 weight ratio of protein to sucrose or maltose. In certain embodiments, the weight ratio of protein to sucrose or maltose can be from 1:2 to 1:5.

[0585] In certain embodiments, the pH of the formulation can be set by adding a pharmaceutically acceptable acid and / or base prior to freeze drying. In certain embodiments, the pharmaceutically acceptable acid can be hydrochloric acid. In certain embodiments, the pharmaceutically acceptable base can be sodium hydroxide.

[0586] Prior to freeze drying, the pH of the solution containing the protein of the present disclosure may be adjusted between 6 and 8. In certain embodiments, the pH of the freeze-dried drug product may range from 7 to 8.

[0587] In certain embodiments, salt or buffer components can be added in an amount of 10mM-200mM. These salts and / or buffers are pharmaceutically acceptable and are derived from various known acids (inorganic and organic acids) with "alkali-forming" metals or amines. In certain embodiments, the buffer can be a phosphate buffer. In certain embodiments, the buffer can be a glycine salt, carbonate, citrate buffer, in which case sodium, potassium or ammonium ions can be used as counter ions.

[0588] In certain embodiments, a "filler" may be added. A "filler" is a compound that increases the mass of the freeze-dried mixture and contributes to the physical structure of the freeze-dried cake (e.g., helps produce a substantially uniform freeze-dried cake that maintains an open pore structure). Illustrative fillers include mannitol, glycine, polyethylene glycol, and sorbitol. The freeze-dried formulations of the present invention may contain such fillers.

[0589] Preservatives may optionally be added to the formulations herein to reduce bacterial action. The addition of a preservative may, for example, aid in the production of multi-use (multi-dose) formulations.

[0590] In certain embodiments, the freeze-dried drug product can be reconstituted with an aqueous carrier. The aqueous carrier of interest herein is pharmaceutically acceptable (e.g., safe and nontoxic to humans) and can be used to prepare an aqueous carrier for liquid formulations after freeze drying. Illustrative diluents include sterile water for injection (SWFI), bacteriostatic water for injection (BWFI), pH buffer solutions (e.g., phosphate buffered saline), sterile saline, Ringer's solution, or dextrose solution.

[0591] In certain embodiments, the freeze-dried drug product of the present disclosure is reconstituted with Sterile Water for Injection, USP (SWFI) or 0.9% Sodium Chloride Injection, USP. During the reconstitution process, the freeze-dried powder dissolves into a solution.

[0592] In certain embodiments, the freeze-dried protein product of the present disclosure is reconstituted with water for injection and diluted with 0.9% saline solution (sodium chloride solution).

[0593] Actual dosage levels of the active ingredients in the pharmaceutical compositions of this invention may be varied so as to obtain an amount of the active ingredient which is effective to achieve the desired therapeutic response for a particular patient, composition, and mode of administration, without being toxic to the patient.

[0594] The specific dosage can be a uniform dosage for each patient, for example, 50-5000 mg protein. Alternatively, the dosage of the patient can be adjusted according to the approximate body weight or surface area of ​​the patient. Other factors determining the appropriate dosage may include the disease or condition to be treated or prevented, the severity of the disease, the route of administration, and the age, sex, and medical condition of the patient. The calculation required to determine the appropriate dosage for treatment can be routinely further refined by those skilled in the art, especially according to the dosage information and determination disclosed herein. The dosage can also be determined by a combination of known determinations for determining dosage and suitable dose-response data. The dosage of a single patient can be adjusted with the monitoring of disease progression. The blood level of the targetable construct or complex of the patient can be measured to understand whether the dosage needs to be adjusted to reach or maintain an effective concentration. Pharmacogenomics can be used to determine which targetable constructs and / or complexes and their dosages are most likely to be effective for a given individual (Schmitz et al., Clinica Chimica Acta 308:43-53, 2001; Steimer et al., Clinica Chimica Acta 308:33-41, 2001).

[0595] Dosage can be given once or multiple times every day, every week, every month or every year, or even every 2 to 20 years. Based on the residence time and concentration of the targetable construct or complex in body fluid or tissue measured, those of ordinary skill in the art can easily estimate the repetition frequency of administration. The administration of the present invention can be intravenous, intraarterial, intraperitoneal, intramuscular, subcutaneous, intrapleural, intrathecal, intracavitary, by catheter perfusion or by direct intralesional injection. This can be used once or multiple times a day, once or multiple times a week, once or multiple times a month and once or multiple times a year.

[0596] The above description describes various aspects and embodiments of the present invention. This patent application specifically contemplates all combinations and permutations of these aspects and embodiments.

[0597] Example

[0598] The invention now generally described will be more readily understood by reference to the following examples, which are included merely for the purpose of illustrating certain aspects and embodiments of the invention and are not intended to limit the invention.

[0599] Example 1. Immunization and hybridoma generation

[0600] BALB / cJ mice were purchased from The Jackson Laboratory (Stock No. 000651) and immunized with one of the following: isogenic Ba / F3 cell lines overexpressing cynomolgus and / or human CEACAM5; hCEACAM5-A3B3-domain; recombinant cynomolgus and / or human CEACAM5; or human NABA constructs (NABA consists of the N, A1, and A2 domains of human CEACAM1 and the B3 domain of human CEACAM5). These mice were used to generate murine mAbs (monoclonal antibodies). H2L2 TM Mice were purchased from Harbour Biomed and immunized with one of the following: isogenic Ba / F3 cell lines overexpressing cynomolgus and / or human CEACAM5; hCEACAM5-A3B3-domain; recombinant cynomolgus and / or human CEACAM5; or human NABA construct (NABA consists of the N, A1 and A2 domains of human CEACAM1 and the B3 domain of human CEACAM5). These mice were used to generate human / rat chimeric mAbs with human variable regions and rat constant regions. Thereafter, spleen cells from immunized mice were fused with mouse myeloma cells to generate hybridoma cells.

[0601] The fused hybridoma cells were cultured in supplemented DMEM medium in a humidified atmosphere of 8% CO2 at 37°C. The supernatants of the hybridomas were evaluated for CEACAM5, CEACAM1, CEACAM6, and CEACAM8 binding by enzyme-linked immunosorbent assay (ELISA) and multiplex surface plasmon resonance (SPR) (Carterra LSA). The CEACAM5 antigen-specific hybridoma was subsequently subcloned. Based on preliminary multiplex SPR (Carterra) binding affinity estimates, binding to cells expressing human and cynomolgus monkey CEACAM5, binding to CEACAM5 + Binding to cancer cell lines as well as epitope diversity, clones were selected for further studies. For clone 16F6.A2-CEACAM5-B.02, cross-reactivity with cynomolgus monkey CEACAM5 was observed.

[0602] Using AmMag TMBalb / cJ murine mAb was purified from hybridoma supernatants by protein A chromatography using protein A magnetic beads (P / N L00695, Genscript Biotech, Piscataway, NJ). The beads were balanced with 1.5M glycine, 3.0M NaCl pH 8.5. The supernatant was diluted 1:1 with 1.5M glycine, 3.0M NaCl pH 8.5 and incubated with ProA magnetic beads for 1 h with gentle shaking. The beads were washed with 1.5M glycine, 3.0M NaCl pH 8.5 to remove unbound protein. The antibody was eluted with 100mM glycine pH 3.0 and immediately neutralized to pH 7.5 with 1.0M Tris, pH 8.3. The protein concentration was determined by A280 using a Nanodrop spectrophotometer. H2L2 source human (variable region) / rat (constant region) mAb was purified from hybridoma supernatant by protein G chromatography (Global Life Sciences Solutions, Marlborough, MA). Protein G chromatographic column was balanced with 50mM sodium acetate, pH 5.0+10mM NaCl. The supernatant was diluted 10 times with 50mM sodium acetate, pH 5.0+10mM NaCl and gently shaken for 1h with balanced protein G medium. The beads were washed with 50mM sodium acetate, pH 5.0+10mM NaCl to remove unbound proteins. The antibody was eluted with 100mM glycine pH 2.5 and immediately neutralized to pH 7.5 with 1.0M Tris, pH 8.3. After purification, protein concentration was determined by A280 using a Nanodrop spectrophotometer. Similarly, H2L2-derived human / rat chimeric mAb (having human variable regions and rat constant regions) was purified from hybridoma supernatants by protein G chromatography (Global Life Sciences Solutions, Marlborough, MA); protein concentration was determined by A280 using a Nanodrop spectrophotometer.

[0603] Balb / cJ mouse mAbs were tested for cell surface binding to CEACAM5 and cross-reactivity to CEACAM1, CEACAM6, and CEACAM8. In addition, Balb / cJ mouse mAbs were tested for in vitro binding to CEACAM1, CEACAM5, CEACAM6, and CEACAM8 using surface plasmon resonance (SPR). Experiments were performed at 37°C using Carterra LSA or Biacore 8K instruments to simulate physiological temperature.

[0604] The cell surface binding of H2L2-derived human / rat chimeric mAbs to CEACAM5 and cross-reactivity with CEACAM1, CEACAM6, and CEACAM8 were tested. In addition, the in vitro binding of H2L2-derived human / rat chimeric mAbs to CEACAM5, CEACAM1, CEACAM6, and CEACAM8 was tested using surface plasmon resonance (SPR). Experiments were performed at 37°C using Carterra LSA or Biacore 8K instruments to simulate physiological temperature.

[0605] Example 2. Generation of binders using yeast display technology

[0606] Additional CEACAM5 binders were generated by constructing an H2L2 immune library using yeast display technology. Briefly, yeast were transfected with a starting construct containing the parental CEACAM5 VH sequence and / or VL sequence. Novel clones were selected and binders were isolated and characterized as described above.

[0607] Yeast strains and plasmids

[0608] MATa auxotrophic Saccharomyces cerevisiae strain EBY100 (Meyen ex ECHansen ( MYA-4941 TM )) was used to construct a scFv yeast display library. EBY100 has a genomic insertion of AGA1 for surface display; its expression is regulated by a galactose promoter with a uracil selectable marker. Each scFv also contains a carboxyl-terminal flag tag also controlled by the galactose promoter.

[0609] Construction of CEACAM5 immune scFv yeast display library

[0610] After the immunization procedure, mouse spleen cells were collected, RNA was extracted with a high-purity RNA isolation kit (Roche, product number 11828665001) according to the manufacturer's instructions, and RT-PCR was performed to generate a cDNA library (Invitrogen, 18090010). The cDNA was used as a template for amplifying variable heavy chain (VH) and light chain (VL; only κ) antibody genes, which were then assembled into single-chain antibody fragments (scFv) libraries in two directions (VH-VL or VL-VH scFv). After transformation (electroporation) and homologous recombination, the scFv library was integrated into a yeast surface display vector.

[0611] Isolation of CEACAM5-specific scFv binders from the library

[0612] After electroporation, yeast cells were grown in a selective medium (Teknova product number C8240). The scFv library was then induced to display scFvs on the surface of yeast cells by switching to galactose medium. Isolation of CEACAM5-specific binders was accomplished by three rounds of selection. The library was first screened for human CEACAM5 hits using magnetic activated cell sorting (MACS), followed by two rounds of selection by fluorescence activated cell sorting (FACS) to generate a test set of CEACAM5-specific scFvs. The CEACAM5 binding affinity and specificity of individual scFv clones were characterized and displayed on yeast, and then transferred to vectors for expression in mammalian cells using molecular biology techniques known in the art.

[0613] Example 3. Humanization and Generation of Sequence Liability Variants

[0614] Balb / cJ hybridoma mAbs were humanized. Humanized variants were generated by grafting the murine CDRs onto human framework regions. The following clones were humanized: murine 1A1.A3-CEACAM5-B.02 and murine 16F6.A2-CEACAM5-B.02. Their sequences and the sequences of their humanized variants are provided in Table 4. The murine VH and VL sequences of clone 16F6.A2-CEACAM5-B.02 were blasted against the human sequence database to identify suitable frameworks to accommodate the CDRs. IGHV1-2*02 (sequence identity: 66.3%) and IGKV4-1*01 (sequence identity: 80.2%) were selected to move forward. Structural models were built to examine and identify potential back mutations. After the hypervariable regions from the murine corresponding antibody were grafted onto the above human framework, the following back mutations were introduced into the variable heavy chain (VH): V67T, M69L, R71A, S76P and A93T (all Chothia numbering). There were no back mutations in the variable light chain (VL).

[0615] The mouse VH and VL sequences of clone 1A1.A3 were blasted against a human sequence database to identify suitable frameworks to accommodate the CDRs. IGHV1-69-2*01 (sequence identity: 63.5%) and IGKV3-11*01 (sequence identity: 64.2%) were selected to move forward. A structural model was built to examine and identify potential back mutations. After the hypervariable regions from the mouse corresponding antibodies were transplanted into the above human frameworks, the following back mutations in VH were introduced: V5Q, K12V, I20L, V24A, Q38T, M48I, V67A, I69M, M80L, A93N and T94V (all Chothia numbering). The following back mutations in VL were also introduced: L13A, Y36F, L47W, I58V and F71Y.

[0616] The potential sequence disadvantages of the clones were examined. The following potential sequence disadvantages were considered: M (potential oxidation site); NG, NS and NT sequence motifs (potential deamidation sites); DG, DS and DT sequence motifs (potential isomerization sites); and DP sequence motifs (potential sites for chemical hydrolysis). The variants of these antibodies were designed to remove the putative sequence disadvantage motifs; the sequences of such variants are provided in Table 4.

[0617] Example 4. Classification of monoclonal antibodies

[0618] As indicated in Tables 3 and 4, mAbs are classified into tiers (tier 1, tier 2, and tier 3). mAbs classified as "tier 1" are CEACAM5-specific high affinity mAbs with KD values ​​less than or equal to about 15 nM (e.g., KD ranges from about 4 nM to about 15 nM). mAbs classified as tier 2 are CEACAM5-specific low to medium affinity mAbs with KD values ​​ranging from about 25 nM to about nM to about 80 nM (for medium affinity), and about 120 nM to about 560 nM (for low affinity). mAbs classified as "tier 3" are mAbs that exhibit low cross-reactivity with CEACAM1, CEACAM6, and CEACAM8 in addition to binding to CEACAM5. For classification purposes, mAbs generated by the method of Example 1 were evaluated in mAb format, while mAbs generated by the method of Example 2 were evaluated in the form of multispecific antibodies containing a scFv that binds CEACAM5 and a Fab fragment that binds a different antigen.

[0619] The clones for each tier are listed below in Table 13. The VH, VL and CDR sequences of these clones are provided in Table 4.

[0620] Table 13. Classification of antibody clones

[0621]

[0622]

[0623]

[0624] Example 5. NKG2D binding domain binding to NKG2D

[0625] Binding affinities of various NKG2D binding domains

[0626] The kinetics and affinities of various NKG2D binding domains were evaluated by surface plasmon resonance using a Biacore 8K instrument (GE Healthcare). Anti-human Fc antibodies were immobilized on a CM5 chip using standard amine coupling chemistry. Human monoclonal antibodies containing various NKG2D binding domains were captured on the anti-human Fc chip at a density of approximately 100 RU. A solution containing 0.411–100 nM soluble mouse Fc-human NKG2D dimer was injected at 30 μl / min over the captured NKG2D antibody and control surfaces at 37°C. The surface was regenerated between cycles by rapid injection of 10 mM glycine, pH 1.8. To obtain kinetic rate constants, dual reference data were fit to a 1:1 interaction model using the Biacore 8K evaluation software (GE Healthcare). The dissociation constant k d and the binding constant k a The ratio (k d / k a ) Determine the equilibrium binding constant K D As shown in Table 14 below, the binding affinity of the NKG2D binding domain to NKG2D is in the range of 10–62 nM.

[0627] Table 14. Binding affinity of NKG2D binding domain

[0628] NKG2D binding domain <![CDATA[k a (1 / Ms)]]> <![CDATA[k d (1 / s)]]> <![CDATA[K D (nM)]]> ADI-27744(A44) 2.95E+05 2.99E-03 10.1 ADI-27749(A49) 3.95E+05 4.89E-03 12.4 ADI-29378(E78) 8.32E+05 4.87E-02 58.5 ADI-29379(E79) 4.43E+05 2.25E-02 50.7 ADI-29463(F63) 1.64E+06 1.01E-01 61.8

[0629] NKG2D binding domain binds to purified recombinant NKG2D

[0630] The nucleic acid sequence of the extracellular domain of human, mouse or cynomolgus monkey NKG2D is fused with the nucleic acid sequence encoding the human IgG1Fc domain and introduced into mammalian cells for expression. After purification, the NKG2D-Fc fusion protein is adsorbed to the wells of the microplate. After the wells are blocked with bovine serum albumin to prevent nonspecific binding, the NKG2D binding domain is titrated and added to the wells of the pre-adsorbed NKG2D-Fc fusion protein. The primary antibody binding is detected using a secondary antibody conjugated to horseradish peroxidase and specifically recognizing human κ light chain to avoid Fc cross-reactivity. The substrate 3,3',5,5'-tetramethylbenzidine (TMB) of horseradish peroxidase is added to the wells to visualize the binding signal, and its absorbance is measured at 450nM and corrected at 540nM. NKG2D binding domain clones, isotype controls or positive controls (comprising a heavy chain variable domain and a light chain variable domain selected from the group consisting of SEQ ID NOs: 526-529, or anti-mouse NKG2D clones MI-6 and CX-5 available at eBioscience) were added to each well.

[0631] The isotype control showed minimal binding to the recombinant NKG2D-Fc protein, while the positive control had the strongest binding to the recombinant antigen. The NKG2D binding domains produced by all clones showed binding between human, mouse, and cynomolgus monkey recombinant NKG2D-Fc proteins, although the affinity varied between clones. Overall, each anti-NKG2D clone bound to human ( Fig.18 ) and cynomolgus monkeys ( Fig.19 ) recombinant NKG2D-Fc, but binds with lower affinity to mouse ( Fig. 20 )Recombinant NKG2D-Fc.

[0632] NKG2D binding domain binds to cells expressing NKG2D

[0633] The EL4 mouse lymphoma cell line was engineered to express human or mouse NKG2D-CD3ζ signaling domain chimeric antigen receptors. NKG2D binding clones, isotype controls or positive controls were used at a concentration of 100 nM to stain extracellular NKG2D expressed on EL4 cells. Antibody binding was detected using a fluorophore-conjugated anti-human IgG secondary antibody. Cells were analyzed by flow cytometry, and the mean fluorescence intensity (MFI) of cells expressing NKG2D compared to parental EL4 cells was used to calculate the multiple relative to background (FOB).

[0634] The NKG2D binding domains produced by all clones bound to EL4 cells expressing human and mouse NKG2D. The positive control antibody (comprising a heavy chain variable domain and a light chain variable domain selected from the group consisting of SEQ ID NOs: 526-529, or anti-mouse NKG2D clones MI-6 and CX-5 available at eBioscience) gave the best FOB binding signal. The NKG2D binding affinity of each clone was expressed in EL4 cells expressing human NKG2D ( Fig.21 ) and mice ( Fig. 22 )NKG2D is similar between cells.

[0635] Example 6. NKG2D binding domain blocks natural ligand binding to NKG2D

[0636] Competition with ULBP-6

[0637] Recombinant human NKG2D-Fc protein is adsorbed to the wells of microplates, and the wells are blocked with bovine serum albumin to reduce nonspecific binding. Saturated concentrations of ULBP-6-His-biotin are added to the wells, followed by the NKG2D binding domain clone. After 2 hours of incubation, the wells are washed and the ULBP-6-His-biotin that remains bound to the wells coated with NKG2D-Fc is detected by conjugating to streptavidin and TMB substrate of horseradish peroxidase. The absorbance is measured at 450nM and corrected at 540nM. After background subtraction, the specific binding of the NKG2D binding domain to the NKG2D-Fc protein is calculated by the percentage of ULBP-6-His-biotin that is blocked from binding to the NKG2D-Fc protein in the wells. The positive control antibody (comprising a heavy chain variable domain and a light chain variable domain selected from the group consisting of SEQ ID NOs: 526-529) and various NKG2D binding domains blocked ULBP-6 binding to NKG2D, while the isotype control showed little competition with ULBP-6 ( Fig.23 ).

[0638] The ULBP-6 sequence is represented by SEQ ID NO:566.

[0639]

[0640] Competing with MICA

[0641] Recombinant human MICA-Fc protein was adsorbed to the wells of microplates, and the wells were blocked with bovine serum albumin to reduce nonspecific binding. NKG2D-Fc-biotin was added to the wells, followed by the NKG2D binding domain. After incubation and washing, streptavidin-HRP and TMB substrates were used to detect the NKG2D-Fc-biotin that remained bound to the wells coated with MICA-Fc. The absorbance was measured at 450nM and corrected at 540nM. After background subtraction, the specific binding of the NKG2D binding domain to the NKG2D-Fc protein was calculated by the percentage of the NKG2D-Fc-biotin that was blocked to bind to the wells coated with MICA-Fc. Positive control antibodies (comprising heavy chain variable domains and light chain variable domains selected from the group consisting of SEQ ID NO: 526-529) and various NKG2D binding domains blocked MICA from binding to NKG2D, while isotype controls showed little competition with MICA ( Fig.24 ).

[0642] Competition with Rae-1δ

[0643] Recombinant mouse Rae-1δ-Fc (purchased from R&D Systems) was adsorbed to the wells of the microplate and the wells were blocked with bovine serum albumin to reduce nonspecific binding. Mouse NKG2D-Fc-biotin was added to the wells, followed by the NKG2D binding domain. After incubation and washing, streptavidin-HRP and TMB substrates were used to detect the NKG2D-Fc-biotin that remained bound to the wells coated with Rae-1δ-Fc. The absorbance was measured at 450nM and corrected at 540nM. After background subtraction, the specific binding of the NKG2D binding domain to the NKG2D-Fc protein was calculated by the percentage of NKG2D-Fc-biotin that was blocked from binding to the wells coated with Rae-1δ-Fc. Positive control antibodies (comprising a heavy chain variable domain and a light chain variable domain selected from the group consisting of SEQ ID NOs: 526-529, or anti-mouse NKG2D clones MI-6 and CX-5 available at eBioscience) and various NKG2D binding domain clones blocked Rae-1δ binding to mouse NKG2D, while isotype controls showed little competition with Rae-1δ ( Fig.25 ).

[0644] Example 7. NKG2D binding domain cloning and activation of NKG2D

[0645] The nucleic acid sequence of human and mouse NKG2D is fused to the nucleic acid sequence encoding the CD3ζ signaling domain to obtain a chimeric antigen receptor (CAR) construct. The NKG2D-CAR construct is then cloned into a retroviral vector using Gibson assembly and transfected into expi293 cells for the production of retroviruses. EL4 cells are infected with a virus containing NKG2D-CAR and 8 μg / mL polybrene. 24 hours after infection, the expression level of NKG2D-CAR in EL4 cells was analyzed by flow cytometry, and clones expressing high levels of NKG2D-CAR on the cell surface were selected.

[0646] To determine whether the NKG2D binding domains activated NKG2D, they were adsorbed to wells of a microtiter plate and NKG2D-CAR EL4 cells were cultured in antibody fragment-coated wells in the presence of brefeldin A and monensin for 4 hours. Intracellular TNF-α production (an indicator of NKG2D activation) was measured by flow cytometry. The percentage of TNF-α positive cells was normalized to cells treated with the positive control. All NKG2D binding domains activated human NKG2D ( Fig.26 ) and mouse NKG2D ( Fig. 27 )Both.

[0647] Example 8. NKG2D binding domain activates NK cells

[0648] Primary human NK cells

[0649] Peripheral blood mononuclear cells (PBMCs) were isolated from human peripheral blood buffy coats using density gradient centrifugation. NK cells (CD3 - CD56 + ), and the purity of the isolated NK cells is usually >95%. The isolated NK cells are then cultured in medium containing 100ng / mL IL-2 for 24-48 hours, and then transferred to the wells of a microplate adsorbed with the NKG2D binding domain and cultured in medium containing fluorophore-conjugated anti-CD107a antibodies, brefeldin A, and monensin. After culture, NK cells are measured by flow cytometry using fluorophore-conjugated antibodies against CD3, CD56, and IFN-γ. In CD3 - CD56 + CD107a and IFN-γ staining were analyzed in the cells to evaluate NK cell activation. The increase in CD107a / IFN-γ double positive cells indicates that NK cells are better activated by engaging two activation receptors rather than one receptor. The NKG2D binding domain and positive controls (e.g., a heavy chain variable domain represented by SEQ ID NO: 526 or SEQ ID NO: 528, and a light chain variable domain represented by SEQ ID NO: 527 or SEQ ID NO: 529) show that NK cells become CD107a + and IFN-γ + The percentage was higher than that of isotype control ( Fig.28 and Fig.29 Representative data from two independent experiments, each using PBMCs from different donors for NK cell preparation).

[0650] Primary mouse NK cells

[0651] Spleens were obtained from C57Bl / 6 mice and crushed through a 70 μm cell strainer to obtain a single cell suspension. The cells were pelleted and resuspended in ACK lysis buffer (purchased from Thermo Fisher Scientific #A1049201; 155 mM ammonium chloride, 10 mM potassium bicarbonate, 0.01 mM EDTA) to remove red blood cells. The remaining cells were cultured with 100 ng / mL hIL-2 for 72 hours and then harvested and prepared for NK cell isolation. NK cells (CD3 - NK1.1 +), typically >90% pure. Purified NK cells were cultured in medium containing 100 ng / mL mIL-15 for 48 hours and then transferred to wells of a microplate adsorbed with the NKG2D binding domain and cultured in medium containing fluorophore-conjugated anti-CD107a antibodies, radofelibactin A, and monensin. After culture in wells coated with the NKG2D binding domain, NK cells were measured by flow cytometry using fluorophore-conjugated antibodies against CD3, NK1.1, and IFN-γ. - NK1.1 + CD107a and IFN-γ staining were analyzed in the cells to assess NK cell activation. The increase in CD107a / IFN-γ double positive cells indicates that NK cells are better activated by engaging two activating receptors rather than one receptor. The NKG2D binding domain and positive controls (selected from the group consisting of: anti-mouse NKG2D clones MI-6 and CX-5 available at eBioscience) showed that NK cells became CD107a + and IFN-γ + The percentage was higher than that of isotype control ( Fig.30 and Fig.31 Representative data from two independent experiments, each using different mice for NK cell preparation).

[0652] Example 9. NKG2D binding domain enables cytotoxicity of target tumor cells

[0653] Human and mouse primary NK cell activation assays demonstrated an increase in cytotoxic markers on NK cells after incubation with the NKG2D binding domain. To address the question of whether this translates into increased tumor cell lysis, a cell-based assay was used in which each NKG2D binding domain was developed as a monospecific antibody. The Fc region was used as one targeting arm, and the Fab fragment region (NKG2D binding domain) as the other targeting arm to activate NK cells. THP-1 cells are derived from humans and express high levels of Fc receptors, used as tumor targets, and the Perkin Elmer DELFIA cytotoxicity kit was used. THP-1 cells were labeled with BATDA reagent and 10 5 The NK cells were then resuspended in culture medium at 37 ° C. The labeled THP-1 cells were combined with NKG2D antibodies and isolated mouse NK cells in the wells of microtiter plates for 3 hours. After incubation, 20 μL of culture supernatant was taken out, mixed with 200 μL of europium solution, and incubated for 15 minutes in dark shaking. The fluorescence was measured over time by a PheraStar plate reader equipped with a time-resolved fluorescence module (excitation 337 nm, emission 620 nm), and specific lysis was calculated according to the kit instructions.

[0654] The positive control ULBP-6 (natural ligand of NKG2D) showed an increase in the specific lysis of THP-1 target cells by mouse NK cells. The NKG2D antibody also increased the specific lysis of THP-1 target cells, while the isotype control antibody showed a decreased specific lysis. The dotted line indicates the specific lysis of THP-1 cells by mouse NK cells without the addition of antibody ( Fig.32 ).

[0655] Example 10. NKG2D antibodies show high thermal stability

[0656] The melting temperature of the NKG2D binding domain was determined using differential scanning fluorimetry. The inferred apparent melting temperature is high relative to typical IgG1 antibodies ( Fig.33 ).

[0657] Example 11. Synergistic activation of human NK cells by cross-linking NKG2D and CD16

[0658] Primary human NK cell activation assay

[0659] Peripheral blood mononuclear cells (PBMCs) were isolated from human peripheral blood buffy coats using density gradient centrifugation. NK cells were purified from PBMCs using negative magnetic beads (StemCell, #17955). NK cells were >90% CD3 as determined by flow cytometry. - CD56 + . The cells were then expanded in medium containing 100 ng / mL hIL-2 (Peprotech, #200-02) for 48 hours before being used in activation assays. Antibodies at a concentration of 2 μg / mL (anti-CD16, Biolegend, #302013) and 5 μg / mL (anti-NKG2D, R&D #MAB139) in 100 μL sterile PBS were coated onto 96-well flat-bottom plates overnight at 4°C, followed by thorough washing of the wells to remove excess antibody. For the assessment of degranulation, IL-2 activated NK cells were plated at 5 × 10 5 1×10 cells / mL were resuspended in culture medium supplemented with 100 ng / mL human IL-2 (hIL2) and 1 μg / mL APC-conjugated anti-CD107a mAb (Biolegend #328619). 5Cells / well were added to the antibody-coated plates. Protein transport inhibitors Brefeldin A (BFA, Biolegend #420601) and Monensin (Biolegend #420701) were added at final dilutions of 1:1000 and 1:270, respectively. The plated cells were incubated for 4 hours at 37°C and 5% CO2. For intracellular staining of IFN-γ, NK cells were labeled with anti-CD3 (Biolegend #300452) and anti-CD56 mAb (Biolegend #318328), followed by fixation, permeabilization and labeling with anti-IFN-γ mAb (Biolegend #506507). In the presence of live CD56 + CD3 - After cell gating, the expression of CD107a and IFN-γ in NK cells was analyzed by flow cytometry.

[0660] To investigate the relative potency of receptor combinations, cross-linking of NKG2D or CD16 as well as co-cross-linking of both receptors was performed by plate-bound stimulation. Figure 34A-Figure 34C ), combined stimulation of CD16 and NKG2D resulted in greatly increased CD107a levels (degranulation) ( Fig.34A ) and / or IFN-γ production ( Fig.34B ). Dashed lines represent the additive effect of individual stimulation of each receptor.

[0661] IL-2 activated NK cells were analyzed for CD107a levels and intracellular IFN-γ production after 4 hours of plate-bound stimulation with anti-CD16, anti-NKG2D or a combination of both monoclonal antibodies. Graphs indicate mean (n=2) ± SD. Fig.34A The levels of CD107a are shown; Fig.34B The levels of IFN-γ are shown; Fig.34C Levels of CD107a and IFN-γ are shown. Figure 34A-Figure 34C The data shown are representative of five independent experiments using five different healthy donors.

[0662] Example 12. Generation of CEACAM5 TriNKET Binding Protein

[0663] Generation of AB0264 and AB0621 scFv from AB0131

[0664] AB0131 is a scFv identified from the BALB / c immunization work described in Example 1. AB0131 is derived from clone 16F6.A2-CEACAM5-B.02-BM. To generate humanized AB0264, five (5) back mutations were introduced into the VH domain of AB0131 and two (2) cysteines were introduced to stabilize disulfide bonds. The scFv polypeptide sequence of AB0264 is set forth below as SEQ ID NO: 703. Back mutations are identified in bold, and introduced cysteines are identified in bold underline.

[0665]

[0666] The proline residue in the VH domain of AB0264 was found to be present in the human framework at a frequency of <1%. This residue, identified in bold italics, was replaced with serine (36% of human framework sequences; Abysis) to generate AB0621. The scFv polypeptide sequence of AB0621 is set forth below as SEQ ID NO: 714.

[0667]

[0668] Generation of AB0411 scFv from AB0100 scFv

[0669] AB0100 is a fully human scFv identified by querying the H2L2 yeast immune library described in Example 2. AB0100 is derived from a variant of clone 1078_C04CEACAM5. The glutamine residue in the VH domain of AB0100 was found to be present in <1% of the human framework. To generate AB0411, the glutamine residue was replaced with leucine (63% human framework; Abysis), and two (2) cysteines were introduced to stabilize the disulfide bonds. The scFv polypeptide sequence of AB0411 is set forth below as SEQ ID NO:707. The stabilizing cysteine ​​is identified in bold underline, while the leucine replacement is shown in bold italics.

[0670]

[0671] Generation of AB0466 scFv from AB0073 scFv

[0672] AB0073 is a fully human scFv identified by querying the H2L2 yeast immune library described in Example 2. AB0073 is derived from a variant of clone PH_420-CEACAM5. The arginine residue in the VL domain of AB0073 was found to be present in <1% of the human framework. To generate AB0466, the arginine residue was replaced with glutamine (15% human framework; Abysis). In addition, the NS motif in the VH domain was replaced with SS, and the NS motif in the VL domain was replaced with NA because the NS motif was shown to deamidate after stress. Finally, two (2) cysteines were introduced to stabilize disulfide bonds. The scFv polypeptide sequence of AB0466 is set forth below as SEQ ID NO:710. Cysteines with stabilizing effects are identified in bold underline, while all other amino acid substitutions are shown in bold italics.

[0673]

[0674] Generation of F3'CEACAM5 TriNKET multispecific binding protein

[0675] The above AB0264, AB0621, AB0411 and AB0466 scFvs were redesigned into F3'TriNKET multispecific binding proteins. F3'TriNKET contains:

[0676] (a) a scFv sequence that binds CEACAM5, comprising a light chain variable domain connected to the C-terminus of a heavy chain variable domain via a (G4S)4 linker (SEQ ID NO: 532), wherein the scFv is connected to an Fc domain, and the Fc domain comprises Q347R, D399V and F405T substitutions for heterodimerization and an S354C substitution for disulfide bond formation; and

[0677] (b) A49MI-derived NKG2D-binding Fab fragment comprising a heavy chain portion comprising a heavy chain variable domain and a CH1 domain, and a light chain portion comprising a light chain variable domain and a light chain constant domain, wherein the CH1 domain is linked to an Fc domain, and wherein the Fc domain comprises K360E and K409W substitutions for heterodimerization and a Y349C substitution for disulfide bond formation. The amino acid sequence of the F3'CEACAM5 TriNKET binding protein is set forth in Table 15 below.

[0678] Table 15. Amino acid sequences of CEACAM5 TriNKET binding proteins

[0679]

[0680]

[0681]

[0682]

[0683]

[0684] Generation and purification of F3'CEACAM5 TriNKET multispecific binding protein

[0685] F3'CEACAM5 TriNKET molecules were transiently expressed in Chinese hamster ovary (CHO cells) using different DNA ratios for different protein chains. Supernatants containing expressed molecules were captured from culture supernatants by overnight incubation with protein A chromatography resin using established methods. Bound TriNKET molecules were eluted from the protein A resin with 0.1 M glycine, pH 3.5.

[0686] After adjusting to pH 7.0, the sample eluate was passed through an anion exchange (AEX) resin and the flow-through was collected. The sample flow-through was then polished by cation exchange (CEX) chromatography using conventional methods to yield >98% purity as determined by capillary gel electrophoresis, analytical size exclusion chromatography, and intact mass determined by mass spectrometry.

[0687] Example 13. Binding of CEACAM5 TriNKET to target proteins

[0688] Binding to human and cynomolgus monkey CEACAM5 at pH 6.0 and 7.4

[0689] To determine the affinity and cross-reactivity of CEACAM5 TriNKET molecules, AB0411, AB0466 or AB0621 TriNKET was captured using an anti-human Fc capture antibody (Cytiva, #BR100839) immobilized on a Biacore CM5 standard surface sensor chip according to the manufacturer's instructions. Human or cynomolgus monkey CEACAM5 was titrated in 2-fold serial dilutions starting from 300 nM on the captured AB0411, AB0466 and AB0621. Association was monitored for 300 seconds and dissociation was monitored for 600 seconds. The assay was performed at 37°C at pH 7.4 and 6.0. The surface of the CM5 chip was regenerated for 20 seconds at 100 μL / min with 10 mM glycine pH 1.5.

[0690] The results show that AB0411 and AB0466 bind to human but not cynomolgus monkey CEACAM5, with AB0411 showing higher affinity than AB0466. AB0621 binds to both human and cynomolgus monkey CEACAM5 at both pH 7.4 and 6.0 with affinity comparable to AB0466, with a slightly increased affinity at lower pH. AB0264 shows affinity comparable to AB0621, but is directed to the K398 allele of human CEACAM5. The results of the in vitro binding assays are shown in Table 16.

[0691] Table 16. CEACAM5 TriNKET in vitro binding assay

[0692]

[0693]

[0694] NB – No BindingND – Not Determined* Using CEACAM5 (K398)

[0695] CEACAM5 TriNKET binds to different domains of human CEACAM5

[0696] To determine which hCEACAM5 protein domain(s) TriNKET binds to, different assay formats were run, depending on whether the protein domain has a His tag or is an Fc fusion. If the protein domain is an Fc fusion, AB0411, AB0466, AB0264, or AB0621 is captured using an anti-human Fab antibody (Cytiva, #28958325) immobilized on a CM5 chip. On the captured TriNKET, various hCEACAM5 protein domains (N-terminal, A1-B1, A2, B2, A3, or B3) were titrated at 50 μL / min, associated for 180 seconds and dissociated for 300 seconds. The assay was run at 25°C in HBS-EP+ running buffer, pH 7.4. The surface of the CM5 chip was regenerated using 10 mL of glycine pH 2.1.

[0697] If the protein domain had a His tag, AB0411, AB0466, AB0264 or AB0621 was captured using an anti-human Fc antibody (Cytiva) immobilized on a CM5 chip. Various hCEACAM5 protein domains were titrated at 50 μL / min on the captured antibody, with association for 180 seconds and dissociation for 300 seconds. The assay was run at 37°C in HBS-EP+ running buffer, pH 7.4. The surface of the CM5 chip was regenerated using 3M MgCl2.

[0698] The results showed that TriNKET bound only to the A1-B1 domain, consistent with the hydrogen-deuterium exchange mass spectrometry (HDX-MS) epitope mapping data (not shown). The affinity for the A1-B1 domain was comparable to that for the extracellular domain of human CEACAM5. TriNKET showed no binding to either the A2-B2 or A3-B3 domains. The results of the domain binding assays are shown in Table 17.

[0699] Table 17. Human CEACAM5 protein domain binding assay

[0700]

[0701]

[0702] Binding of CEACAM5 TriNKET to human CEACAM5 SNPs

[0703] Several notable single nucleotide polymorphisms (SNPs) of human CEACAM5 were identified and generated and evaluated for binding to AB0411, AB0466, and AB0621 using a Biacore surface plasmon resonance (SPR) assay developed for the major CEACAM5 SNP E398. The affinity loss fold (KD / KD(reference)) was calculated for comparison. As shown in Table 18 below, for various SNPs, AB0621 showed up to about 2-fold affinity loss, AB0411 showed up to about 5-fold affinity loss, and AB0466 showed up to about 9-fold affinity loss.

[0704] Table 18. hCEACAM5 domain in vitro binding assay

[0705]

[0706]

[0707] CEACAM5 TriNKET binds to human and cynomolgus monkey CEACAM1, 6, and 8

[0708] Binding to human and cynomolgus CEACAM1, 6 and 8 was assessed by SPR using the same protocol as for CEACAM5, but using up to 600 and 1200 nM for titration of each protein. AB0621 had no detectable binding signal at these concentrations. In contrast, reproducible weak binding signals were detected for AB0411 and AB0466, but both TriNKETs also bound to one or more different epitopes at the N-terminus of CEACAM5. Typical weak binding sensorgrams of AB0411 and AB0466 for human CEACAM1 and 6 indicated weak but significant signals compared to AB0621. The steady-state approximation did not reach a high enough concentration to approach saturation for reliable KD determination, but the KD estimate is likely to be a value greater than about half of the maximum analyte concentration used. The results of the CEACAM1, 6 and 8 binding assays are shown in Table 19.

[0709] Table 19. CEACAM1, 6, 8 binding assay

[0710]

[0711] Simultaneous binding to CEACAM5, NKG2D, and CD16 target proteins

[0712] Simultaneous binding of the three arms of AB0264 (which differs from AB0621 only by having a proline instead of a serine at position 77) was demonstrated using SPR on a CM5 chip immobilized with human CEACAM5. This surface was used to first stably capture AB0264 onto the chip surface. Fig.35A and Fig.35B It was demonstrated that subsequent injection of saturating levels (2 μM) of CD16A (V158 / V176) alone, followed by a 2 μM CD16A and 2 μM NKG2D-His master mix, displayed a stepwise binding signal on the chip surface indicating the formation of a heterotetrameric complex. The ratio of the binding signal (proportional to MW) was approximately 5:1:1, which is consistent with a 1:1:1 molar stoichiometry of the complex bound on the CEACAM5 surface.

[0713] Binding to NKG2D at pH 7.4 and 6.0

[0714] The binding to NKG2D was assessed by SPR using a mouse Fc capture kit (Cytiva, #BR100838) fixed to a CM5 chip. Human and cynomolgus monkey NKG2D fused to mouse Fc were captured and titrated from 600 nM to AB0411, AB0466 or AB0621. The results shown in Table 20 below demonstrate that the affinity of all three TriNKETs for human and cynomolgus monkey NKG2D is comparable.

[0715] Table 20. NKG2D binding assay

[0716]

[0717]

[0718] Binding to CD16A (FcγR3a)

[0719] The binding to human and cynomolgus CD16A was assessed using biotinylated CD16A captured on a streptavidin (SA) sensor chip (Cytiva, #BR100531). At 25°C, AB0411, AB0466 or AB0621 were titrated at 30 μL / min starting from 1500 nM, associated for 150 seconds, and then dissociated for 300 seconds. The surface of the chip was regenerated for 5 seconds at 30 μL / min with 2 mM sodium hydroxide. The running buffer was HBS-EP+. As shown in Table 21 below, the affinity between all TriNKETs was comparable. As expected, the V158 isotype is about 2 to 3 times tighter than the F158 isotype. Cynomolgus FcγRIIIA showed slightly tighter binding than human FcγRIIIAV158.

[0720] Table 21. CD16A (FcγR3a) binding assay

[0721]

[0722]

[0723] Binding of AB0621 to human FcγRs other than CD16a (FcγR3a)

[0724] AB0621 was captured onto a protein A chip (Cytiva, #29127556). On the captured AB0621, human FcγRs were titrated with 3-fold serial dilutions starting from 300 nM for FcγRI, 1000 nM for FcγRIIA R131 and H131, and 3000 nM for FcγRIIB and FcγRIIIB. Association was monitored for 120 seconds and dissociation was monitored for 180 seconds. The assay was run at 25°C at pH 7.4. The surface of the chip was regenerated for 20 seconds at 30 μL / min with 10 mM glycine pH 1.5. The results shown in Table 22 demonstrate that the affinity of AB0621 binding to FcγR is comparable to that of a typical IgG1 antibody.

[0725] Table 22. FcγR binding assay

[0726] AB0621, pH 7.4, 25℃ ka(1 / Ms) kd(1 / M) KD(M) KD / KD Reference FcγRI 2.6E+06 3.6E-02 1.4E-08 0.64 FcγRIIA H131 -- -- 6.0E-07 0.68 FcγRIIA R131 -- -- 3.3E-07 0.68 FcγRIIB -- -- 3.3E-06 0.94 FcγRIIIB -- -- No combination n / a

[0727] Binding of AB0621 to human FcRn

[0728] Anti-κ light chain antibodies were immobilized onto a CM5 chip following a standard amine coupling protocol. AB0621 was captured onto the chip and titrated against human or cynomolgus FcRn in 2-fold serial dilutions starting at 2000 nM. Association was monitored for 120 seconds and dissociation was monitored for 180 seconds. The assay was run at both pH 7.4 and pH 6.0 at 25°C. The chip surface was regenerated with two pulses of 10 mM glycine pH 1.5 at 30 μL / min for 20 seconds followed by 10 mM NaOH at 30 μL / min for 20 seconds.

[0729] The results shown in Table 23 demonstrate that, as expected, at pH 6.0 AB0621 binds to both human and cynomolgus monkey FcRn with comparable affinity, whereas at pH 7.4 AB0621 does not bind to either human or cynomolgus monkey FcRn.

[0730] Table 23. FcRn binding assay

[0731]

[0732] Example 14. CEACAM5 TriNKET Cell Binding Assay

[0733] Quantification of cell surface CEACAM5 molecules by flow cytometry

[0734] Various cancer cell lines were diluted into FACS buffer and 200,000 cells of each cell type were seeded per well in duplicate into 96-well plates for FACS staining. The monovalent mouse Fc variant of the anti-CEACAM5 antibody labetuzumab was diluted to 200 nM in FACS buffer and used to resuspend the cells. The plate was incubated at 4°C for 120 minutes, washed with FACS buffer, and resuspended with the second detection reagent of the commercially available receptor quantification kit QIFIKIT from Agilent. The FITC anti-mouse second detection reagent was diluted 1:50 into FACS buffer and incubated on cells at 4°C for 60 minutes. The calibration beads were washed with FACS buffer, resuspended with the FITC anti-mouse detection reagent as prepared for the cells, and incubated at 4°C for 60 minutes.

[0735] The cells and beads were washed with FACS buffer, resuspended with 70 μl of fixation buffer, and incubated at 4°C for 20 minutes. The cells and beads were washed again with FACS buffer, and data were collected using Thermo Fisher Attune NxT. Cells of interest were identified using the FSC versus SSC curve, and gates of appropriate shapes were drawn around these cells. Within the gated cells, doublet events were removed by observing the FSC-H versus FSC-A curve. Within the single cell population, live cells were gated. Within the live gate, the MFI of each sample was calculated. The background was subtracted from the MFI of the cells and calibration beads using wells containing only the second detection reagent and converted to log(MFI). The log(MFI) of the calibration beads was plotted against the log(number of receptors, as provided by the manufacturer) and fitted using linear regression using GraphPad Prism. The log(MFI) of the cells was then used to interpolate the log(number of receptors) of the cells; these data were reported as antibody binding capacity (ABC) or antibodies bound per cell. As shown in Table 24, widespread expression was observed among cancer cell lines, with the highest number of CEACAM5 molecules per cell observed in MKN-45 cells.

[0736] Table 24. Number of CEACAM5 molecules per cell

[0737] Cell lines Cell surface CEACAM5 MKN-45 460335 SK-CO-1 118984 ZR-75-30 80569 HPAF-II 63997 LS-174T 44027 LoVo 35055 BxPC-3 21691 Kato-III 12638 AsPC-1 8503 HT-29 660

[0738] Quantification of TriNKET binding to CEACAM5 on tumor cell lines

[0739] MKN-45 and HPAF-II human cancer cell lines were diluted into FACS buffer and 100,000 cells of each cell type were seeded per well in duplicate into 96-well plates for FACS staining. Cells were washed with PBS, incubated in a live / dead dye at a dilution of 1:2000 in PBS for 15 minutes, and then washed with FACS buffer. AB0264, AB0411, AB0466, and AB0621 were diluted into FACS buffer, and 50 μl of each diluted TriNKET was added to the cells. After incubation on ice for 30-120 minutes, the cells were washed with FACS buffer. Anti-human IgG-Fc secondary antibody was diluted into FACS buffer and 50 μl was added to each well for detection of bound TriNKET. Cells were incubated on ice for 30-60 minutes and then washed with FACS buffer. 50 μl of fixation buffer was added to each well and the cells were incubated at room temperature for 10 minutes. Cells were washed with FACS buffer and resuspended in FACS buffer for analysis with a ThermoFisher Attune NxT, BD FACS Celesta SN #H66034400085, or BD FACS Celesta SN #H66034400160.

[0740] Use FSC to SSC curve to identify cells of interest, and draw a gate of appropriate shape around these cells. In the gated cells, remove doublet cells by observing the curve of FSC-H to FSC-A. In the single cell population, gates are set for living cells. In the living gate, the median fluorescence intensity (MFI) of each sample and only the second antibody control is calculated. The multiple relative to background (FOB) is calculated as the ratio of the test object MFI to the second antibody background MFI only. Use GraphPad Prism 7.0 to fit the data to a four-parameter nonlinear regression curve.

[0741] The binding potency (EC50) and maximum load (maximum FOB) of CEACAM5 TriNKET on MKN-45 and HPAF-II human cancer cell lines are shown in Table 25. For AB0261, the binding potency of MKN-45 and HPAF-II was similar, with EC50 of 20.6 nM and 18.1 nM, respectively. The maximum load was 51.32 and 24.92 FOB, respectively, consistent with the high and moderate expression of CEACAM5 on these cell lines.

[0742] Table 25. EC50 and maximum FOB values ​​for TriNKET binding to human cancer cell lines

[0743]

[0744] Comparison of TriNKET and mAb binding to CEACAM5 on tumor cell lines

[0745] Using the above method, the binding of CEACAM5 TriNKET to its respective corresponding monoclonal antibody was compared between five human cancer cell lines (Table 26). MKN-45 and SK-CO-1 cells represent higher CEACAM5 expression levels, while LoVo and BxPC-3 represent moderate expression, and KATO-III represents very low CEACAM5 expression. Similar binding patterns were observed between the five cell lines. AB0264 and AB0411 showed reduced EC50s compared to AB0755 and AB0509, respectively. AB0755 is a humanized mAb for CEACAM5, and its Fab sequence corresponds to the scFV anti-CEACAM5 present in AB0264. AB0509 is a humanized mAb for CEACAM5, and its Fab sequence corresponds to the scFV anti-CEACAM5 present in AB0411. However, both TriNKETs were always loaded to a maximum FOB higher than their corresponding mAbs.

[0746] Table 26. EC50 and maximum FOB values ​​of TriNKET and corresponding mAbs

[0747]

[0748] CEACAM5 TriNKET binds to Ba / F3 cells expressing CEACAM family proteins

[0749] To examine the binding specificity of TriNKET for CEACAM5 compared to other CEACAM family members, Ba / F3 cells were engineered to express either human CEACAM1, CEACAM6, CEACAM8 or cynomolgus monkey CEACAM5. CEACAM5 TriNKET was evaluated by flow cytometry starting at 1600 nM followed by 7 5-fold dilutions.

[0750] AB0264 and AB0621 bound to cynomolgus monkey CEACAM5, while AB0466 and AB0411 did not ( Fig.36D AB0264 and AB0621 did not cross-react with Ba / F3 cells expressing human CEACAM1, CEACAM6, or CEACAM8 ( Figure 36A-36C ), while both AB0411 and AB0466 showed cross-reactive binding to Ba / F3 cells expressing human CEACAM1 ( Fig.36A ), and AB0411 also showed cross-reactive binding to Ba / F3 cells expressing human CEACAM6 ( Fig.36B). CEACAM5 TriNKET did not bind to parental BA / F3 cells lacking expression of CEACAM family member proteins ( Fig.36E ).

[0751] Example 15. Cell-based cytotoxicity and activity assays

[0752] Isolation and preparation of primary PBMC and NK cells

[0753] Human blood was obtained from Stanford Blood Bank or Biological Specialty Corporation (#225-11-04). Cynomolgus monkey whole blood from three animals was obtained from BioIVT (#NHP01WBNHUZN). Both human and cynomolgus monkey PBMCs were isolated by density gradient centrifugation. After purification, PBMCs were used immediately or frozen for future use. EasySep was used by negative depletion according to the manufacturer's protocol. TM (StemCell, #17955) or RosetteSep TM (StemCell, #15065) purified human primary NK cells. Alternatively, frozen NK cells were purchased from BioIVT (#HUMAN-HL65-U-200429). Primary NK cells were cultured overnight in RPMI primary cell culture medium and then used in assays, such as DELFIA assays.

[0754] Transduction of KHYG-1-CD16V cells

[0755] KHYG-1 cells (DSMZ, #ACC-725) were retrovirally transduced to express human CD16a variant 158V (UniProt P08637). Cells were selected under puromycin and the tolerized population positive for human CD16 was confirmed by FACS analysis. KHYG-1-CD16V cells were cultured at 0.2 × 10 in RPMI medium in the presence of 10 ng / mL recombinant human IL-2. 6 -1.0×10 6 The density is routinely maintained between 100 and 100 mL.

[0756] CD8 + Preparation of T cells

[0757] Frozen PBMCs were thawed and incubated at 37°C in 25 cm 2 The culture medium in the culture flask (each culture flask has 20-25×10 6 1 μg / mL ConA was then removed and PBMCs were plated at 37°C in 25 cm 2The cells were cultured in flasks with 25 units / mL IL-2 for 4 days. Negative selection was performed using magnetic beads (EasySep TM Human CD8 + T cell isolation kit, StemCell) to purify CD8 + Finally, CD8 T cells were cultured at 37°C + T cells were cultured at 100,000 cells / 200 μL / well in 96-well round-bottom plates in medium containing 10 ng / mL IL-15 for 6-10 days and then used, for example, for cytolysis assays. + CD3 of T cells + CD8 + Cell purity and NKG2D and CD16 expression. CD8 + T cell activity.

[0758] DELFIA cytotoxicity assay

[0759] Target cancer cells expressing CEACAM-5 were detached from the culture vessel, pelleted, washed with 1× HBS, and incubated at 10 6 Resuspend in pre-warmed cell culture medium at 1:400 cells / mL. Dilute BATDA (bis(acetoxymethyl) 2,2':6',2"-tripyridine-6,6"-dicarboxylate) reagent to the cell suspension at 1:400. Mix the cells and incubate at 37°C and 5% CO2 for 15-20 minutes. Wash the labeled target cells 3x with 1x HBS and resuspend in cell culture medium to the final desired concentration.

[0760] Resting effector cells (such as human NK cells, KHYG1-CD16V or activated CD8 + T cells) were removed from culture and pelleted, and the cells were resuspended in RPMI primary cell culture medium. TriNKET was titrated in RPMI primary cell culture medium. The assay was set up in round-bottom TC 96-well plates with the desired amount of labeled target cells, effector cells, and TriNKET.

[0761] Control wells for background were prepared using 100 μl of supernatant from the precipitated labeled target cells and an additional 100 μl of RPMI primary cell culture medium. Spontaneous release wells were prepared by adding 100 μl of labeled target cells to wells containing 100 μl of RPMI primary cell culture medium. Maximum release wells were prepared by adding 100 μl of labeled target cells to wells containing 80 μl of RPMI primary cell culture medium and 20 μl of 10% TritonX-100 solution. The assay plate was incubated at 37°C and 5% CO2 for 2-3 hours.

[0762] At the end of the assay, 20 μl of supernatant was removed from each well and transferred to a clean 96-well DELFIA assay plate. 200 μl of Europium solution was added to each well and incubated for a further 15 minutes at room temperature on a plate shaker at 250 RPM. The HTRF box in the was used to read the assay plate. The average of the background samples was calculated and subtracted from the values ​​from all sample wells. The calculation of specific lysis was performed using the following formula:

[0763] Specific lysis % = (sample - spontaneous) / (maximum - spontaneous) * 100%.

[0764] Potency of CEACAM5 TriNKET in cytotoxicity assays using NK cells and tumor cell lines

[0765] Human NK cells were rested overnight. The next day, resting NK cells were co-cultured with BATDA-labeled CEACAM5-expressing target cancer cells at a ratio of 10:1 (SK-CO-1) or 5:1 (LS-174T, ZR-75-30, and HPAF-II) for DELFIA assay. Data were fitted to a 4-parameter nonlinear regression model to generate potency values.

[0766] In a 2.5-hour short-term DELFIA assay using SK-CO-1 target cells, CEACAM5 TriNKET elicited efficient lysis of target cells by resting primary NK cells from healthy human donors ( Fig.37A ). Similar results for LS-174T, HPAF-II, and ZR075-30 are shown in Figure 37B-Figure 37D RSV targeting TriNKET (F3'-TriNKET-palivizumab) and human IgG1 isotype control (palivizumab-IgG1) resulted in minimal target cell death, indicating that cytolysis was dependent on the engagement of the anti-CEACAM5 arm with the target cells. The concentrations required for AB0411, AB0466, and AB0621 to produce half of their maximal killing (EC50) were 4.55 nM, 9.07 nM, and 1.02 nM, respectively (Table 27).

[0767] Table 27. Potency of CEACAM5 TriNKET in cytolysis of CEACAM5+ cancer cells

[0768]

[0769] AB0264 enhances the activity of IL-2-stimulated human NK cells

[0770] The ability of AB0264 to activate primary human NK cells was characterized. Purified frozen human NK cells were thawed and rested or activated overnight in culture with IL-2. The next day, NK cells were co-cultured with labeled ZR-75-30 target cells for DELFIA determination. Dose titration of AB0264 or AB0755 (a humanized mAb for CEACAM5 whose Fab sequence corresponds to the scFV anti-CEACAM5 present in AB0264) was prepared from 50nM and added to the co-culture of resting human NK cells or activated human NK cells and ZR-75-30 target cells. The curve of specific lysis to concentration was drawn, and the data were fitted to a 4-parameter nonlinear regression model to generate a potency value.

[0771] The lysis of ZR-75-30 target cells by resting hNK cells and IL-2 activated hNK cells derived from the same healthy donor was compared (Figure 38). EC50 and maximum lysis values ​​are summarized in Table 28. Compared with resting NK cells, IL-2 activated NK cells show more potent killing of ZR-75-30 cancer cells. Larger differences were observed in the maximum lysis of ZR-75-30 target cells, where AB0264 showed that IL-2 activated NK cells were 4 times higher than resting NK cells in the lysis of target cells.

[0772] The same analysis was applied to two other human cancer cell lines expressing CEACAM-5, HPAF-II and LS-174T. Both human cancer cell lines showed improved EC50 and higher maximum killing when using IL-2 activated NK cells as effectors compared to resting NK cells. Killing EC50 and maximum lysis values ​​are summarized in Table 28.

[0773] Table 28. EC50 and % maximum lysis of resting NK cells and activated NK cells against human cancer cell lines expressing CEACAM5

[0774]

[0775] ND = Not Determined

[0776] Compared with the corresponding mAb, TriNKET exhibits stronger cytolytic activity

[0777] The ability of AB0264 and AB0411 CEACAM-TriNKET to lyse CEACAM5-expressing human cancer cells was compared to AB0755 and AB0509, respectively, using a short-term primary NK cell cytotoxicity assay. AB0755 is a humanized mAb against CEACAM5 whose Fab sequence corresponds to the scFV anti-CEACAM5 present in AB0264. AB0509 is a humanized mAb against CEACAM5 whose Fab sequence corresponds to the scFV anti-CEACAM5 present in AB0411. Purified human NK cells were thawed and rested overnight. The next day, resting NKs were co-cultured with labeled (a) MKN-45, (b) SK-CO-1, (c) LS-174T, (d) ZR-75-30, and (E) HPAF-II target cells for DELFIA assay. Dose titrations of AB0264 and AB0411 TriNKET or their corresponding mAbs were prepared starting from 20 nM and added to co-cultures of human NK cells and target cells. Specific lysis versus concentration of each TriNKET or mAb was plotted and the data were fitted to a 4-parameter nonlinear regression model to generate potency values. AB0264 ( Figure 39A-Figure 39E ) and AB0411( Figure 40A-40E ) showed high potency in killing five cancer cell lines and was superior to its corresponding mAb. The results are summarized in Table 29.

[0778] Table 29. EC50 and % maximum killing of AB0264, AB0411 and corresponding mAbs

[0779]

[0780] ND = Not Determined

[0781] NK-mediated killing of tumor cells expressing CEACAM5 depends on co-engagement of the TriNKET binding arm with CD16, NKG2D, and CEACAM5

[0782] Variants of AB0264 were generated with mutations in different binding arms. AB0754 is a CD16-silent variant of AB0264 designed to eliminate FcγR binding by introducing mutations into the CH2 domain. AB0752 is a NKG2D-dead variant with mutations in the NKG2D binding arm. AB0444, a F3'-TriNKET with the CEACAM5 binding arm replaced with a palivizumab-based scFv, was generated to eliminate binding to CEACAM5 on target cells.

[0783] KHYG-1-CD16V cells were rested overnight. The next day, KHYG-1-CD16V cells were co-cultured with labeled MKN-45 target cells for DELFIA assay. Dose titrations of AB0264 or loss-of-function variants were prepared starting from 20 nM and added to co-cultures of KHYG-1-CD16V cells and MKN-45 target cells. Specific lysis versus concentration was plotted and the data were fitted to a 4-parameter nonlinear regression model to generate potency values.

[0784] In the absence of target cell binding, no activity was observed for AB0444, demonstrating the contribution of CEACAM5 binding to NK-mediated target cell lysis. AB0754 and AB0752 showed little or no activity, demonstrating the further importance of CD16 and NKG2D binding for the killing activity of AB0264 ( Fig.41 ).

[0785] IFNγ ELISA assay

[0786] The assay plate was set up as in the DELFIA assay, but with a longer incubation time of 48 to 72 hours. Freshly isolated human NK cells were rested overnight. The next day, resting NKs were co-cultured with SK-CO-1 at a 10:1 ratio. A dose titration of CEACAM5 TriNKET was prepared in a 1:5 dilution series starting at 133 nM final concentration. After incubation, the assay plate was briefly centrifuged and the cells were diluted with hIFNγ following the manufacturer's protocol. The kit (R&D, #SIF50) quantifies IFNγ in the supernatant from the assay wells. The data were fitted to a 4-parameter nonlinear regression model to generate titer values. EC50 and maximum IFNγ release levels were generated by averaging the results from three independent NK donors.

[0787] At 48 hours after treatment, the ability of CEACAM5 TriNKET to trigger IFNγ production in the co-culture system was evaluated. Although the control palivizumab-TriNKET did not trigger significant amounts of IFNγ, CEACAM5 TriNKET induced significant amounts of IFNγ in a dose-dependent manner ( Fig.42A ). EC50 and maximum induction of IFNγ by TriNKET are summarized in Table 30.

[0788] Table 30. EC50 and IFNγ maximum release in human primary NK

[0789]

[0790]

[0791] IFNγ and CD107a activation assay

[0792] TriNKET or hIgG1 control were diluted in culture medium. Human cancer cells expressing CEACAM-5, resting human primary NK cells or PBMC were harvested from culture and resuspended in culture medium to 1 × 10 6 cells / mL. Recombinant hIL-2 and fluorophore-conjugated anti-CD107a antibodies were added to NK cells or PBMCs for activation culture. For intracellular cytokine staining, Brefeldin A (BFA) and monensin were diluted into the culture medium to block protein transport out of the cells. MKN-45 tumor cells expressing CEACAM5 and primary NK or PBMC effector cells were mixed at a 1:1 ratio. The assay plates were incubated for 4 hours to allow NK cell activation, and then the cells were stained and analyzed by flow cytometry. In the CD3 - CD56 + CD107a and IFNγ staining were analyzed in the population to assess activation of human NK cells. Induced IFNγ was plotted + CD107 + The percentage of NK cells was plotted against the concentration and the data were fitted to a 4-parameter nonlinear regression model to generate titer values.

[0793] After four hours, the human IgG1 isotype showed essentially no basal induction of CD107a degranulation or intracellular IFNγ accumulation. Addition of AB0264 to the co-culture resulted in a strong induction of IFNγ production and CD107a degranulation in a dose-responsive manner ( Fig.42B ).

[0794] The ability of CEACAM5 TriNKET molecules to enhance cynomolgus monkey NK cell activation was evaluated in a co-culture assay using human cancer cell lines with primary cynomolgus monkey PBMCs. The assay using cynomolgus monkey PBMCs was set up in a similar manner to the human assay described above. Frozen cynomolgus monkey PBMCs were thawed and rested in culture medium at 37°C and 5% CO2. MKN-45 or SK-CO-1 human cancer cell lines were mixed with resting cynomolgus monkey PBMCs at a 5:1 effector to target cell ratio, along with the addition of the desired concentration of TriNKET or hIgG1 control. BFA, monensin, rhIL-2, and fluorophore-conjugated anti-CD107a were added to the PBMCs for activation cultures. Plates were incubated at 37°C and 5% CO2 for 4 hours, and samples were then prepared for flow cytometric analysis to measure NK cell CD107a degranulation in the NK and tumor cell co-culture system. CD8 + CD107a in NK cells +The percentage of NKET or control concentration was plotted and the data were fitted to a 4-parameter nonlinear regression model to generate potency values.

[0795] Cynomolgus monkey NKG2D expression is always only present in CD8 + On NK cells, not CD8 - Therefore, the use of CD45 + CD14 - CD20 - CD3 - CD8 + The gating strategy was used to define the cynomolgus monkey CD8 + NK cells. Using each of the three cynomolgus macaque PBMC samples tested, all tested CEACAM5-TriNKETs showed enhanced CD8 + Dose-responsive activity of NK cell degranulation ( Figure 42C-42D ). In contrast, the hIgG1 isotype control showed similar levels of CD107a staining to untreated samples at all concentrations evaluated. Table 31 summarizes the potency and maximum percentage of CEACAM-5 TriNKET triggering CD107a degranulation in cynomolgus monkeys and human NK cells.

[0796] Table 31. NK cells targeting CEACAM5 + EC50 and maximum CD107a degranulation in cancer cells

[0797]

[0798] Potency in short-term killing assays using resting primary human NK and patient-derived primary lung tumor organoid lines

[0799] NSCLC 10910 and NSCLC 3222 are two tumor organoid lines derived from primary non-small cell lung cancer (NSCLC) patients. Flow cytometric analysis with unconjugated anti-CEACAM5 mAb (labetuzumab) and PE-conjugated secondary antibody demonstrated that CEACAM5 was expressed at much lower levels on these two lines compared to SK-CO-1 ( Fig.43A). The short-term DELFIA assay was used to quantify the ability of CEACAM5 TriNKET to trigger NK-mediated cytolysis of these two primary NSCLC organoid lines. Freshly isolated human NK cells were rested overnight. The next day, resting NKs were co-cultured with NSCLC 10910 or NSCLC3222 at a 10:1 ratio. Dose titrations of TriNKET and control molecules were prepared in a 1:5 dilution series starting from a final concentration of 133nM. The data were fit to a 4-parameter nonlinear regression model to generate potency values. Despite differences in CEACAM5 expression levels, human primary NK cells triggered potent tumor cell lysis in both lines ( Figure 43B-Figure 43C , Table 32).

[0800] Table 32. EC50 and % maximal lysis of resting NK cells against primary human cancer organoid lines expressing CEACAM5

[0801]

[0802] When using activated CD8 + T cells and the potency of MKN45 in short-term killing assays

[0803] In addition to NK cells, NKG2D is also expressed on cytotoxic T cells. + T cells can be directly triggered by NKG2D stimulation. Fig.44A The described protocol generates cytokine stimulated CD8 cells. + T cells were co-cultured with MKN-45 cells at a 20:1 E:T ratio. Specific lysis versus concentration was plotted and the data were fitted to a 4-parameter nonlinear regression model to generate efficacy values. Activated T cells did not show basal lysis of target cells. Addition of the corresponding mAb or NKG2D silent variant (which could not excite NKG2D) did not show any triggering of T cell activity. In contrast, AB0264 had a significant effect on CD8 + T cell-mediated cytolysis of MKN-45 target cells showed a dose-dependent induction ( Fig.44B ).

[0804] Example 16. Antitumor activity of mouse surrogate TriNKET mAB0621 in hCEACAM5Tg mice bearing B16F10-hCEACAM5 tumors

[0805] Transgenic mice

[0806] B6.CG-Tg (hCEACAM5) 2682Wzm / Ieg mice express human CEACAM5 under the control of the human CEACAM5 promoter (Eades-Perner, 1994). Female heterozygous mice, approximately 7 to 12 weeks old, were obtained from a breeding colony maintained at Taconic Laboratory (Germantown, NY) and weighed an average of 21.7 g. Based on a recent re-evaluation of copy number by quantitative PCR, these mice contain approximately 2.5 copies of a 33 kb cosmid clone insert per haploid genome that contains the entire human CEACAM5 gene and flanking sequences.

[0807] The expression distribution of human CEACAM5 in these mice was comparable to that in humans, with low levels of mRNA detected in the colon, ileum, cecum, and stomach. Human CEACAM5 protein expression was observed throughout the full thickness of the colonic mucosa of hCEACAM5 transgenic mice, with the vast majority of colonic epithelial cells staining positively. In contrast, CEACAM5 expression in the human colon was observed to be primarily localized to the upper mucosa, particularly along the luminal surface.

[0808] Antibody reagents and preparations

[0809] A mouse surrogate duobody TriNKET was generated, designated mAB0621, in which the human anti-CEACAM5 Fab arm of the human TriNKET molecule AB0621 was constructed as a heterodimeric antibody (duobody), in which the mouse anti-mouse NKG2D binder clone 13 formed the second Fab arm, linked together on the mouse IgG2a isotype. Mutations from Genmab DuoBody were used in the CH3 domain of mouse IgG2a to form a bispecific duobody TriNKET molecule. An isotype control mouse surrogate duobody TriNKET was similarly made using the synagis anti-human RSV Fab sequence to replace the AB0621 Fab. The mouse surrogate duobody TriNKET was produced by a recombinant cell line, formulated in 20mM Na acetate, 9% sucrose, pH 5.5, and stored as a frozen (-80°C) stock.

[0810] Generation of tumor cell lines

[0811] The B16F10 mouse melanoma tumor cell line was engineered to stably express human hCEACAM5 using the pRG-RV 2-5 retroviral vector without selection. B16F10-hCEACAM5 clone 7-2B11 has been shown by IHC to express high levels of hCEACAM5 in tumors grown subcutaneously (SC) in mice, and soluble CEACAM5 levels are elevated in the serum of tumor-bearing mice. The binding of mAB0621 to the hCEACAM5-B16F10 cell line was evaluated by flow cytometry. mAB0621 showed an EC50 value of 21.8 nM.

[0812] Preparation and implantation of tumor cell lines

[0813] B16F10-hCEACAM5 clone 7-2B11 cells from frozen stocks were maintained in vitro as monolayer cultures in DMEM medium supplemented with 10% heat-inactivated fetal bovine serum (FBS), 1X glutamine, and 1X MEM non-essential amino acids at 37°C and 5% CO2 in air. Cells grown in exponential growth phase at 80% confluence were harvested and washed. 1.5×10 50 μL of DMEM basal medium was added to the culture medium. 6 Each mouse was injected subcutaneously into the right flank of the back.

[0814] Tumor and body weight measurements

[0815] Tumors were measured one day before the first dose and twice weekly thereafter. Tumor length and width were measured using electronic calipers, and tumor volume was determined using the following formula: Volume (mm 3 )=0.5×length×width 2 , where length is the longer dimension. Mice were weighed regularly to monitor overall health. Prior to treatment, mice were weighed and tumors from individual mice were measured. To prevent bias, any outliers in weight or tumor volume were removed and the remaining mice were divided into treatment groups with equal mean tumor size. When the mean tumor volume in B16F10-hCEACAM5 tumor-bearing mice reached approximately 104mm 3 (Range 80-120mm 3 ), i.e., 8 days after implantation, dosing was started. Duobody TriNKET was administered to the animals as follows.

[0816] Preparation, administration and analysis of dosing solutions

[0817] The frozen stock of duobody TriNKET to be tested in the animal model was thawed and transferred to wet ice. The stock solution of each duobody TriNKET was diluted to the nominal concentration in a suitable diluent and administered immediately. mAB0621 duobody TriNKET or isotype control duobody TriNKET was administered to B16F10-hCEACAM5 tumor-bearing hCEACAM5-Tg mice, with a dose of 15, 5, 1.5 or 0.5 mg / kg every 3-4 days SC administration, a total of 6 doses. Each treatment group included 15 animals. After administration, the animals were continuously monitored and the tumor volume was measured twice a week. The anti-tumor activity of mAB0621 was evaluated by two parameters; the percentage of animals remaining on the 26th day after treatment with mAB0621 (determined by Kaplan-Meyer analysis), and the tumor volume measured after grouping. Statistical analysis was performed using the log-rank test on day 26 (*: p<0.05, ***: p<0.001, ****: p<0.0001; ns=not significant).

[0818] Fig.45 Kaplan-Meier curves evaluating the percentage of animals remaining over time are shown. 3 On day 26, the number of animals remaining was significantly greater in mice treated with mAB0621 at 15 mg / kg (p<0.0001), 5 mg / kg (p=0.0004), and 1.5 mg / kg (p=0.0274) compared to the isotype control group.

[0819] Figure 46 shows the volume of a single B16F10-hCEACAM5 tumor measured for each animal in the five treatment groups. Tumor volume was measured twice a week. An overall comparison of tumor volume between the different treatment groups was performed at all time points using the area under the curve (AUC) as a summary measure for each tumor. Differences between two treatment groups were assessed by Wilcoxon-type nonparametric tests of growth curves under dependent right censoring as proposed by Vardi et al., 2001.

[0820] The isotype control ( Fig.46A ) or 15 mg / kg ( Fig.46B )、5mg / kg( Fig.46C )、1.5mg / kg( Fig.46D ) or 0.5mg / kg( Fig.46EFigure 22 shows a single curve of tumor volume in B16F10-hCEACAM5 tumor-bearing mice in the hCEACAM5 transgenic model after 15 mg / kg (p=0.00050) and 5 mg / kg (p=0.00145) and 0.5 mg / kg (p=0.02870) of mAB0621. Statistically significant tumor regression was observed in the different groups treated with mAB0621 at 15 mg / kg (p=0.00050) and 5 mg / kg (p=0.00145) and 0.5 mg / kg (p=0.02870) compared to the control group. The group treated with 1.5 mg / kg mAB0261 did not show significant tumor regression (p=0.07865). On day 26, complete tumor regression (CR) was observed in the 5 mg / kg (4 mice) and 15 mg / kg (2 mice) mAB0621 treated groups.

[0821] Example 17. Antitumor activity of the combination of mouse surrogate TriNKET mAB0621 and anti-PD-1 antibody in hCEACAM5 Tg mice bearing B16F10-hCEACAM5 tumors

[0822] Transgenic mice

[0823] B6.Cg-Tg(hCEACAM5)2682Wzm / Ieg mice are described in Example 16. Female heterozygous mice approximately 8 to 12 weeks old were obtained from a breeding colony maintained at Taconic Laboratory (Germantown, NY) and weighed an average of 23.2 g.

[0824] Antibody reagents and preparations

[0825] mAB0621 mouse surrogate duobody TriNKET and isotype control TriNKET are described in Example 16. Mouse surrogate duobody TriNKET and anti-murine PD-1 mouse IgG1 antibody (muDX400) were produced by recombinant cell lines, formulated in 20 mM Na-acetate, 9% sucrose, pH 5.5, and stored as frozen (-80°C) stocks.

[0826] Tumor cell lines, preparation, and implantation

[0827] B16F10-hCEACAM5 clone 7-2B11 and the culture, preparation and injection of these cells into mice are described in Example 16.

[0828] Tumor measurements and body weight

[0829] Tumors were measured one day before the first dose and twice weekly thereafter. Tumor length and width were measured using electronic calipers, and tumor volume was determined using the following formula: Volume (mm 3 )=0.5×Length×Width 2, where length is the longer dimension. Mice were weighed regularly to monitor overall health. Prior to treatment, mice were weighed and tumors from individual mice were measured. To prevent bias, any outliers in weight or tumor volume were removed and the remaining mice were divided into treatment groups with equal mean tumor size. When the mean tumor volume of B16F10-hCEACAM5 tumor-bearing mice reached approximately 237mm 3 (Range 220-270mm 3 ), i.e., 7 days after implantation, dosing was started. Duobody TriNKET was administered to the animals as follows.

[0830] Preparation and administration of dosing solutions

[0831] Frozen stocks of duobody TriNKET or muDX400 anti-PD-1 antibodies to be tested in animal models were thawed and transferred to wet ice. The stock solutions of each duobody TriNKET were diluted to the nominal concentration in the appropriate diluent and administered immediately.

[0832] Administration and outcomes

[0833] B16F10-hCEACAM5-bearing hCEACAM5-Tg mice were administered a 5 mg / kg dose of control isotype antibody or mAB0621 or anti-PD1 muDX400, as a single agent or as a combination treatment, SC every 3-4 days for a total of 6 doses. Each treatment group included 15 animals. After dosing, animals continued to be monitored and tumor volume was measured twice a week. The antitumor activity of mAB0621 was evaluated by two parameters; the percentage of animals remaining on day 37 after treatment (determined by Kaplan-Meier analysis) and the tumor volume measured after grouping. Statistical analysis was performed using the log-rank test on day 37.

[0834] Fig.47 Kaplan-Meier curves evaluating the percentage of animals remaining over time are shown. 3 On day 37, the number of remaining animals was significantly greater in mice treated with the combination of mAB0621 and anti-PD-1 muDX400 compared to isotype controls (p<0.0001), compared to anti-PD-1 single agent treatment (p=0.0356), and compared to mAB0621 single agent treatment (p=0.0368).

[0835] Figure 48 shows the volume of individual B16F10-hCEACAM5 tumors measured for each animal in the four treatment groups. Tumor volumes were measured twice a week. An overall comparison of tumor volumes between the different treatment groups was performed at all time points using AUC as a summary measure for each tumor. Differences between two treatment groups were assessed by Wilcoxon-type nonparametric tests of growth curves under dependent right censoring as proposed by Vardi et al., 2001.

[0836] The combination treatment showed statistically significant tumor regression compared to isotype control (p<0.0001), and compared to the single agent treatment groups of mAB0621 (p=0.0270) and anti-PD-1 (p=0.00490), respectively. Fig.48A ) or 5 mg / kg of anti-PD-1 muDX400 ( Fig.48B ) or 5 mg / kg mAB0621 ( Fig.48C ) or the combination of mAB0621 and anti-PD-1 muDX400 at 5 mg / kg ( Fig.48D Figure 3. Individual curves of tumor volume in B16F10-hCEACAM5 tumor-bearing mice in the hCEACAM5 transgenic model from day 1 to day 37. Complete tumor regression was observed on day 37 for treatment with mAB0621 in combination with anti-PD-1 (11 mice), anti-PD-1 single agent treatment (2 mice), and mAB0621 single agent treatment (3 mice).

[0837] Example 18. Pharmacokinetics of AB0264 and AB0411 in cynomolgus monkeys

[0838] In the absence of biologics ) were studied in cynomolgus monkeys. AB0264 and AB0411 were obtained from internal sources as frozen (-80°C) stocks. The dosing solution was transferred from nominal -80°C to nominal 4°C the night before dosing. Prior to dosing, the dosing solution was allowed to reach room temperature for at least 1 hour and inverted 5-10 times to ensure uniform mixing of the formulation before transferring from the tube to the syringe.

[0839] Due to the cross-reactivity of AB0264 with cynomolgus monkey CEACAM5, the PK of AB0264 at 0.1, 1, and 10 mg / kg doses was studied to understand target-mediated drug disposal. Six cynomolgus monkeys were divided into three dose groups, with 1 male and 1 female in each group. AB0411 does not cross-react with cynomolgus monkey CEACAM5. Therefore, the PK of AB0411 was studied only with two female and two male cynomolgus monkeys at a dose of 10 mg / kg. On day 0, these cynomolgus monkeys were administered with AB0264 or AB0411, respectively, by intravenous administration.

[0840] Serum samples were collected at the indicated time points up to 14 days after dosing. Drug concentrations were measured by two ligand binding assays. Free drug was measured using a recombinant human CEACAM5 and anti-hNKG2D arm anti-IDS antibody pair. Total drug was measured using an anti-human Fc and anti-human IgG antibody pair.

[0841] The concentration-time PK profiles of free and total drug were plotted as the mean concentration versus time in each group (AB0264 PK profiles were plotted in Fig.49A In the PK spectrum of AB0411 Fig.49B Both AB0264 and AB0411 showed antibody-like PK in cynomolgus monkeys. They were stable in vivo with similar exposures of free and total drug over the duration of the study. No gender differences in exposure were observed. For AB0264, serum exposure was linear between 0.1 mg / kg and 10 mg / kg doses.

[0842] Example 19. Pharmacokinetics in hCEACAM5 Transgenic Mice

[0843] The PK of AB0621, AB0411 and AB0466 was studied. To allow the evaluation of the effect of human CEACAM5 expression on TriNKET PK, studies were performed in female heterozygous B6.CG-Tg(hCEACAM5)2682Wzm / Ieg mice as described in Example 16. Dosing solutions of AB0621, AB0411 and AB0466 were prepared following the same procedure as described for the PK study in cynomolgus monkeys above.

[0844] On day 0, AB0621, AB0411 and AB0466 were administered to hCEACAM5 transgenic mice at doses of 1 mg / kg and 10 mg / kg, respectively, by intravenous administration. Serum samples were collected at the indicated time points for up to 14 days after dosing. Drug concentrations were measured by two ligand binding assays. Free drug was measured using recombinant human CEACAM5 and anti-hNKG2D arm anti-ids antibody pairings. Total drug was measured using anti-human Fc and anti-human IgG antibody pairings.

[0845] AB0621( Fig.50A )、AB0411( Fig.50B ) and AB0466( Fig.50C ) were plotted as the mean concentration versus time in each group. AB0621, AB0411, and AB0466 showed antibody-like PK profiles in hCEACAM5 transgenic mice. Total and free drug exposures were consistent for all three TriNKETs tested, indicating that they are stable in vivo. AB0621 PK was linear from 1 mg / kg to 10 mg / kg. AB0411 and AB0466 showed nonlinear PK from 1 mg / kg to 10 mg / kg, likely due to hCEACAM5-mediated drug disposition at 1 mg / kg doses.

[0846] Incorporated by Reference

[0847] Unless otherwise indicated to the contrary, the entire disclosure of each patent document and scientific article mentioned herein is incorporated by reference for all purposes.

[0848] Equivalent

[0849] The present invention may be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the foregoing embodiments should be considered to be illustrative in all aspects, rather than limiting the invention described herein. Therefore, the scope of the present invention is indicated by the appended claims rather than the foregoing description, and all changes within the meaning and equivalent range of the claims are intended to be included therein.

Claims

1. A protein comprising: (a) binds to the first antigen binding site of NKG2D; (b) a second antigen binding site that binds to CEACAM5; and (c) a third antigen binding site that binds to CD16, or an antibody Fc domain or a portion thereof, wherein: The second antigen binding site that binds to CEACAM5 comprises a heavy chain variable domain (VH) comprising complementarity determining region (CDR) 1 (CDRH1), CDRH2 and CDRH3, and a light chain variable domain (VL) comprising CDR 1 (CDL1), CDRL2 and CDRL3, wherein: (i) CDRH1 comprises an amino acid sequence selected from the group consisting of: SEQ ID NO: 3 and 102; (ii) CDRH2 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 37, 104 and 718; (iii) CDRH3 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 6, 38 and 105; (iv) CDRL1 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 7, 40 and 107; (v) CDRL2 comprises an amino acid sequence selected from the group consisting of: SEQ ID NO: 8, 41 and 108; and (vi) CDRL3 comprises an amino acid sequence selected from the group consisting of: SEQ ID NO: 9, 42 and 109.

2. The protein according to claim 1, wherein: (a) the CDRH1, CDRH2 and CDRH3 of the second antigen binding site are respectively: (i) SEQ ID NOs: 3, 37 and 38; (ii) SEQ ID NO: 3, 718 and 6; or (iii) SEQ ID NOs: 102, 104 and 105; and (b) the CDRL1, CDRL2 and CDRL3 of the second antigen binding site are respectively: (i) SEQ ID NOs: 7, 8 and 9; (ii) SEQ ID NOs: 40, 41 and 42; or (iii) SEQ ID NO: 107, 108 and 109.

3. The protein according to claim 2, wherein the CDRH1, CDRH2, CDRH3, CDRL1, CDRL2 and CDRL3 are respectively: (i) SEQ ID NOs: 3, 37, 38, 40, 41 and 42; (ii) SEQ ID NO: 3, 718, 6, 7, 8 and 9; or (iii) SEQ ID NO: 102, 104, 105, 107, 108 and 109.

4. The protein according to any one of claims 1 to 3, wherein: The VH comprises an amino acid sequence selected from the group consisting of: SEQ ID NOs: 704, 708, 711 and 715; and The VL comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 591, 705, 712 and 716.

5. The protein according to any one of claims 1 to 4, wherein the VH and VL are: (i) SEQ ID NOs: 704 and 705; (ii) SEQ ID NOs: 708 and 591; (iii) SEQ ID NOs: 711 and 712; or (iv) SEQ ID NO:715 and 716.

6. The protein of any one of claims 1-5, wherein the second antigen binding site is a single chain variable fragment (scFv), wherein the scFv comprises an amino acid sequence selected from the group consisting of: SEQ ID NO: 703, 707, 710 and 714.

7. The protein of any one of claims 1-6, wherein the second antigen binding site binds to a human CEACAM5 variant comprising the amino acid sequence of SEQ ID NO:

391.

8. The protein according to any one of claims 1 to 7, wherein the protein comprises an antibody Fc domain or a portion thereof that binds to CD16.

9. The protein according to any one of claims 1-8, wherein the first antigen binding site that binds NKG2D is a Fab fragment, and the second antigen binding site that binds CEACAM5 is a scFv.

10. The protein of any one of claims 1-8, wherein the first antigen binding site that binds NKG2D is a scFv, and the second antigen binding site that binds CEACAM5 is a Fab fragment.

11. The protein of any one of claims 1-10, further comprising an additional antigen binding site that binds CEACAM5.

12. The protein of claim 11, wherein the first antigen binding site that binds NKG2D is a scFv, and the second antigen binding site that binds CEACAM5 and the additional antigen binding site are each Fab fragments.

13. The protein of claim 11, wherein the first antigen binding site that binds NKG2D is a scFv, and the second antigen binding site that binds CEACAM5 and the additional antigen binding site are each a scFv.

14. The protein according to any one of claims 6-13, wherein the scFv that binds to CEACAM5 and / or the scFv that binds to NKG2D comprises a heavy chain variable domain and a light chain variable domain.

15. The protein of any one of claims 6-14, wherein the scFv is linked to an antibody Fc domain or a portion thereof that binds CD16 via a hinge comprising Ala-Ser or Gly-Ser.

16. The protein of claim 15, wherein the hinge further comprises the amino acid sequence Thr-Lys-Gly.

17. The protein of any one of claims 6-16, wherein the heavy chain variable domain of the scFv forms a disulfide bridge with the light chain variable domain of the scFv.

18. The protein of claim 17, wherein the disulfide bridge is formed between C44 of the heavy chain variable domain and C100 of the light chain variable domain, numbered according to the Kabat numbering scheme.

19. The protein of any one of claims 6-18, wherein the heavy chain variable domain of the scFv is connected to the light chain variable domain of the scFv via a flexible linker.

20. The protein of claim 19, wherein the flexible linker comprises (G4S)4 (SEQ ID NO: 532).

21. The protein of any one of claims 14-20, wherein within the scFv, the heavy chain variable domain is located C-terminally to the light chain variable domain.

22. The protein of any one of claims 14-20, wherein within the scFv, the heavy chain variable domain is located N-terminally to the light chain variable domain.

23. The protein of any one of claims 9-12 and 14-22, wherein the Fab is not located between the antigen binding site and the antibody Fc domain or portion thereof.

24. The protein of any one of claims 1-23, wherein the first antigen binding site that binds to NKG2D comprises a VH comprising an amino acid sequence that is at least 90% identical to a VH sequence selected from Table 1; and a VL comprising an amino acid sequence that is at least 90% identical to a VL sequence selected from Table 1; wherein the VH sequence and VL sequence selected from Table 1 are from the same clone.

25. The protein of any one of claims 1-23, wherein the first antigen binding site that binds to NKG2D comprises: VH comprising CDRH1, CDRH2 and CDRH3 comprising the amino acid sequences of SEQ ID NOs: 495, 496 and 510, respectively; and VL comprising CDRL1, CDRL2 and CDRL3 comprising the amino acid sequences of SEQ ID NOs: 530, 224 and 499, respectively.

26. The protein of any one of claims 24-25, wherein the VH of the first antigen binding site comprises an amino acid sequence at least 90% identical to SEQ ID NO:508, and the VL of the first antigen binding site comprises an amino acid sequence at least 90% identical to SEQ ID NO:

493.

27. The protein of any one of claims 24-26, wherein the VH of the first antigen binding site comprises the amino acid sequence of SEQ ID NO: 508, and the VL of the first antigen binding site comprises the amino acid sequence of SEQ ID NO:

493.

28. The protein of any one of claims 1-27, wherein the antibody Fc domain is a human IgG1 antibody Fc domain.

29. The protein of claim 28, wherein the antibody Fc domain or portion thereof comprises an amino acid sequence that is at least 90% identical to SEQ ID NO:

531.

30. The protein of claim 28 or 29, wherein at least one polypeptide chain of the antibody Fc domain or portion thereof comprises one or more mutations at one or more positions selected from the group consisting of: Q347, Y349, L351, S354, E356, E357, K360, Q362, S364, T366, L368, K370, N390, K392, T394, D399, S400, D401, F405, Y407, K409, T411, and K439 relative to SEQ ID NO: 531 as numbered by the EU numbering system.

31. The protein of any one of claims 28-30, wherein at least one polypeptide chain of the antibody Fc domain or portion thereof comprises one or more mutations selected from the group consisting of Q347E, Q347R, Y349S, Y349K, Y349T, Y349D, Y349E, Y349C, L351K, L351D, L351Y, S354C, E356K, E357Q, E357L, E357W, K360E, K360W, Q362E, S364K, S364E, S364H, S364D, T366V, T366I, T366L, T366M, T366K, T366V, 90S, N390D, N390E, K392L, K392M, K392V, K392F, K392D, K392E, T394F, D399R, D399K, D399V, S400K, S400R, D401K, F405A, F405T, Y407A, Y407I, Y407V, K409F, K409W, K409D, T411D, T411E, K439D, and K439E.

32. The protein of any one of claims 29-31, wherein one polypeptide chain of the antibody Fc domain or portion thereof comprises one or more mutations at one or more positions selected from the group consisting of: Q347, Y349, L351, S354, E356, E357, K360, Q362, S364, T366, L368, K370, K392, T394, D399, S400, D401, F405, Y407, K409, T411, and K439 relative to SEQ ID NO: 531 according to the EU numbering system, and the other polypeptide chain of the antibody Fc domain or portion thereof comprises one or more mutations at one or more positions selected from the group consisting of: Q347, Y349, L351, S354, E356, E357, K360, Q362, S364, T366, L368, K370, K392, T394, D399, S400, D401, F405, Y407, K409, T411, and K439 relative to SEQ ID NO: NO:531 comprises one or more mutations at one or more positions selected from the group consisting of: Q347, Y349, L351, S354, E356, E357, S364, T366, L368, K370, N390, K392, T394, D399, D401, F405, Y407, K409, T411 and K439.

33. The protein of claim 32, wherein one polypeptide chain of the antibody Fc domain or portion thereof comprises K360E and K409W substitutions relative to SEQ ID NO: 531; and another polypeptide chain of the antibody Fc domain or portion thereof comprises Q347R, D399V and F405T substitutions relative to SEQ ID NO: 531, according to the EU numbering system.

34. The protein of claim 32 or 33, wherein one polypeptide chain of the antibody heavy chain constant region comprises a Y349C substitution relative to SEQ ID NO: 531 according to the EU numbering system; and the other polypeptide chain of the antibody heavy chain constant region comprises a S354C substitution relative to SEQ ID NO:

531.

35. A protein comprising: (a) a first polypeptide comprising the amino acid sequence of SEQ ID NO: 549; (b) a second polypeptide comprising the amino acid sequence of SEQ ID NO: 550; and (c) a third polypeptide comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 702, 706, 709 and 713.

36. An isolated nucleic acid molecule or a plurality of isolated nucleic acid molecules encoding the protein of any one of claims 1-35.

37. An expression vector comprising the isolated nucleic acid molecule or a plurality of isolated nucleic acid molecules according to claim 36.

38. A plurality of expression vectors collectively comprising a plurality of isolated nucleic acid molecules according to claim 36.

39. A host cell comprising the expression vector according to claim 37 or a plurality of expression vectors according to claim 38.

40. The host cell of claim 39, wherein the host cell is a Chinese Hamster Ovary (CHO) cell.

41. A method of producing a protein comprising: (a) binds to the first antigen binding site of NKG2D; (b) binds to the second antigen binding site of CEACAM5; and (c) a third antigen binding site that binds to CD16, or an antibody Fc domain or a portion thereof; The method comprises: (i) providing a host cell according to claim 39; (ii) culturing the host cell in a culture medium under conditions suitable for expressing the protein; and (iii) isolating the protein from the culture medium.

42. A method of producing a protein comprising a first polypeptide, a second polypeptide, and a third polypeptide, wherein the method comprises: (a) providing one or more host cells, wherein the one or more host cells comprise an expression vector or a plurality of expression vectors comprising: (i) a first isolated nucleic acid molecule encoding the first polypeptide comprising the amino acid sequence of SEQ ID NO: 549; (ii) a second isolated nucleic acid molecule encoding the second polypeptide comprising the amino acid sequence of SEQ ID NO: 550; and (iii) a third nucleic acid molecule encoding a third polypeptide comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 702, 706, 709 and 713; (b) culturing the one or more cells in a culture medium under conditions suitable for expressing the first polypeptide, the second polypeptide, and the third polypeptide; (c) recovering the polypeptide from the host cell and / or culture medium; and (d) purifying the recovered polypeptide under conditions whereby the protein is produced.

43. A pharmaceutical composition comprising the protein according to any one of claims 1-35 and a pharmaceutically acceptable carrier.

44. A method of increasing tumor cell death in a subject having cancer, the method comprising exposing the tumor cells and natural killer cells to an effective amount of a protein according to any one of claims 1-35 or a pharmaceutical composition according to claim 43.

45. A method for treating cancer, comprising administering an effective amount of a protein according to any one of claims 1 to 35 or a pharmaceutical composition according to claim 43 to a patient in need thereof.

46. ​​Use of a protein according to any one of claims 1-35 in the preparation of a medicament for treating cancer in a human subject in need thereof.

47. Use of a protein in the preparation of a medicament for treating cancer in a human subject in need thereof, wherein the protein comprises: (a) a first polypeptide comprising the amino acid sequence of SEQ ID NO: 549; (b) a second polypeptide comprising the amino acid sequence of SEQ ID NO: 550; and (c) a third polypeptide comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 702, 706, 709 and 713.

48. The method of claim 44 or claim 45 or the use of claim 46 or claim 47, wherein the cancer is selected from the group consisting of gastrointestinal cancer, colorectal cancer, pancreatic cancer, non-small cell lung cancer and esophageal cancer.

49. The method of claim 44 or claim 45 or the use of claim 46 or claim 47, wherein the cancer expresses CEACAM5.

50. A combination therapy for treating cancer, comprising administering to a subject in need thereof a therapeutically effective amount of a protein according to any one of claims 1-35 or a pharmaceutical composition according to claim 43 and a therapeutically effective amount of a checkpoint inhibitor.

51. The combination therapy of claim 50, wherein the checkpoint inhibitor is an anti-PD1 antibody or an anti-PD-L1 antibody.

52. The combination therapy of claim 50, wherein the checkpoint inhibitor is pembrolizumab.

53. A protein comprising an antigen binding site that binds CEACAM5, wherein the antigen binding site comprises a VH comprising CDRH1, CDRH2, and CDRH3, and a VL comprising CDRL1, CDRL2, and CDRL3, wherein: (i) CDRH1 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 3 and 102; (ii) CDRH2 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 37, 104 and 718; (iii) CDRH3 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 6, 38 and 105; (iv) CDRL1 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 7, 40 and 107; (v) CDRL2 comprises an amino acid sequence selected from the group consisting of: SEQ ID NO: 8, 41, and 108; and (vi) CDRL3 comprises an amino acid sequence selected from the group consisting of: SEQ ID NO: 9, 42 and 109.

54. The protein of claim 53, wherein: (a) the CDRH1, CDRH2 and CDRH3 are respectively: (i) SEQ ID NOs: 3, 37 and 38; (ii) SEQ ID NO: 3, 718 and 6; or (iii) SEQ ID NOs: 102, 104 and 105; and (b) CDRL1, CDRL2 and CDRL3 are respectively: (i) SEQ ID NOs: 7, 8 and 9; (ii) SEQ ID NOs: 40, 41 and 42; or (iii) SEQ ID NO: 107, 108 and 109.

55. The protein of claim 54, wherein the CDRH1, CDRH2, CDRH3, CDRL1, CDRL2 and CDRL3 are respectively: (i) SEQ ID NOs: 3, 37, 38, 40, 41 and 42; (ii) SEQ ID NO: 3, 718, 6, 7, 8 and 9; or (iii) SEQ ID NO: 102, 104, 105, 107, 108 and 109.

56. The protein of any one of claims 53-55, wherein: The VH comprises an amino acid sequence selected from the group consisting of: SEQ ID NOs: 704, 708, 711 and 715; and The VL comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 591, 705, 712 and 716.

57. The protein of any one of claims 53-56, wherein the VH and VL are: (i) SEQ ID NOs: 704 and 705; (ii) SEQ ID NOs: 708 and 591; (iii) SEQ ID NOs: 711 and 712; or (iv) SEQ ID NO:715 and 716.

58. The protein of any one of claims 53-57, wherein the antigen binding site is a Fab fragment or a scFv.

59. The protein of claim 58, wherein the antigen binding site is a scFv comprising an amino acid sequence selected from the group consisting of: SEQ ID NO: 703, 707, 710 and 714.

60. The protein of any one of claims 53-59, wherein the antigen binding site binds to a human CEACAM5 variant comprising the amino acid sequence of SEQ ID NO:

391.

61. The protein of any one of claims 53-60, wherein the protein comprises an antibody Fc domain or a portion thereof that binds CD16.

62. The protein of claim 61, wherein the antibody Fc domain or portion thereof that binds CD16 is linked to the antigen binding site via a hinge comprising Ala-Ser or Gly-Ser.

63. The protein of claim 62, wherein the hinge further comprises the amino acid sequence Thr-Lys-Gly.

64. The protein of claim 59, wherein the VH domain of the scFv forms a disulfide bridge with the VL domain of the scFv.

65. The protein of claim 64, wherein the disulfide bridge is formed between C44 of the VH and C100 of the VL, numbered according to the Kabat numbering scheme.

66. The protein of any one of claims 58-65, wherein the VH of the scFv is linked to the VL of the scFv via a flexible linker.

67. The protein of claim 66, wherein the flexible linker comprises (G4S)4 (SEQ ID NO: 532).

68. The protein of any one of claims 58-67, wherein within the scFv, the VH is located at the C-terminus of the VL.

69. The protein of any one of claims 58-67, wherein within the scFv, the VH is located at the N-terminus of the VL.

70. The protein of any one of claims 61-69, wherein the antibody Fc domain is a human IgG1 antibody Fc domain.

71. The protein of claim 70, wherein the antibody Fc domain or portion thereof comprises an amino acid sequence at least 90% identical to SEQ ID NO:

531.

72. The protein of claim 70 or 71, wherein at least one polypeptide chain of the antibody Fc domain or portion thereof comprises one or more mutations at one or more positions selected from the group consisting of: Q347, Y349, L351, S354, E356, E357, K360, Q362, S364, T366, L368, K370, N390, K392, T394, D399, S400, D401, F405, Y407, K409, T411, and K439 relative to SEQ ID NO: 531, numbered according to the EU numbering system.

73. The protein of any one of claims 70-72, wherein at least one polypeptide chain of the antibody Fc domain or portion thereof comprises one or more mutations selected from the group consisting of Q347E, Q347R, Y349S, Y349K, Y349T, Y349D, Y349E, Y349C, L351K, L351D, L351Y, S354C, E356K, E357Q, E357L, E357W, K360E, K360W, Q362E, S364K, S364E, S364H, S364D, T366V, T366I, T366L, T366M, T366K, T366V, 90S, N390D, N390E, K392L, K392M, K392V, K392F, K392D, K392E, T394F, D399R, D399K, D399V, S400K, S400R, D401K, F405A, F405T, Y407A, Y407I, Y407V, K409F, K409W, K409D, T411D, T411E, K439D, and K439E.

74. The protein of any one of claims 71-73, wherein one polypeptide chain of the antibody Fc domain or portion thereof comprises one or more mutations at one or more positions selected from the group consisting of: Q347, Y349, L351, S354, E356, E357, K360, Q362, S364, T366, L368, K370, K392, T394, D399, S400, D401, F405, Y407, K409, T411, and K439 relative to SEQ ID NO: 531 as numbered by the EU numbering system, and the other polypeptide chain of the antibody Fc domain or portion thereof comprises one or more mutations at one or more positions selected from the group consisting of: NO:531 comprises one or more mutations at one or more positions selected from the group consisting of: Q347, Y349, L351, S354, E356, E357, S364, T366, L368, K370, N390, K392, T394, D399, D401, F405, Y407, K409, T411 and K439.

75. The protein of claim 74, wherein one polypeptide chain of the antibody Fc domain or portion thereof comprises K360E and K409W substitutions relative to SEQ ID NO: 531; and another polypeptide chain of the antibody Fc domain or portion thereof comprises Q347R, D399V and F405T substitutions relative to SEQ ID NO: 531, according to the EU numbering system.

76. The protein of claim 74 or claim 75, wherein one polypeptide chain of the antibody heavy chain constant region comprises a Y349C substitution relative to SEQ ID NO: 531, according to the EU numbering system; and the other polypeptide chain of the antibody heavy chain constant region comprises a S354C substitution relative to SEQ ID NO:

531.

77. The protein of any one of claims 53-76, further comprising a second antigen binding site that binds CEACAM5.

78. An isolated nucleic acid molecule encoding the protein of any one of claims 53-77.

79. An expression vector comprising the isolated nucleic acid molecule of claim 78.

80. A host cell comprising the expression vector according to claim 79.

81. The host cell of claim 80, wherein the host cell is a Chinese Hamster Ovary (CHO) cell.

82. A method for producing a protein, comprising: (a) providing a host cell according to claim 80 or claim 81; (b) culturing the host cell in a culture medium under conditions suitable for expressing the protein; and (c) isolating the protein from the culture medium.

83. A pharmaceutical composition comprising the protein according to any one of claims 53-77 and a pharmaceutically acceptable carrier.

84. A method of increasing tumor cell death in a subject having cancer, the method comprising exposing the tumor cells and natural killer cells to an effective amount of a protein according to any one of claims 53-77 or a pharmaceutical composition according to claim 83.

85. A method for treating cancer, comprising administering an effective amount of a protein according to any one of claims 53-77 or a pharmaceutical composition according to claim 83 to a patient in need thereof.

86. Use of a protein according to any one of claims 53-77 in the preparation of a medicament for treating cancer in a human subject in need thereof.

87. The method of claim 83 or claim 84 or the use of claim 86, wherein the cancer is selected from the group consisting of gastrointestinal cancer, colorectal cancer, pancreatic cancer, non-small cell lung cancer and esophageal cancer.

88. The method of claim 83 or claim 84 or the use of claim 86, wherein the cancer expresses CEACAM5.

89. A combination therapy for treating cancer, comprising administering to a subject in need thereof a therapeutically effective amount of a protein according to any one of claims 53-77 or a pharmaceutical composition according to claim 83 and a therapeutically effective amount of a checkpoint inhibitor.

90. The combination therapy of claim 89, wherein the checkpoint inhibitor is an anti-PD1 antibody or an anti-PD-L1 antibody.

91. The combination therapy of claim 89, wherein the checkpoint inhibitor is pembrolizumab.

Images