Bispecific binding proteins and uses thereof

Abstract

To provide bispecific binding proteins and methods of using the same.SOLUTION: A protein comprises: a first binding domain (BD1) that binds to a first epitope; a second binding domain (BD2) that binds to a second epitope; and an Fc region comprising CH2 and CH3 domains; where the Fc region comprises BD2 at a solvent exposed loop in the CH2 domain, the CH3 domain, or at the interface between the CH2 and CH3 domains; and where the protein is bivalent for binding to each of the first and second epitopes. The disclosure also provides specific binding proteins, including antibodies, which bind to a target protein. The disclosure also provides methods of inducing an immune response in a subject, and methods of treating or preventing cancer in a subject by administering the proteins, nucleic acid molecules, and / or compositions to the subject.SELECTED DRAWING: Figure 1-1

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority to and the benefit of U.S. Provisional Patent Application No. 62 / 332,788, filed May 6, 2016, the disclosure of which is incorporated herein by reference in its entirety.

[0002] Sequence Listing This application is provided with a sequence listing in electronic format. The sequence listing is provided as a file named "IOBS_100_ST25.txt", created on May 2, 2017, and 244kb in size. The information in the electronic format of the sequence listing is incorporated herein by reference in its entirety.

[0003] Technical Field The present invention relates to bispecific binding proteins and uses thereof. [Background technology]

[0004] Cancer remains a major global health burden. Despite advances in the treatment of cancer, there continues to be an unmet need for more effective and less toxic therapies, especially for patients with advanced disease or cancers resistant to existing therapeutic agents.

[0005] The role of the immune system, particularly T cell-mediated cytotoxicity, in tumor control is well recognized. There is increasing evidence that T cells control tumor growth and survival in cancer patients at both early and late stages of the disease. However, it is difficult to enhance and sustain tumor-specific T cell responses in cancer patients. The continued advancement and success of cancer immunotherapy drugs that stimulate or enhance the innate immune response against cancer makes these therapeutic agents an attractive treatment option compared to therapies that use non-specific chemotherapy drugs and / or radiation.

[0006] Several molecular targets have been identified for their potential usefulness as cancer immunotherapy (IO) agents against cancer. Some molecular targets that are being investigated for their therapeutic potential in the field of cancer immunotherapy include cytotoxic T-lymphocyte antigen 4 (CTLA-4 or CD152), programmed death-ligand 1 (PD-L1 or B7-H1 or CD274), programmed death-1 (PD-1), OX40 (CD134 or TNFRSF4) and T-cell inhibitory receptor T-cell immunoglobulin and mucin-domain containing 3 (TIM3). Some of these targets have been successfully exploited for treatment (e.g., PD-1 and CTLA-4), but many patients are unresponsive to the therapeutic agents developed. Treatment regimens involving higher doses and / or combinations of immunotherapy agents can also be considered, but such treatments may be associated with an increased risk of side effects, which tend to increase with higher doses and cumulative exposure, and may be additive when used with combination immunotherapy. Some common side effects include hypophysitis, thyroiditis, adrenal insufficiency, enteritis, dermatitis, pneumonitis, hepatitis, pancreatitis, motor and sensory neuropathy, and arthritis. Furthermore, immunotherapy drugs typically have high costs, so treatments that include combinations of immunotherapy drugs can be cost-prohibitive for patients.

[0007] Thus, there remains a need to identify candidate targets for IO therapeutics, develop novel therapeutics against existing targets, and develop treatment strategies that avoid the shortcomings of currently used immunotherapeutic agents, such as lack of patient responsiveness and increased risk of side effects associated with combination treatments. IO therapeutics (e.g., binding proteins) with dual specificity for combinations of target molecules, especially those that exhibit superior binding affinity for the target molecules when compared to the binding affinity for the combination of individual monospecific binding proteins, provide a particularly desirable class of molecules for therapeutic efficacy. Summary of the Invention [Means for solving the problem]

[0008] The present invention provides bispecific molecules or proteins that bind two epitopes (e.g., a first and a second epitope) and are bivalent for binding to each of the first and second epitopes. The present invention also provides methods of inducing an immune response in a subject, as well as methods of treating or preventing cancer in a subject (e.g., a human subject) by administering the proteins, nucleic acid molecules and / or compositions to the subject.

[0009] In one aspect, the present invention provides a method for the production of a mAb comprising a first binding domain (BD1) that binds to a first epitope, a second binding domain (BD2) that binds to a second epitope, and a C H 2 domain and C H and an Fc region having three domains, H 2 domain, C H 3 domain, or C H 2 and C H The inclusion of BD2 in a solvent exposed loop at the interface of the three domains provides a protein that is bivalent for binding to each of a first and a second epitope.

[0010] In another aspect, the invention provides a composition comprising a protein or antibody according to any aspect described herein and a pharma- ceutically acceptable carrier.

[0011] In another aspect, the invention provides a method of treating or preventing cancer in a subject, the method comprising administering to a subject (e.g., a human subject) a protein or antibody according to any aspect described herein. In various embodiments, the cancer is one or more of ovarian cancer, breast cancer, colon cancer, prostate cancer, cervical cancer, uterine cancer, testicular cancer, bladder cancer, head and neck cancer, melanoma, pancreatic cancer, renal cell carcinoma, and lung cancer.

[0012] In another aspect, the invention provides a method of inducing an immune response in a subject, the method comprising administering to the subject (e.g., a human subject) a protein or antibody according to any aspect described herein.

[0013] In another aspect, the invention provides a nucleic acid molecule having a nucleotide sequence encoding a protein or antibody according to any aspect described herein.

[0014] In another aspect, the present invention provides a vector containing a nucleic acid molecule according to any aspect described herein.

[0015] In another aspect, the invention provides a host cell containing a vector according to any aspect described herein.

[0016] In one aspect, the invention provides a bispecific binding protein that binds to PD-1 and CTLA-4 having a first peptide having the amino acid sequence of SEQ ID NO:1 and a second peptide having the amino acid sequence of SEQ ID NO:2.

[0017] In another aspect, the invention provides a bispecific binding protein that binds to PD-1 and CTLA-4 having a first peptide having the amino acid sequence of SEQ ID NO:3 and a second peptide having the amino acid sequence of SEQ ID NO:4.

[0018] In another aspect, the invention provides a bispecific binding protein that binds to PD-1 and CTLA-4 having a first peptide having the amino acid sequence of SEQ ID NO:5 and a second peptide having the amino acid sequence of SEQ ID NO:6.

[0019] In one aspect, the invention provides a bispecific binding protein that binds PD-1 and CTLA-4 having a first heavy chain having the amino acid sequence of SEQ ID NO:9, a first light chain having the amino acid sequence of SEQ ID NO:7, a second heavy chain having the amino acid sequence of SEQ ID NO:12, and a second light chain having the amino acid sequence of SEQ ID NO:4.

[0020] In one aspect, the invention provides a bispecific binding protein that binds PD-L1 and CTLA-4 having a first peptide having the amino acid sequence of SEQ ID NO:14 and a second peptide having the amino acid sequence of SEQ ID NO:15.

[0021] In another aspect, the invention provides a bispecific binding protein that binds PD-L1 and CTLA-4 having a first peptide having the amino acid sequence of SEQ ID NO:16 and a second peptide having the amino acid sequence of SEQ ID NO:17.

[0022] In another aspect, the invention provides a bispecific binding protein that binds PD-L1 and CTLA-4 having a first peptide having the amino acid sequence of SEQ ID NO:18 and a second peptide having the amino acid sequence of SEQ ID NO:19.

[0023] In one aspect, the invention provides a bispecific binding protein that binds to PD-1 and TIM3 having a first peptide having the amino acid sequence of SEQ ID NO:22 and a second peptide having the amino acid sequence of SEQ ID NO:23.

[0024] In another aspect, the invention provides a bispecific binding protein that binds to PD-1 and TIM3 having a first peptide having the amino acid sequence of SEQ ID NO:24 or SEQ ID NO:91 and a second peptide having the amino acid sequence of SEQ ID NO:23 or SEQ ID NO:92.

[0025] In one aspect, the invention provides a bispecific binding protein that binds PD-1 and TIM3 having a first heavy chain having the amino acid sequence of SEQ ID NO:9, a first light chain having the amino acid sequence of SEQ ID NO:7, a second heavy chain having the amino acid sequence of SEQ ID NO:27 or SEQ ID NO:30, and a second light chain having the amino acid sequence of SEQ ID NO:26 or SEQ ID NO:28.

[0026] In one aspect, the invention provides a bispecific binding protein that binds to OX40 and PD-L1 having a first peptide having the amino acid sequence of SEQ ID NO:34 and a second peptide having the amino acid sequence of SEQ ID NO:32.

[0027] In another aspect, the invention provides a bispecific binding protein that binds to OX40 and PD-L1 having a first peptide having the amino acid sequence of SEQ ID NO:35 and a second peptide having the amino acid sequence of SEQ ID NO:32.

[0028] In another aspect, the invention provides a bispecific binding protein that binds OX40 and PD-L1 having a first peptide having the amino acid sequence of SEQ ID NO:36 or SEQ ID NO:94, and a second peptide having the amino acid sequence of SEQ ID NO:32 or SEQ ID NO:93.

[0029] In another aspect, the invention provides an antibody or antigen-binding fragment thereof that binds to TIM3 having a heavy chain having CDR1, CDR2 and CDR3 and a light chain having CDR1, CDR2 and CDR3, wherein the heavy chain CDR1 comprises SEQ ID NO: 88, the heavy chain CDR2 comprises SEQ ID NO: 80, the heavy chain CDR3 comprises SEQ ID NO: 81, and the light chain CDR1 comprises SEQ ID NO: 82, the light chain CDR2 comprises SEQ ID NO: 83 and the light chain CDR3 comprises SEQ ID NO: 84.

[0030] In other aspects, the invention provides compositions having the bispecific binding proteins and a pharma- ceutically acceptable carrier, nucleic acid molecules having a nucleotide sequence encoding the bispecific binding proteins, methods of treating or preventing cancer in a subject by administering the bispecific binding proteins, and methods of enhancing an immune response in a subject by administering the bispecific binding proteins.

[0031]

[0023] In various embodiments of any aspect delineated herein, the Fc region comprises H 2 domain, C H 3 domain, or C H 2 and CH The solvent-exposed loop in the amino acid sequence at the interface of the three domains contains BD2.

[0032]

[0023] In various embodiments of any aspect delineated herein, the solvent exposed loop comprises a C H In certain embodiments, the solvent exposed loop comprises the amino acid sequence ISRTP (SEQ ID NO: 39).

[0033]

[0023] In various embodiments of any aspect delineated herein, the solvent exposed loop comprises a C H In certain embodiments, the solvent exposed loop comprises the amino acid sequence SNG.

[0034]

[0023] In various embodiments of any aspect delineated herein, the solvent exposed loop comprises a C H 2 domain and C H The amino acid sequence of claim 7 is derived from the interface of the three domains. In certain embodiments, the solvent exposed loop comprises the amino acid sequence AKGQP (SEQ ID NO: 40).

[0035] In various embodiments of any aspect delineated herein, BD2 is or comprises a single chain variable fragment (scFv).

[0036] In various embodiments of any of the aspects delineated herein, BD1 is or comprises a binding domain that is one or more of a Fab domain, an scFv, a single domain antibody, and an antibody variable domain. In certain embodiments, BD1 comprises a Fab domain.

[0037] In various embodiments of any aspect described herein, the Fab domain is linked to the Fc region via an antibody hinge region. In certain embodiments, the Fc region is or comprises one or more domains of Fc regions from IgG1, IgG2, IgG3, IgG4, IgA, IgM, IgE, or IgD. In certain embodiments, the Fc region comprises a mutant Fc region. In some embodiments, the Fc region is non-glycosylated, deglycosylated, and / or non-fucosylated, or has low fucosylation.

[0038] In various embodiments of any of the aspects described herein, the protein further comprises a protein linker L1 between BD2 and the Fc region. In various embodiments of any of the aspects described herein, the protein further comprises a first protein linker L1 and a second protein linker L2 between BD2 and the Fc region. In various embodiments of any of the aspects described herein, BD2 is linked to the Fc region via a protein linker L1. In various embodiments of any of the aspects described herein, BD2 is linked to the Fc region via two protein linkers L1 and L2. In an embodiment, L1 and L2 are independently selected from (G4S)2 (SEQ ID NO: 41), (G4S)3 (SEQ ID NO: 42), and (G4S)4 (SEQ ID NO: 43).

[0039]

[0023] In various embodiments of any of the aspects delineated herein, the protein comprises from N-terminus to C-terminus the following polypeptide domains: H 1-C H 1-C H 2(N-terminus)-BD2-C H 2(C-terminus)-C H 3, and BD1 comprises a Fab domain, H 1 contains a heavy chain variable domain of a Fab domain, and C H 1 contains the heavy chain constant domain 1 of the Fab.

[0040]

[0023] In various embodiments of any of the aspects delineated herein, the protein comprises from N-terminus to C-terminus the following polypeptide domains: H 1-C H 1-C H 2-BD2-C H 3, and BD1 comprises a Fab domain, H 1 contains a heavy chain variable domain of a Fab domain, and C H 1 contains the heavy chain constant domain 1 of the Fab.

[0041]

[0023] In various embodiments of any of the aspects delineated herein, the protein comprises from N-terminus to C-terminus the following polypeptide domains: H 1-C H 1-C H 2-C H 3(N-terminus)-BD2-C H 3 (C-terminus), and BD1 comprises a Fab domain, V H 1 contains a heavy chain variable domain of a Fab domain, and C H 1 contains the heavy chain constant domain 1 of the Fab.

[0042] In various embodiments of any aspect delineated herein, BD2 is or comprises an scFv. In certain embodiments, the scFv comprises, from N-terminus to C-terminus, H 2-Polypeptide Linker-V L 2 or V L 2 Polypeptide Linker-V H 2, including V H 2 comprises the heavy chain variable domain of the scFv, and V L 2 comprises the light chain variable domain of the scFv.

[0043] In various embodiments of any aspect delineated herein, the protein further comprises a protein linker L1 between BD2 and the Fc region. In various embodiments of any aspect delineated herein, the protein further comprises a first protein linker L1 and a second protein linker L2 between BD2 and the Fc region.

[0044]

[0023] In various embodiments of any of the aspects described herein, BD2 is linked to the C of the Fc region via a linker (L1). H 2 domain, C H 2 domain, or C H 2 and C H It is bound to the interface of the three domains.

[0045]

[0023] In various embodiments of any of the aspects described herein, BD2 is linked to the C of the Fc region via two protein linkers L1 and L2. H 2 domain, C H 3 domain, or C H 2 and C H The linker is attached to the interface of the three domains. In various embodiments, L1 and L2 are independently selected from protein linkers having a length of 1 to 25 amino acids. In certain embodiments, L1 and L2 are independently selected from (G4S)2 (SEQ ID NO: 41), (G4S)3 (SEQ ID NO: 42), and (G4S)4 (SEQ ID NO: 43).

[0046] In various embodiments of any aspect delineated herein, the first and second epitopes are different. In various embodiments of any aspect delineated herein, the first and second epitopes are the same.

[0047] In describing the present disclosure, certain aspects of the present disclosure are shown in the drawings. However, the present disclosure is not limited to the precise arrangements and instrumentalities of the aspects shown in the drawings. [Brief description of the drawings]

[0048] [Figure 1-1]

[0023] Figure 1 shows a general schematic of some exemplary proteins described herein. The CH2 and CH3 regions are shown in Figures 1A-1C using PyMOL to reveal the solvent exposed surface loop regions as spheres. Figure 1A depicts loops in the CH2 region; Figure 1B depicts loops in the CH2-CH3 interface; Figure 1C depicts loops in the CH3 region. Exemplary constructs are illustrated in Figures 1D-1F, which include representative BD1 and BD2 domains as Fab and scFv domains, respectively. Figure 1D depicts BD1 bound at the hinge region and BD2 bound at a solvent exposed loop in the CH2 region. Figure 1E depicts BD1 bound at the hinge region and BD2 bound at a solvent exposed loop in the CH2-CH3 interface. Figure 1F depicts BD1 bound at the hinge region and BD2 bound at a solvent exposed loop in the CH3 region. [Figure 1-2] This is a continuation of Figure 1-1. [Figure 2-1] As described herein, a close-up view of the solvent accessible loop sequences in CH2, the CH2-CH3 interface, and CH3 is provided. An example of a construct incorporating BD2 (scFv) is included in each of Figures 2A-2C. Figure 2A illustrates a representative loop sequence, ISRTP (SEQ ID NO: 39), identified in the CH2 loop upstream of the CH2-CH3 interface. Figure 2B illustrates a representative loop sequence, AKGQP (SEQ ID NO: 40), in the CH2-CH3 interface. Figure 2C illustrates a representative loop sequence, SNG, in the CH3 region downstream of the CH2-CH3 interface. [Figure 2-2] This is a continuation of Figure 2-1. [Figure 2-3] This is a continuation of Figure 2-2. [Diagram 3] 1 demonstrates simultaneous binding of BiS2, BiS3, and BiS5 constructs targeting PD-1 and CTLA-4. Trace A9 shows BiS2 PD-1 / CTLA-4, trace B9 shows BiS3 PD-1 / CTLA-4, and trace C9 shows BiS5 PD-1 / CTLA-4. [Figure 4] Schematic diagram of the proposed mechanism for PD-1 / CTLA-4 blockade. [Diagram 5]1 shows the results of an octet binding assay demonstrating simultaneous binding of DuetMab constructs targeting PD-1 and CTLA-4. [Figure 6-1] Figure 6 shows PD-1 / CTLA-4 bispecific binding proteins inhibiting PD-1 and CTLA-4 pathways in reporter gene assays. Figure 6A depicts T cell activation by PD-1 blockade. Figure 6B depicts T cell activation by CTLA-4 blockade. Figure 6C depicts the results of a PD-1 reporter assay. Figure 6D depicts the results of a CTLA-4 reporter assay. [Figure 6-2] This is a continuation of Figure 6-1. [Figure 7] 1 shows the results of a Staphylococcal Enterotoxin B (SEB) assay demonstrating that PD-1 / CTLA-4 DuetMab and BiS5Ab have comparable activity in the SEB assay. [Figure 8] Shows activity of PD-1 / CTLA-4 DuetMabs compared to isotype and parental mAb controls in the SEB assay. [Figure 9] Shows the activity of PD-1 / CTLA-4 BiS5Ab compared to PD-1 / CTLA-4 DuetMab in the SEB assay. [Figure 10-1] Figure 10A shows that PD-1 / CTLA-4 DuetMab and BiS5Ab have comparable activity in a mixed leukocyte reaction assay (MLR). Figure 10A is a schematic of the assay (n=4 donors; 2 independent experiments). [Figure 10-2] This is a continuation of Figure 10-1. [Figure 11-1] Shows activity of PD-1 / CTLA-4 DuetMab compared to isotype control in MLR assay (n=2 donors; 1 experiment). [Figure 11-2] This is a continuation of Figure 11-1. [Figure 12-1] Shows activity of PD-1 / CTLA-4 DuetMab compared to parental mAb control in MLR assay (n=2 donors; 1 experiment). [Figure 12-2] This is a continuation of Figure 12-1. [Figure 13-1] Shows activity of PD-1 / CTLA-4 DuetMab compared to competing antibodies in the MLR assay (n=2 donors; 1 experiment). [Figure 13-2] This is a continuation of Figure 13-1. [Figure 14] 1 shows the study design for a single-dose pharmacokinetic / pharmacology (PK / PD) study in cynomolgus monkeys. [Figure 15] To demonstrate that PD-1 / CTLA-4 DuetMab exhibits distinct pharmacological effects (PD) in cynomolgus monkeys. [Figure 16] Shows T cell dependent antibody responses (TDAR) in cynomolgus monkeys administered PD-1 / CTLA-4 DuetMab (MEDI5752) and PD-1 / CTLA-4 BiS5Ab (MEDI8500). [Figure 17] A model system for testing PD-1 / CTLA-4 bispecific molecules in which stable CHO cells express various levels of human PD-1 and / or CTLA-4 is presented. [Figure 18] We demonstrate that the PD-1 / CTLA-4 DuetMab binds simultaneously to PD-1 and CTLA-4 on the surface of the same cells. [Figure 19] 13 shows an experiment to determine whether cooperative binding differs for combinations of anti-PD-1 and CTLA-4 antibodies at saturation of CTLA-4 on cells expressing excess PD-1. [Figure 20] We demonstrate that PD-1 and CTLA-4 parental monoclonal antibodies bind to and occupy their target receptors without having a measurable effect on non-target receptors. [Figure 21] We demonstrate that the PD-1 / CTLA-4 DuetMab saturates CTLA-4 on CHO cells expressing excess PD-1 at approximately 250-fold lower concentrations compared to the monoclonal antibody combination. [Figure 22]We demonstrate that the PD-1 / CTLA-4 DuetMab saturates CTLA-4 on CHO cells expressing excess PD-1 at approximately 500-fold lower concentrations compared to cells expressing CTLA-4 alone. [Figure 23-1] We demonstrate that the PD-1 / CTLA-4 DuetMab preferentially binds in cis to PD-1 and CTLA-4 on the surface of the same cells. Treme in Figure 23B is a CTLA-4 mAb. [Figure 23-2] This is a continuation of Figure 23-1. [Figure 24] Binding and internalization of PD-1 / CTLA-4 DuetMab and parental monoclonal antibody to cultured T cells. PD-1 / CTLA-4 DuetMab has the internalization properties of tremelimumab. [Diagram 25] Figure 25A: Schematic of the internalization assay. Figure 25B: Demonstration that PD-1 / CTLA-4 DuetMab adopts the internalization properties of tremelimumab in stable CHO cells expressing 10-fold excess PD-1. [Figure 26] Demonstrating simultaneous binding of BiS2, BiS3, and BiS5 constructs targeting PD-L1 and CTLA-4. Trace A11 shows BiS2 PD-L1 CTLA-4, trace B11 shows BiS3 PD-L1 CTLA-4, and trace C11 shows BiS5 PD-L1 CTLA-4. [Figure 27A] Demonstrating simultaneous binding of BiS3 constructs targeting PD-1 and TIM3 (clone 62, wild type). [Figure 27B] Demonstrating simultaneous binding of DuetMab constructs targeting PD-1 and TIM3 (clone 62, wild type). [Figure 28-1]Provides a summary of cell killing activity of monospecific TIM3 and bispecific PD-1_TIM3PD-1 bispecific constructs in cell killing assays. Figure 28A shows brightfield images of co-cultured wells at 18 hours. Combining anti-TIM-3+anti-PD1 or TIM-3 / PD-1 bispecific formats enhances tumor cell death as assessed by reduction of adherent cells and increased T cell blasting (aggregation) and increases T cell activation. Figure 28B shows assessment of vital dye uptake by tumor cells after 18 hours of co-culture with melanoma-specific CD8+ T cells. Figure 28C shows IFNγ secretion after 18 hours of co-culture. Bispecific constructs generally show superior killing activity, with the DuetMab format exerting the most robust killing activity. [Figure 28-2] This is a continuation of Figure 28-1. [Figure 29] Demonstrates simultaneous binding of PD-1 / TIM3 DuetMab constructs bearing TIM3 arm sequences that are affinity matured variants of clone 62. [Diagram 30] Figure 2 shows the binding of PD-1 / TIM3 bispecific antibodies, including BiS3, BiS5, and DuetMab, to CHO cells overexpressing human TIM3 or human PD1. PD-1 and TIM3 expression data are shown in the inset. [Diagram 31] Figure 2 shows TIM3 and PD-1 expression data on an activated T cell clone (DMF4). [Diagram 32] We present results from a CMV antigen recall assay demonstrating that PD-1 / TIM3 bispecific antibodies, including BiS3, BiS5, and DuetMab, exhibit enhanced activity compared to isotype treatment (3 donors (1-2 replicates per treatment / donor), 1 experiment). [Diagram 33] We demonstrate that PD-1 / TIM3 bispecific antibodies, including BiS3, BiS5, and DuetMab, enhanced interferon (IFNγ) at concentrations greater than 8 nM in a mixed lymphocyte reaction (MLR) assay (2–4 replicates per treatment / donor pair / one of two independent experiments). [Diagram 34] Shown are the results of a PD-1 reporter assay (dual cell line) using PD-1 / TIM3 bispecific antibodies including BiS3, BiS5, and DuetMab. All bispecific formats show similar activity to the parental LO115 IgG1 (collection of 3-5 independent experiments per treatment / 3 biological replicates). [Diagram 35] 1 shows the results of an octet assay using BiS2 and BiS3 OX40 / PD-L1 bispecific molecules. [Figure 36-1] 1 shows the results of an SEB assay using BiS2 and BiS3 OX40 / PD-L1 bispecific molecules. [Figure 36-2] This is a continuation of Figure 36-1. [Figure 37] Figure 1 shows the results of a PD-L1 reporter assay using BiS2 and BiS3 OX40 / PD-L1 bispecific molecules. [Figure 38] 1 shows the results of a CMV Ag recall assay using BiS2 and BiS3 OX40 / PD-L1 bispecific molecules. [Figure 39] 1 shows the results of an octet binding assay demonstrating simultaneous binding of the OX40(SLR) / PD-L1 BiS5 construct targeting PD-L1 and OX40. [Figure 40-1] Shows binding of bispecific constructs targeting PD-L1 and OX40 to CHO cells expressing human or cynomolgus OX40 and PD-L1 / B7H1. [Figure 40-2] This is a continuation of Figure 40-1. [Diagram 41] Binding of bispecific constructs targeting PD-L1 and OX40 to Jurkat OX40 reporter cells, NCI H358, CHOK1 B7H1(PD-L1) / OKT3 cells, and HEK CD32a cells as measured by flow cytometry (HyperCyt). [Diagram 42] 1 shows the results of a PD-L1 reporter assay using OX40 / PD-L1 bispecific molecules. [Diagram 43]1 shows the results of an OX40 reporter assay in HEK CD32a cells using OX40 / PD-L1 bispecific molecules. [Diagram 44] Figure 1 shows the results of an OX40 reporter assay in CHOK PD-L1 overexpressing cells using OX40 / PD-L1 bispecific molecules. [Diagram 45] 1 shows PD-L1-mediated OX40 agonism on tumor cells using OX40 / PD-L1 bispecific molecules. [Figure 46] 1 shows results showing that in the OX40 / tumor cell assay control, no agonism was detected in NCI H358PD-L1 KO cells with the OX40 / PD-L1 bispecific molecule. [Figure 47-1] 1 shows the results of an SEB assay using OX40 / PD-L1 bispecific molecules. [Figure 47-2] Continuation of Figure 47-1. [Figure 48-1] Figure 48A: Schematic diagram of Treg suppression assay experiment for testing OX40 / PD-L1 bispecific molecules. Figure 48B-C: Treg suppression based on binding of bispecific molecules. [Figure 48-2] This is a continuation of Figure 48-1. [Figure 49] 1 shows the results of a Treg suppression assay using OX40 / PD-L1 bispecific molecules. [Figure 50] 1 shows the results of a Treg suppression assay using OX40 / PD-L1 bispecific molecules. [Figure 51] 1 shows the mixed leukocyte reaction (MLR) assay experimental design for testing OX40 / PD-L1 bispecific molecules. [Figure 52-1] 1 shows the results of an MLR assay using OX40 / PD-L1 bispecific molecules. [Figure 52-2] Continuation of Figure 52-1. [Diagram 53]We demonstrate that BiS2 and BiS5 OX40 / PD-L1 bispecific molecules mediate antibody-dependent cell-mediated cytotoxicity (ADCC) of natural killer (NK) cells against PD-L1- or OX40-expressing CHO cells. [Figure 54] We demonstrate that BiS2 and BiS5 OX40 / PD-L1 bispecific molecules mediate antibody-dependent cell-mediated cytotoxicity (ADCC) of NK cells against PD-L1- and OX40-expressing CHO cells. [Figure 55] We demonstrate that BiS2 and BiS5 OX40 / PD-L1 bispecific molecules enhance CD107a recruitment of NK cells against PD-L1 and OX40-expressing CHO cells in antibody-dependent cell-mediated cytotoxicity (ADCC). [Figure 56] We demonstrate that BiS2 and BiS5 OX40 / PD-L1 bispecific molecules mediate antibody-dependent cell-mediated cytotoxicity (ADCC) of NK cells against activated allogeneic T cells. [Figure 57] We demonstrate that BiS5 OX40 / PD-L1 enhances CD107a recruitment of NK cells from two different donors against activated allogeneic T cells in antibody-dependent cell-mediated cytotoxicity (ADCC). [Figure 58] 1 shows the study design for comparing the PK / PD of OX40 / PD-L1 bispecific molecules. [Figure 59] Figure 1 shows a comparison of serum concentration-time profiles of PD-L1 / OX40 bispecific molecules in cynomolgus monkeys. [Figure 60] 1 shows depletion of soluble PD-L1 in serum by PD-L1 / OX40 bispecific molecules. [Figure 61-1] Provides an overview of pharmacodynamic data for PD-L1 / OX40 bispecific molecules. Baseline is defined as the mean of pre-dose days -5 and 0. [Figure 61-2] This is a continuation of Figure 61-1. [Figure 62] 1 depicts a schematic diagram of PD-1 / OX40 BiS2 IgG4P monoclonal antibody (mAb). [Figure 63] Schematic representation of the potential mechanism of action of PD-1 / OX40 BiS2 mAb. [Figure 64] 1 depicts the simultaneous binding activity of two different lots of PD-1(LO115) / OX40 BiS2 mAb to PD1-His and human OX40-Fc. [Figure 65] Figure 65A: Schematic of the OX40 reporter assay. Figure 65B: Results of the OX40 reporter assay using PD1 LO115mAb, OX40 mAb, control mAb and PD1(LO115) / OX40 BiS2 mAb. [Figure 66] Figure 66A: Schematic of the PD-1 / PD-L1 reporter assay. Figure 66B: Results of the PD1 / PD-L1 reporter assay using PD1 LO115mAb, OX40 mAb, control mAb and PD1(LO115) / OX40 BiS2 mAb. [Figure 67] 1 shows the results of an SEB assay using the BiS2 variants of the PD-1(LO115) / OX40 bispecific molecule and controls. [Figure 68] 1 shows the results of a CMV antigen recall assay using the BiS2 variant of the PD-1(LO115) / OX40 bispecific molecule and a control. [Figure 69] 1 shows the results of a CMV antigen recall assay using the BiS2 and BiS3 variants of the PD-1 (AMP514) / OX40 bispecific molecule and controls. [Figure 70] Figure 1 shows serum concentration-time profile of PD1(LO115) / OX40 BiS2mAb after a single IV dose in cynomolgus monkeys. Data represent mean ± standard deviation of 3 males / group. LLOQ (5 ng / mL is indicated by dotted line. PKC = pharmacokinetic curve; LLOQ = lower limit of quantification. [Figure 71]Figure 1 shows the percentage of Ki67 positive CD4+ and CD8+ memory T cells after a single IV dose of PD-1(LO115) / OX40 BiS2mAb in cynomolgus monkeys. Data represent the mean ± standard deviation of 3 males / group. Left panel A shows CD4+ memory T cells, right panel A shows CD8+ memory T cells. IV = intravenous. [Figure 72] A representative standard curve for PD-1 / OX40 quantification in cynomolgus monkey serum is shown. [Figure 73-1] Provided are exemplary DSC thermograms of a bispecific binding protein disclosed herein ("BiS5") compared to a different BiS format ("BiS4") at various pH values. Figure 73A shows the effect of pH value on the thermal stability of BiS4. Figure 73B shows the effect of pH value on the thermal stability of BiS5. Figure 73C depicts a representative curve-fitted DSC thermogram of BiS4, including Tonset, Tm1, Tm2, ​​and Tm3. Figure 73D depicts a representative curve-fitted DSC thermogram of BiS5, including Tonset, Tm1, Tm2, ​​and Tm3. Figure 73E depicts a plot representing the effect of pH on Tonset, Tm1, Tm2, ​​and Tm3 for BiS4 and BiS5 formats. [Figure 73-2] Continuation of Figure 73-1. [Figure 74-1] FIG. 74A depicts HP-SEC analysis of samples at pH 7.5 before and after 4 weeks of storage at 40° C. FIG. 74A shows overlaid plots of SEC chromatograms of BiS4 and BiS5 before and after heat stress, with the solid lines corresponding to BiS4 and BiS5 samples at time zero (unstressed) and the dotted lines corresponding to BiS4 and BiS5 samples incubated at 40° C. for 4 weeks (stressed). FIG. 74B provides bar graphs depicting the effect of pH 7.5 on the various species (monomers, fragments, and aggregates) of BiS4 and BiS5 measured at day zero and after 4 weeks at 40° C. [Figure 74-2] This is a continuation of Figure 74-1. [Figure 75]Figure 75 provides kinetic plots showing the effect of pH 7.5 on accelerated and short-term storage stability at 40°C for BiS4 (triangle trace) and BiS5 (circle trace). Figure 75A shows the percentage of residual monomer measured by HP-SEC over a 4-week period. Figure 75B shows the percentage of fragmentation measured by HP-SEC over a 4-week period. Figure 75C shows the percentage of aggregation measured by HP-SEC over a 4-week period. Data shown is from a single vial analysis. [Figure 76] The pH-rate profile plots of BiS4 (triangles) and BiS5 (circles) are depicted. Figure 76A shows the effect of various pH conditions on the rate of monomer loss at 40°C. Figure 76B shows the effect of various pH conditions on the rate of fragmentation at 40°C. Figure 76C shows the effect of various pH conditions on the rate of aggregation at 40°C. [Figure 77-1] Further analysis of BiS4 and BiS5 fragmentation is depicted. Figure 77A shows that at pH 7.5 and 40°C (time = 0) neither molecule exhibits obvious fragmentation. Figure 77B shows that under the same conditions as Figure 77A, but after 2 weeks of storage at 40°C, obvious fragmentation is observed for BiS4 and slight fragmentation for BiS5. [Figure 77-2] Continuation of Figure 77-1. [Figure 78] Analysis of BiS4 and BiS5 fragmentation (left panel) and aggregation (right panel) as a function of pH is depicted. Both formats have reduced fragmentation and aggregation at low (5.5) pH, but BiS5 has superior performance for both fragmentation and aggregation at both pH values. [Figure 79]DSC thermograms of several BiS5 bispecific binding proteins disclosed herein are depicted for constructs (left panel) A, B, C and D, and (right panel) E, F, G and H. Constructs A and E contain scFvs in IS-RTP; B and F contain scFvs in AK-GQP; C and G contain scFvs in S-NG; D and H contain scFvs in SN-G. The scFvs in constructs A, B, C and D are 2F4 (IgG is LC10) and E, F, G and H are LC10 (IgG is 2F4). The various TM values ​​relate to the following domains: TM1=CH2 / scFv; TM2=Fab; TM3=CH3. [Figure 80] A representative data set for FcRn binding with the bispecific binding proteins disclosed herein (constructs D and H) is shown. The scFv positioning of the CH2-CH3 domain (i.e., within the ISTRP loop) may have an effect on FcRn binding activity. [Figure 81] A representative data set for FcγR binding with the bispecific binding proteins disclosed herein (constructs E, G, and H, including FcγRIIIa-158V) is shown. The inset represents the same data as the main figure, re-normalized to the time of FcγRIIIa-158V injection. All constructs were able to bind FcγR, but there were observable differences in affinity based on the position of the scFv domain. [Figure 82] This demonstrates that an intact ISRTP loop is important for FcRn binding of exemplary bispecific binding constructs (e.g., A and E). Introduction of the N3 loop does not compensate for the disruption of the ISRTP loop (BiS5E+N3). The scFv introduced into the N3 loop inserted into the IgG1 Fc renders the IgG unable to bind to FcRn. Time on the x-axis is measured in seconds. [Figure 83] The general schematic structural format of each of the BiS1, BiS2, BiS3, BiS4, and BiS5 constructs is depicted. The "k1" and "k2" designations refer to the fragmentation patterns used in the kinetic analysis discussed in Example 3. [Figure 84] The fragmentation rates of BiS1, BiS2, BiS3, BiS4, and BiS5 as a function of pH are shown. [Figure 85] 1 shows the aggregation rate of each of BiS1, BiS2, BiS3, BiS4, and BiS5 as a function of pH. [Figure 86] 1 shows the rate of monomer loss for BiS1, BiS2, BiS3, BiS4, and BiS5 as a function of pH. [Figure 87] The fragmentation patterns and their correspondence with HPSEC chromatogram peaks for BiS1, BiS2, BiS3, BiS4, and BiS5, respectively, are depicted. [Figure 88] A representative analysis of fragmentation patterns under reducing conditions is depicted. [Figure 89] 1 depicts the structural arrangement of BiS5 under reducing and non-reducing conditions. [Figure 90] 1 depicts the SEB assay format. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0049] Before continuing to describe the present disclosure in greater detail, it is to be understood that the present disclosure is not limited to particular compositions or process steps, which may vary. It should be noted that as used in this specification and the appended claims, the singular forms "a," "an," and "the" include plural meanings unless the context clearly dictates otherwise.

[0050] Unless otherwise stated, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present invention pertains.For example, Concise Dictionary of Biomedicine and Molecular Biology, Juo, Pei-Show, 2nd ed., 2002, CRC Press; The Dictionary of Cell and Molecular Biology, 3rd ed., 1999, Academic Press; and Oxford Dictionary Of Biochemistry And Molecular Biology, Revised, 2000, Oxford University Press provide a general dictionary for those skilled in the art for many of the terms used in the present invention.

[0051] Amino acids may be referred to herein by either their commonly known three letter symbols or by the one-letter symbols recommended by the IUPAC-IUB Biochemical Nomenclature Commission. Nucleotides, likewise, may be referred to by their commonly accepted one-letter symbols.

[0052] The numbering of amino acids in the variable domains, complementarity determining regions (CDRs) and framework regions (FRs) of antibodies follows the Kabat definition as set forth in Kabat et al. Sequences of Proteins of Immunological Interest, 5th Ed. Public Health Service, National Institutes of Health, Bethesda, MD. (1991) unless otherwise stated. Using this numbering system, the actual linear amino acid sequence may contain fewer or additional amino acids corresponding to a shortening of the FR or CDR of the variable domain or an insertion into the FR or CDR. For example, the heavy chain variable domain may contain an insertion of one amino acid after residue 52 of H2 (residue 52a according to Kabat) and an inserted residue after heavy chain FR residue 82 (e.g., residues 82a, 82b and 82c, etc. according to Kabat). The Kabat numbering of residues can be determined for an individual antibody by aligning the regions of homology of the antibody sequence with the "standard" Kabat numbered sequences. Maximal alignment of framework residues often requires the insertion of "spacer" residues in this numbering system used for the Fv region. Furthermore, the identity of each particular residue at any particular Kabat site number may vary between antibody chains due to species or allelic differences.

[0053] As used herein, the term "antibody", also referred to as an immunoglobulin, encompasses monoclonal antibodies (including full-length monoclonal antibodies), polyclonal antibodies, multispecific antibodies formed from at least two different epitope-binding fragments (e.g., multispecific antibodies, e.g., WO 2009 / 018386, PCT / US2012 / 045229, which are incorporated herein by reference in their entireties), BiSAbs, human antibodies, humanized antibodies, camelized antibodies, single-chain Fvs (scFvs), single-chain antibodies, single domain antibodies, domain antibodies, Fab fragments, F(ab')2 fragments, antibody fragments (e.g., antigen-binding portions) exhibiting a desired biological activity, disulfide-linked Fvs (dsFvs), and anti-idiotypic (anti-Id) antibodies (such as, for example, anti-Id antibodies against an antibody of the invention), intrabodies, as well as epitope-binding fragments of any of the above. In particular, antibodies include immunoglobulin molecules and immunologically active fragments of immunoglobulin molecules, i.e., molecules that contain at least one antigen-binding site. Antibodies also include peptide fusions with antibodies or portions thereof, such as fusion proteins with Fc domains. Immunoglobulin molecules may be of any isotype (e.g., IgG, IgE, IgM, IgD, IgA, and IgY), subisotype (e.g., IgG1, IgG2, IgG3, IgG4, IgA1, and IgA2), or allotype (e.g., Gm, e.g., G1m(f, z, a, or x), G2m(n), G3m(g, b, or c), Am, Em, and Km(1, 2, or 3)). Antibodies may be derived from any mammal, including but not limited to, human, monkey, pig, horse, rabbit, dog, cat, mouse, etc., or other animals, such as birds (e.g., chicken).

[0054] CTLA-4 Cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) is expressed on activated T cells and acts as a co-inhibitor that blocks T-cell responses after CD28-mediated T-cell activation. CTLA-4 is thought to be part of a central inhibitory pathway that regulates the magnitude of early activation of naive and memory T cells after TCR engagement as well as affecting both antitumor immunity and autoimmunity. CTLA-4 is expressed exclusively on T cells, and expression of its ligands CD80 (B7.1) and CD86 (B7.2) is primarily restricted to antigen-presenting cells, T cells, and other immune-mediating cells. Antagonist anti-CTLA-4 antibodies that block the CTLA-4 signaling pathway have been reported to enhance T-cell activation. One such antibody, ipilimumab, was approved by the FDA in 2011 for the treatment of metastatic melanoma. The use of anti-CTLA-4 antibodies to treat infectious diseases and tumors and to upmodulate adaptive immune responses has been proposed (see U.S. Pat. Nos. 6,682,736; 7,109,003; 7,132,281; 7,411,057; 7,824,679; 8,143,379; 7,807,797; 8,491,895; 8,883,984; and U.S. Patent Publication No. 20150104409, which are incorporated by reference in their entireties).

[0055] PD-L1 Programmed death-ligand 1 (PD-L1) is also part of a complex system of receptors and ligands involved in the control of T-cell activation. In normal tissues, PD-L1 is expressed on T cells, B cells, dendritic cells, macrophages, mesenchymal stem cells, bone marrow-derived mast cells, and various non-hematopoietic cells. Its normal function is to regulate the balance between T-cell activation and tolerance through the interaction with its two receptors: programmed death 1 (also known as PD-1 or CD279) and CD80 (also known as B7-1 or B7.1). PD-L1 is also expressed by tumors and acts at many sites to promote tumor evasion from detection and elimination by the host immune system. PD-L1 is frequently expressed in a variety of cancers. In some cancers, PD-L1 expression is associated with poor survival and poor prognosis. Antibodies that block the interaction of PD-L1 with its receptor can reduce the PD-L1-dependent immunosuppressive effect and enhance the cytotoxic activity of antitumor T cells in vitro. Durvalumab is a human monoclonal antibody against human PD-L1 that can block the binding of PD-L1 to both the PD-1 and CD80 receptors. The use of anti-PD-L1 antibodies to treat infections and tumors and to increase adaptive immune responses has been reported (see U.S. Patent Nos. 8,779,108 and 9,493,565, which are incorporated herein by reference in their entireties).

[0056] PD-1 Programmed Death-1 ("PD-1") is an approximately 31 kD type I membrane protein member of the extended CD28 / CTLA-4 family of T cell regulators (see Ishida, Y. et al. (1992) Induced Expression Of PD-1, A Novel Member Of The Immunoglobulin Gene Superfamily, Upon Programmed Cell Death," EMBO J. 11:3887-3895.

[0057] PD-1 is expressed on activated T cells, B cells, and monocytes (Agata, Y. et al. (1996) “Expression of the PD-1 Antigen on the Surface of Stimulated Mouse T and B Lymphocytes,” Int. Immunol. 8(5):765-772; Martin-Orozco, N. et al. (2007) “Inhibitory Costimulation and Anti-Tumor Immunity,” Semin. Cancer Biol. 17(4):288-298). PD-1 is a receptor responsible for downregulation of the immune system after activation by binding PDL-1 or PDL-2 (Martin-Orozco,N.et al.(2007)“Inhibitory Costimulation and Anti-Tumor Immunity,”Semin.Cancer Biol.17(4):288-298) and acts as a cell death inducer (Ishida,Y.et al.(1992)“Induced Expression of PD-1,A Novel Member of The Immunoglobulin Gene Superfamily,Upon Programmed Cell Death,”EMBO J.11:3887-3895;Subudhi,SKet al.(2005)“The Balance of Immune Responses:Costimulation Verse Coinhibition,”J.Molec.Med.83:193-202). This process is exploited by many tumors through overexpression of PD-L1, leading to suppression of immune responses.

[0058] PD-1 is a well-validated target for immune-mediated therapy in oncology, with promising results from clinical trials for the treatment of melanoma and non-small cell lung cancer (NSCLC), among others. Competitive inhibition of PD-1 / PD-L-1 interaction enhances T cell activation and enhances the recognition and elimination of tumor cells by the host immune system. The use of anti-PD-1 antibodies to treat infections and tumors and to increase adaptive immune responses has been proposed (see U.S. Patent Nos. 7,521,051; 7,563,869; and 7,595,048).

[0059] OX40 OX40 (CD134; TNFRSF4) is primarily involved in the regulation of activated CD4 + and CD8 + It is a tumor necrosis factor receptor found on T cells, regulatory T (Treg) cells and natural killer (NK) cells (Croft et al., 2009, Immunol Rev. 229:173-91). OX40 has one known endogenous ligand, OX40 ligand (OX40L; CD152; TNFSF4), which exists in a trimeric form and can cluster OX40, thereby resulting in potent cell signaling events in T cells. Ibid. Activated CD4 + and CD8 + Signaling by OX40 on T cells leads to increased cytokine production, granzyme and perforin release, and expansion of effector and memory T cell pools (Jensen et al., 2010, Semin Oncol. 37:524-32). In addition, OX40 signaling on Treg cells inhibits Treg expansion, abolishes Treg induction, and blocks Treg suppressive function (Voo et al., 2013, J Immunol. 191:3641-50; Vu et al., 2007, Blood. 110:2501-10).

[0060] Immunohistochemical studies and early flow cytometric analysis demonstrated that OX40 is expressed by T cells infiltrating a variety of human cancers (Baruah et al., 2011, Immunobiology 217:668-675; Curti et al, 2013, Cancer Res. 73:7189-98; Ladanyi et al, 2004, Clin Cancer Res. 10:521-30; Petty et al, 2002, Am J Surg. 183:512-8; Ramstad et al, 2000, Am J Surg. 179:400-6; Sarff et al, 2008, Am J Surg. 195:621-5; discussion 625; Vetto et al, 1997, Am J Surg. 174:258-65). Without intending to be bound by any particular theory, OX40 expression in tumor-infiltrating lymphocytes correlates with longer survival in several human cancers, suggesting that OX40 signaling may play a role in establishing anti-tumor immune responses (Ladanyi et al., 2004, Clin Cancer Res. 10:521-30; Petty et al., 2002, Am J Surg. 183:512-8).

[0061] Agonists of OX40, including antibodies and OX40-ligand fusion proteins, have been used successfully in a variety of preclinical mouse tumor models with encouraging results (Kjaergaard et al., 2000, Cancer Res. 60:5514-21; Ndhlovu et al., 2001, J Immunol. 167:2991-9; Weinberg et al., 2000, J Immunol. 164:2160-9). Co-stimulation of T cells with OX40 promoted persistent antitumor activity in some cases, resulting in long-lasting protection against subsequent tumor challenge (Weinberg et al., 2000, J Immunol. 164:2160-9). Treg cell inhibition and co-stimulation of effector T cells were found to be necessary for OX40 agonists to inhibit tumor growth (Piconese et al., 2008, J Exp Med. 205:825-39). In order to enhance the antitumor effect of OX40 agonist therapy, many strategies and techniques have been explored by combining with vaccines, chemotherapy, radiation therapy, and immunotherapy (Jensen et al., 2010, Semin Oncol. 37:524-32; Melero et al., 2013, Clin Cancer Res. 19:997-1008). The use of anti-OX40 antibodies has been proposed to treat infections and tumors and upmodulate adaptive immune responses (see US Patent Publication No. 20160137740, the entirety of which is incorporated herein by reference).

[0062] TIM3 The T cell inhibitory receptor Tim-3 (T cell immunoglobulin and mucin-domain containing 3) plays a role in regulating antitumor immunity as it is expressed on IFNγ-producing CD4+ helper 1 (Th1) and CD8+ T cytotoxic 1 (Tc1) T cells. It was initially identified as a T cell inhibitory receptor acting as an immune checkpoint receptor that functions to limit the duration and magnitude of Th1 and Tc1 T cell responses in particular. Further studies identified that the Tim-3 pathway may cooperate with the PD-1 pathway to promote the development of a severe dysfunctional phenotype in CD8+ T cells in cancer. It is also expressed in regulatory T cells (Tregs) in certain cancers. Given the involvement of the TIM3 pathway in major immune cell populations that are immunosuppressed in several cancers, it represents an attractive candidate for cancer immunotherapy. See Anderson, AC, Cancer Immunol Res., (2014) 2:393-398; and Ferris, RL, et al., J Immunol. (2014) 193:1525-1530.

[0063] A. Bispecific Binding Proteins The addition of multiple binding sites to a molecule with specificity for a single binding domain can greatly enhance the capabilities of the molecule (e.g., therapeutic, diagnostic, etc.). For example, bispecific antibodies can bind to two or more regions of the same target biomolecule, providing greater specificity than monospecific polypeptides that bind only one epitope of the target. Alternatively, bispecific antibodies can bind to multiple target biomolecules, such as targets that are present in a complex or where sequestering and / or clustering is desired. In a third scenario, the same bispecific antibody can perform different functions at any one time depending on the localization and / or expression of its target molecule.

[0064] Described herein are novel binding proteins. One such structure of these novel binding proteins is referred to as "DuetMab." DuetMab has the following basic structure: a modified heavy chain, where the CH1 region of the modified heavy chain has a substitution of a native cysteine ​​for a non-cysteine ​​amino acid, and a substitution of a native non-cysteine ​​amino acid for a cysteine ​​amino acid; a corresponding modified light chain, where the CL region of the modified light chain also has a substitution of a native cysteine ​​for a non-cysteine ​​amino acid, and a substitution of a native non-cysteine ​​amino acid for a cysteine ​​amino acid; a second Fc region with a second heavy chain; and an Fc region with a corresponding second modified light chain, where the modified heavy chain is directly linked to the corresponding modified light chain, and on a separate target binding arm, the second heavy chain is directly connected to a second corresponding light chain, and the substituted cysteine ​​of the modified heavy chain, obtained by the substitution of the native non-cysteine ​​amino acid for a cysteine ​​amino acid, and the substituted cysteine ​​of the corresponding modified light chain, obtained by the substitution of the native non-cysteine ​​amino acid for a cysteine ​​amino acid, can form a disulfide bond. Disclosure regarding DuetMab can be found, for example, in US Pat. No. 9,527,927, which is incorporated herein by reference in its entirety.

[0065] Another exemplary structure of these novel binding proteins is referred to as "BiSAb" or "BiSAbs". A schematic diagram of an exemplary BiSAb, as well as specific examples of specific BiSAbs, is provided herein. More generally, a BiSAb is a polypeptide containing two binding units, each of which binds to an epitope (e.g., binding unit 1 binds to a first epitope and binding unit 2 binds to a second epitope). A basic BiSAb is bivalent for binding to each of the two epitopes (e.g., a polypeptide includes two binding units 1 ("BD1" or "BU1") and two binding units 2 ("BD2" or "BU2"). Thus, when binding units 1 and 2 bind to different epitopes, a BiSAb has the multispecificity of a conventional bispecific antibody and the bivalency of a conventional antibody molecule. In embodiments where binding units 1 and 2 bind to the same epitope, a BiSAb has the monospecificity of a conventional antibody but is tetravalent. In addition to the binding unit, BiSAb also comprises a linker polypeptide and an Fc portion. The present disclosure relates to a broad set of bispecific binding proteins, such as BiSAbs and proteins comprising a BiSAb core, which target molecules that modulate immune responses. In general, the novel binding protein platform and exemplary bispecific binding proteins (BiSAbs) described herein comprise a binding unit / domain, a linker polypeptide and an Fc portion. The present disclosure also provides nucleic acid molecules encoding such BiSAbs, as well as vectors and host cells that contain such nucleic acids and can be used in methods of producing and using such BiSAbs. BiSAbs, binding proteins comprising a BiSAb core, and various portions of BiSAbs are described in further detail herein.

[0066] In some embodiments, a BiSAb may comprise two heavy-light chain pairs derived from a specific binding protein (i.e., an antibody), where the heavy and light chains each comprise a variable region (e.g., VL and VH) that together form a first binding unit, and the heavy chains each further comprise a second binding unit (e.g., an scFv domain linked to an Fc or Fab). If the first and second binding units bind to different epitopes, each heavy-light chain pair is bispecific, and the two pairs are both bivalent for each of the epitopes. If the first and second binding units bind to the same epitope, each heavy-light chain pair is monospecific, and the two pairs are both tetravalent for that epitope. In some embodiments, the two heavy-light chain pairs are identical. In some embodiments, the two heavy-light chain pairs are not identical.

[0067] In certain embodiments, the domains of BiSAb may be based on known immunoglobulin domains. Immunoglobulin molecules such as monoclonal antibodies (mAbs) are widely used as diagnostic and therapeutic agents, and methods for generating binding fragments of mAbs are known in the art. Monoclonal antibodies, such as any immunoglobulin molecule, are composed of heavy and light peptide subunits, which contain variable and constant domains that confer binding specificity (variable domain) and isotype (constant domain), respectively.

[0068] The BiSAbs disclosed herein have a similar overall structure to conventional antibodies, but are distinguishable by the presence of additional binding units attached at positions within the Fab domain, attached at positions distant from the Fab domain, and attached at the hinge or Fc region (e.g., at the CH2, CH3, or CH4 regions or at the interface of such regions, such as the CH2-CH3 interface). Thus, unlike conventional antibodies that are bivalent for binding to a single epitope, BiSAbs are bivalent for binding to two epitopes. However, as described herein, BiSAbs can still maintain many of the desirable properties of conventional antibodies, such as the ability to bind C1q and FcRn, and the ability to bind Fcγ receptors (e.g., exhibiting the ability to mediate antibody and complement-dependent cytotoxicity).

[0069] The binding domains described herein may include antigen-binding fragments that contain only portions of the mAb molecule, such as Fab, F(ab')2, Fab', scFv, di-scFv, sdAb fragments, since these fragments have been found to be useful as diagnostic or therapeutic agents. In addition, certain residues in the variable domains can also be modified to improve the binding specificity and / or stability of antibodies and antibody fragments. Other residues not directly involved in antigen binding can be substituted to "humanize" regions of non-human antibodies and reduce the immunogenicity of the mAb.

[0070] BiSAbs are different from conventional antibodies, e.g., bivalent for binding to two different epitopes (or tetravalent for binding to one epitope), but many moieties of BiSAbs are derived from or similar to conventional antibody moieties. Any mAb domain and / or fragment known in the art may be used in the BiSAbs described herein. In particular, BiSAbs may include Fab and / or scFv fragments or variants thereof. Exemplary, non-limiting variants of scFv include, but are not limited to, tandem di-scFv, tandem tri-scFv, diabodies and tribodies or triabodies.

[0071] The present disclosure generally relates to novel binding proteins, of which BiSAb is illustrative. Another example is a binding protein that includes a BiSAb core as well as one or more additional binding units and / or a binding protein that includes an extended BiSAb core. Whenever a BiSAb or a BiSAb feature is described herein, it should be understood that such description applies generally to the novel binding proteins of the present disclosure, regardless of whether such binding proteins include two binding units or three or more binding units. Thus, the term BiSAb exemplifies the binding proteins described herein, and any such reference to BiSAb may be used to describe a binding protein that includes a BiSAb core, where the context permits.

[0072] A novel BiSAb structural platform In one aspect, the disclosure provides BiSAb binding proteins having a structural platform generally illustrated by the schematic diagrams in Figures 1A-1F. These diagrams are exemplary, and therefore insertions between additional residues are also within the scope of the disclosed binding proteins. Figures 1A-1C depict the Fc region of an antibody at the CH2-CH3 interface of IgG1 modeled using PyMOL, showing several exemplary BiSAbs of the disclosure. Three surface-exposed loops were identified at or near the CH2-CH3 interface, which could potentially tolerate the insertion of a second binding moiety (e.g., scFv) without compromising the structural integrity or stability of the IgG or the second binding moiety. Figure 1A is a schematic diagram of one such representative loop, ISRTP (SEQ ID NO: 39), identified in the CH2 region near the CH2-CH3 interface. Figure 1D also shows a representative construct IS-scFv-RTP, where the scFv is inserted between the S and R of the ISRTP loop. Figure 1B is a schematic diagram of a representative loop AKGQP (SEQ ID NO: 40) identified at the CH2-CH3 interface. Figure 1E shows a representative construct AK-scFv-GQP, where the scFv is inserted between K and G of the AKGQP loop. Figure 1C is a schematic diagram of a representative loop SNG, identified in the CH3 region downstream of the CH2-CH3 interface. Figure 1F also shows a representative construct S-scfv-NG, where the scFv is inserted between S and N of the SNG loop. The examples described herein provide a description of a construct oriented like SN-scFv-G, but in this case the scFv is inserted between N and G of the SNG loop.

[0073] Thus, one aspect of the disclosure relates to a BiSAb comprising two identical heavy-light chain pairs, where each heavy-light chain pair is bispecific and the two identical pairs are both bivalent for each epitope. Each heavy-light chain pair comprises a binding domain (BD) (binding unit 1), which may comprise a Fab domain that binds a first epitope, a second binding domain (BD2) (or binding unit 2, which may be, for example, an scFv) that binds a second epitope, and an Fc region. In some embodiments, the second binding domain may be associated with a Fab domain. In some embodiments, the Fc region of the BiSAb may be associated with a second binding domain (BD2) (binding unit 2; depicted as an scFv in Figures 1D-1F) that binds a second epitope.

[0074] In some embodiments, the disclosure provides BiSAbs with a common platform structure comprising two chimeric heavy chains, each of which comprises a heavy chain variable region (VH1), a heavy chain constant region (CH1), a hinge or polypeptide linker region, an Fc region comprising a CH2 domain and a CH3 domain, where a second binding domain (BD2), optionally flanked on one or both sides by a polypeptide linker (L1 and / or L2), is linked to a solvent-exposed loop in the Fc region in the order of (i) the CH2 region, (ii) the CH2 and CH3 regions, or (iii) the CH3 region. BiSAbs of this aspect of the disclosure also comprise two conventional antibody light chains, each of which comprises a light chain variable region (VL1) and a light chain constant region (CL), forming part of the first binding domain (BD1). The binding domain (BD2) of the particular BiSAb shown in Figures 1D-1F is an scFv.

[0075] 1D-1F provide a useful schematic diagram of a BiSAb, sometimes referred to herein as the BiSAb "core." As shown in FIG. 1D, the polypeptide chain comprises a heavy chain having a VH1 domain, a CH1 domain, a hinge / linker, a partial N-terminal CH2 domain, an optional linker (referred to herein as L1 or the first polypeptide linker), binding unit 2 (such as VL2 and VH2 of an scFv), another optional linker (e.g., L2 or the second polypeptide linker), the remaining C-terminal CH2 domain, and a CH3 domain. This heavy chain is referred to herein as a chimeric heavy chain, since it may comprise a BD2 with an alternative binding protein and / or a conventional light chain region. A BiSAb comprises two such chimeric heavy chains, which may be the same or different. Note that the variable heavy chain domain (VH) of binding unit 1 is VH1. In one embodiment, this is the variable heavy chain of a Fab that binds to the first epitope. Similarly, the variable light domain (VL) of binding unit 1 is designated VL1. In some embodiments, this is the variable light chain of a Fab that binds to a first epitope, while in embodiments where binding unit 2 is an scFv that binds to a second epitope, the domains of binding unit 2 are designated with the number "2", such as VH2 and VL2.

[0076] Similarly, the polypeptide chain comprises a heavy chain having a VH1 domain, a CH1 domain, a hinge / linker, a CH2 domain, an optional linker (referred to herein as L1 or the first polypeptide linker), binding unit 2 (such as VL2 and VH2 of an scFv), another optional linker (e.g., L2 or the second polypeptide linker), and a CH3 domain, as shown in Figure IE. In this embodiment, the BiSAb comprises a second binding domain, shown as an scFv bound to Fc with a sequence at the interface of the CH2 and CH3 regions.

[0077] The polypeptide chain comprises a heavy chain having a VH1 domain, a CH1 domain, a hinge / linker, a CH2 domain, a partial CH3 domain, an optional linker (referred to herein as L1 or a first polypeptide linker), binding unit 2 (such as VL2 and VH2 of an scFv), another optional linker (e.g., L2 or a second polypeptide linker), and a CH3 domain, as shown in FIG. 1F.

[0078] In these embodiments, BiSAbs typically comprise a typical or modified antibody hinge region in the chimeric heavy chain sequence. Non-limiting examples of amino acid sequences that comprise a hinge region include EPKSCDKTHTCPPCP (SEQ ID NO: 44); EPKSCDKT (SEQ ID NO: 45); EPKSCGKT (SEQ ID NO: 46); EPKSC (SEQ ID NO: 47).

[0079] Having described a general format for aspects relating to the particular structural platform of some of the BiSAb molecules disclosed herein, the various portions and exemplary functional properties of the disclosed BiSAbs are described in further detail below. In other embodiments, the present disclosure contemplates and provides other BiSAb binding proteins that include alternative structural formats and configurations as briefly described herein and in other disclosures, each of which is incorporated herein by reference (see, e.g., U.S. Patent Publication No. 20090155275 and U.S. Patent No. 9,580,509).

[0080] 1. Bonding Unit The BiSAb of the present disclosure comprises at least two binding units or binding domains (binding unit / domain 1 and binding unit / domain 2). In one embodiment, each binding unit binds to a different epitope, whether different epitopes on the same target molecule or on different targets. Since the binding units of a BiSAb each exist in pairs (there are two binding units 1 and two binding units 2), the BiSAb exhibits bivalent binding to each epitope. It will be understood from the teachings of this specification that if each binding unit binds to the same epitope, the BiSAb exhibits tetravalent binding to the epitope.

[0081] In one embodiment, the first binding unit is a Fab fragment, such as a Fab fragment of a conventional monoclonal antibody, or a recombinantly produced antigen-binding fragment that includes a variable light chain (VL1), a constant light chain (CL), a variable heavy chain (VH1) and a constant heavy chain portion (CH1). Optionally, the light and heavy chains of the Fab may be linked to each other via one or more disulfide bonds, such as, for example, a suitable antibody hinge region. The Fab binds to a first epitope.

[0082] In certain embodiments, the Fab is derived from or based on the sequence of a conventional monoclonal antibody, such as a conventional murine, humanized, or human antibody. In certain embodiments, BiSAbs containing a Fab derived from or based on the sequence of a conventional monoclonal antibody retain one or more functional activities of the conventional antibody (e.g., retain at least 80% or more (80%, 85%, 90%, 95%, 97%, 98%, 99% or 100%) of the functional activity). For example, in certain embodiments, BiSAbs containing such Fabs retain one or more of the affinity for antigen, inhibitory activity, immune system modulating activity, activation or induction of an immune response, and / or cell (e.g., cancer cell) killing activity of the conventional antibody.

[0083] In certain aspects, the BiSAbs of the present disclosure comprise a binding unit 2, which comprises a binding domain that binds to a second epitope. The binding unit 2 (or binding domain 2 (BD2)) may be attached to the BiSAb using any suitable strategy. As used herein, BD2 "attached" to the BiSAb (e.g., in some embodiments, in the Fc region, and in other embodiments, in the Fab region) means that these two molecules have interactions between them such that they maintain the orientation for target binding and binding to the Fc or Fab portion of the BiSAb structure. Examples of such interactions include covalent attachment via an amino acid linker, covalent attachment at the CH2, CH3, or CH2 and CH3 interface, or by recombinant expression of BD2 in the Fab region, hinge region, or Fc region in the CH4 region, and non-covalent interactions such as van der Waals forces and hydrogen bond interactions within these same regions. Non-limiting examples of binding domains (or "BDs" or "binding units") within the scope of the present disclosure include antibody variable regions, antibody fragments, scFvs, single chain diabodies, or other binding domains known in the art. Binding domains also include bispecific single chain diabodies, or single chain diabodies designed to bind two different epitopes. In one aspect, epitope binding domains useful for constructing the multispecific epitope binding domains of the present disclosure are exemplified in US Patent Publication Nos. 20100298541 and 20130079280, which are incorporated herein by reference for all purposes.

[0084] In some embodiments, a BiSAb may comprise a binding domain comprising an scFv. Thus, in some embodiments, binding unit 2 comprises an scFv. It will be understood that an scFv encompasses a polypeptide chain comprising a variable heavy domain (VH) linked to a variable light domain (VL) via a flexible polypeptide linker. Figures 1D-1F show schematic diagrams of an exemplary BiSAb in which BD (herein designated as binding unit 2) is an scFv having domains as described herein, which may be designated VL2 and VH2. In some embodiments, the polypeptide linker between VH2 and VL2 comprises a protease cleavage site. The VH and VL domains of the scFv may be derived from the same antibody or from different antibodies. In some embodiments, the VH or VL of the scFv may comprise one or more CDRs that bind to a target of interest, and the remainder of the VH or VL domain is derived from a different antibody or is a synthetic domain. In some embodiments, the scFv comprises at least one CDR of an antibody, e.g., an antibody known in the art that binds to a target of interest. In some embodiments, an scFv comprises at least two CDRs of an antibody. In some embodiments, an scFv comprises at least three CDRs of an antibody. In some embodiments, an scFv comprises at least four CDRs of an antibody. In some embodiments, an scFv comprises at least five CDRs of an antibody. In some embodiments, an scFv comprises at least six CDRs of an antibody.

[0085] In some aspects, the BD may comprise a ligand binding domain of a receptor or a receptor binding domain of a ligand. In some aspects, the BD comprises a sequence having binding affinity for one or more epitopes on a target selected from the group consisting of CTLA-4, PD-1, PD-L1, OX40, and TIM3, as described above. In some embodiments, the binding domain exhibits specific binding activity for a target selected from the group consisting of CTLA-4, PD-1, PD-L1, OX40, and TIM3. The BiSAbs disclosed herein may comprise any combination of binding domains having binding affinity or specific binding activity for the molecular targets disclosed herein. For example, the BiSAbs disclosed herein may comprise a combination of binding domains that allow bispecific binding to targets including CTLA-4 and PD-1; CTLA-4 and PD-L1; CTLA-4 and TIM3; PD-1 and PD-L1; PD-L1 and OX40; PD-1 and TIM3; PD-L1 and TIM3; and TIM3. BiSAbs that contain binding domains that bind to specific target combinations are exemplified in the Examples, including non-limiting combinations of PD-1 / CTLA-4; PD-L1 / CTLA-4; PD-1 / OX40; PD-L1 / OX40; and PD-1 / TIM3.

[0086] In some further embodiments, the BiSAb exhibits greater binding activity (e.g., binding affinity and / or binding specificity) for at least one of the target molecules than the binding activity of the parent monospecific binding sequences used to generate the BiSAb. In similar embodiments, the BiSAb may exhibit greater binding activity (e.g., binding affinity and / or binding specificity) for both of the target molecules than the binding activity of both of the parent monospecific binding sequences used to generate the BiSAb. In yet another embodiment, the BiSAb may exhibit greater binding activity (e.g., binding affinity and / or binding specificity) for both of the target molecules than the binding activity of the combination of the parent monospecific binding sequences used to generate the BiSAb. The enhanced binding properties of the BiSAb relative to the parent monospecific binding sequences, either alone or in combination, provide unexpected advantages over the use of monospecific therapeutics targeting the same molecules, even when used in combination.

[0087] In some embodiments, the disclosure relates to an antibody or antigen-binding fragment thereof that binds to a target selected from the group consisting of CTLA-4, PD-1, PD-L1, OX40, and TIM3. In such embodiments, the antibody or antigen-binding fragment thereof may comprise a heavy chain sequence and a light chain sequence, or a portion of a heavy chain sequence and a light chain sequence comprising the CDR1, CDR2, and CDR3 sequences of the heavy chain and light chain sequences. In other embodiments, the antibody or antigen-binding fragment thereof may comprise a heavy chain variable (HCv) region sequence and a light chain variable (LCv) region sequence, or a portion of a HCv and a LCv comprising the CDR1, CDR2, and CDR3 sequences of the heavy chain and light chain sequences. In yet another embodiment, the antibody or antigen-binding fragment thereof may comprise the CDR1, CDR2, and CDR3 sequences of the heavy chain and light chain sequences. In some embodiments, the antibody may be a chimeric, humanized, or human antibody. In some embodiments, the antibody may be a polyclonal or monoclonal antibody. In another embodiment, the antibody is a monoclonal antibody.

[0088] In some embodiments, all or a domain comprising the antigen-binding region of a "parent" antibody, as described above, can be used to generate a bispecific binding protein (BiSAb) as disclosed herein. The non-limiting embodiments illustrated in the Examples provide a description of how antibody sequences can be identified and combined to generate BiSAbs that exhibit bispecific binding to a combination of molecular targets.

[0089] Several methods may be used alone or in combination to increase the stability of BiSAbs that contain scFv molecules. One method that could be used alone or in combination with one or more other methods described herein is by manipulating the length and / or composition of the linker that links the scFv domains to stabilize the scFv portion.

[0090] Another method that may be used is to introduce at least two amino acid substitutions (also called modifications or mutations) in the VH and / or VL domains of the scFv to promote disulfide bond formation (see, e.g., Brinkmann et al., 1993, PNAS, 90:7538-42; Zhu et al., 1997, Prot. Sci. 6:781-8; Reiter et al., 1994, Biochem. 33:5451-9; Reiter et al., 1996, Nature 14:1239-45; Luo et al., 1995, J. Biochem. 118:825-31; Young et al., 1995, FEBS Let. 377:135-9; Glockshuber et al., 1990, Biochem. 29:1362-7). This method may be used alone or in combination with one or more of the other methods described herein.

[0091] In some embodiments, one or more mutations can be introduced into the VH and VL domains of the scFv to promote the formation of interchain disulfide bonds between the VH and VL domains upon expression of a BiSAb comprising the scFv. In another embodiment, two mutations are introduced into the same domain of a chain. In some embodiments, two mutations are introduced into different chains. In some embodiments, multiple complementary mutations are introduced to promote the formation of multiple disulfide bonds or other stabilizing interactions. In some embodiments, cysteines are introduced to promote the formation of disulfide bonds. Exemplary amino acids that may be mutated to cysteines include amino acids 43, 44, 45, 46, 47, 103, 104, 105 and 106 of VH2, and amino acids 42, 43, 44, 45, 46, 98, 99, 100 and 101 of VL2. The numbering above is based on Kabat numbering, which identifies the position only for VH2 and VL2 of scFv (not for the amino acid positions of the full-length sequence of BiSAb or SEQ ID NO: provided herein). Exemplary combinations of amino acid positions that may be mutated to cysteine ​​residues include VH44-VL100, VH105-VL43, VH105-VL42, VH44-VL101, VH106-VL43, VH104-VL43, VH44-VL99, VH45-VL98, VH46-VL98, VH103-VL43, VH103-VL44 and VH103-VL45. In some embodiments, amino acid 44 of VH and amino acid 100 of VL are mutated to cysteine.

[0092] Another method, which may be used alone or in combination with one or more other methods described herein, is to select the order of the scFv. In some embodiments, the orientation of the VH domain relative to the VL domain is optimized for stability. In some embodiments, the scFv is in a VH-linker-VL orientation. In some embodiments, the scFv is in a VL-linker-VH orientation. In embodiments of the novel BiSAb formats disclosed herein, the orientation of the domains within the scFv can determine how the scFv binds to the Fc portion of the BiSAb. This is described in more detail below with respect to polypeptide linkers. However, briefly, since the BD (e.g., scFv) is interconnected with the CH2, CH3, or CH2 and CH3 interface by optional polypeptide linkers (L1) and (L2), the order of the domains determines which part of the scFv is interconnected with L1 and which part of the scFv is interconnected with L2.

[0093] A further method, which may be used alone or in combination with other methods, is to mutate one or more surface residues of the scFv to introduce one or more stabilizing mutations. In some embodiments, one, two, three, four, five, six or more residues in one or both of the VH and / or VL domains of the scFv are mutated. In some embodiments, only the VH domain of the scFv is altered. In some embodiments, only the VL domain of the scFv is altered. In some embodiments, alterations are made in both the VH and VL domains of the scFv. The same number of alterations may be made in each domain, or a different number of alterations may be made in each domain. In some embodiments, one or more alterations are conservative amino acid substitutions of a residue present in the unmodified parent scFv. In other embodiments, one or more alterations are non-conservative amino acid substitutions of a residue present in the unmodified parent scFv. When multiple substitutions are made in one or both of the VH or VL domains of the scFv, each substitution is independently a conservative or non-conservative substitution. In some embodiments, the substitutions are all conservative substitutions. In some embodiments, the substitutions are all non-conservative. In some embodiments, at least one of the substitutions is conservative. In some embodiments, at least one or the substitutions are non-conservative.

[0094] Yet another method, which may be used by itself or in combination with other methods, is to mutate one or more residues in the VH and / or VL domains of the scFv to introduce one or more amino acid substitutions to match the most frequent residues at said particular positions of the consensus sequences of the VH and / or VL domains of known antibodies. In one embodiment, substitutions are introduced at one, two, three, four, five, six or more positions in one or both of the VH and / or VL domains of the scFv. The same number of changes may be made in each domain, or a different number of changes may be made in each domain. In one embodiment, the one or more changes in the sequence that matches a consensus sequence are conservative amino acid substitutions of the residues present in the unmodified VH and / or VL sequence. In other embodiments, the one or more changes are non-conservative amino acid substitutions of the residues present in the unmodified VH and / or VL sequence. When multiple substitutions are made in one or both of the VH and / or VL domains of the scFv, each substitution is independently a conservative or non-conservative substitution. In some embodiments, the substitutions are all conservative substitutions. In some embodiments, the substitutions are all non-conservative substitutions. In some embodiments, at least one of the substitutions is conservative. In some embodiments, at least one or the substitutions are non-conservative.

[0095] It should be noted that any of the modifications described as useful for modifying or stabilizing the scFv portion can also be applied to modify the Fab portion. For example, the variable domain of the Fab portion of a BiSAb can be modified to improve stability, antigen binding, and the like. In addition, the Fab portion or the scFv portion can also be modified to reduce immunogenicity.

[0096] In one embodiment, binding unit 2 (BD) is an scFv, e.g., an scFv derived from a conventional monoclonal antibody comprising a variable light chain (VL2) and a variable heavy chain (VH2) linked together by a flexible linker, such as a glycine-serine linker. Optionally, the variable light chain and the variable heavy chain of the scFv may be further linked together via one or more disulfide bonds and may include one or more mutations or modifications as described above. The scFv binds to a second epitope. In one embodiment, the second epitope is different from the first epitope bound by binding unit 1. In other embodiments, the second epitope is the same as the first epitope bound by binding unit 1. In one embodiment, the scFv is derived from or based on the sequence of a conventional monoclonal antibody, such as a conventional mouse antibody, a humanized antibody, or a human antibody. In certain embodiments, BiSAbs comprising scFvs derived from or based on the sequence of a conventional monoclonal antibody retain one or more functional activities of the conventional antibody (e.g., retain at least 80% or more (80%, 85%, 90%, 95%, 97%, 98%, 99% or 100%) of the functional activity). For example, in certain embodiments, BiSAbs comprising such scFvs retain one or more of the affinity for antigen, inhibitory activity or cell killing activity of the conventional antibody.

[0097] In certain embodiments, a BiSAb comprises any of binding unit 1 and binding unit 2 described herein, including any combination of binding unit 1 and binding unit 2. For example, in certain embodiments, the disclosure provides polypeptides comprising a Fab that binds to a specific target (e.g., binds to an epitope on a specific target), e.g., a Fab that comprises a specific amino acid sequence or encoded by a specific nucleotide sequence, and / or a scFv that binds to a specific target (e.g., binds to an epitope on a specific target), e.g., an scFv that comprises a specific amino acid sequence or encoded by a specific nucleotide sequence.

[0098] As detailed above, binding unit 1 and binding unit 2 may be covalently linked to the BiSAb via linker polypeptide 1 (L1, L2). Generally, this linkage is achieved by linking the heavy chain C H 2 domain, heavy chain C H 3 domain or heavy chain C H 2 domain and C H The L1 and L2 may be linked to each other via the chimeric heavy chain of the BiSAb, such as at the interface of the three domains, or in some embodiments within the hinge region or Fab domain.L1 and L2 may vary in length and sequence independently of each other, and exemplary structures are described herein.The present disclosure contemplates BiSAbs that include any combination of binding units and linker polypeptides, such as any combination of specific binding units that bind to desired targets and specific L1 and L2 polypeptide linkers described herein.

[0099] 2.Fc area As used herein, "Fc region" encompasses domains derived from the constant region of an immunoglobulin, preferably a human immunoglobulin, including fragments, analogs, variants, mutants, or derivatives of the constant region. Suitable immunoglobulins include IgG1, IgG2, IgG3, IgG4, and other classes, such as IgA, IgD, IgE, and IgM. The Fc region may be a native sequence Fc region or a modified Fc region. The Fc region of an immunoglobulin generally consists of two constant domains, the C and C-terminal constant domains, and the C and D-terminal constant domains. H 2 domain and C H 3 domains, optionally including C H The BiSAbs of this disclosure comprise an Fc domain of the same class as one or both of the L1 or L2 hinge portions.

[0100] a. Modified Fc region Modified Fc regions (also referred to herein as "mutated Fc regions") may be used to alter the effector function and / or half-life of the BiSAbs of the present disclosure. The Fc region may have one or more modifications to alter the functional and / or pharmacokinetic properties of the molecule. Such modifications result in reduced or enhanced IgG C1q binding and complement dependent cytotoxicity (CDC), or FcγR binding and antibody-dependent cellular cytotoxicity (ADCC), or antibody-dependent phagocytosis (ADCP). The present disclosure encompasses BiSAbs with modifications to fine-tune effector function by enhancing or attenuating function or conferring a desired effector function. Thus, in one embodiment of the present disclosure, the BiSAb comprises a mutant Fc region (i.e., an Fc region with modifications as discussed below). BiSAbs with mutant Fc regions are also referred to herein as "Fc-mutated BiSAbs." As used herein, "native" refers to the unmodified parent sequence, and BiSAbs with native Fc regions are referred to herein as "native Fc BiSAbs." Fc mutant BiSAbs can be generated by many methods well known to those skilled in the art. Non-limiting examples include isolating an antibody coding region (e.g., from a hybridoma) and making one or more desired substitutions in the Fc region. Alternatively, the antigen-binding portion of the BiSAb (e.g., variable region) may be subcloned into a vector encoding the mutant Fc region. In one embodiment, the mutant Fc region has a similar level of effector function induction compared to the native Fc region. In another embodiment, the mutant Fc region has a higher induction of effector function compared to the native Fc. In another embodiment, the mutant Fc region has a lower induction of effector function compared to the native Fc. Some specific embodiments of mutant Fc regions are detailed below. Methods for measuring effector function are well known in the art.

[0101] Generally, effector function is modified by alterations to the Fc region, including, but not limited to, amino acid substitutions, amino acid additions, amino acid deletions, and alterations in post-translational modifications of Fc amino acids (e.g., glycosylation). Using the methods described below, the effector function of BiSAbs of the disclosure, the binding ratios (e.g., affinity and specificity) of the Fc region to FcR, may be fine-tuned to obtain BiSAbs with desired properties.

[0102] As used herein, Fc region is understood to include the polypeptides comprising the constant region of an antibody molecule, excluding the first constant region immunoglobulin domain. Thus, Fc refers to the last two constant region immunoglobulin domains of IgA, IgD and IgG, and the last three constant region immunoglobulin domains of IgE and IgM, and optionally all or part of the N-terminus of the flexible hinge of these domains. For IgA and IgM, Fc may include the J chain. For IgG, Fc includes immunoglobulin domains Cgamma2 and Cgamma3 (Cγ2 and Cγ3), and optionally the lower hinge portion between Cgamma1 (Cγ1) and Cgamma2 (Cγ2). Although the boundaries of the Fc region may vary, as used herein, the Fc region of a human IgG heavy chain includes residues A231 to its carboxyl terminus, numbering according to the EU index as described in Kabat. Fc may refer to this region in isolation or in the context of an antibody, antibody fragment, or Fc fusion protein. Polymorphisms have been observed at a number of different Fc positions, including but not limited to, 270, 272, 312, 315, 356, and 358 of IgG1 as numbered by the EU index, and therefore some differences may exist between the present and prior art sequences.

[0103] In one aspect, the disclosure encompasses Fc mutant BiSAbs that have altered binding to Fc ligands (e.g., Fc receptors, C1q) compared to native Fc BiSAbs. Examples of binding properties include binding specificity, equilibrium dissociation constant (K d), dissociation rate and association rate (k off and k on ), binding affinity and / or avidity. d ) is k off / k on It is known in the art that the term "low K" is used to refer to a compound having a low K d BiSAbs containing Fc mutant regions with high K d However, in some cases, it may be preferable to use a BiSAb having k on or k off The value of K d The value of the equilibrium dissociation constant (K) may be more important than the value of the kinetic parameters. One of skill in the art can determine which kinetic parameters are most important for a particular application. For example, modifications that decrease binding to one or more positive regulators (e.g., FcγRIIIA) and / or increase binding to inhibitory Fc receptors (e.g., FcγRIIB) may be suitable for decreasing ADCC activity. Thus, the ratio of binding affinities to various receptors (e.g., the equilibrium dissociation constant (K)) may be more important than the value of the equilibrium dissociation constant (K). d The ratio of ADCC activity of the Fc mutant BiSAbs of the disclosure to C1q may indicate whether the ADCC activity of the Fc mutant BiSAbs of the disclosure is enhanced or reduced. In addition, modifications that reduce binding to C1q are also believed to be suitable for reducing or eliminating CDC activity.

[0104] In one embodiment, the binding affinity of the Fc mutant BiSAb to one or more Fc receptors, including but not limited to, FcRn, FcγRI (CD64) (including isoforms FcγRIA, FcγRIB and FcγRIC); FcγRII (CD32, e.g., isoforms FcγRIIA, FcγRIIB and FcγRIIC); and FcγRIII (CD16, e.g., isoforms FcγRIIIA and FcγRIIIB), is altered compared to a native Fc BiSAb.

[0105] In some embodiments, the Fc variant BiSAb has increased affinity for an Fc ligand, while in other embodiments, the Fc variant BiSAb has decreased affinity for an Fc ligand compared to a native Fc BiSAb.

[0106] In a particular embodiment, the Fc mutant BiSAb has enhanced binding to the Fc receptor FcγRIIIA. In another particular embodiment, the Fc mutant BiSAb has enhanced binding to the Fc receptor FcγRIIB. In a further particular embodiment, the Fc mutant BiSAb has enhanced binding to both the Fc receptors FcγRIIIA and FcγRIIB. In an embodiment, the Fc mutant BiSAb with enhanced binding to FcγRIIIA does not simultaneously enhance binding to the FcγRIIB receptor compared to the native Fc BiSAb. In a particular embodiment, the Fc mutant BiSAb has reduced binding to the Fc receptor FcγRIIIA. In a further particular embodiment, the Fc mutant BiSAb has reduced binding to the Fc receptor FcγRIIB. In another particular embodiment, the Fc mutant BiSAb has enhanced binding to the Fc receptor FcRn. In yet another specific embodiment, the Fc mutant BiSAb with altered affinity for FcγRIIIA and / or FcγRIIB has enhanced binding to the Fc receptor FcRn. In yet another specific embodiment, the Fc mutant BiSAb with altered affinity for FcγRIIIA and / or FcγRIIB has altered binding to C1q compared to the native Fc BiSAb.

[0107] In another embodiment, the Fc mutant BiSAb has increased or decreased affinity for C1q compared to the native Fc BiSAb. In yet another particular embodiment, the Fc mutant BiSAb with altered affinity for C1q has enhanced binding to the Fc receptor FcRn. In yet another particular embodiment, the Fc mutant BiSAb with altered affinity for C1q has altered binding to FcγRIIIA and / or FcγRIIB compared to the native Fc BiSAb.

[0108] It is recognized that antibodies can induce the attack and destruction of target antigens through several processes collectively referred to in the art as antibody effector functions. One of these processes is called "antibody-dependent cellular cytotoxicity" or "ADCC" and refers to a form of cytotoxicity in which secretory Ig binds to Fcγ receptors (FcγR) present on certain cytotoxic cells (e.g., natural killer (NK) cells, neutrophils and macrophages) and allows these cytotoxic effector cells to specifically bind to antigen-bearing target cells and then kill the target cells with cytotoxins. High affinity IgG antibodies specific to the surface of target cells "arm" the cytotoxic cells and are essential for such killing. Lysis of target cells is extracellular and requires direct cell-to-cell contact but does not require complement. Another process encompassed by the term effector function is complement-dependent cytotoxicity (hereinafter referred to as "CDC"), which refers to the biochemical phenomenon of antibody-mediated destruction of target cells by the complement system. The complement system is a complex system of proteins found in normal plasma that combines with antibodies to destroy pathogens and other foreign cells. Yet another process encompassed by the term effector function is antibody-dependent phagocytosis (ADCP), which refers to a cellular reaction in which nonspecific cytotoxic cells expressing one or more effector ligands recognize antibody bound on a target cell and subsequently cause phagocytosis of the target cell.

[0109] It is intended that Fc mutant BiSAbs be characterized by in vitro functional assays for effector cell function via one or more FcγRs. In some embodiments, Fc mutant BiSAbs have similar binding and effector cell functions (such as those described and disclosed herein) in in vivo models as in in vitro assays. However, the present disclosure does not exclude Fc mutant BiSAbs that do not exhibit the desired phenotype in in vitro assays, but do exhibit the desired phenotype in vivo.

[0110] The serum half-life of a protein containing an Fc region can be extended by increasing the binding affinity of the Fc region to FcRn. The term "antibody half-life" as used herein refers to a pharmacokinetic property of an antibody that is a measure of the average survival time of an antibody molecule after administration. Antibody half-life can be expressed as the time required to clear 50 percent of a known amount of immunoglobulin from a patient's body (or other mammal) or from a specific compartment thereof, for example, in serum (i.e., circulating half-life) or in other tissues. Half-life may vary for different immunoglobulins or classes of immunoglobulins. In general, extending antibody (or BiSAb) half-life increases the mean residence time (MRT) of an administered BiSAb in the circulation.

[0111] Prolonging the half-life may allow for less frequent administration as well as less drug doses to be administered to patients. To prolong the serum half-life of BiSAbs, salvage receptor binding epitopes, such as those described in U.S. Pat. No. 5,739,277, may be incorporated into the BiSAb (particularly antibody fragments). As used herein, the term "salvage receptor binding epitope" refers to an epitope in the Fc region of an IgG molecule (e.g., IgG1, IgG2, IgG3, or IgG4) that is involved in the in vivo serum half-life of the IgG molecule. Alternatively, the half-life-prolonged BiSAbs of the present disclosure may be generated by modifying amino acid residues identified to be involved in the interaction between Fc and the FcRn receptor (see, e.g., U.S. Pat. Nos. 6,821,505 and 7,083,784). Additionally, the half-life of the BiSAbs of the present disclosure may be prolonged by conjugation to PEG or albumin using techniques widely available in the art.

[0112] It is contemplated that either the insertion of another binding domain into the Fc region as described herein and / or subsequent binding by antigen can affect Fc activity. For example, bound antigen can increase or decrease binding affinity and activity to FcgR, Clq, and FcRn. This can form an antigen-dependent switch to modulate various antibody-dependent processes. In one embodiment, antigen binding can reduce interaction with FcRn, allowing free BiSAb to interact with FcRn and have normal half-life, but rapid clearance / internalization of BiSAb-Ag complexes. Furthermore, this can allow BD2-antigen mediated interactions to affect the clearance of antigen bound to BD1. In another embodiment, BiSAb can include an Fc region (Fc-BD2) directly inserted into BD2.

[0113] In one embodiment, the disclosure provides a method for the preparation of a mAb having an Fc region comprising administering to a patient a mAb having an Fc region selected from the group consisting of 221, 225, 228, 234, 235, 236, 237, 238, 239, 240, 241, 243, 244, 245, 247, 250, 251, 252, 254, 255, 256, 257, 262, 263, 264, 265, 266, 267, 268, 269, 279, 280, 284, 292, 296, 297, 298, 299, 305, 306, 307, 308, 309, 310, 311, 312, 313, 314, 315, 316, 317, 318, 319, 320, 321, 322, 323, 324, 325, 326, 327, 328, 329, 330, 331, 332, 333, 334, 335, 336, 337, 338, 339, 340, 341, 342, 343, 344, 345, 346, 347, 348, 349, 350, 351, 352, 353, 354, 355, 356, 357, 358, 359, 36 , 308, 313, 316, 318, 320, 322, 325, 326, 327, 328, 329, 330, 331, 332, 333, 334, 339, 341, 343, 370, 373, 378, 392, 416, 419, 421, 428, 433, 434, 435, 436, 440 and 443. Optionally, the Fc region may contain modifications in additional and / or alternative positions known to those of skill in the art (see, e.g., U.S. Pat. Nos. 5,624,821; 6,277,375; 6,737,056; 7,083,784; 7,317,091; 7,217,797; 7,276,585; 7,355,008). Further, useful amino acid positions and specific substitutions are exemplified in Tables 2 and 6-10 of U.S. Pat. No. 6,737,056; the table shown in Figure 41 of U.S. Patent Application Publication No. 2006 / 024298; the tables shown in Figures 5, 12 and 15 of U.S. Patent Application Publication No. 2006 / 235208; the tables shown in Figures 8, 9 and 10 of U.S. Patent Application Publication No. 2006 / 0173170; and the tables shown in Figures 8-10, 13 and 14 of WO 09 / 058492.

[0114] In certain aspects, the disclosure provides a method for the preparation of a polypeptide comprising an Fc region numbered according to the EU index as set forth in Kabat, comprising administering to the patient a polypeptide comprising the polypeptide of the present invention, the polypeptide of the present invention being one of the polypeptides of the present invention, the polypeptide of the present invention being one of the polypeptides of the present invention, the polypeptide of the present invention being one of the polypeptides of the present invention 9T, 239H, 239Y, 240I, 240A, 240T, 240M, 241W, 241L, 241Y, 241E, 241R, 243W , 243L, 243Y, 243R, 243Q, 244H, 245A, 247L, 247V, 247G, 250E, 250Q, 251F, 2 52L, 252Y, 254S, 254T, 255L, 256E, 256F, 256M, 257C, 257M, 257N, 262I, 262 A, 262T, 262E, 263I, 263A, 263T, 263M, 264L, 264I, 264W, 264T, 264R, 264F, 2 64M, 264Y, 264E, 265A, 265G, 265N, 265Q, 265Y, 265F, 265V, 265I, 265L, 265 H, 265T, 266I, 266A, 266T, 266M, 267Q, 267L, 268E, 269H, 269Y, 269F, 269R, 270E, 280A, 284M, 292P, 292L, 296E, 296Q, 296D, 296N, 296S, 296T, 296L, 29 6I, 296H, 296G, 297S, 297D, 297E, 298A, 298H, 298I, 298T, 298F, 299I, 299L, 299A, 299S, 299V, 299H, 299F, 299E, 305I, 308F, 313F, 316D, 318A, 318S, 32 0A, 320S, 322A, 322S, 325Q, 325L, 325I, 325D, 325E, 325A, 325T, 325V, 325H, 326A, 326D, 326E, 326G, 326M, 326V, 327G, 327W, 327N, 327L, 328S, 328M, 32 8D, 328E, 328N, 328Q, 328F, 328I, 328V, 328T, 328H, 328A, 329F, 329H, 329Q,330K, 330G, 330T, 330C, 330L, 330Y, 330V, 330I, 330F, 330R, 330H, 331G, 331A, 331L, 331M, 331F, 331W, 331K, 331Q, 331E, 331 S, 331V, 331I, 331C, 331Y, 331H, 331R, 331N, 331D, 331T, 332D, 332S, 332W, 332F, 332E, 332N, 332Q, 332T, 332H, 332Y, 332A, 3 Fc variants are provided which comprise at least one substitution selected from the group consisting of 33A, 333D, 333G, 333Q, 333S, 333V, 334A, 334E, 334H, 334L, 334M, 334Q, 334V, 334Y, 339T, 370E, 370N, 378D, 392T, 396L, 416G, 419H, 421K, 428L, 428F, 433K, 433L, 434A, 424F, 434W, 434Y, 436H, 440Y and 443W. Optionally, the Fc region may include additional and / or alternative amino acid substitutions known to one of skill in the art, including but not limited to those exemplified in Tables 2 and 6-10 of U.S. Patent No. 6,737,056; the table shown in Figure 41 of U.S. Patent Application Publication No. 2006 / 024298; the tables shown in Figures 5, 12, and 15 of U.S. Patent Application Publication No. 2006 / 235208; the tables shown in Figures 8, 9, and 10 of U.S. Patent Application Publication No. 2006 / 0173170; and the tables shown in Figures 8, 9, and 10 of U.S. Patent Application Publication No. 20090041770, all of which are incorporated herein by reference.

[0115] In certain aspects, the disclosure provides Fc mutated BiSAbs, in which the Fc region comprises at least one modification (e.g., an amino acid substitution, an amino acid insertion, an amino acid deletion) at one or more positions selected from the group consisting of 228, 234, 235 and 331 as numbered by the EU index as set forth in Kabat. In one aspect, the modification is at least one substitution selected from the group consisting of 228P, 234F, 235E, 235F, 235Y and 331S as numbered by the EU index as set forth in Kabat.

[0116] In another specific embodiment, the disclosure provides Fc mutated BiSAbs, wherein the Fc region is an IgG4 Fc region and comprises at least one modification at one or more positions selected from the group consisting of 228 and 235, as numbered by the EU index as set forth in Kabat. In yet another specific embodiment, the Fc region is an IgG4 Fc region and the unnatural amino acid is selected from the group consisting of 228P, 235E and 235Y, as numbered by the EU index as set forth in Kabat.

[0117] In another specific embodiment, the disclosure provides Fc mutant BiSAbs, wherein the Fc region comprises at least one unnatural amino acid at one or more positions selected from the group consisting of 239, 330 and 332, as numbered by the EU index as set forth in Kabat. In one embodiment, the modification is at least one substitution selected from the group consisting of 239D, 330L, 330Y and 332E, as numbered by the EU index as set forth in Kabat. See U.S. Patent No. 7,317,091, which is incorporated herein by reference in its entirety.

[0118] In certain aspects, the disclosure provides Fc mutant BiSAbs, wherein the Fc region comprises at least one unnatural amino acid at one or more positions selected from the group consisting of 252, 254 and 256, as numbered by the EU index as set forth in Kabat. In one aspect, the modification is at least one substitution selected from the group consisting of 252Y, 254T and 256E, as numbered by the EU index as set forth in Kabat. See U.S. Patent No. 7,083,784, which is incorporated herein by reference in its entirety.

[0119] In one embodiment, the disclosure provides an Fc mutated BiSAb, whose Fc region comprises an unnatural amino acid at position 428 as numbered by the EU index as described in Kabat. In one embodiment, the modification at position 428 is selected from the group consisting of 428T, 428L, 428F and 428S as numbered by the EU index as described in Kabat. See U.S. Patent No. 7,670,600, which is incorporated herein by reference in its entirety. In another embodiment, the Fc mutated BiSAb may further comprise an unnatural amino acid at position 434 as numbered by the EU index as described in Kabat. In one embodiment, the modification at 434 is selected from the group consisting of 434A, 434S and 434F as numbered by the EU index as described in Kabat. In another embodiment, the disclosure provides an Fc mutated BiSAb, whose Fc region comprises unnatural amino acids at positions 428 and 434 as numbered by the EU index as described in Kabat. In a particular embodiment, the Fc region comprises 428L, 434S. See U.S. Patent No. 8,088,376.

[0120] In certain embodiments, the effector functions elicited by IgG antibodies are highly dependent on the sugar moieties attached to the Fc region of the protein (Claudia Ferrara et al., 2006, Biotechnology and Bioengineering 93:851-861). Thus, glycosylation modifications can be made to the Fc region to enhance or reduce effector function (e.g., Umana et al, 1999, Nat. Biotechnol 17:176-180; Davies et al., 2001, Biotechnol Bioeng 74:288-294; Shields et al, 2002, J Biol Chem 277:26733-26740; Shinkawa et al., 2003, J Biol Chem 277:26733-26740). 278:3466-3473; U.S. Patent Nos. 6,602,684; 6,946,292; 7,064,191; 7,214,775; 7,393,683; 7,425,446; 7,504,256; POTELLIGENT™ technology (Biowa, Inc. Princeton, NJ); GLYCOMAB™ glycosylation engineering technology (GLYCART biotechnology AG, Zurich, Switzerland). Thus, in one embodiment, the Fc region of a BiSAb of the disclosure comprises altered glycosylation of amino acid residues. In another embodiment, the altered glycosylation of amino acid residues results in decreased effector function. In another embodiment, the altered glycosylation of amino acid residues results in enhanced effector function. In certain embodiments, the Fc region is suppressed from fucosylation, hi other embodiments, the Fc region is hypofucosylated (see, e.g., U.S. Patent Application Publication No. 2005 / 0226867).In one embodiment, such BiSAbs with enhanced effector function, particularly ADCC, are expressed in host cells (e.g., CHO cells, Lemna minor) that have been engineered to produce highly defucosylated polypeptides that have more than 100-fold ADCC compared to polypeptides produced by parental cells (Mori et al., 2004, Biotechnol Bioeng 88:901-908; Cox et al., 2006, Nat Biotechnol., 24:1591-7).

[0121] Addition of sialic acid to oligosaccharides on IgG molecules can enhance their anti-inflammatory activity and alter their cytotoxicity (Keneko et al., Science, 2006, 313:670-673; Scallon et al., Mol. Immuno. 2007 Mar; 44(7):1524-34). The above studies establish that IgG molecules with increased sialylation have anti-inflammatory properties, while IgG molecules with decreased sialylation have enhanced immunostimulatory activity (e.g., enhanced ADCC activity). Thus, BiSAbs may be modified with sialylation profiles appropriate for specific applications (U.S. Patent Application Publication No. 2009 / 0004179 and WO 2007 / 005786).

[0122] In one embodiment, the Fc region of a BiSAb of the disclosure comprises an altered sialylation profile compared to a native Fc region. In one embodiment, the Fc region of a BiSAb of the disclosure comprises an increased sialylation profile compared to a native Fc region. In another embodiment, the Fc region of a BiSAb of the disclosure comprises a decreased sialylation profile compared to a native Fc region.

[0123] In one embodiment, the Fc variants of the disclosure may be combined with other known Fc variants, e.g., Ghetie et al., 1997, Nat Biotech. 15:637-40; Duncan et al, 1988, Nature 332:563-564; Lund et al., 1991, J. Immunol 147:2657-2662; Lund et al, 1992, Mol Immunol 29:53-59; Alegre et al, 1994, Transplantation 57:1537-1543; Hutchins et al., 1995, Proc Natl. Acad Sci USA 92:11980-11984; Jefferis et al, 1995, Immunol Lett. 44:111-117; Lund et al., 1995, Faseb J 9:115-119;Jefferis et al,1996,Immunol Lett 54:101-104;Lund et al,1996,J Immunol 157:4963-4969;Armour et al.,1999,Eur J Immunol 29:2613-2624;Idusogie et al,2000,J Immunol 164:4178-4184;Reddy et al,2000,J Immunol 164:1925-1933;Xu et al.,2000,Cell Immunol 200:16-26;Idusogie et al,2001,J Immunol 166:2571-2575;Shields et al. al.,2001,J Biol Chem 276:6591-6604;Jefferis et al,2002,Immunol Lett 82:57-65;Presta et al., 2002, Biochem Soc Trans 30:487-490); U.S. Patent Nos. 5,624,821; 5,885,573; 5,677,425; 6,165,745; 6,277,375; 5,869,046; 6,121,022; 5,624,821; 5,648,260; 6,528,624; 6,194,5 51; 6,737,056; 7,122,637; 7,183,387; 7,332,581; 7,335,742; 7,371,826; 6,821,505; 6,180,377; 7,317,091; and 7,355,008. Other modifications and / or substitutions and / or additions and / or deletions of the Fc domain will be readily apparent to one of skill in the art.

[0124] It should be noted that the polypeptides depicted in the BiSAb format that contain native Fc, as shown in the examples, retain the ability to bind FcRn and C1q and mediate ADCC. Thus, in certain embodiments, the BiSAb retains the ability to bind FcRn and / or C1q and / or one or more Fcγ receptors (FcγR). For example, in certain embodiments, the BiSAb retains at least 70%, 75%, 80%, 85%, 90%, 95%, 96%, 97%, 98%, 99% or 100% of the ability to bind FcRn and / or C1q and / or one or more FcγRs compared to a conventional antibody that binds to one of the epitopes to which the BiSAb binds. In certain embodiments, the BiSAb is made from the binding domains of one or two conventional antibodies and the activity is compared to one or both of these conventional antibodies.

[0125] Heavy chain heterodimers may also be generated using engineered Fc regions, resulting in BiSAbs that contain two different heavy-light chain pairs. To favor heterodimer formation, the interface between a pair of Fc regions is engineered to maximize the percentage of heterodimers recovered from recombinant cell culture. In one embodiment, the interface comprises at least a portion of the CH3 domain. In this method, a "protrusion" is created by replacing one or more small amino acid side chains at the interface of a first antibody molecule with larger side chains (e.g., tyrosine or tryptophan). By replacing the large amino acid side chain with a smaller amino acid side chain (e.g., alanine or threonine), a "cavity" of identical or similar size to the large side chain is created at the interface of the second antibody molecule at the expense of the large side chain. This provides a mechanism for increasing the yield of heterodimers over other unwanted end-products such as homodimers. Modifications of CH3 include, for example, Y407V / T366S / L368A in one heavy chain and T366W in the other heavy chain; S354C / T366W in one heavy chain and Y349C / Y407V / T366S / L368A in the other heavy chain. Other modifications that provide a protuberance on one chain and a cavity on the other chain are described in U.S. Pat. No. 7,183,076; U.S. Patent Application Publication No. 2014 / 0348839; and Merchant et al., 1998, Nat. Biotech 16:677-681. Some non-limiting examples of modifications that can result in a protuberance-cavity configuration are shown in Table 1a. Other modifications that can be used to form heterodimers include, but are not limited to, modifications that change the charge polarity across the Fc dimer interface such that heterodimerization occurs when electrostatically matched Fc regions are co-expressed. Modifications that change charge polarity include, but are not limited to, those shown in Table 1b below (see also WO20090182127; Gunasekaran et al., 2010, JBC 285:19637-46).Additionally, Davis et al. (2010, Prot. Eng. Design & Selection 23:195-202) describe a heterodimeric Fc platform using a strand-exchanged engineered domain (SEED) CH3 region that is a derivative of the human IgG domain and the IgA CH3 domain (see also WO 2007 / 110205).

[0126] [Table 1]

[0127] [Table 2]

[0128] One of skill in the art will appreciate that in some embodiments, Fc fusion proteins may form dimers due to the homodimerization of molecules that contain an Fc region. In some embodiments, the Fc region of a binding protein (e.g., a BiSAb) may be differentially engineered with mutations to promote and / or maintain heterodimerization (e.g., chimeric mutations, complementary mutations, dock-and-lock mutations, knob-into-hole mutations, strand-exchange engineered domain (SEED) mutations, etc., see, e.g., U.S. Pat. No. 7,183,076; Merchant et al. (1998) Nat. Biotech 16:677-681; Ridgway et al. (1996) Protein Engineering 9:617-621; Davis et al. (2010) Prot. Eng. Design & Selection 23:195-202; WO 2007 / 110205; WO 2007 / 147901; Gunasekaran et al. (2010) JBC 285:19637-46, all of which are incorporated herein by reference.) Thus, binding proteins can be engineered to form heterodimers comprising, for example, a first binding protein, binding domain, or BiSAb fused to a first Fc region or fragment thereof, and a second (i.e., different) binding protein, binding domain, or BiSAb fused to a second Fc region or fragment thereof, where the first and second Fc regions, or fragments thereof, have been engineered to heterodimerize.

[0129] 3. Glycosylation Not only can glycosylation alter the effector function of a polypeptide, but glycosylation modifications of the variable region can also alter the affinity of an antibody (or BiSAb) for a target antigen. In one embodiment, the glycosylation pattern of the variable region of the present BiSAb is modified. For example, an aglycosylated BiSAb can be generated (i.e., the BiSAb lacks glycosylation). Glycosylation can be altered, for example, to increase the affinity of the BiSAb for a target antigen. Such glycosylation can be achieved, for example, by altering one or more glycosylation sites within the BiSAb sequence. For example, one or more variable region framework glycosylation sites can be removed, thereby making one or more amino acid substitutions that eliminate glycosylation at that site. Such aglycosylation can increase the affinity of the BiSAb for the antigen. Such approaches are described in further detail in U.S. Pat. Nos. 5,714,350 and 6,350,861. One or more amino acid substitutions may also be made to eliminate glycosylation sites present in the Fc region (e.g., asparagine 297 in IgG). Additionally, aglycosylated BiSAbs may be produced using bacterial cells that lack the necessary glycosylation machinery.

[0130] 4. Polypeptide Linker Linkers can be used to link the domains / regions of a BiSAb chimeric heavy chain to adjacent molecules. As described herein, BiSAbs can include one, two, or more linker polypeptides (e.g., L1 and L2). In addition, BiSAbs can include other linkers, such as flexible linkers that link the variable heavy and light chains of an scFv to each other. Furthermore, BiSAbs can include other linkers, such as flexible linkers that link the variable heavy and light chains of an scFv to each other, and other linkers that link other binding units to the BiSAb core structure.

[0131] Illustrative, non-limiting examples of linkers include polypeptide chains that include at least four residues. Some of these linkers may be flexible, hydrophilic, and may have little or no secondary structure of their own (linker or flexible linker portion). Linkers of at least four amino acids may be used to interconnect adjacent domains and / or regions after the molecule is assembled. Longer or shorter linkers may also be used. Thus, linkers may be about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 25, 30, 35, 40, 45, or about 50 residues in length. When multiple linkers are used to interconnect portions of the molecule, the linkers may be the same or different (e.g., the same or different length and / or amino acid sequence).

[0132] The linker may be a cleavable linker that includes at least one bond that is selectively cleaved by a cleavage reagent. A cleavable linker may be used to facilitate removal of all or part of the linker sequence. The linker may be engineered to include a protease cleavage site such that cleavage occurs within the linker or at least one end of the linker. For example, a thrombin site may be engineered into each of the two ends of the linker. Depending on the type of linker used, cleavage may also be mediated by drugs such as TCEP, TFA and DTT. The linker may be designed such that the cleavage reagent removes all of the residues from the linker from the cleavage product. Other exemplary and non-limiting linkers include prodrug linkers whose bonds may be selectively cleaved under in vivo conditions, for example, in the presence of endogenous enzymes or other endogenous factors, or simply in aqueous fluids present in the body or within cells in the body. When a BiSAb includes more than one polypeptide linker, the linkers may be different from each other, or at least one of the linkers may be different from the others. In some embodiments, the BiSAb includes a cleavable linker. In certain embodiments, the BiSAb comprises an scFv that comprises a cleavable linker between VH2 and VL2.

[0133] The linker facilitates the formation of the desired structure. The linker is (Gly-Ser). n The linker may contain a cysteine ​​residue, with some Glu or Lys residues dispersed throughout to enhance solubility. Alternatively or additionally, the linker does not contain any serine residues, which may be preferred when the linker is subject to O-linked glycosylation. In some embodiments, the linker may contain cysteine ​​residues, for example, when linker dimerization is used to force the domains of the BiSAb into the properly folded domain structure. In some embodiments, the BiSAb contains at least two polypeptide linkers connecting the polypeptide domains. In other embodiments, the BiSAb contains at least three polypeptide linkers. In other embodiments, the BiSAb contains four or more polypeptide linkers.

[0134] In some embodiments, the polypeptide linker comprises a portion of the Fc portion. For example, in some embodiments, the polypeptide linker may comprise a portion of an immunoglobulin hinge domain of an IgG1, IgG2, IgG3 and / or IgG4 antibody. In some embodiments, the polypeptide linker comprises a portion of a mutated immunoglobulin hinge domain of an IgG1, IgG2, IgG3 and / or IgG4 antibody. In some embodiments, the polypeptide linker comprises at least 5, 7, 8 or 15 amino acid residues of an immunoglobulin hinge region / domain of an IgG1, IgG2, IgG3 and / or IgG4 antibody. In some embodiments, the polypeptide linker comprises at least 5, 7, 8 or 15 amino acid residues of a modified immunoglobulin hinge region / domain of an IgG1, IgG2, IgG3 and / or IgG4 antibody.

[0135] The polypeptide linker may include all or a portion of a hinge region that naturally contains three cysteines. In certain embodiments, the selected hinge region is truncated or otherwise modified or substituted such that only one or two of the cysteine ​​residues remain relative to the intact and / or native hinge region. Similarly, in certain other embodiments, the polypeptide linker may include a mutated or otherwise modified portion of a hinge region in which the number of cysteine ​​residues is reduced by amino acid substitution or deletion, e.g., the mutated or otherwise modified hinge region contains zero, one, or two cysteine ​​residues as described herein.

[0136] Thus, the mutated or otherwise modified hinge domain may be derived from or constructed with a wild-type immunoglobulin hinge domain that contains one or more cysteine ​​residues. In some embodiments, the mutated or otherwise modified portion of the hinge region may contain no cysteine ​​residues or only one cysteine ​​residue, and the mutated or otherwise modified hinge region is or is derived from a wild-type immunoglobulin hinge region that contains one or more cysteine ​​residues or two or more cysteine ​​residues, respectively. In the mutated or otherwise modified portion of the hinge region, the cysteine ​​residues of the wild-type immunoglobulin hinge region are preferably deleted or replaced with amino acids that cannot form disulfide bonds. In some embodiments, the mutated or otherwise modified portion of the hinge region is or is derived from a human IgG wild-type hinge region that may contain any of the four human IgG isotype subclasses IgG1, IgG2, IgG3 or IgG4.

[0137] In some embodiments, the polypeptide linker comprises a portion of the hinge region that includes a cysteine ​​residue that forms a disulfide bond with an immunoglobulin light chain (EU residue 220). In some embodiments, the polypeptide linker comprises a modified portion of the hinge region that includes an amino acid substitution at EU residue C220. In some embodiments, the polypeptide linker includes the amino acid substitution C220V.

[0138] In some embodiments, the polypeptide linker comprises an amino acid substitution that prevents spontaneous hinge-related self-cleavage. In some embodiments, the polypeptide linker comprises an amino acid substitution at EU position D221. In some embodiments, the polypeptide linker comprises the amino acid substitution D221G. In some embodiments, the polypeptide linker lacks amino acid D221.

[0139] As previously mentioned, some embodiments include one or more polypeptide linkers that include or consist of a gly-ser linker. As used herein, the term "gly-ser linker" refers to a peptide that consists of glycine and serine residues. An exemplary gly-ser linker includes an amino acid sequence of the formula (Gly4Ser)n, where n is a positive integer (e.g., 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10). Some preferred non-limiting examples of gly-ser linkers include (Gly4Ser)2 (SEQ ID NO:41) and (Gly4Ser)4, (SEQ ID NO:42), and (Gly4Ser)3 (SEQ ID NO:43). In still other aspects, the polypeptide linker incorporates two or more gly-ser linkers in tandem. In some embodiments, the polypeptide linker comprises at least a portion of a hinge region (e.g., from an IgG1, IgG2, IgG3, or IgG4 molecule) and a series of gly-ser amino acid residues (e.g., a gly-ser linker such as (Gly4Ser)n, where n is 2, 3, or 4).

[0140] In certain embodiments, the linker (e.g., L1 and / or L2 and / or L3, etc.) comprises both a hinge portion and a linker portion, e.g., a linker portion comprising a gly-ser linker. In other embodiments, L1 and / or L2 comprise only a hinge portion, or only a linker portion, e.g., only a gly-ser linker. In other embodiments, L1 and L2 comprise a gly-ser linker portion. In certain embodiments, the gly-ser linkers in the BiSAb are the same length, whereas in other embodiments, the gly-ser linker portions in the BiSAb (e.g., L1 and L2) are different lengths. When the BiSAb comprises an scFv, the heavy and light chains of the scFv may be linked to the BiSAb (e.g., BD1, Fab, Fc, etc.) by a flexible linker. This flexible linker generally does not comprise a hinge portion, but rather is a gly-ser linker or other flexible linker. The length and amino acid sequence of the flexible linker that interconnects the domains of the scFv can be easily selected and optimized (e.g., (Gly4Ser)n, (SEQ ID NO:48), where n is 2, 3, or 4 or more).

[0141] Independent of the polypeptide linker used to interconnect the various binding units and domains (e.g., binding domains / units (e.g., Fab-scFv), or binding domains / units to the Fc (e.g., scFv via L1 and L2), BiSAbs may optionally comprise additional polypeptide linkers. The length and sequence of such additional polypeptide linkers are independently selected. For example, BiSAbs may further comprise a flexible polypeptide linker that interconnects the variable heavy and light chains of the scFv. This flexible polypeptide linker may comprise a gly-ser linker. Generally, this linker does not comprise a hinge portion.

[0142] It is considered here that changing the length of the linkers flanking BD2 can affect the orientation of the BD2 antigen binding site and its spacing relative to the rest of the BiSAb molecule. For example, a short N-terminal linker and a long C-terminal linker can generate an orientation in which the binding site is conformed in one direction, while a long N-terminal linker and a short C-terminal linker can affect the opposite conformational orientation. Thus, the length of the linker can be modulated to orient the BD2 antigen binding site, which can have important effects on creating or avoiding steric interactions between BD1 and BD2 and / or between BD2 and other entities that bind to the antibody molecule in the Fc or other domains.

[0143] 5. Specific Structure of BiSAbs As previously mentioned, one aspect of the disclosure relates to a BiSAb structural arrangement (platform) comprising two heavy-light chain pairs (shown in Figures 1A-1F). In some embodiments of this aspect, the polypeptide sequence of the BiSAb chimeric heavy chain may comprise a polypeptide sequence comprising an antibody heavy chain variable domain (VH1), a polypeptide sequence comprising an antibody heavy chain constant domain 1 (CH1), a portion of an Fc domain, a polypeptide sequence comprising a first polypeptide linker (L1), a polypeptide sequence comprising a binding domain (BD2), a polypeptide sequence comprising a second polypeptide linker (L2), and a polypeptide sequence comprising the remaining Fc domain. In some aspects, the Fc domain comprises a polypeptide sequence comprising a C H 2 domain and C H Thus, one embodiment includes three domains in the following orientation from N-terminus to C-terminus: VH1-C H 1-C H 2(N-terminus)-L1-BD2-L2-C H 2(C-terminus)-C H 3;VH1-C H 1-C H 2-L1-BD2-L2-C H 3; and VH1-C H 1-C H 2-C H 3(N-terminus)-L1-BD2-L2-C HThe BiSAb chimeric heavy chain may comprise a polypeptide sequence at 3 (C-terminus). The polypeptide sequence of the BiSAb light chain may comprise a light chain variable domain (VL1) and a light chain constant domain (CL). Thus, the BiSAb light chain may comprise a polypeptide sequence from N-terminus to C-terminus in the following orientation: VL1-CL. Note that VH1, VL1, and CL are used to denote the portion of "binding unit 1" (BD1) that binds to the first epitope. BD2 is used to denote the portion of "binding unit 2" that binds to the second epitope.

[0144] In embodiments in which the binding domain is an scFv, the BiSAb chimeric heavy chain may comprise a polypeptide sequence comprising an antibody heavy chain variable domain (VH1), a polypeptide sequence comprising an antibody heavy chain constant domain 1 (CH1), a polypeptide sequence comprising a first polypeptide linker (L1), a polypeptide sequence comprising an antibody light chain variable domain (VL2), a polypeptide sequence comprising a flexible linker, a polypeptide sequence comprising an antibody heavy chain variable domain (VH2), a polypeptide sequence comprising a second polypeptide linker (L2), and a polypeptide sequence comprising an antibody Fc domain. For this reason, the chimeric heavy chain of a BiSAb containing an scFv as BD2 has the following orientation from N-terminus to C-terminus: VH1-CH1-CH2 (N-term)-L1-VL2-L3-VH2-L2-CH2 (C-term)-CH3; VH1-CH1-CH2-L1-VL2-L3-VH2-L2-CH3; VH1-CH1-CH2-CH3 (N-term)-L1-VL2-L3-VH VH1-CH1-CH2(N-terminus)-L1-VH2-L3-VL2-L2-CH2(C-terminus)-CH3; VH1-CH1-CH2-L1-VH2-L3-VL2-L2-CH3; and VH1-CH1-CH2-CH3(N-terminus)-L1-VH2-L3-VL2-L2-CH3(C-terminus).

[0145] A chimeric heavy chain is a polypeptide chain comprising an amino acid sequence (e.g., an amino acid sequence of each polypeptide domain). A chimeric heavy chain is a polypeptide chain comprising an amino acid sequence (e.g., an amino acid sequence of each polypeptide domain). Note that VH1, VL1 and CL are used to denote a portion of binding unit 1, with VH1 and VL1 denoting the portion that binds to a first epitope. VH2 and VL2 are used to denote a portion of binding unit 2 that binds to a second epitope. In certain embodiments, additional scFv binding domains are present at the N-terminus and / or C-terminus of the polypeptides that make up the BiSAb core (wherein the BiSAb further comprises binding units (BD) 3 and / or 4 and / or 5), see. In certain embodiments, two or more scFv binding domains are present within the BiSAb core. The additional scFvs comprise antibody heavy chain variable regions, designated VH3, VH4, VH5, and corresponding antibody light chain variable regions, designated VL3, VL4, VL5, respectively.

[0146] 6. Labels, Conjugates and Moieties In one aspect, drugs and other molecules may be targeted to BiSAbs by site-specific conjugation. For example, BiSAbs may contain cysteine ​​engineered domains (including cysteines introduced into the binding units and / or Fc domains) that provide free thiol groups for conjugation reactions. In some embodiments, BiSAbs may be engineered to contain specific conjugation sites. In some embodiments, the disclosure provides Fc mutant BiSAbs in which the Fc region contains amino acid substitutions at one or more of positions 239, 282, 289, 297, 312, 324, 330, 335, 337, 339, 356, 359, 361, 383, 384, 398, 400, 440, 422, and 442, as numbered by the EU index. In some embodiments, the Fc region comprises substitutions at one or more of the following positions: a) 289 and 440; b) 330 and 440; c) 339 and 440; d) 359 and 440; e) 289 and 359; f) 330 and 359; g) 339 and 359; h) 289 and 339; i) 330 and 339; j) 289 and 330; k) 339 and 442; l) 289, 339 and 442; m) 289, 330 and 339; n) 330, 339 and 442; and o) 289, 330 and 442. In another aspect, the present disclosure provides a BiSAb in which the CH1 domain of the Fab arm comprises a substitution at one or more of positions 131, 132, 134, 135, 136 and 139, as numbered by the EU index. In one aspect, the substitution comprises a substitution of an amino acid selected from cysteine, lysine, tyrosine, histidine, selenocysteine ​​and selenomethionine. In a particular aspect, the substitution is cysteine. Methods for generating stable cysteine ​​engineered antibodies are described in U.S. Patent No. 7,855,275, U.S. Patent Publication No. 20110033378 and U.S. Patent Publication No. 20120213705, the contents of which are incorporated herein in their entirety.

[0147] 7. Exemplary Targets Although the various DuetMab and BiSAb platform aspects and embodiments described herein can be made to bind any desired target or targets, the BiSAbs disclosed herein are preferably targeted to a specific pair of target molecules (e.g., binding unit 1 binds to one of the targets and binding unit 2 binds to the other target). As described above and illustrated in the illustrative examples below, the antibodies, DuetMabs and BiSAbs disclosed herein target molecules that modulate immune responses in recipient subjects or immune cells in culture. In some embodiments, the binding domains exhibit specific binding activity for targets selected from the group consisting of CTLA-4, PD-1, PD-L1, OX40, and TIM3. DuetMabs and BiSAbs can include combinations of different binding domains in various orders and orientations, which have binding affinity for or specifically bind to the targets disclosed herein. For example, the DuetMabs and BiSAbs disclosed herein may comprise combinations of binding domains that allow bispecific binding to targets such as CTLA-4 and PD-1; CTLA-4 and PD-L1; and CTLA-4; CTLA-4 and TIM3; PD-1 and PD-L1; PD-L1 and OX40; PD-1 and TIM3; PD-L1 and TIM3. DuetMabs and BiSAbs that comprise binding domains that bind to specific target combinations are illustrated in the Examples, including the non-limiting combinations of PD-1 / CTLA-4; PD-L1 / CTLA-4; PD-1 / TIM3; and PD-L1 / OX40. In certain embodiments, the BiSAbs have enhanced binding properties compared to the binding properties of the individual monospecific binding protein combinations used to generate the BiSAb.

[0148] In certain embodiments, a DuetMab or BiSAb of the disclosure binds to two different epitopes on the same target (e.g., binding unit 1 binds to a first epitope on a target and binding unit 2 binds to a second epitope on the same target).

[0149] In some embodiments, the DuetMab or BiSAb of the present disclosure is a multimer, allowing the label or therapeutic agent to target a specific cell type or molecular target.For example, one functional domain of DuetMab or BiSAb can bind to a target on the cell surface, while another functional domain of the same DuetMab or BiSAb can bind to a hapten or labeling agent useful for detection.Similarly, one functional domain can bind to a cellular target, while a second functional domain can bind to a toxin.Since both binding reactions are mediated by a single molecule, the toxin can be placed in the vicinity of the cellular target where it exerts its cytotoxic function.

[0150] B. Nucleic Acid Molecules Encoding BiSAbs The present disclosure provides nucleic acid molecules encoding BiSAbs. One aspect of the present disclosure provides nucleic acid molecules encoding any of the BiSAbs of the present disclosure. The nucleic acid molecules can encode any of the BiSAb molecules disclosed herein, as well as the heavy and / or light chains of any of the individual binding domains (e.g., scFvs) disclosed herein. Those skilled in the art will recognize that the nucleotide sequence of such polynucleotide molecules can vary, taking into account nucleic acid codon degeneracy and codon frequency for a particular organism, as is well known in the art.

[0151] C. Vectors and host cells for producing BiSAbs and subsequent purification The present disclosure relates to a method for making BiSAbs. In one embodiment, a recombinant nucleic acid molecule encoding all or part of a BiSAb disclosed herein may be operably linked to one or more regulatory nucleotide sequences in an expression construct. The nucleic acid sequences encoding the light chain and chimeric heavy chain of a BiSAb may be cloned into the same expression vector in any orientation (e.g., light chain before heavy chain or vice versa) or may be cloned into two different vectors. When expression is performed using one vector, the two coding genes may have their own genetic elements (e.g., promoter, RBS, leader, stop, polyA, etc.) or may be cloned in one set of genetic elements but linked to a cistronic element. The regulatory nucleotide sequence will generally be appropriate for the host cell used for expression. A large variety of expression vectors and suitable regulatory sequences suitable for various host cells are known in the art. Typically, the one or more regulatory nucleotide sequences may include, but are not limited to, promoter sequences, leader or signal sequences, ribosome binding sites, transcription initiation and termination sequences, translation initiation and termination sequences, and enhancer or activator sequences. The present disclosure contemplates constitutive or inducible promoters known in the art. The promoter may be a natural promoter or a hybrid promoter that combines elements of two or more promoters. The expression construct may be present on a cell episome, such as a plasmid, or the expression construct may be inserted into a chromosome.

[0152] In some embodiments, the expression vector contains a selectable marker gene that allows for the selection of transformed host cells. Selectable marker genes are well known in the art and will vary depending on the host cell used. In some embodiments, the present disclosure relates to an expression vector that contains a nucleotide sequence encoding a polypeptide and is operably linked to at least one regulatory sequence. Regulatory sequences are known in the art and are selected to induce expression of the encoded polypeptide. Thus, the term regulatory sequence includes promoters, enhancers and other expression control elements. Exemplary, non-limiting regulatory sequences are described in Goeddel; Gene Expression Technology: Methods in Enzymology, Academic Press, San Diego, CA (1990). It should be understood that the design of the expression vector may vary depending on such factors as the choice of the host cell to be transformed and / or the type of protein desired to be expressed. Additionally, the copy number of the particular vector, the ability to control the copy number, and the expression of any other proteins encoded by the vector, such as antibiotic markers, should also be considered.

[0153] The present disclosure further relates to a method for producing the BiSAbs of the present disclosure. For example, a host cell transfected with one or more expression vectors encoding the BiSAb (e.g., a single vector encoding a chimeric heavy and light chain, or two vectors, one encoding a chimeric heavy chain and one encoding a light chain) may be cultured under appropriate conditions to allow expression of the polypeptide. The BiSAb may be secreted and isolated from a mixture of cells and medium containing the polypeptide. Alternatively, the BiSAb may be retained in the cytoplasm or membrane fraction, and the cells may be harvested and lysed to isolate the protein. A cell culture comprises host cells, medium and other by-products. Media suitable for cell culture are well known in the art. To purify the protein, the BiSAb may be isolated from the cell culture medium, the host cells, or both using techniques known in the art, such as ion exchange chromatography, gel filtration chromatography, ultrafiltration, electrophoresis, and immunoaffinity purification. In some embodiments, the BiSAb is produced as a fusion protein that includes a domain that facilitates its purification.

[0154] The cloned gene or a portion thereof may be ligated into a vector suitable for expression in prokaryotic cells, eukaryotic cells (yeast, avian, insect or mammalian) or both to produce a recombinant nucleic acid. Expression vehicles for producing recombinant polypeptides include plasmids and other vectors. For example, suitable vectors include the following types of plasmids for expression in prokaryotic cells, such as E. coli: pBR322-derived plasmids, pEMBL-derived plasmids, pEX-derived plasmids, pBTac-derived plasmids and pUC-derived plasmids. In some embodiments, a mammalian expression vector contains both prokaryotic sequences that facilitate the vector's propagation in bacteria and one or more eukaryotic transcription units that are expressed in eukaryotic cells. Vectors derived from pcDNAI / amp, pcDNAI / neo, pRc / CMV, pSV2gpt, pSV2neo, pSV2-dhfr, pTk2, pRSVneo, pMSG, pSVT7, pko-neo and pHyg are examples of mammalian expression vectors suitable for transfection of eukaryotic cells. Some of these vectors are modified with sequences derived from bacterial plasmids, such as pBR322, to facilitate replication and drug resistance selection in both prokaryotic and eukaryotic cells. Alternatively, derivatives of viruses, such as bovine papilloma virus (BPV-1) or Epstein-Barr virus (pHEBo, pREP and p205), may be used for transient expression of proteins in eukaryotic cells. The various methods used for preparation of plasmids and transformation of host organisms are known in the art. For other suitable expression systems for both prokaryotic and eukaryotic cells, as well as general recombinant procedures, see Molecular Cloning A Laboratory Manual, 2nd Ed., ed. by Sambrook, Fritsch and Maniatis (Cold Spring Harbor Laboratory Press, 1989) Chapters 16 and 17. In some cases, it may be desirable to express recombinant polypeptides through the use of a baculovirus expression system.Examples of such baculovirus expression systems include pVL-derived vectors (such as pVL1392, pVL1393 and pVL941), pAcUW-derived vectors (such as pAcUW1), and pBlueBac-derived vectors (such as pBlueBac III containing β-gal).

[0155] Once produced, the molecules may be purified by any method known in the art for purifying proteins, immunoglobulin molecules, or other multimeric molecules, for example, using chromatography (e.g., ion exchange chromatography, affinity chromatography (particularly by affinity for specific antigens Protein A or Protein G), and size column chromatography), centrifugation, differential solubility, or any other multimeric molecule technique for purifying proteins. Additionally, the molecules disclosed herein may be fused to heterologous polypeptide sequences (e.g., affinity tags) routinely used to facilitate purification.

[0156] Regardless of how the BiSAbs are made and purified, binding assays, such as double ELISA assays, may be performed (before and / or after purification) to confirm the functional binding activity of the BiSAbs. Such binding assays are generally well known in the art.

[0157] D. Pharmaceutical Preparations In one embodiment, the present disclosure provides a pharmaceutical composition. Such a pharmaceutical composition may be a composition comprising a nucleic acid molecule encoding a BiSAb. Also, such a pharmaceutical composition may be a composition comprising a DuetMab, a BiSAb, a combination of DuetMabs, or a combination of BiSAbs and a pharma- ceutical acceptable excipient. In one embodiment, the pharmaceutical composition of the present disclosure is used as a drug.

[0158] E. Use As described herein, DuetMabs and BiSAbs can be used to bind targets associated with cancer and cell proliferation diseases or disorders that may be responsive to immunotherapy, for example, by inhibiting immunosuppressive activity and / or inducing immune responses associated with target molecules. For example, aberrant signaling and / or inhibited immune responses may contribute to undesired cell proliferation and cancer. Thus, the DuetMabs, BiSAbs and antibodies disclosed herein can be used to treat undesired cell proliferation and / or cancer associated with inhibited, reduced or insufficient immune responses targeting cancer. In particular, tumor growth curves of tumors and / or tumor volumes may be reduced by administration of DuetMabs, BiSAbs that induce and / or stimulate immune responses in subjects, such as human patients, with cancer.

[0159] Thus, the present disclosure also relates to various methods that include administering the binding proteins disclosed herein to a subject in need thereof. In one aspect, the present disclosure relates to a method of inducing an immune response in a subject having or at risk of developing cancer, comprising administering to the subject a binding protein disclosed herein. In some embodiments, the method activates an immune response against cancer in the subject. In some embodiments, the method enhances an immune response against cancer in the subject. In some embodiments, the method activates an inhibited immune response pathway in the subject, which activation increases an immune response targeting cancer in the subject. In some embodiments, the method enhances an immune response pathway targeting cancer in the subject.

[0160] In another aspect, the present disclosure relates to a method of treating cancer in a subject in need thereof, comprising administering to the subject a binding protein disclosed herein. In one embodiment, the method of treating cancer comprises stopping or slowing the growth of cancer in the subject. In one embodiment, the method of treating cancer comprises stopping or slowing the metastasis of cancer to other sites in the subject. In one embodiment, the method of treating cancer comprises killing cancer cells in the subject. In one embodiment, the method of treating cancer comprises stopping the proliferation and / or spread of cancer cells in the subject.

[0161] In various embodiments of the foregoing aspects, the method relates to treating a patient for a neoplastic and / or cancerous disease. In several embodiments, the cancer is selected from the group of cancers that are susceptible to an induced immune response in the subject. In some embodiments, the cancer is one or more of ovarian cancer, breast cancer, colon cancer, prostate cancer, cervical cancer, uterine cancer, testicular cancer, bladder cancer, head and neck cancer, melanoma, pancreatic cancer, renal cell carcinoma, or lung cancer. In some embodiments, the cancer is one or more of digestive tract or gastrointestinal cancer (e.g., anal cancer; bile duct cancer; extrahepatic bile duct cancer; appendix cancer; carcinoid tumors, colon cancer; colorectal cancer, including pediatric colorectal cancer; esophageal cancer, including pediatric esophageal cancer; bladder cancer; gastric (stomach) cancer, including pediatric gastric cancer; hepatocellular carcinoma, including adult (primary) hepatocellular carcinoma and pediatric hepatocellular carcinoma (e.g., hepatocellular carcinoma); pancreatic cancer, including pediatric pancreatic cancer; sarcoma, rhabdomyosarcoma; pancreatic islet cell carcinoma; rectal cancer; and small intestine cancer); lung cancer (including For example, selected from small cell lung cancer (NSCLC) and small cell lung cancer (SCLC); head and neck cancer (e.g., lip and oral cavity cancer; oral cancer, including pediatric oral cavity cancer; hypopharyngeal cancer; pharyngeal cancer, including pediatric laryngeal cancer; metastatic squamous cell neck cancer of occult primary; oral cavity cancer; nasal cavity and paranasal sinus cancer; nasopharyngeal cancer, including pediatric nasopharyngeal cancer; oropharyngeal cancer; parathyroid cancer; pharyngeal cancer; salivary gland cancer, including pediatric salivary gland cancer; laryngeal cancer; and thyroid cancer); ovarian cancer, and breast cancer.

[0162] In the foregoing methods, the amount of binding protein administered to a subject is effective to induce an immune response in the subject, increase an immune response, stop or slow the growth of cancer, stop or slow the metastasis of cancer, kill cancer cells, and / or stop or slow the proliferation and / or spread of cancer cells.

[0163] In embodiments of the foregoing methods, the binding protein comprises a Duetmab and a BiSAb as disclosed herein. In some embodiments of the foregoing methods, the binding protein comprises an antibody or an antigen-binding fragment thereof as disclosed herein.

[0164] As used herein, the term "subject" is intended to include humans and non-human animals, particularly mammals. Examples of subjects include human patients with disorders, such as those described herein, such as cancer, or normal patients. "Non-human animals" include all vertebrates, such as mammals, including non-mammals (e.g., chickens, amphibians, reptiles, etc.) and non-human primates, domestic and / or farm animals (e.g., sheep, dogs, cats, cows, pigs, etc.) and rodents (e.g., mice, rats, hamsters, guinea pigs, etc.). In certain embodiments, the subject is a human patient.

[0165] "Treatment" or "treating" refers to both therapeutic treatment and preventive or prophylactic measures. Subjects in need of treatment include those who already have the disorder and those who are prone to have the disorder or those who should be prevented from having the disorder. When used in reference to a disease or subject in need of treatment, the term therefore includes, but is not limited to, stopping or slowing the progression of the disease, remission of the disease, prevention of symptoms, reduction of the severity of the disease and / or symptoms, or shortening the length of the disease compared to untreated subjects. In some embodiments, the method of treatment can alleviate one or more clinical indications of the particular disease to be treated.

[0166] The examples set forth below are provided to illustrate certain aspects and embodiments of the disclosure described above, and should not be construed as limiting the scope of the described or appended claimed subject matter. EXAMPLES

[0167] material and method Immune Response Modulation Assays To assess the potential immune response induced by some of the immunotherapeutic molecules described herein, a cytomegalovirus (CMV) antigen recall assay was used. Reagents for the assay included: - CMV-reactive frozen peripheral blood mononuclear cells (PBMCs); - AIM V® Medium (Life Technologies, cat♯12055-091); - Phosphate Buffered Saline (PBS, Life Technologies, cat♯20012-043); - PepTivator®, CMVpp65 peptide pool, (Miltenyi Biotec, cat♯130-093-438, 50 μg / ml); - Ovalbumin, (Thermo scientific cat#77120, 1mg / ml); - Coatar, 96-well plates non-TC treated (Corning, cat#3788); and - Immunotherapy drug molecule.

[0168] General assay protocol: The day before performing the assay, frozen PBMCs were thawed in warm AIM V medium. Cells were washed twice in Coater round well plates. The concentration of cells was adjusted to 1×10 6 Adjusted to cells / mL.

[0169] Aliquots of cells (100 μL) were dispensed into individual wells, leaving the outer matrix of the plate empty. The cells were allowed to sit overnight.

[0170] The next day, 100 μL of AIMV media containing 2× PepTivator CMV peptide pools (0.1 μg / ml to 0.05 μg / ml final) and 2× immunotherapeutic molecules was added to the wells.

[0171] After 72 hours, 25 μL of supernatant from each well was transferred to a previously blocked and washed MSD plate (anti-human IFNγ). After adding the standard solution, the plate was incubated for 2 hours at room temperature. After the incubation period, the MSD plate was washed three times. After washing, 25 μL of SULFO-TAG detection antibody was added and reacted for 1 hour at room temperature. The plate was washed again and 150 μL of 2×MSD read buffer was added before reading.

[0172] Staphylococcal Enterotoxin A / B (SEA / SEB) Assay Protocol To determine the effect of DuetMabs or BiSAbs on IL-2 immune responses, reagents used in either the SEB or SEA assay protocols include the following: - Leukocyte cone (NHSBT code NC24; from Addenbrookes Hospital); - 50ml Falcon tube (BD352070); - Ficoll-Paque PLUS(GE Healthcare 17-1440-02); - anti-CD3 (clone OKT3; 1mg / ml; eBioscience; cat no: 16-0037-85); - Ammonium chloride solution (Stemcell Technologies 07850); - Staphylococcal enterotoxin A (SEA; Sigma, S-9399) or Staphylococcal enterotoxin B (SEB; Sigma, S-4881) stock solution at 1 mg / ml, stored at -20°C: - Media (all from Life Technologies): RPMI 1640 with glutamax (61870) supplemented with 10% v / v heat-inactivated FCS (90005M) and 100 U / mL penicillin + 100 ug / ml streptomycin (15140-122); - V-bottom plates (Greiner BioOne 651201); - 96-well flat bottom plate (Corning Costar 7107).

[0173] Reagents for the IL-2 DELFIA ELISA include: - FLUONUNC Maxisorp ELISA plates (Nunc 437958); - Europium-labeled streptavidin, SA-EU (Perkin-Elmer 1244-360); - DELFIA® Assay Buffer (Perkin-Elmer, #4002-0010); - DELFIA® Strengthening Solution (Perkin-Elmer 4001-0010); RT before use; - Assay Diluent: DELFIA Wash Buffer (0.05% Tween®-20, 20 mM Tris, 150 mM NaCl; pH 7.2-7.4) supplemented with 0.1% BSA and sterile filtered; - Milk powder (Marvel; Premier Foods): - Sample diluent (PRMI1640 + 10% FCS + 1% penicillin / streptomycin (as above)); - PBS (ThermoFisher 14190235); - PBS-Tween® (0.01% Tween®-20 in PBS); - Human IL-2 ELISA kit (Duoset DY202, R&D Systems); - Biotek plate washer with automatic plate loader (EL406).

[0174] General Assay Protocol PBMCs were isolated from human blood leukocyte cones (NHS Blood and Transplant Service code NC24) using density gradient centrifugation (Ficoll-Paque PLUS; GE Healthcare), followed by lysis of red blood cells in ammonium chloride solution (Stemcell Technologies). Anti-human CD3 (clone OKT3 at 0.5ug / ml in PBS; eBioscience) was coated in flat-bottom 96-well plates (Corning Costar 7107) for 2 hours at 37°C. PBMCs were then added at 0.2 × 10 per well in culture medium (RPMI1640-Glutamax supplemented with 10% v / v heat-inactivated bovine serum and 100U / 100ug / ml streptomycin / penicillin (respectively) (Life Technologies)). 6 Cells were added. PBMCs were further stimulated by addition of Staphylococcal Enterotoxin A or B (SEB; Sigma Aldrich) in the range of 0.0088-0.1 ug / ml, followed by addition of candidate DuetMab or BiSAb to the final test concentration. After 3 days of incubation at 37°C and 5% CO2, supernatants were removed from the cells and IL-2 secretion was determined using a commercial ELISA according to the manufacturer's instructions (R&D Systems Duoset product code DY202). See Figure 90.

[0175] Mixed Leukocyte Reaction (MLR) Assay Protocol (Fresh Blood) Additionally, an MLR cell assay was also used to obtain an in vitro correlate of T cell function in response to the DuetMabs and BiSAbs disclosed herein. Reagents used to perform the MLR assay from fresh blood samples include: - 8mL CPT heparin tube; - AIM-V Medium (serum-free) Gibco #12055-091, no additives; - 50ml conical tube; - 2ml cryopreservation container; - ACK lysis buffer (Gibco #A10492-01); - 96-well tissue culture treated round bottom plates BDfalcon #3077; - PHA (Roche) 1mg / ml (10ug / mL final concentration) as a positive control.

[0176] General Assay Protocol PBMCs were prepared from blood samples introduced into CTP heparin tubes. The tubes were centrifuged at 2700 rpm for 20 minutes at 25°C without brake. The top layer of serum was aspirated. The remaining material was gently pipetted to collect everything above the CPT tube plug and introduced into a 50 ml conical tube. AIM-V medium was added to the cells and the cells were washed (three times 5 minutes at 1500 rpm, 25°C with brake on). Any remaining red blood cells were lysed using red blood cell lysis buffer (e.g., about 3 ml of buffer for about 5 minutes). The remaining cells were washed twice with AIM-V medium (5 minutes at 1500 rpm, 25°C with brake on). If necessary, the pellet was compacted in a single tube and resuspended in AIM-V medium to perform a cell count.

[0177] To perform the MLR assay, cells were plated at 200,000 cells / donor / well in 50ul AIM-V medium per donor (total 400,000 / 100ul) in 96-well plates. Candidate molecules were added at 50ul per well (4x) and diluted in serum-free AIM-V medium. After 72 hours, plates were imaged and 30ul of supernatant was removed for human TH1 / TH2 (MSD) cytokine assay.

[0178] Human TH1 / TH2 MSD10-plex protocol This assay was used to determine the amount of cytokines present in culture supernatants in response to administration of DuetMabs and BiSAbs disclosed herein. To perform this assay, blocking agents were prepared by dissolving 200 mg of blocking agent B in 20 ml of PBS per plate. 150 ul of the dissolved blocking agent was added to each well. The plates were sealed and then shaken at room temperature for 2 hours or at 4°C overnight. The wells were washed three times with PBST buffer. Calibrators were prepared by diluting 10 ul of frozen calibrator blend with 1 ml of diluent, which was further serially diluted 4-fold. To separate wells, 25 ul of calibrator (standard) and 25 ul of sample were added. The wells were incubated at room temperature for 2 hours with shaking. After incubation, the wells were washed three times with PBST.

[0179] Detection antibodies were prepared and diluted to the required concentration and then added to each well. After 2 hours of incubation at room temperature with shaking, the wells were washed (3 times) with PBST. Read buffer was added to each well before reading on the MSD instrument.

[0180] Tumor-specific killing assay protocol Human gp100 209-217A human CD8+ T cell line (JR6C12) with reactivity against the peptide was kindly provided by Dr. Steven Rosenberg (National Cancer Institute, Bethesda, MD). JR6C12 cells were co-cultured with a CFSE (CellTrace CFSE proliferation kit, ThermoFisher)-labeled human melanoma cell line (Mel624) at a 1:1 ratio (20,000 JR6C12 + 20,000 Mel624) in 96-well flat-bottom plates for 18 hours at 37°C. Candidate molecules were added at a concentration of 69 nM at time 0 of the co-culture. After 18 hours, wells were visualized by bright-field microscopy. Supernatants were collected for MSD analysis, and adherent cells were trypsinized and then washed (2x) with PBS before being subjected to vital staining (Zombie UV Fixable Viability kit, Biolegend). Uptake of the vital dye into CFSE-labeled cells was assessed by flow cytometry on an LSRFortessa (BD).

[0181] Example 1. Determination of candidate Fc positions for binding domain attachment Using the open source software PyMOL molecular visualization system, the antibody structure was examined in the CH2 and CH3 regions, as well as at or near the CH2-CH3 interface, to find candidate regions for binding domain binding, such as exposed surface loops. Such regions would be amenable to insertion of a second binding domain (e.g., scFv) without compromising the structural integrity or stability of the IgG or the second binding domain itself. Three regions were found from this analysis (represented as spheres in Figures 1A-1C). Figures 1D, 1E, and 1F depict embodiments of binding domains and show the binding of a second binding domain (with scFv for illustrative purposes) in each of the loops found in Figures 1A, 1B, and 1C, respectively.

[0182] Figure 2A provides a more detailed schematic of the amino acid sequence of one of the representative loops found in the CH2 region near the CH2-CH3 interface, which contains the sequence ISRTP (SEQ ID NO: 39). A binding domain may be inserted into this amino acid sequence to generate any number of representative constructs, such as an inserted scFv domain as illustrated in the Examples (e.g., I-scFv-SRTP, IS-scFv-RTP, ISR-scFv-TP, or ISRT-scFv-P, scFv-ISRTP, and ISRTP-scFv, where "-scFv-" identifies the point in the native loop sequence where a binding domain may be attached. Figure 2B is a similar schematic depicting a loop found at the CH2-CH3 interface, which contains the amino acid sequence AKGQP (SEQ ID NO: 40). The representative constructs described herein contain a binding domain ( For example, scFv domains) including A-scFv-KGQP, AK-scFv-GQP, AKG-scFv-QP, AKGQ-scFv-P, scFv-AKGQ, AKGQ-scFv, where "-scFv-" identifies the point in the native loop sequence where the binding domain may be attached. Figure 2C provides a schematic diagram of a representative loop found downstream of the CH2-CH3 interface in the CH3 region and containing the amino acid sequence SNG. Representative constructs of this loop sequence are discussed with respect to the exemplary embodiments for the other two loop regions above and include scFv-SNG, S-scFv-NG, SN-scFv-G, and SNG-scFv.

[0183] Example 2. Generation and characterization of a series of parent antibodies and bispecific binding proteins containing combinations of binding units A series of monoclonal antibodies were generated and characterized. A series of bispecific binding proteins were generated using combinations of antigen-binding sequences (e.g., CDR, HCv, LCv, HC, LC) derived from these "parent" antibodies, which were found to have bispecific binding activity for the combined target antigens. The bispecific binding proteins were designed to have a specific structural platform motif (i.e., "BiS5") disclosed herein.

[0184] The parent antibody sequences are listed in the table below.

[0185] [Table 3]

[0186] [Table 4]

[0187] [Table 5]

[0188] [Table 6]

[0189] [Table 7]

[0190] [Table 8]

[0191] [Table 9]

[0192] Example 2(a) PD-1 / CTLA-4 Bispecific Binding Proteins Without wishing to be bound by any particular theory, there is strong clinical and preclinical rationale for combining PD-1 and CTLA-4 blockade. Therefore, it would be desirable to maximize the risk / benefit ratio of combining PD-1 with CTLA-4 (Figure 4).

[0193] The following bispecific binding proteins that bind to PD-1 and CTLA-4 were generated using the parental sequences identified above in Table 2. Proteins identified as BiS2, BiS3, and BiS5 were generated using the sequences identified below and assessed for simultaneous antigen binding activity using the Octet binding assay described below.

[0194] [Table 10]

[0195] [Table 11]

[0196] Octet binding assay (BiS2, BiS3, and BiS5) To assess the binding of the bispecific binding molecules disclosed herein, an Octet QK with Ni-NTA biosensor chip and 10x kinetics buffer was used (ForteBio, Menlo Park, CA). For this particular series of bispecific binding proteins, His-tagged PD-L1-Fc, his-tagged PD-1-Fc and CTLA-4-Fc (human recombinant proteins) were purchased from R&D Systems (Minneapolis, MN). All binding assays were performed at 25°C.

[0197] The sample plate was agitated at 1000 rpm before analysis. The Ni-NTA biosensor chip was pre-wetted with 1× kinetic buffer for 5 min. The 1× kinetic buffer also served as the running buffer for baseline determination and the dilution buffer for the antigen and bispecific antibody. The Ni-NTA biosensor chip was immersed in 100 nM his-tagged PD-L1-Fc (see (b) below) or his-tagged PD-1-Fc for approximately 1 min for antigen capture. The antigen-coated biosensor chip was immersed in 10 μg / ml of bispecific antibody each for approximately 5 min and then transferred into a column of wells containing 100 nM CTLA-4-Fc antigen for 2 min. The binding results are shown in Figure 3.

[0198] Bispecific binding proteins in the DuetMab format were generated that bind to PD-1 and CTLA-4 using the parental sequences identified above in Table 2. PD-1 / CTLA-4 DuetMabs were generated using the sequences shown in Table 4 below and evaluated as described below, including comparison to PD-1 / CTLA-4 BiS5.

[0199] [Table 12]

[0200] [Table 13]

[0201] [Table 14]

[0202] [Table 15]

[0203] Octet binding assay (DuetMab) Simultaneous binding studies against two separate antigens were performed by Octet analysis. Biotinylated human PD-1 was loaded onto a streptavidin sensor followed by sequential interactions first with DuetMab PD-1 / CTLA-4 and then with soluble CTLA-4 antigen. Streptavidin (SA) biosensor (ForteBio) was used to capture 5 μg / mL biotinylated human PD-1 in PBS pH 7.2, 3 mg / ml BSA, 0.05% (v / v) Tween® 20 (assay buffer). After a washing step, the loaded biosensor was subjected to sequential association and dissociation interactions first with sample wells carrying 133 nM DuetMab PD-1 / CTLA-4 bispecific antibody and then with wells carrying 200 nM human CTLA-4 antigen. The binding results are shown in Figure 5.

[0204] The intrinsic kinetics of the PD-1 / CTLA-4 DuetMab bispecific antibody was also evaluated by BiaCore. Binding experiments were performed using a BIAcore T200 instrument (BIAcore). Mouse anti-huIgG-Fab was immobilized on a CM5 chip to a target response of 2000 RU to capture the antibody. To achieve approximately 100 response units of captured antibody, 100 nM DuetMab or mAb was flowed at 20 μL / min for 5 min. Antigen was then continuously injected for 5 min at a flow rate of 50 μl / min. Kinetic parameters (k on and k off ) as well as the dissociation constant (KD) were calculated from non-linear fits using BIAevaluation4.1 software. Binding results are shown in Table 5.

[0205] [Table 16]

[0206] Reporter gene assays Results from reporter gene assays show that the PD-1 / CTLA-4 bispecific binding proteins inhibited the PD-1 and CTLA-4 pathways (FIGS. 6A-D). The BiS5 binding protein maintained PD-1 titers compared to the parent, but had a 3-fold reduction in titers compared to anti-CTLA-4 IgG. The DuetMab antibody showed a 9-fold reduction in PD-1 titers and a 16-fold reduction in CTLA-4 (compared to IgG4P). There is a need in the art for molecules that have low CTLA-4 targeting but maintain functional activity (e.g., as shown in the SEB assay). Thus, the PD-1 / CTLA-4 bispecific binding proteins have the potential to provide safety benefits to patients.

[0207] Staphylococcal Enterotoxin B (SEB) Assay Results from the SEB assay revealed that DuetMab and BiS5Ab had comparable activity in the SEB primary immune cell assay (Figure 7), with DuetMab showing higher activity compared to the DummyDuet control arm (Figure 8A). DuetMab showed nearly equivalent activity to the parent molecule combination, as well as higher activity compared to LO115 or the CTLA-4 antibody MEDI1123 (tremelimumab) (Figure 8B). Finally, BiS5 and DuetMab showed higher activity compared to the novel isotype control combination (Figure 9A-B). Data were obtained from 4 donors across two independent experiments, and these specific assays required the use of IFNγ.

[0208] Mixed Leukocyte Reaction (MLR) Assay MLR assays were performed to test the PD-1 / CTLA-4 bispecific molecules. PD-1 / CTLA-4 DuetMab and BiS5Ab had comparable activity in the mixed leukocyte reaction (MLR) assay (Figure 10A-C). PD-1 / CTLA-4 DuetMab had higher activity compared to the DummyDuet / isotype control arm combination (Figure 11A-D). PD-1 / CTLA-4 DuetMab had nearly equivalent activity compared to the parent antibody control combination (Figure 12A-D). Finally, PD-1 / CTLA-4 DuetMab had nearly equivalent activity compared to the competitive PD-1 / CTLA-4 combination and higher activity than anti-PD-1 alone (e.g., pembrolizumab and nivolumab) (Figure 13A-D).

[0209] Pharmacokinetic and pharmacodynamic (PK / PD) studies A study was performed to investigate single-dose pharmacokinetics / pharmacodynamics (PK / PD) in cynomolgus monkeys. The study design is shown in Figure 14. DuetMab showed clear pharmacodynamics (PD) in cynomolgus monkeys, with robust PD responses observed for both molecules (Figure 15A-B), thus confirming the viability of the PD-1 / CTLA-4 bispecific binding protein in an in vivo environment.

[0210] T cell-dependent antibody response (TDAR) Cynomolgus monkeys were administered DuetMab or BiS5 bispecific molecules intravenously (saphenous or cephalic vein) at the indicated doses (0.5, 5, 50 mg / kg). Keyhole limpet hemocyanin (KLH) protein was reconstituted with an appropriate volume of sterile injection solution under sterile conditions. Low-dose KLH solution was administered subcutaneously twice (days 1 and 29) on the back of each animal. Blood samples were collected from all animals for further analysis. Evaluation of KLH-specific IgM and IgG antibody titers was performed. ELISA was used to detect anti-KLH antibodies in monkey serum.

[0211] T cell-dependent antibody responses (TDAR) were observed in cynomolgus monkeys administered PD-1 / CTLA-4 DuetMab (Figure 16A) and PD-1 / CTLA-4 BiS5Ab (Figure 16B).

[0212] CHO cells expressing various levels of human PD-1 and / or CTLA-4 To test the PD-1 / CTLA-4 bispecific molecule, a model system was created using stable CHO cells expressing various levels of human PD-1 and / or CTLA-4 (Figure 17). Fluorescently labeled soluble PD-1 and CTLA-4 proteins were used to detect the free antigen-binding arms on the cell-bound DuetMab by flow cytometry. The results of this assay reveal that the PD-1 / CTLA-4 DuetMab simultaneously binds to PD-1 and CTLA-4 on the surface of the same cells (Figure 18A-C).

[0213] CTLA-4 is constantly internalized into clathrin-coated pits, representing only a small fraction of receptors expressed on the cell surface at any given time. Recycling of cell surface CTLA-4 is rapid, with less than 80% of surface CTLA-4 being internalized within 5 min. Therefore, experiments were performed to determine whether cooperative binding would discriminate between PD-1 / CTLA-4 DuetMab versus anti-PD-1 and anti-CTLA-4 antibody combinations in the presence of CTLA-4 saturation on cells expressing excess PD-1 (Figure 19A-C). Receptor occupancy for each target antigen was determined independently using targeted anti-PD-1 and anti-CTLA-4 mAbs.

[0214] The parental monoclonal antibodies were found to bind and occupy their target receptors without measurable effects on non-target receptors (Figure 20A-D). The PD-1 / CTLA-4 DuetMab saturated CTLA-4 on CHO cells expressing excess PD-1 at approximately 250-fold lower concentrations compared to the monoclonal antibody combination (Figure 21A-D). The PD-1 / CTLA-4 DuetMab saturated CTLA-4 on CHO cells expressing excess PD-1 at approximately 500-fold lower concentrations compared to cells expressing CTLA-4 alone (Figure 22A-F). As determined by quantification of doublet formation within premixed total CHO populations, the PD-1 / CTLA-4 DuetMab preferentially bound in cis to PD-1 and CTLA-4 on the surface of the same cells (Figure 23A-B). However, the PD-1 / CTLA-4 DuetMab can also bind in trans to single expressing cells. The PD-1 / CTLA-4 DuetMab adopted the internalization properties of the parent anti-CTLA-4 antibody, tremelimumab (Figure 24A-D). Without being bound to any particular theory, the action of this molecule may be to induce downregulation of PD-1. The internalization properties of the PD-1 / CTLA-4 DuetMab were also observed in stable CHO cells expressing a 10-fold excess of PD-1 (Figure 25B).

[0215] Example 2(b) PD-L1 / CTLA-4 Bispecific Binding Proteins The parent sequences identified above in Table 2 were used to generate the following bispecific binding proteins that bind to PD-1 and CTLA-4. Proteins identified as BiS2, BiS3, and BiS5 were generated using the sequences shown in Table 6 below, and the sequences identified below were evaluated for simultaneous antigen binding activity using the Octet binding assay as previously described in section 2(a) (Figure 26).

[0216] [Table 17]

[0217] [Table 18]

[0218] Example 2(c) PD-1 / TIM3 Bispecific Binding Proteins The parent sequences identified above in Table 2 were used to generate the following bispecific binding proteins (Table 7) that bind to PD-1 and TIM3. Proteins identified as BiS3, BiS5, and DuetMab were generated using the sequences identified below and evaluated for simultaneous binding studies by Octet analysis. Briefly, biotinylated human TIM3-IgV domain was captured using a streptavidin (SA) biosensor (ForteBio) at 2 μg / ml in PBS pH 7.2, 3 mg / ml BSA, 0.05% (v / v) Tween® 20 (assay buffer). After a washing step, the loaded biosensor was subjected to sequential binding and dissociation interactions first with a sample well carrying 200 nM of bispecific antibody, then with a well carrying 200 nM of PD-1 antigen. Biotinylated human TIM3-IgV domain was loaded onto a streptavidin sensor and then subjected to sequential interactions first with the bispecific molecule and then with the PD-1 antigen. The binding results are shown in Figures 27A-27B.

[0219] [Table 19]

[0220] [Table 20]

[0221] [Table 21]

[0222] Tumor-specific killing activity assay Rosenberg clonal melanoma killing assay The general cell killing activity of the TIM3 / PD-1 bispecific binding molecule and the parental TIM3 antibody was tested using the Rosenberg Clone: ​​JR6C12 and the melanoma cell line: Me1324.

[0223] General Assay Protocol JR6C12 is a human CD8+ T cell line that functions as an effector, expanded from a melanoma patient, and specific for gp100-melanoma antigen. To evaluate therapeutic potential, Mel624 tumor cells were fluorescently labeled and then added together with the effector (JR6C12) and candidate antibodies that bind to TIM3 and / or PD-1. The cells were co-cultured for 16 hours. The panels shown in Figure 28A provide a visual illustration that adding TIM3 62 either in combination with anti-PD1 or as a PD-1 / TIM3 bispecific molecule (as described in Table 7) enhances T cell activation and tumor killing.

[0224] Furthermore, as shown in Figures 28B-28C, the PD-1 / TIM3 bispecific molecule exhibits the greatest tumor killing efficacy compared to anti-TIM3, anti-PD-1 or isotype control monotherapy, as assessed by (b) tumor cell vital dye uptake and (c) IFNγ secretion.

[0225] In addition to Clone 62, the parental sequences identified above in Table 2 were used to generate additional bispecific binding proteins in a DuetMab format that bind PD-1 and TIM3. The PD-1 / TIM3 DuetMab was generated using the sequences shown in Table 8 below. The sequence of the TIM3 arm was taken from O13-1, an affinity matured variant of Clone 62, and the sequence of the anti-PD-1 arm was taken from LO115, which is identical to the PD-1 arm used in the PD-1 / CTLA-4 DuetMab bispecific antibody described above. The PD-1(LO115) / TIM3(O13-1) bispecific antibody was evaluated as discussed below, including comparison to PD-1 / TIM BiS3 and BiS5.

[0226] [Table 22]

[0227] [Table 23]

[0228] [Table 24]

[0229] [Table 25]

[0230] Octet binding assay (DuetMab, TIM3 arm affinity matured mutants) Simultaneous binding studies for two separate antigens, PD-1 and TIM3, were performed by Octet assay. Biotinylated human TIM3 was loaded onto a streptavidin sensor and then sequentially interacted with first the PD-1 / TIM3 DuetMab and then the soluble PD-1 antigen. Biotinylated human TIM3 was captured using a streptavidin (SA) biosensor (ForteBio) at 5 μg / ml in PBS pH 7.2, 3 mg / ml BSA, 0.05% (v / v) Tween® 20 (assay buffer). After a washing step, the loaded biosensor was subjected to sequential association and dissociation interactions first with sample wells carrying DuetMab PD-1 / CTLA-4 bispecific antibody with TIM3 arm (O13-1), an affinity matured variant of the clone 62 TIM antibody, loaded at 200 nM, and then with wells carrying 200 nM human PD-1 antigen. The binding results are shown in FIG.

[0231] The intrinsic kinetics of the PD-1 / TIM3 DuetMab bispecific antibody was also evaluated by BiaCore. Binding experiments were performed using a BIAcore T200 instrument (BIAcore). Mouse anti-huIgG-Fab was immobilized on a CM5 chip to a target response of 2000 RU to capture the antibody. Approximately 100 response units of captured antibody were achieved by flowing 100 nM DuetMab or mAb at 20 μL / min for 5 min. Antigen was then injected sequentially at a flow rate of 50 μL / min for 5 min. Kinetic parameters (k on and k off ) as well as the dissociation constant (KD) were calculated from non-linear fits using BIAevaluation4.1 software. Binding results are shown in Table 9.

[0232] [Table 26]

[0233] PD-1 / TIM3 bispecific antibodies, including BiS3, BiS5, and DuetMab, bind to CHO cells overexpressing human TIM3 or human PD-1 (Figure 30 and Table 25), and PD-1 and TIM3 expression (DMF4) is shown in Figure 31.

[0234] [Table 27]

[0235] CMV Ag recall assay PD-1 / TIM3 bispecific antibodies, including BiS3, BiS5, and DuetMab, increased CD8+ T cell proliferation in a CMV antigen recall assay compared to isotype treatment (Figure 32A-C).

[0236] Mixed Leukocyte Reaction (MLR) Assay PD-1 / TIM3 bispecific antibodies, including BiS3, BiS5, and DuetMab, increased interferon (IFNγ) secretion in a mixed leukocyte reaction (MLR) assay with activity exceeding that of single agent and combination therapy (Figure 33A-D). PD-1 / TIM3 bispecific antibodies, including BiS3, BiS5, and DuetMab, showed activity similar to the parental LO115 IgG1 in a jurkat NFκB reporter line that predominantly expresses PD-1 (87% PD-1 single positive) (Figure 34A-C).

[0237] In summary, three bispecific formats (DuetMab, BiS3, and BiS5) were generated for PD-1 / TIM3. The bispecific formats all demonstrated comparable or superior in vitro functionality to anti-PD-1, suggesting that these molecules may offer superior advantages over existing cancer immunotherapy strategies.

[0238] Example 2(d) OX40 / PD-L1 Bispecific Binding Protein The following bispecific binding proteins were generated that bind to PD-1 and OX40 using the parent sequences identified above in Table 2. Proteins identified as BiS2, BiS3, and BiS5 were generated using the sequences shown in Table 10 below and evaluated for simultaneous binding studies using the Octet binding assay described below.

[0239] [Table 28]

[0240] [Table 29]

[0241] Octet binding assay To assess the binding of the bispecific binding molecules disclosed herein, an Octet QK with a Ni-NTA biosensor chip and 10x kinetics buffer was used (ForteBio, Menlo Park, CA). For this particular series of bispecific binding proteins, His-tagged PD-L1-Fc, his-tagged PD-1-Fc and hOX40-Fc (human recombinant protein) were purchased from R&D Systems (Minneapolis, MN). All binding assays were performed at 25°C.

[0242] The sample plate was agitated at 1000 rpm before analysis. The Ni-NTA biosensor chip was pre-wetted with 1× kinetic buffer for 5 min. The 1× kinetic buffer also served as the running buffer for baseline determination and the dilution buffer for the antigen and bispecific antibody. The Ni-NTA biosensor chip was immersed in 100 nM his-tagged PD-L1-Fc (see (b) below) or his-tagged PD-1-Fc for approximately 1 min for antigen capture. The antigen-coated biosensor chip was immersed in 10 μg / ml of bispecific antibody each for approximately 5 min and then transferred into the column of wells containing 100 nM hOX40-Fc antigen for 2 min. The binding results show that BiS2 / BiS3 OX40Ab / PD-L1 molecules bind to both PD-L1-His and hOX40-Fc, and that BiS2 OX40Ab / PD-L1 binds with higher affinity than BiS3 OX40Ab / PD-L1. BiS2 OX40Ab / PD-1 was used as a control (Figure 35).

[0243] Staphylococcal Enterotoxin B (SEB) Assay SEB assays using the protocol described above demonstrated that the OX40 / PD-L1 bispecific molecule was active in both BiS2 and BiS3 formats (Figure 36A-B).

[0244] PD-L1 reporter assay material: - Cell lines and culture conditions: - Human PD-1 Jurkat NFAT Luciferase Clone 2 Receptor - PD-L1 expressing CHO scFv OKT3(UBC) (all cells were maintained in RPMI1640 medium containing 10% FBS and 1x pen / strep antibody (RPMI complete medium) in a humidified tissue culture incubator at 37℃). - RPMI-1640, LifeTechnologies cat♯A1049101 - Heat-inactivated newborn bovine serum (FBS), LifeTechnologies cat♯26010074 - Complete RPMI medium: RPMI-1640 containing 10% FBS - 100x Penicillin / Streptomycin, LifeTechnologies cat♯15140-122 - 96-well TC-treated flat bottom culture plate, Costar 3903, VWR cat♯29444-010 - SteadyGlo Luciferase Assay System,Promega,cat#E2510 - Test antibody - EnVision Multilabel Plate Reader, Perkin Elmer

[0245] method For the two-cell bioactivity assay for neutralization of PD-1 inhibition, PD-L1 expressing CHO scFv OKT3 cells were trypsinized and neutralized with warm RPMI complete medium, then collected in a 50mL conical tube. Cells were pelleted at 380g for 5 minutes at RT, then suspended in fresh RPMI complete medium and counted on a Vi cell counter. PD-L1 expressing CHO scFv OKT3 cells were adjusted to 0.4e6 / mL and 25μL (10,000 cells) per well were plated as shown in the plate layout. Cells were allowed to attach to the plate for 3 hours. Then, 50μL of RPMI containing test reagent (2x final concentration) was added in aliquots onto the CHO cells and incubated for an additional hour. This incubation allows the test reagent time to bind to PD-L1 on the CHO cell surface. After 1 hour, PD-1 expressing Jurkat NFAT luciferase reporter cells were collected into 50mL conical tubes, pelleted at 380g for 5 minutes at RT, and resuspended in fresh warm RPMI complete medium. Cells were adjusted to 1.2e6 / mL and 25μL (30,000) cells were plated into wells containing PD-L1 expressing CHO scFv OKT3 cells and test article.

[0246] Cells and test reagents were further incubated for 18 hours for PD-1 Jurkat reporter cell activation. SteadyGlo luciferase reagent was then prepared and 100 μL was aliquoted and added to each well. Complete lysis was achieved by gentle shaking (200 rpm orbital shaker) at RT for 15 minutes. After lysis, luciferase activity was measured on an Envision Multilabel Plate Reader using the US96 luminescence protocol. EC values ​​of PD-L1 antagonism were calculated using Graphpad Prism software by plotting luciferase RLU against log [test reagent] and using nonlinear regression analysis, a 4-parameter fit of a sigmoidal dose-response curve. 50 value was determined.

[0247] result: OX40 / PD-L1 BiS2 / 3 was tested against the PD-L1 / PD-1 parent and NIP228 (G4P) controls using a 5-point dose titration starting at 100 nM (PD-L1). Both OX40-PD-L1 BisAbs were found to be active, with stronger agonism than the PD-L1 (4736) parent (Figure 37). The BiS2 and BiS3 formats performed similarly.

[0248] CMV Ag recall assay In a CMV Ag recall assay (using the protocol described above), the BiS2 and BiS3 molecules showed comparable activity compared to the combination (Figure 38).

[0249] All of the above-mentioned binding and immune response assays provide illustrative data that the bispecific binding molecules disclosed herein exhibit specific binding to both target molecules, and in some cases exhibit higher activity than the combination of the individual monospecific parent binding molecules (antibodies), and thus can induce or enhance immune responses.Furthermore, they are also shown to have cell killing activity against cancer cell lines.Thus, the data reveal that these molecules and bispecific platform structures are excellent candidates for cancer immunotherapy drugs.

[0250] Octet binding assay (OX40(SLR) / PD-L1 BiS5) To assess the binding of the bispecific binding molecules disclosed herein, an Octet QK with a Ni-NTA biosensor chip and 10x kinetics buffer was used (ForteBio, Menlo Park, CA). For this particular series of bispecific binding proteins, His-tagged PD-L1-Fc, his-tagged PD-1-Fc and hOX40-Fc (human recombinant protein) were purchased from R&D Systems (Minneapolis, MN). All binding assays were performed at 25°C. The binding results demonstrate that the BiS5 OX40Ab / PD-L1 molecule binds to both PD-L1-His and hOX40-Fc (Figure 39).

[0251] PD-L1 / OX40 BiS5 bound to CHO cells expressing human or cynomolgus OX40 and PD-L1 / B7H1 (Figure 40A-F). Binding of PD-L1 / OX40 BiS5 constructs was also measured by flow cytometry (HyperCyt) (Figure 42). OX40 IgG4P and OX40 / PD-L1 bispecific molecules bound to Jurkat OX40 receptor cells. PD-L1 IgG and OX40 / PD-L1 bispecific molecules bound to NCI H358 and CHOK1 B7H1(PD-L1) / OKT3 cells. IgG and bispecific molecules all bound to HEK CD32a cells.

[0252] PD-L1 and OX40 receptor assays In the PD-L1 receptor assay (using the protocol described above), the PD-L1scFv-containing bispecific molecules and the positive control IgG all demonstrated activity (Figure 42A-B). The single-arm OX40 control and the isotype control showed no activity in this assay. EC 50 and Hill slopes are consistent with values ​​obtained in previous assays for the anti-PD-L1 parental control and the PD-L1 Bis2, Bis3 and Bis5 constructs.

[0253] In an OX40 reporter gene assay using HEK CD32a cells, the bispecific constructs had comparable activity to each other and Fc-mediated agonism was observed (Figure 43A-B). OX40 / PD-L1 Bis5 N434A IgG1 had an EC2 activity comparable to that of OX40 IgG4P and MEDI0562 (OX40 IgG1). 50 It had activity.

[0254] In an OX40 reporter gene assay using CHOK1 PD-L1 on expressing cells, the OX40 / PD-L1 bispecific molecules show comparable agonism (Figure 44A-B). OX40 / PD-L1 Bis5 N434A IgG1 has comparable EC20 values ​​to other Fc variants of the OX40 / PD-L1 Bis5 bispecific Mab tested. 50 No agonism was detected by OX40 IgG or PD-L1 IgG, thus demonstrating PD-L1-mediated OX40 agonism.

[0255] PD-L1-mediated OX40 agonism with tumor cells was detected using the OX40 / PD-L1 bispecific molecule (Figure 45A-B). The OX40 / PD-L1 bispecific molecule showed equal agonism-normal curves in this assay. No agonism was observed with OX40 IgG, demonstrating the benefit of using the bispecific molecule versus the combination of OX40 IgG and PD-L1 IgG. No agonism was detected with NCI H358 PD-L1 KO cells (Figure 46A-D), indicating that the NCI H358 agonism observed in the cells is PD-L1 specific.

[0256] Staphylococcal Enterotoxin B (SEB) Assay In the SEB assay, the OX40 / PD-L1 bispecific molecule had higher activity than the combination of individual antibodies against OX40 and PD-L1 (Figure 47A-D). Notably, the G4P construct had higher activity than the G1 construct. The wild type, YTE-containing mutant, and N434A mutant had comparable activity.

[0257] Treg suppression assay A Treg suppression assay was performed to test the OX40 / PD-L1 bispecific molecule (Figure 4A-D). The OX40 / PD-L1 bispecific molecule suppressed CD4+ Treg suppression only in the presence of PD-L1. effThe OX40 / PD-L1 bispecific molecule was active on T reg The inhibitory effect was suppressed, but only when cross-linked by binding to plate-immobilized PD-L1.

[0258] Mixed Leukocyte Reaction (MLR) Assay To test the OX40 / PD-L1 bispecific molecule, an MLR assay was performed (Figure 51A-B). The OX40 / PD-L1 bispecific molecule had higher activity than the combination of individual antibodies against OX40 and PD-L1 (Figure 52A-E).

[0259] Antibody-dependent cell-mediated cytotoxicity (ADCC) assay ADCC assays were performed to test the OX40 / PD-L1 bispecific molecules. Freshly isolated NK cells were used as effector cells and CHOK1 PD-L1 B7H1 and CHOK1 OX40 overexpressing cells as target cells, respectively, at an effector:target molecule (E:T) ratio of 20:1. Target cell lysis was analyzed using europium release from labeled target cells after 5 hours. In the ADCC assays, OX40 / PD-L1 BiS2 and BiS5 mediated ADCC against PD-L1 or OX40 expressing CHO cells (Figures 53A-B and 54).

[0260] CD107a mobilization assay was performed using freshly isolated NK cells as effector cells and PD-L1 and OX40 overexpressing CHO K1 as target cells, respectively, at an E:T ratio of 10:1. CD107a mobilization to the cell surface of NK cells was analyzed by flow cytometry after 4 hours. BiS2 and BiS5 OX40 / PD-L1 bispecific molecules enhanced CD107a mobilization of NK cells to PD-L1 and OX40 expressing CHO cells in an antibody-dependent cell-mediated cytotoxicity (ADCC) assay (Figure 55). BiS2 and BiS5 OX40 / PD-L1 bispecific molecules enhanced CD107a mobilization of NK cells to activated allogeneic T cells that upregulated OX40 and PD-L1 (Figure 56). BiS5 OX40 / PD-L1 enhanced CD107a recruitment of NK cells from two different donors to activated allogeneic T cells (Figure 57A-B).

[0261] Pharmacokinetic and pharmacodynamic (PK / PD) studies A study was designed to compare the PK / PD of the OX40 / PD-L1 bispecific molecules (Figure 58). The serum concentration-time profiles of the PD-L1 / OX40 bispecific molecules were compared in cynomolgus monkeys (Figure 59 and Table 11). The mean T 1 / 2 clearance rate was lower for the Bis5 OX40 / PD-L1 IgG1 N434A molecule compared to the WTBis5 molecule. Both molecules similarly reduced soluble PD-L1 in serum and induced significant increases in the percentage of Ki67+ total memory CD4+ T cells, total memory CD8+ T cells, and NK cells.

[0262] [Table 30]

[0263] The OX40 / PD-L1 bispecific molecule reduced serum soluble PD-L1 concentrations below the assay LLOQ (Figure 60). The N434A mutation improved the pharmacokinetics of Bis5 OX40 / PD-L1-G1. Notably, CL was reduced by nearly half; there was a corresponding 2-fold increase in T1 / 2 and AUCinf; and Cmax and Vss were unaffected. This was consistent with the previously reported effect of this mutation on monoclonal antibody PK. Thus, progress toward mAb-like PK of the Bis5 OX40 / PD-L1-G1 IO BisAb was achieved. Serum concentrations of Bis5 OX40 / PD-L1-G1 BisAb were below the lower limit of quantification (BLOQ) at 2 weeks, which may be relevant for ADA.

[0264] A substantial and statistically significant increase was observed in total memory CD4, total memory CD8, and NK cell proliferation (percentage of Ki67+ cells) comparing the PD-L1 OX40 Bis5 group with the control (anti-PcrV-Psl control) Ab group (Figure 61A-F). A trend toward significant differences was observed between the PD-1 LO115 and PD-L1 OX40 Bis5 groups in total memory CD4, total memory CD8, and NK cell proliferation (Ki67+). There was no statistically significant difference in proliferation between the PD-L1 OX40 Bis5 N434A (half-life extended) version and the G1 version. The PD-L1 OX40 Bis5 N434A and IgG1 versions are biologically active bispecific molecules.

[0265] Example 2(e). OX40 / PD-1 Bispecific Binding Protein The following bispecific binding proteins were generated that bind to PD-1 and OX40 using the parent sequences identified above in Table 2. Proteins identified as BiS2 and BiS3 were generated using the sequences shown in Table 24 below and evaluated for simultaneous antigen binding studies using the Octet binding assay described below.

[0266] [Table 31]

[0267] The PD-1 / OX40 BiS2 monoclonal antibody (mAb) is a bispecific antibody engineered to simultaneously bind human and cynomolgus PD-1 and human and cynomolgus OX40 (Figure 62; PD-1 binding proteins are shown in grey and OX40 binding proteins in light grey). Without being bound to a particular theory, the proposed mechanism of action suggests a dual signaling effect on T cells following cis binding to both OX40 and PD-1, agonism of the T cell costimulatory surface receptor OX40, and blockade of immunosuppressive PD-1 (Figure 63).

[0268] Octet binding assay Shows simultaneous binding activity for two different lots of PD-1(LO115) / OX40 BiS2 mAb to PD1-His and human OX40-Fc (Figure 64).

[0269] OX40 reporter assay PD-1(LO115) / OX40 BiS2 mAb showed comparable activity to other OX40 agonists (Figure 65A-B). Protein was stored at 4°C, used immediately, frozen / thawed three times, and stored at 4°C for 7 days, followed by storage at 40°C for 7 days. Activity was recorded as relative light output versus concentration of mAb. EC of PD1(LO115) / OX40 BiS2 mAb at day 0 and 4°C 50 was approximately 2 nM.

[0270] PD-1 / PD-L1 reporter assay PD-1(LO115) / OX40 BiS2 mAb showed comparable activity to other PD-1 agonists (Figure 66A-B). Protein was stored at 4°C, used immediately, frozen / thawed three times, and stored at 4°C for 7 days, followed by storage at 40°C for 7 days. Activity was recorded as relative light output versus concentration of mAb. EC 50was approximately 1 nM. Two sets of primary human in vitro potency assays were performed; antigen recall T cell assay and T cell costimulation with Staphylococcal enterotoxin B (SEB).

[0271] Staphylococcal Enterotoxin B (SEB) Assay In a SEB assay, PD-1(LO115) / OX40 BiS2 mAb induced increased levels of IL-2 detected in the supernatant of cells after 3 days of culture (Figure 67). Thus, PD-1 / OX40 BiS2 mAb can simultaneously bind its human target antigen and costimulate T cells in vitro.

[0272] In an antigen recall assay, PD-1 / OX40 BiS2 mAb drove increased levels of interferon (IFN) gamma compared to the parental mAb and the parental mAb combination (Figures 68 and 69).

[0273] CMV Ag recall assay In a CMV Ag recall assay (using the protocol described above), the BiS2 and BiS3 molecules did not show equal activity compared to the combination (Figure 70). The data show that the PD-1 / OX40 BiS2 IgG4P mAb is active in vitro and in vivo. PD-1 / OX40 BiS3, which differs in structure from PD-1 / OX40 BiS2, was not detectably active. Thus, PD-1 / OX40 BiS3 (not active) differs from BiS2 (active).

[0274] Pharmacokinetic and pharmacodynamic (PK / PD) studies Cynomolgus monkeys were considered a pharmacologically relevant non-clinical species to test the functional activity of PD-1 / OX40 BiS2 mAb. The pharmacokinetics (PK) and pharmacodynamics (PD) of PD-1 / OX40 BiS2 mAb were evaluated in a non-GLP (Good Laboratory Practices) study in cynomolgus monkeys. PD-1(LO115) / OX40 BiS2 mAb PK and PD (% of Ki67 positive CD4+ and CD8+ total memory T cells) were evaluated in cynomolgus monkeys (n=3; males) after a single intravenous (IV) dose over a dose range of 0.1mg / kg to 30mg / kg. PBMCs were collected pre-dose, 1 day, 8 days, 11 days, and 15 days after dosing, cryopreserved and thawed, and then analyzed by flow cytometry. In summary, PD-1(LO115) / OX40 BiS2 mAb demonstrated near-linear PK with a short half-life of 0.6-1.7 days (Figure 70; Table 12).

[0275] [Table 32]

[0276] Average peak concentration (C max ) increased approximately dose-proportional from 2.0 μg / mL at 0.1 mg / kg to 607 μg / mL at 30 mg / kg. AUC∞ increased approximately dose-proportional from 1.7 μg·day / mL at 0.1 mg / kg to 577 μg·day / mL at 30 mg / kg. Mean serum clearance ranged from 41.8 mL / day / kg to 60.2 mL / day / kg. Steady-state volume of distribution ranged from 43.2 mL / kg to 85.6 mL / kg. PD results (Figure 71) showed a dose-dependent increase in CD4+ total memory T cell proliferation (Ki67), and an increase in CD8+ total memory T cell proliferation (Ki67). A representative standard curve for quantification of PD-1 / OX40 in cynomolgus monkey serum is shown (Figure 72).

[0277] Example 3. Physical and chemical stability of BiSAb constructs A series of experiments were performed to evaluate the physical and chemical stability of the BiSAb constructs described herein relative to other bispecific binding protein structural strategies and platforms. In particular, a series of stability studies described below revealed and analyzed the effect of different pH ranges on BiSAb stability (e.g., hydrolysis, fragmentation, aggregation, thermal stability). For different exemplary embodiments of various BiSAb formats, the data shows that the BiSAb disclosed herein (identified as "BiS5" in the studies below, which is in D / H format as shown in Table 13) exhibited unexpected and surprising physical and chemical stability relative to all other BiSAb structural motifs.

[0278] [Table 33]

[0279] Example 3.1 A comparison was further performed between the BiS format disclosed herein ("BiS5") and another BiS format, identified as "BiS4", which includes two binding domains (scFv domains) linked at a hinge region (e.g., between the Fc and Fab regions). BiS4 and BiS5 proteins were expressed in Chinese Hamster Ovary (CHO) and purified by conventional chromatographic methods. As noted above, these two formats have similar Fab and scFv sequences, with the primary difference between them being the location of the scFv domain (as noted herein, for BiS4, the scFv is located within the hinge region; for this particular BiS5, the scFv is located within the C H The purified BiS molecule was added to PBS buffer at 1.54 M. -1 cm -1 Protein concentrations were determined using a NanoDrop ND-1000 (Thermo Scientific, Wilmington, Delaware) using an extinction coefficient of 0.01 μg / ml.

[0280] pH Screen and Short-Term Stability Testing For pH screen testing, BiS4 and BiS5 antibodies were concentrated to approximately 12 mg / mL and dialyzed against six different pH conditions, 20 mM sodium succinate (pH 5.0), histidine / histidine HCl (pH 5.5, 6.0, and 6.5), and sodium phosphate (pH 7.0 and 7.5), all containing 240 mM sucrose. Dialysis was performed using Slide-A-Lyzer dialysis cassettes (10 kDa molecular weight cut-off (MWCO), Thermo-Fisher, Rockford, Illinois). After completion of dialysis, 0.02% polysorbate 80 was added to adjust the final protein concentration to approximately 10 mg / mL. BiS4 and BiS5 formulations were sterilized using 0.22 μm filters (Millipore, Billerica, Massachusetts) in a pre-sterilized clean bench. One milliliter aliquots were dispensed into 3 mL borosilicate glass type I vials (West Pharmaceutical Services, Exton, Pennsylvania). Samples were stored at 40° C. and analyzed by SEC at time zero and after 1, 2, 3, and 4 weeks of storage.

[0281] Differential Scanning Calorimetry (DSC) Differential scanning calorimetry thermograms were acquired for the zero time samples using a VP-Capilary DSC connected to a temperature-regulated autosampler (Malvern Instruments Ltd., Westborough, Massachusetts). A protein concentration of 1 mg / mL was used to acquire the thermograms along with a scan rate of 90 °C / h over a temperature range of 20 °C to 100 °C. Thermograms of BiS4 and BiS5 at different pH conditions ranging from 5.0 to 7.5 were buffer subtracted and then baseline corrected. Data analysis was performed using the DSC plug-in for the Origin 7 SR4 software package. Experimental results were fitted to a multistate model with three transitions to determine the melting temperatures (T m The heat capacity (C) of the first temperature transition was calculated. p) value is 500 cal mol -1 ℃ -1 The point at which the temperature reaches the start temperature (T onset ) was considered.

[0282] High Performance Size Exclusion Chromatography (HP-SEC) To separate aggregate and fragment species from monomers based on size, a 7.8 x 30 cm 2 Stability samples were analyzed using an Agilent high-performance liquid chromatography system equipped with a photodiode array detector capable of recording 200-400 nm UV absorption spectra on a 5 μm, 250 Å, Tosho TSKgel G3000SWxl (TOSOH Biosciences, King of Prussia, Pennsylvania) and corresponding guard columns. A mobile phase containing 0.1 M anhydrous disodium hydrogen phosphate, 0.1 M sodium sulfate, 0.01% sodium azide, pH 6.8, and a flow rate of 1 mL / min was used to separate the species. The amount of protein injected was approximately 250 μg. The separation of BiS4 and BiS5 was monitored using absorption spectra at 280 nm. The peak areas of soluble aggregates (multimers and dimers), monomers, and fragments were quantified. A kinetic plot was then created by calculating the percentage of each of these species and plotting against the incubation time. pH profile curves for the monthly rates of monomer loss, fragmentation, and aggregation were generated by calculating the slope of each kinetic plot.

[0283] Thermal stability of BiS4 and BiS5 The effect of pH on the thermal stability of BiS4 and BiS5 was evaluated by analysis of thermograms acquired using capillary DSC and generated at six different pH conditions. Figures 74A and 74B show overlaid DSC thermograms of BiS4 and BiS5, respectively, from pH 5.0 to 7.5. As shown in Figure 73, each thermogram exhibits a transition temperature T m 1. T m 2, and T mThe three thermal denaturation events from 3 are shown. The first transition (T m 1) is C H 2 and scFv domains, and the second (T m 2) and 3rd (T m 3) Transition is C H 3 and F ab For both formats, the T onset , T m 1. T m 2, and T m An increase of 3 was observed (Figures 73A, 73B, 73E and Table 14 below). For BiS4 and BiS5, the T onset , T m 2, and T m No differences were observed in the hinge region or C3 (Figure 73E and Table 15). H The presence of scFv in any of the three domains H 3 and F ab Interestingly, the T m A slight increase of 1 was observed for BiS5, which is due to the scFv being C H If located within the 3 domain, scFv, C H These results show increased thermal stability of either or both of these compounds.

[0284] [Table 34]

[0285] Physical and chemical stability of BiS4 and BiS5 The physical and chemical stability of BiS4 and BiS5 formats at different pH values ​​(5.0-7.5) was evaluated at 40°C for up to 4 weeks. The HP-SEC chromatogram at "time zero" was used to compare the total area, monomer, aggregate, and fragment content of the HP-SEC chromatograms at other time points. Representative chromatograms of BiS4 and BiS5 at pH 7.5 time zero compared to 4 weeks are shown in Figure 74A. All samples contain primarily monomer with low levels of soluble aggregates and with or without fragments. At time zero (solid line), the majority of the sample is monomer with no notable differences other than slight differences in peak height between the two samples likely due to slight concentration differences (Figure 74A). The dotted line shows an overlay of HP-SEC chromatograms of both formats at the same pH conditions after 4 weeks of storage at 40°C. Under accelerated temperature stress conditions, both formats show additional peaks, early eluting peaks (multimeric species), reduced monomer, and increased fragment levels (Figure 74A). The loss of monomers due to fragmentation was more pronounced for BiS4 compared to BiS5, indicating that BiS5 is more chemically stable. Based on their structures, possible fragmentation sites, and retention times, it is speculated that the small fragment peak (RT 10.8 min) is Fab, and the large fragment peak (RT 9.8 min) and shoulder peak (RT 8.7 min) are Fab containing scFv and its corresponding high molecular weight fragment (HMWF) containing Fab, scFv, and Fc, respectively.

[0286] To better evaluate the effect of scFv position on the physical and chemical stability of BiS4 and BiS5, the percentage of total area of ​​each species was plotted in a bar graph for zero time and 4 weeks at pH 7.5, 40°C (Figure 74B). As shown in Figure 74B, at zero time, the monomer purity of BiS4 and BiS5 is similar. Samples incubated at 40°C for up to 4 weeks showed significant differences in the type and extent of fragments formed. In the case of BiS4, 11.8%, 7.2% and 3.5% of the shoulder peak (RT ~ 8.7 min), large fragments (RT ~ 9.8 min) and small fragments (RT ~ 10.8 min) were formed, respectively (Figure 74B). Surprisingly, the BiS5 sample showed only 1.4% small fragments (RT ~ 10.8 min), probably due to scFv tethering to Fc from both sides of the domain.

[0287] Figures 75A-75C show the kinetics of aggregation, fragmentation and monomer loss for BiS4 and BiS5 incubated at 40°C for pH 7.5 samples. The BiS4 sample showed a faster rate of monomer loss compared to BiS5 (Figure 75A). The rates of monomer loss for BiS4 and BiS5 at pH 7.5 were 27.4% and 4.5% per month, respectively (Figure 75A). For BiS4, the majority of the monomer loss was due to fragmentation, which was 23.9% per month, and to a lesser extent aggregation, which was 3.5% per month (Figures 75B and 75C). Interestingly, for BiS5, aggregation appears to be at a slightly higher rate (2.8% per month) compared to the fragmentation rate (1.7% per month) (Figures 75B-75C).

[0288] Further analysis of the effect of pH on the physical and chemical stability, monthly monomer loss, fragmentation and aggregation rates of the BiS4 and BiS5 formats was performed by plotting the above values ​​against the six pH conditions (Figures 76A-76C). Throughout all six pH conditions ranging from pH 5.0 to 7.5, the monomer loss rate was lower for the BiS5 format than for BiS4 (Figure 76A), suggesting that the BiS5 format disclosed herein has unexpectedly superior physical and chemical stability than other bispecific protein formats. In the case of BiS4, the majority of the monomer degradation was due to fragmentation at even lower pH conditions (Figure 76B). BiS5 showed lower fragmentation rates than Bis4 in all pH conditions tested. Surprisingly, the fragmentation rate of BiS5 is flat over a wide pH range and appears to be lower than Bis4. Without being bound to any particular theory, the relatively low fragmentation rate observed for BiS5 may be due to the G4S linkers at either end of the scFv that connect it to the Fc. Fragmentation of one of the G4S linkers linking the Fc would not release the scFv, as it may still be linked to the Fc via the other G4S linker. In BiS4 and BiS5, aggregation rates appeared similar across all pH conditions tested (Figure 76C), indicating that the position of the scFv has minimal effect on aggregation kinetics, as measured using capillary DSC, with similar T between the two formats at all pH conditions. onsetThis is also supported by the fact that no change was observed in the solubility of BiS4 at pH 7.5 and 40° C. (time=0) (FIG. 73E and Table 14 above). At pH 7.5 and 40° C. (time=0), none of the molecules showed any obvious fragmentation (FIG. 77A), while obvious fragmentation was observed for BiS4 and slight fragmentation was observed for BiS5 under the same conditions after 2 weeks of storage at 40° C. (FIG. 77B). Both BiS4 and BiS5 showed reduced fragmentation and aggregation at low (5.5) pH, while BiS5 has superior performance in both fragmentation and aggregation at both pH values ​​(FIG. 78). This series of experiments demonstrates that BiS5 disclosed herein has better chemical stability than BiS4 and similar physical stability to BiS4.

[0289] Example 3.2 Further testing was performed to assess the physical and chemical stability of various embodiments of the bispecific binding proteins disclosed herein and identified as Constructs A-H (e.g., Table 13 above and associated examples). These constructs were analyzed using DSC, accelerated storage stability, and FcRn and FcgR binding assays, as described below.

[0290] Differential scanning calorimetry analysis DSC experiments for this data set were performed using a Microcal VP-DSC Scanning Microcalorimeter (Microcal). All solutions and samples used for DSC were filtered using a 0.22 μm filter and degassed before loading into the calorimeter. Antibodies used for DSC studies were >98% monomeric as determined by analytical SEC. All samples were extensively dialyzed (at least 3 buffer exchanges) in 25 mM histidine-HCl (pH 6.0) prior to DSC analysis. The buffer from this dialysis was used as the standard buffer for the following DSC experiments. A baseline measurement (buffer vs. buffer) was subtracted from the sample measurement prior to sample measurement. Dialyzed samples (at a concentration of 1 mg / ml) were added to the sample wells and DSC measurements were performed at a scan rate of 1° C. / min. Data analysis and deconvolution were performed using the Origin™ DSC software provided by Microcal. Deconvolution analysis was performed using a non-two-state model and a best fit was obtained using 100 iterative cycles. T onset is defined as the qualitative temperature at which the thermogram is found to have a non-zero slope, T m is defined as the temperature at which half of the molecules in the set are denatured, which is calculated as the temperature value corresponding to each peak maximum in the thermogram.

[0291] The results of the different constructs are shown in Figure 79. In general, constructs A, C, and D containing 2F4 as the scFv showed lower T than constructs E, G, and H containing LC10 as the scFv. M 1. Without being bound by theory, it is believed that the difference in TM1 values ​​is due to the inherently superior thermostability of the LC10 scFv domain relative to the 2F4 variable domain. These data suggest that constructs A-D, which have 2F4 as the scFv, will be less thermostable than constructs E-H, which have LC10 as the scFv.

[0292] Accelerated storage stability analysis Construct concentrations were normalized to 1 mg / mL. 1 mL of each bispecific construct or IgG control was aliquoted into a 1.5 ml Eppendorf tube. Samples were incubated at 45°C in a static incubator for 2 weeks. Samples were analyzed at 3, 7, and 14 days to assess stability. Visual inspection was performed at each time point to note any turbidity or increased precipitation. Samples were filtered using a 0.2 um spin column and 120 ul of sample was aliquoted and introduced into the HPLC to ensure there were no air bubbles at the bottom of the vial. The samples were then run on a TSK-GEL G3000SW with 0.1 M sodium phosphate, 0.1 M sodium sulfate (pH 6.8) as running buffer. XL Samples were tested on an Agilent 1100 series HPLC-SEC to check for aggregation and degradation using a (300 x 7.8 mm) Tosoh Bioscience column. 60 μL of sample was injected and run at a flow rate of 1 mL / min. Monomer retention time (min), total peak area, % monomer, % aggregates, % fragments, % monomer loss were obtained and used for analytical SEC analysis. The results are summarized in Table 15.

[0293] [Table 35]

[0294] As described herein, the positions of the scFv domains in the above constructs are as follows ("-" indicates scFv): A and E are IS-RTP; B and F are AK-GQP; C and G are S-NG; D and H are SN-G. M The values ​​are associated with the following domains: M 1=CH2 / scFv;T M 2=Fab;T M3=CH3. The data tend to indicate that constructs A and C, with 2F4 scFv inserted into the ISRTP (A) and SNG (C) loops, are more prone to aggregation than constructs E and G, with LC10 scfv inserted into the same positions. This observation suggests that the sequence identity and behavior of the scFv domains may play a role in the stability of bispecific binding protein constructs. Furthermore, from the above, construct D, containing 2F4 scFv, can be predicted to have a similar low stability as A and C, but the insertion of 2F4 scFv into the SNG loop appears to stabilize the molecule and reduce its tendency to form aggregates. Taken together, this accelerated stability study indicates that the scFv sequence and location within the Fc region may play a significant role in the stability of BiSAb constructs.

[0295] FcRn and FcγR binding analysis Binding experiments were performed using a BIAcore3000 instrument (BIAcore). 1000RU IsdH(Fab) antigen was immobilized on a CM5 chip to capture antibodies. Antibodies were captured by flowing 100nM BiSAb constructs or mAb control at 20μL / min for 5 minutes. 5uM huFcRn or FcγRI, IIa, IIb, IIIa-158V or IIIA158F were flowed at 5μL / min for 20 minutes. FcRn binding was performed in PBS pH 6.0 + 5uM EDTA as well as in PBS pH 7.4 + 5uM EDTA.

[0296] Constructs A, C, D, E, G and H were evaluated for FcRn binding. Representative data is shown in Figure 80 for each of the bispecific constructs E and H, as well as for 2F4 IgG binding to FcRn. Constructs with scFv downstream of the CH2-CH3 interface are found to retain FcRn binding (e.g., D and H constructs). Constructs with scFv located in the ISRTP loop upstream of the CH2-CH3 interface appear to eliminate detectable FcRn binding (e.g., A and E constructs). The ISRTP loop is within the region of the known half-life extending YTE mutation in the Fc (M252Y / S254T / T256E) known to be important for FcRn binding.

[0297] Constructs A, C, D, E, G, and H were tested for binding to FcγRI, FcγRIIa, FcγRIIb, FcγRIIIa-158F, and FcγRIIIa-158V. Representative data for binding of constructs E, G, and H to FcγRIIIa-158V is shown in Figure 81. All constructs tested retained binding to FcγR, but with different affinities (Figure 81, inset). Table 16 shows the observed binding trends of various constructs to FcγR.

[0298] [Table 36]

[0299] The difference observed in FcγR binding with constructs with an scFv inserted in the ISRTP loop upstream of the CH2-CH3 interface (A and E) compared to other constructs with an scFv inserted in the SNG loop downstream of the CH2-CH3 interface (C, D, G, and H) shows consistently lower FcγR binding.

[0300] Attempts to evaluate whether FcRn binding of constructs A and E could be improved or restored were performed by introducing a half-life extension loop (N3) into the Fc region. Figure 82 shows representative data, showing that for the E construct, neither BiS5Ab E nor construct E with the N3 loop introduced (BiS5Ab E+N3) were able to bind FcRn. Furthermore, inserting LC10 scFv into the N3 loop (N3 scFv) and keeping the ISRTP loop intact reduced FcRn binding to below detectable levels. These data indicate that, at least for the E construct, and otherwise for each of the constructs disclosed herein, both the ISRTP loop and the N3 loop (if present) must be kept intact and unmodified to retain FcRn binding.

[0301] Example 3.3 In addition to comparing BiS4 with the bispecific binding construct disclosed herein (BiS5), studies were performed to evaluate three other BiS structural motif platforms, identified as BiS1, BiS2, and BiS3 (see Figure 83). As can be seen with reference to Figure 83, these platforms vary with respect to the location of one binding domain (shown as an scFv domain). Of the five motifs, only BiS4 and BiS5 contain two linker moieties as attachment points to large proteins, while the others (BiS1, BiS2, and BiS3) are attached by a single linker.

[0302] Briefly, representative molecules of each construct were analyzed for stability using the techniques described in Examples 3.1 and 3.2 above. Samples of each construct were added to buffers of pH 5.0, 5.5, 6.0, 6.5, 7.0, and 7.5 and stored at 40° C. for a period of two months. The samples were then analyzed for fragmentation rates (FIG. 84), aggregation rates (FIG. 85), and monomer loss rates (FIG. 86) using HP-SEC. Under these conditions, the analysis showed that the bispecific binding protein format disclosed herein ("BiS5"; and the D / H format shown above in Table 13) has superior physical and chemical stability to all other formats at all pH conditions.

[0303] The SEC data was also used to map the various peaks to the corresponding fragments of the BiS molecule (Figure 87). The mapping was based on assumptions that fragmentation occurs in the hinge and linker regions of the molecule, the size of the fragments, the theoretical fragmentation, and how the expected fragment species adjust with respect to the fragment species observed in other formats. While there is good resolution between the low molecular weight fragments (LMWF) in each format, there is poor or no resolution between the monomers and the high molecular weight fragments (HMWF) in all formats. Based on the information in Table 17, we conclude that the HP-SEC technique underestimates the fragmentation of the BiS format much more than the monoclonal antibodies. An alternative analysis was developed as described below.

[0304] [Table 37]

[0305] An alternative analysis was developed to calculate the fragmentation rates of HMWF using molar extinction coefficients based on the assumption that (i) during degradation, if a small fragment is detected, then a corresponding large fragment should also be present; (ii) secondary fragmentation (fragment of fragments) does not occur significantly during the stability study; and (iii) fragmentation occurs in the linker and / or hinge regions. The fragmentation rates were determined based on the following relationship:

number

[0306] [Table 38]

[0307] Furthermore, to determine whether disulfide bonds influenced fragmentation and stability, fragmentation rate analysis was performed with the constructs under reducing conditions. Representative data from this assay is shown in (Figure 86). Under reducing conditions, higher fragmentation rates were observable for all BiS formats except BiS1 (Table 19). It was concluded that the higher fragmentation rates under reducing conditions support the tethering of the scFv moiety to the CH3 region in the BiS5 construct (Figure 89).

[0308] [Table 39]

[0309] [Table 40]

[0310] The characteristics of the stabilizing disulfide bond disclosed above were further investigated. The results are shown in Tables 23 and 24 below. Bispecific antibodies corresponding to two different specificities were generated in BiS4 and BiS5 (scFv inserted in the SN-G loop) formats with or without the stabilizing disulfide bond in the scFv. Accelerated stability tests showed that the BiS4 construct without the stabilizing disulfide bond had substantial monomer loss due to degradation, which was prevented by the introduction of the stabilizing VL-VH disulfide bond. These results indicate that the removal of the stabilizing disulfide bond in the scFv of the BiS5 construct did not have a significant effect on its stability.

[0311] [Table 41]

[0312] [Table 42]

[0313] Based on all the data above, the bispecific binding protein format disclosed herein appears to be the most stable of all formats tested. Furthermore, BiSAb5 is found to be the most stable in terms of minimizing both fragmentation and aggregation at the low pH values ​​tested (e.g., 5.0, 5.5, and 6.0). Thus, the unexpected and surprising stability properties of BisAb disclosed herein provide additional advantages over other structural platforms and formats used to generate bispecific binding molecules.

[0314] Incorporation by Reference All publications and patents mentioned in this specification are herein incorporated by reference in their entirety to the same extent as if each individual publication or patent was specifically and individually indicated to be incorporated by reference.

[0315] Although specific embodiments of the present disclosure have been discussed, the above description is illustrative and not limiting.Many variations of the present disclosure will become apparent to those skilled in the art upon review of this specification and the following claims.The full scope of the present disclosure should be determined by reference to the claims, together with their full scope of equivalents, and the present specification, together with its variations. An embodiment of the present invention will be described below. (1) a first binding domain (BD1) that binds to a first epitope; a second binding domain (BD2) that binds to a second epitope; C H 2 and C H The Fc region contains three domains A protein comprising: The Fc region includes the C H 2 domain, C H 3 domain, or the C H 2 and C H The solvent-exposed loop at the interface of the three domains contains BD2. The protein is bivalent for binding to each of the first and second epitopes. (2) The Fc region is the C H 2 domain, C H 3 domain, or the C H 2 and C H The protein described in (1), which includes BD2 in a solvent-exposed loop in the amino acid sequence at the interface of the three domains. (3) The solvent exposure loop is H The protein according to (2), comprising an amino acid sequence derived from two domains. (4) The protein according to (3), wherein the solvent-exposed loop comprises the amino acid sequence ISRTP (sequence number 39). (5) The solvent exposure loop is H The protein according to (2), comprising an amino acid sequence derived from three domains. (6) The protein according to (5), wherein the solvent exposed loop comprises the amino acid sequence SNG. H2 domain and the C H The protein described in (2), comprising an amino acid sequence derived from the interface of the three domains. (8) The protein according to (7), wherein the solvent-exposed loop comprises the amino acid sequence AKGQP (sequence number 40). (9) The protein according to any one of (1) to (8), wherein BD2 comprises a single-chain variable fragment (scFv). (10) The protein according to any one of (1) to (8), wherein BD1 comprises a binding domain selected from the group consisting of a Fab domain, an scFv, a single domain antibody, and an antibody variable domain. (11) The protein according to any one of (1) to (8), wherein BD1 includes a Fab domain. (12) The protein according to (11), wherein the Fab domain is linked to the Fc domain via an antibody hinge region. (13) The protein according to any one of (1) to (8), wherein the Fc region comprises a domain selected from the group consisting of Fc regions derived from IgG1, IgG2, IgG3, IgG4, IgA, IgM, IgE, and IgD. (14) The protein described in (13), wherein the Fc region includes a mutant Fc region. (15) The protein according to (13), wherein the Fc region is non-glycosylated. (16) The protein according to (13), wherein the Fc region is deglycosylated. (17) The protein according to (13), wherein the Fc region has low fucosylation or is non-fucosylated. (18) The protein described in any one of (1) to (8), further comprising a protein linker L1 between BD2 and the Fc region. (19) The protein described in any one of (1) to (8), further comprising a first protein linker L1 and a second protein linker L2 between BD2 and the Fc region. (20) The protein according to any one of (1) to (8), wherein BD2 is linked to the Fc region via a protein linker L1. (21) The protein according to any one of (1) to (8), wherein BD2 is linked to the Fc region via two protein linkers L1 and L2. (22) The protein according to any one of (18) to (21), wherein L1 and L2 are independently selected from (G4S)2 (sequence number 41), (G4S)3 (sequence number 42), and (G4S)4 (sequence number 43). (23) a polypeptide domain having the following structure from the N-terminus to the C-terminus: V H 1-C H 1-C H 2(N-terminus)-BD2-C H 2(C-terminus)-C H 3 and BD1 contains the Fab domain, V H 1 comprises the heavy chain variable domain of the Fab domain, and C H 1 comprises heavy chain constant domain 1 of the Fab. (24) a polypeptide domain having the following structure from the N-terminus to the C-terminus: V H 1-C H 1-C H 2-BD2-C H 3 and BD1 contains the Fab domain, V H 1 comprises the heavy chain variable domain of the Fab domain, and C H 1 comprises heavy chain constant domain 1 of the Fab. (25) a polypeptide domain having the following structure from the N-terminus to the C-terminus: V H 1-C H 1-C H 2-C H 3(N-terminus)-BD2-C H 3(C-terminus) and BD1 contains the Fab domain, V H1 comprises the heavy chain variable domain of the Fab domain, H 1 comprises heavy chain constant domain 1 of the Fab. (26) The protein according to any one of (23) to (25), wherein BD2 comprises an scFv. (27) The scFv comprises, from the N-terminus to the C-terminus: V H 2-Polypeptide Linker-V L 2 or V L 2-Polypeptide Linker-V H 2, including V H 2 comprises the heavy chain variable domain of said scFv, and V L 2 comprises the light chain variable domain of the scFv. (28) The protein according to any one of (23) to (27), further comprising a protein linker L1 between BD2 and the Fc region. (29) The protein described in any one of (23) to (27), further comprising a first protein linker L1 and a second protein linker L2 between BD2 and the Fc region. (30) BD2 binds the C of the Fc region via a linker (L1). H 2 domain, C H 2 domain, or the C H 2 and C H The protein according to any one of (23) to (25), which is bound to the interface between the three domains. (31) BD2 connects the C of the Fc region to the C of the Fc region via two protein linkers L1 and L2. H 2 domain, C H 2 domain, or the C H 2 and C H The protein according to any one of (23) to (25), which is bound to the interface between the three domains. (32) The protein according to any one of (28) to (31), wherein L1 and L2 are independently selected from protein linkers having a length of 1 to 25 amino acids. (33) The protein according to any one of (28) to (31), wherein L1 and L2 are independently selected from (G4S)2 (sequence number 41), (G4S)3 (sequence number 42), and (G4S)4 (sequence number 43). (34) The protein according to any one of (1) to (33), wherein the first and second epitopes are different. (35) The protein according to any one of (1) to (34), wherein the first and second epitopes are the same. (36) A bispecific binding protein that binds to PD-1 and CTLA-4, comprising a first peptide comprising the amino acid sequence of SEQ ID NO:1 and a second peptide comprising the amino acid sequence of SEQ ID NO:2. (37) A bispecific binding protein that binds to PD-1 and CTLA-4, comprising a first peptide comprising the amino acid sequence of SEQ ID NO:3 and a second peptide comprising the amino acid sequence of SEQ ID NO:4. (38) A bispecific binding protein that binds to PD-1 and CTLA-4, comprising a first peptide comprising the amino acid sequence of SEQ ID NO:5 and a second peptide comprising the amino acid sequence of SEQ ID NO:6. (39) A bispecific binding protein that binds to PD-1 and CTLA-4, comprising a first heavy chain comprising the amino acid sequence of SEQ ID NO:9, a first light chain comprising the amino acid sequence of SEQ ID NO:7, a second heavy chain comprising the amino acid sequence of SEQ ID NO:12, and a second light chain comprising the amino acid sequence of SEQ ID NO:4. (40) A bispecific binding protein that binds to PD-L1 and CTLA-4, comprising a first peptide comprising the amino acid sequence of SEQ ID NO: 14 and a second peptide comprising the amino acid sequence of SEQ ID NO: 15. (41) A bispecific binding protein that binds to PD-L1 and CTLA-4, comprising a first peptide comprising the amino acid sequence of SEQ ID NO: 16 and a second peptide comprising the amino acid sequence of SEQ ID NO: 17. (42) A bispecific binding protein that binds to PD-L1 and CTLA-4, comprising a first peptide comprising the amino acid sequence of SEQ ID NO: 18 and a second peptide comprising the amino acid sequence of SEQ ID NO: 19. (43) A bispecific binding protein that binds to PD-1 and TIM3, comprising a first peptide comprising the amino acid sequence of SEQ ID NO:22 or SEQ ID NO:89 and a second peptide comprising the amino acid sequence of SEQ ID NO:23 or SEQ ID NO:90. (44) A bispecific binding protein that binds to PD-1 and TIM3, comprising a first peptide comprising the amino acid sequence of SEQ ID NO:24 or SEQ ID NO:91 and a second peptide comprising the amino acid sequence of SEQ ID NO:23 or SEQ ID NO:92. (45) A bispecific binding protein that binds to PD-1 and TIM3, comprising a first heavy chain comprising the amino acid sequence of SEQ ID NO:9, a first light chain comprising the amino acid sequence of SEQ ID NO:7, a second heavy chain comprising the amino acid sequence of SEQ ID NO:27 or SEQ ID NO:30, and a second light chain comprising the amino acid sequence of SEQ ID NO:26 or SEQ ID NO:28. (46) A bispecific binding protein that binds to OX40 and PD-L1, comprising a first peptide comprising the amino acid sequence of SEQ ID NO: 34 and a second peptide comprising the amino acid sequence of SEQ ID NO: 32. (47) A bispecific binding protein that binds to OX40 and PD-L1, comprising a first peptide comprising the amino acid sequence of SEQ ID NO: 35 and a second peptide comprising the amino acid sequence of SEQ ID NO: 32. (48) A bispecific binding protein that binds to OX40 and PD-L1, comprising a first peptide comprising the amino acid sequence of SEQ ID NO: 36 or SEQ ID NO: 94, and a second peptide comprising the amino acid sequence of SEQ ID NO: 32 or SEQ ID NO: 93. (49) An antibody or antigen-binding fragment thereof that binds to TIM3, comprising a heavy chain comprising CDR1, CDR2 and CDR3, and a light chain comprising CDR1, CDR2 and CDR3, wherein the heavy chain CDR1 comprises SEQ ID NO: 79, the heavy chain CDR2 comprises SEQ ID NO: 80, the heavy chain CDR3 comprises SEQ ID NO: 81, and the light chain CDR1 comprises SEQ ID NO: 82, the light chain CDR2 comprises SEQ ID NO: 83, and the light chain CDR3 comprises SEQ ID NO: 84. (50) The antibody or antigen-binding fragment thereof according to (49), comprising a heavy chain variable region and a light chain variable region, wherein the heavy chain variable region comprises SEQ ID NO: 85, and the light chain variable region comprises SEQ ID NO: 86. (51) The antibody or antigen-binding fragment thereof according to (49), wherein the heavy chain comprises SEQ ID NO: 87 and the light chain comprises SEQ ID NO: 88. (52) A composition comprising the protein or antibody according to any one of (1) to (51) and a pharma- ceutically acceptable carrier. (53) A nucleic acid molecule comprising a nucleotide sequence encoding the protein or antibody according to any one of (1) to (51). (54) A vector comprising the nucleic acid molecule according to (53). (55) A host cell comprising the vector according to (54). (56) A method for treating or preventing cancer in a subject, the method comprising the step of administering to the subject a protein or antibody according to any one of (1) to (51). (57) The method according to (56), wherein the cancer is one or more of ovarian cancer, breast cancer, colon cancer, prostate cancer, cervical cancer, uterine cancer, testicular cancer, bladder cancer, head and neck cancer, melanoma, pancreatic cancer, renal cell carcinoma, and lung cancer. (58) A method for enhancing an immune response in a subject, the method comprising the step of administering to the subject a protein or antibody according to any one of (1) to (51).

[0316] [Sequence table] TIFF0007651022000044.tif199141TIFF0007651022000045.tif199141TIFF0007651022000046.tif199141TIFF0007651022000047.tif199141TIFF0007651022000048.tif199141TIFF0007651022000049.tif199141TIFF0007651022000050.tif199141TIFF0007651022000051.tif199141TIFF0007651022000052.tif199141TIFF0007651022000053.tif199141TIFF0007651022000054.tif199141TIFF0007651022000055.tif199141TIFF0007651022000056.tif199141TIFF0007651022000057.tif199141TIFF0007651022000058.tif199141TIFF0007651022000059.tif199141TIFF0007651022000060.tif199141TIFF0007651022000061.tif199141TIFF0007651022000062.tif199141TIFF0007651022000063.tif199141TIFF0007651022000064.tif199141TIFF0007651022000065.tif199141TIFF0007651022000066.tif199141 TIFF0007651022000067.tif199141TIFF0007651022000068.tif199141TIFF0007651022000069.tif199141TIFF0007651022000070.tif199141TIFF0007651022000071.tif199141TIFF0007651022000072.tif199141TIFF0007651022000073.tif199141TIFF0007651022000074.tif199141TIFF0007651022000075.tif199141TIFF0007651022000076.tif199141TIFF0007651022000077.tif199141TIFF0007651022000078.tif199141TIFF0007651022000079.tif199141TIFF0007651022000080.tif199141TIFF0007651022000081.tif199141TIFF0007651022000082.tif199141TIFF0007651022000083.tif199141TIFF0007651022000084.tif199141TIFF0007651022000085.tif199141TIFF0007651022000086.tif199141TIFF0007651022000087.tif199141TIFF0007651022000088.tif199141TIFF0007651022000089.tif199141TIFF0007651022000090.tif199141TIFF0007651022000091.tif199141TIFF0007651022000092.tif199141TIFF0007651022000093.tif199141TIFF0007651022000094.tif199141TIFF0007651022000095.tif199141TIFF0007651022000096.tif199141TIFF0007651022000097.tif199141TIFF0007651022000098.tif199141TIFF0007651022000099.tif199141TIFF0007651022000100.tif199141TIFF0007651022000101.tif199141TIFF0007651022000102.tif199141TIFF0007651022000103.tif199141TIFF0007651022000104.tif199141TIFF0007651022000105.tif199141TIFF0007651022000106.tif199141TIFF0007651022000107.tif199141TIFF0007651022000108.tif199141TIFF0007651022000109.tif199141TIFF0007651022000110.tif199141TIFF0007651022000111.tif199141TIFF0007651022000112.tif199141TIFF0007651022000113.tif199141TIFF0007651022000114.tif199141TIFF0007651022000115.tif199141TIFF0007651022000116.tif199141TIFF0007651022000117.tif199141TIFF0007651022000118.tif199141TIFF0007651022000119.tif199141TIFF0007651022000120.tif199141TIFF0007651022000121.tif199141TIFF0007651022000122.tif199141TIFF0007651022000123.tif199141TIFF0007651022000124.tif199141TIFF0007651022000125.tif199141TIFF0007651022000126.tif199141TIFF0007651022000127.tif199141TIFF0007651022000128.tif199141TIFF0007651022000129.tif199141TIFF0007651022000130.tif199141TIFF0007651022000131.tif199141TIFF0007651022000132.tif199141TIFF0007651022000133.tif199141TIFF0007651022000134.tif199141TIFF0007651022000135.tif199141TIFF0007651022000136.tif199141TIFF0007651022000137.tif199141TIFF0007651022000138.tif199141TIFF0007651022000139.tif199141TIFF0007651022000140.tif199141TIFF0007651022000141.tif199141TIFF0007651022000142.tif199141TIFF0007651022000143.tif199141TIFF0007651022000144.tif199141TIFF0007651022000145.tif199141TIFF0007651022000146.tif199141TIFF0007651022000147.tif199141TIFF0007651022000148.tif199141TIFF0007651022000149.tif199141TIFF0007651022000150.tif199141TIFF0007651022000151.tif199141TIFF0007651022000152.tif199141TIFF0007651022000153.tif199141TIFF0007651022000154.tif199141TIFF0007651022000155.tif199141TIFF0007651022000156.tif199141TIFF0007651022000157.tif199141TIFF0007651022000158.tif199141TIFF0007651022000159.tif199141TIFF0007651022000160.tif199141TIFF0007651022000161.tif199141TIFF0007651022000162.tif199141TIFF0007651022000163.tif199141TIFF0007651022000164.tif199141TIFF0007651022000165.tif199141TIFF0007651022000166.tif199141TIFF0007651022000167.tif199141TIFF0007651022000168.tif199141TIFF0007651022000169.tif199141TIFF0007651022000170.tif199141TIFF0007651022000171.tif199141TIFF0007651022000172.tif199141TIFF0007651022000173.tif199141TIFF0007651022000174.tif199141TIFF0007651022000175.tif199141TIFF0007651022000176.tif199141TIFF0007651022000177.tif199141TIFF0007651022000178.tif199141TIFF0007651022000179.tif199141TIFF0007651022000180.tif199141TIFF0007651022000181.tif199141TIFF0007651022000182.tif199141TIFF0007651022000183.tif199141.

Claims

1. (i) a first heavy chain comprising VH1-CDR1, VH1-CDR2 and VH1-CDR3, wherein said VH1-CDR1 comprises the amino acid sequence of residues 26 to 35 of SEQ ID NO:9, said VH1-CDR2 comprises the amino acid sequence of residues 50 to 66 of SEQ ID NO:9, and said VH1-CDR3 comprises the amino acid sequence of residues 99 to 111 of SEQ ID NO:9; (ii) a first light chain comprising VL1-CDR1, VL1-CDR2 and VL1-CDR3, wherein said VL1-CDR1 comprises the amino acid sequence of residues 24 to 41 of SEQ ID NO:7, said VL1-CDR2 comprises the amino acid sequence of residues 55 to 61 of SEQ ID NO:7, and said VL1-CDR3 comprises the amino acid sequence of residues 94 to 100 of SEQ ID NO:7; (iii) a second heavy chain comprising VH2-CDR1, VH2-CDR2 and VH2-CDR3, wherein the VH2-CDR1 comprises the amino acid sequence of residues 26-35 of SEQ ID NO:12, the VH2-CDR2 comprises the amino acid sequence of residues 50-66 of SEQ ID NO:12, and the VH2-CDR3 comprises the amino acid sequence of residues 99-114 of SEQ ID NO:12; and (iv) a second light chain comprising VL2-CDR1, VL2-CDR2 and VL2-CDR3, wherein the VL2-CDR1 comprises the amino acid sequence of residues 24 to 34 of SEQ ID NO: 4, the VL2-CDR2 comprises the amino acid sequence of residues 50 to 56 of SEQ ID NO: 4, and the VL2-CDR3 comprises the amino acid sequence of residues 89 to 95 of SEQ ID NO:

4. A bispecific binding protein that binds to PD-1 and CTLA-4, comprising:

2. The protein of claim 1 , comprising an unglycosylated Fc region.

3. The protein of claim 1, comprising an Fc region that is deglycosylated.

4. The protein of claim 1, comprising an Fc region that has hypofucosylation or is non-fucosylated.

5. A composition comprising the protein of claim 1 and a pharma- ceutically acceptable carrier.

6. A nucleic acid molecule comprising a nucleotide sequence encoding the protein of claim 1.

7. A vector comprising the nucleic acid molecule of claim 6.

8. An isolated host cell comprising the vector of claim 7.

9. A pharmaceutical composition for treating cancer in a subject, comprising the protein of claim 1.

10. 10. The pharmaceutical composition of claim 9, wherein the cancer is one or more of ovarian cancer, breast cancer, colon cancer, prostate cancer, cervical cancer, uterine cancer, testicular cancer, bladder cancer, head and neck cancer, melanoma, pancreatic cancer, renal cell carcinoma and lung cancer.

11. A pharmaceutical composition for enhancing an immune response in a subject, comprising the protein of claim 1.

Images