BR122024005562A2 - BISPECIFIC BINDING PROTEINS AND THEIR USES, ANTIBODY, COMPOSITION, NUCLEIC ACID MOLECULE, VECTOR AND HOST CELL - Google Patents

Abstract

"BISPECIFIC BINDING PROTEINS AND USES THEREOF, ANTIBODY, COMPOSITION, NUCLEIC ACID MOLECULE, VECTOR AND HOST CELL". The disclosure generally provides proteins that bind to two epitopes (e.g., a first and a second epitope) and that are bivalent for binding to each of the first and second epitopes. The disclosure also provides specific binding proteins, including antibodies, that bind to a target protein. The disclosure also provides compositions comprising such proteins, nucleic acid molecules encoding such proteins, and methods of producing such proteins. The disclosure provides methods for inducing an immune response in a subject, as well as methods for treating or preventing cancer in a subject by administering the proteins, nucleic acid molecules, and/or compositions to the subject.

Description

CROSS-REFERENCE TO RELATED ORDERS

[001] This application claims the benefit of U.S. Provisional Application No. 62/332,788, filed May 6, 2016, the disclosure of which is incorporated herein in its entirety by reference.

SEQUENCE LISTING

[002] This application is being filed together with a Sequence Listing in electronic format. The Sequence Listing is provided as a file titled IOBS_100_ST25.txt created on May 2, 2017, which is 244 kb in size. The information in the electronic format of the sequence listing is incorporated herein by reference in its entirety.

BACKGROUND

[003] Cancer continues to have a major negative impact on global health. Despite progress in cancer treatment, there remains an unmet medical need for more effective and less toxic therapies, especially for those patients with advanced disease or cancers that are resistant to existing therapeutics.

[004] The role of the immune system, in particular 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, in both early and advanced stages of the disease. However, tumor-specific T-cell responses are difficult to develop and maintain in cancer patients. The continued advances and successes of cancer immunotherapies, which stimulate or enhance innate immune responses against cancer, make these therapies an attractive treatment option when compared with therapies using non-specific chemotherapeutics and/or radiation.

[005] A number of molecular targets have been identified for their potential utility as immuno-oncology (IO) therapeutics against cancer. Some molecular targets that are being investigated for their therapeutic potential in the area of immuno-oncology therapy 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 receptor-3 (TIM3). Although some of these targets have been successfully explored (e.g., PD-1 and CTLA-4), many patients have not responded to the therapeutics that have been developed. And although a treatment regimen that includes higher doses and/or a combination of immunotherapies may be considered, these therapies may be associated with an increased risk of side effects, which tend to increase with higher doses and cumulative exposure, and appear to be additive when used with combination immunotherapies. Some common side effects include pituitary, thyroiditis, adrenal insufficiency, enterocolitis, dermatitis, pneumonitis, hepatitis, pancreatitis, motor and sensory neuropathies, and arthritis. Furthermore, because immunotherapeutics are typically associated with high costs, a therapy that includes a combination of immunotherapeutics may be cost-prohibitive for patients.

[006] As such, there remains a need to continue to identify candidate targets for IO therapeutics, to develop new therapeutics for existing targets, and to develop therapeutic strategies that avoid disadvantages of immunotherapies that are currently in use, including lack of patient response and the increased risk of side effects involved with combination treatment. IO therapeutics (e.g., binding proteins) that are bispecific for a combination of target molecules, particularly those that exhibit greater binding affinity for the target molecules when compared to the binding affinity for a combination of individual monospecific binding proteins, represent a particularly desirable class of molecules for therapeutic potential.

SUMMARY OF THE INVENTION

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

[008] In one aspect, the invention provides a protein, containing: 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 having a CH2 and a CH3 domain; wherein the Fc region includes BD2 in a solvent-exposed loop in the CH2 domain, the CH3 domain, or the interface of the CH2 and CH3 domains; and wherein the protein is bivalent for binding to each of the first and second epitopes.

[009] In another aspect, the invention provides a composition containing a protein or antibody according to any aspect of the present document and a pharmaceutically acceptable carrier.

[0010] In another aspect, the invention provides a method of treating or preventing cancer in a subject, the method involving administering to the subject the protein or antibody according to any aspect referred to herein (e.g., a human). In various embodiments, the cancer is one or more of ovarian cancer, breast cancer, colorectal cancer, prostate cancer, cervical cancer, uterine cancer, testicular cancer, bladder cancer, head and neck cancer, melanoma, pancreatic cancer, renal cell carcinoma, and lung cancer.

[0011] In another aspect, the invention provides a method of inducing an immune response in a subject, the method involving administering to the subject the protein or antibody according to any aspect referred to herein (e.g., a human).

[0012] In another aspect, the invention provides a nucleic acid molecule having a nucleotide sequence encoding a protein or an antibody according to any aspect herein.

[0013] In another aspect, the invention provides a vector containing a nucleic acid molecule according to any aspect of the present document.

[0014] In another aspect, the invention provides a host cell containing a vector according to any aspect of the present document.

[0015] 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.

[0016] 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.

[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:5, and a second peptide having the amino acid sequence of SEQ ID NO:6.

[0018] In one aspect, the invention provides a bispecific binding protein that binds to 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.

[0019] In one aspect, the invention provides a bispecific binding protein that binds to 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.

[0020] In another aspect, the invention provides a bispecific binding protein that binds to 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.

[0021] In another aspect, the invention provides a bispecific binding protein that binds to 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.

[0022] 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.

[0023] 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.

[0024] In one aspect, the invention provides a bispecific binding protein that binds to PD-1 and TIM3 having a first heavy chain having the amino acid sequence of SEQ ID NO:9, and 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.

[0025] 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.

[0026] 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.

[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: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.

[0028] In one 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, the light chain CDR3 comprises SEQ ID NO:84.

[0029] In other aspects, the invention provides a composition having a bispecific binding protein and a pharmaceutically acceptable carrier; a nucleic acid molecule having a nucleotide sequence encoding a bispecific binding protein; methods of treating or preventing cancer in a subject by administering a bispecific binding protein; and methods of enhancing an immune response in a subject by administering a bispecific binding protein.

[0030] In various embodiments of any aspect outlined herein, the Fc region comprises BD2 in a solvent-exposed loop at the amino acid sequence in the CH2 domain, the CH3 domain, or the interface of the CH2 and CH3 domains.

[0031] In various embodiments of any aspect outlined herein, the solvent-exposed loop includes an amino acid sequence of the CH2 domain. In particular embodiments, the solvent-exposed loop includes the amino acid sequence ISRTP (SEQ ID NO: 39).

[0032] In various embodiments of any aspect outlined herein, the solvent-exposed loop includes an amino acid sequence of the CH3 domain. In particular embodiments, the solvent-exposed loop includes the amino acid sequence SNG.

[0033] In various embodiments of any aspect outlined herein, the solvent-exposed loop includes an amino acid sequence of the CH2 domain and the CH3 domain. In particular embodiments, the protein of claim 7, wherein the solvent-exposed loop comprises the amino acid sequence AKGQP (SEQ ID NO: 40).

[0034] In various embodiments of any aspect outlined herein, BD2 is or includes a single chain variable fragment (scFv).

[0035] In various embodiments of any aspect outlined herein, BD1 is or includes a binding domain that is one or more of a Fab domain, a scFv, a single domain antibody, and an antibody variable domain. In particular embodiments, BD1 includes a Fab domain.

[0036] In various embodiments of any aspect outlined herein, the Fab domain is linked to the Fc region via an antibody hinge region. In certain embodiments, the Fc region is or includes a domain that is one or more of an Fc region from an IgG1, IgG2, IgG3, IgG4, IgA, IgM, IgE, or IgD. In particular embodiments, the Fc region comprises a variant Fc region. In some embodiments, the Fc region is glycosylated, deglycosylated, and/or is afucosylated or has reduced fucosylation.

[0037] In various embodiments of any aspect outlined herein, the protein further includes a protein linker L1 between BD2 and the Fc region. In various embodiments of any aspect outlined herein, the protein further includes a first protein linker, L1, and a second protein linker, L2, between BD2 and the Fc region. In various embodiments of any aspect outlined herein, BD2 is associated with the Fc region via a protein linker L1. In various embodiments of any aspect outlined herein, BD2 is associated with the Fc region via two protein linkers, L1 and L2. 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).

[0038] In various embodiments of any aspect outlined herein, the protein includes a chimeric heavy chain having the following polypeptide domains, from the N-terminus to the C-terminus: VH1-CH1-CH2(N-term)-BD2-CH2(C-term)-CH3; and BD1 includes a Fab domain; wherein VH1 includes a heavy chain variable domain of the Fab domain and CH1 includes the heavy chain constant domain 1 of the Fab.

[0039] In various embodiments of any aspect outlined herein, the protein includes a chimeric heavy chain having the following polypeptide domains, from the N-terminus to the C-terminus: VH1-CH1-CH2-BD2-CH3; and BD1 includes a Fab domain; wherein VH1 comprises a heavy chain variable domain of the Fab domain and includes the heavy chain constant domain 1 of the Fab.

[0040] In various embodiments of any aspect outlined herein, the protein includes a chimeric heavy chain having the following polypeptide domains, from the N-terminus to the C-terminus: VH1-CH1-CH2-CH3(N-term)-BD2-CH3(C-term); and BD1 includes a Fab domain; wherein VH1 includes a heavy chain variable domain of the Fab domain, and CH1 includes the heavy chain constant domain 1 of the Fab.

[0041] In various embodiments of any aspect outlined herein, BD2 is or includes a scFv. In particular embodiments, the scFv includes, from N-terminus to C-terminus: the VL2-linker of the VH2-polypeptide or the VH2-linker of the VL2-polypeptide; wherein VH2 includes the heavy chain variable domain of the scFv and VL2 includes the light chain variable domain of the scFv.

[0042] In various embodiments of any aspect outlined herein, the protein further includes a protein linker L1 between BD2 and the Fc region. In various embodiments of any aspect outlined herein, the protein further includes a first protein linker, L1, and a second protein linker, L2, between BD2 and the Fc region.

[0043] In various embodiments of any aspect outlined herein, BD2 is associated via a linker (L1) to the CH2 domain, the CH2 domain, or the interface of the CH2 and CH3 domains of the Fc region.

[0044] In various embodiments of any aspect outlined herein, BD2 is associated via two protein linkers, L1 and L2 to the CH2 domain, the CH2 domain, or the interface of the CH2 and CH3 domains of the Fc region. In various embodiments, L1 and L2 are independently selected from protein linkers having a length of 1-25 amino acids. In particular 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).

[0045] In various embodiments of any aspect outlined herein, the first and second epitopes are different. In various embodiments of any aspect outlined herein, the first and second epitopes are the same.

BRIEF DESCRIPTION OF THE DRAWINGS

[0046] For the purpose of illustrating the disclosure, certain aspects of the disclosure are represented in the drawings. However, the disclosure is not limited to the precise arrangements and apparatus of the aspects represented in the drawings.

[0047] Figures 1A-1F depict a general schematic diagram of certain exemplary proteins described herein. The CH2 and CH3 regions are illustrated in Figures 1A-1C using PyMOL and identify the loop regions of the solvent-exposed surface as spheres. Figure 1A depicts the loop in the CH2 region; Figure 1B depicts the loop at the CH2-CH3 interface; and Figure 1C depicts the loop at the CH3 region. Illustrative constructs are illustrated in Figures 1D-1F that include representative BD1 and BD2 domains as Fab and scFv domains, respectively. Figure 1D depicts BD1 bound at the hinge region and BD2 bound to solvent-exposed loops in the CH2 region. Figure 1E depicts BD1 bound at the hinge region and BD2 bound to solvent-exposed loops at the CH2-CH3 interface. Figure 1F depicts BD1 bound at the hinge region and BD2 bound to solvent-exposed loops in the CH3 region.

[0048] Figures 2A-2C provide an expanded view of the solvent accessible loop sequences in CH2, at the CH2-CH3 interface, and in CH3 as described herein. Examples of constructs incorporating a BD2 (scFv) are included in each of Figures 2A-2C. Figure 2A illustrates the representative loop sequence ISRTP (SEQ ID NO:39) identified in the CH2 loop upstream of the CH2-CH3 interface. Figure 2B illustrates the representative loop sequence AKGQP (SEQ ID NO:40) at the CH2-CH3 interface. Figure 2C illustrates the representative loop sequence SNG in the CH3 region downstream of the CH2-CH3 interface.

[0049] Figure 3 demonstrates the concurrent 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.

[0050] Figure 4 shows a schematic of the proposed mechanism for PD-1/CTLA-4 blockade.

[0051] Figure 5 shows the results of an octet binding assay demonstrating concurrent binding of a DuetMab construct targeting PD-1 and CTLA-4.

[0052] Figures 6A-D [show that PD-1/CTLA-4 bispecific binding proteins inhibit both the 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 shows the results of the PD-1 reporter assay. Figure 6D shows the results of the CTLA-4 reporter assay.

[0053] Figure 7 shows the results of a SEB assay showing that PD-1/CTLA-4 DuetMab and BiS5Ab have equivalent activity in the staphylococcal enterotoxin B (SEB) assay.

[0054] Figures 8A-B show the activity of PD-1/CTLA-4 DuetMab in SEB assays compared to isotype controls and parental mAb.

[0055] Figures 9A-B show the activity of PD-1/CTLA-4 BiS5Ab compared to PD-1/CTLA-4 DuetMab in SEB assays.

[0056] Figures 10A-C [show that PD-1/CTLA-4 DuetMab and BiS5Ab have equivalent activity in the mixed lymphocyte reaction (MLR) assay. Figure 10A is a schematic of the assay (n = 4 donors; 2 independent experiments).

[0057] Figures 11A-D show the activity of PD-1/CTLA-4 DuetMab in MLR assays compared to isotype controls (n = 2 donors; 1 experiment).

[0058] Figures 12A-D show the activity of PD-1/CTLA-4 DuetMab in MLR assays compared to parental mAb controls (n = 2 donors; 1 experiment).

[0059] Figures 13A-D show the activity of PD-1/CTLA-4 DuetMab in MLR assays compared to competitor antibodies (n = 2 donors; 1 experiment).

[0060] Figure 14 shows the study design for a single-dose pharmacokinetic/pharmacodynamic (PK/PD) study in cynomolgus monkeys.

[0061] Figures 15A-B show that PD-1/CTLA-4 DuetMab showed clear pharmacodynamics (PD) in cynomolgus monkeys.

[0062] Figure 16A-B shows the T cell-dependent antibody response (TDAR) in cynomolgus monkeys dosed with PD-1/CTLA-4 DuetMab (MEDI5752) and PD-1/CTLA-4 BiS5Ab (MEDI8500).

[0063] Figure 17 shows a model system for studying PD-1/CTLA-4 bispecific molecules in which stable CHO cells express varying levels of human PD-1 and/or CTLA-4.

[0064] Figures 18A-C show that PD-1/CTLA-4 DuetMab binds concurrently to PD-1 and CTLA-4 on the surface of the same cell.

[0065] Figures 19A-C show an experiment to determine whether cooperative binding differs relative to a combination of anti-PD-1 and anti-CTLA-4 antibodies at CTLA-4 saturation in cells overexpressing PD-1.

[0066] Figures 20A-D show that the parental PD-1 and CTLA-4 monoclonal antibodies bind to and occupy their target receptor without a measurable effect on the untargeted receptor.

[0067] Figures 21A-D show that PD-1/CTLA-4 DuetMab saturates CTLA-4 on CHO cells overexpressing PD-1 at ~250-fold lower concentrations compared to a monoclonal antibody combination.

[0068] Figures 22A-F show that PD-1/CTLA-4 DuetMab saturates CTLA-4 in CHO cells overexpressing PD-1 at ~500-fold lower concentrations compared to cells expressing CTLA-4 alone.

[0069] Figures 23A-B show that PD-1/CTLA-4 DuetMab preferentially binds in cis to PD-1 and CTLA-4 on the surface of the same cell. Treme in Figure 23B is a CTLA-4 mAb.

[0070] Figures 24A-D show binding and internalization of PD-1/CTLA-4 DuetMab and parental monoclonal antibodies to cultured T cells. PD-1/CTLA-4 DuetMab has tremelimumab internalization properties.

[0071] Figure 25A shows a schematic of the internalization assay. Figure 25B shows that PD-1/CTLA-4 DuetMab assumes internalization properties of tremelimumab in stable CHO cells expressing 10-fold excess PD-1.

[0072] Figure 26 demonstrates the concurrent binding of BiS2, BiS3, and BiS5 constructs targeting PD-L1 and CTLA-4. Trace A11 shows Bis2 PD-L1CTLA-4, trace B11 shows Bis3 PD-L1/CTLA-4, and trace C11 shows Bis5 PD-L1/CTLA-4.

[0073] Figure 27A demonstrates the concurrent binding of the BiS3 construct targeting PD-1 and TIM3 (clone 62, wild type). Figure 27B demonstrates the concurrent binding of a DuetMab construct targeting PD-1 and TIM3 (clone 62, wild type).

[0074] Figures 28A-28C provide a summary of the cell killing activity of TIM3 monospecific and PD-1_TIM3PD-1 bispecific constructs in a cell killing assay. Figure 28A shows brightfield images of wells co-cultured at 18 h. The combination of anti-TIM-3+anti-PD1 or TIM-3/PD-1 bispecific formats enhances tumor cell killing and increases T cell activation, as assessed by reduced adherent cells and improved T cell detachment (clustering). Figure 28B shows assessment of the viability of dye uptake by tumor cells after 18 h of co-culture with melanoma-specific CD8+ T cells. Figure 28C shows IFNγ secretion after 18 h co-culture. Bispecific constructs generally demonstrate better killing activity, with the DuetMab format exhibiting the most robust killing activity.

[0075] Figure 29 demonstrates concurrent binding of a PD-1/TIM3 DuetMab construct having a TIM3 arm sequence that is an affinity mature variant of clone 62.

[0076] Figure 30 shows 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.

[0077] Figure 31 shows TIM3 and PD-1 expression data in the activated T cell clone (DMF4).

[0078] Figures 32A-C depict the results of a CMV antigen retrieval assay showing that PD-1/TIM3 bispecific antibodies, including BiS3, BiS5, and DuetMab, demonstrate improved activity compared to isotype treatment (3 donors (1-2 replicates per treatment/per donor), 1 experiment).

[0079] Figures 33A-D show that PD-1/TIM3 bispecific antibodies, including BiS3, BiS5, and DuetMab, enhanced interferon (IFNY) at concentrations above 8 nM in the mixed lymphocyte reaction (MLR) assay (2-4 replicate wells per treatment/1 donor pair/1 of 2 independent experiments).

[0080] Figures 34A-C show the results of a PD-1 reporter assay (dual cell system) using PD-1/TIM3 bispecific antibodies, including BiS3, BiS5, and DuetMab. All bispecific formats demonstrated similar activity to the parental IgG1 LO115 (compilation of 3-5 independent experiments/3 biological replicates per treatment).

[0081] Figure 35 shows the results of an octet assay using BiS2 and BiS3 OX40/PD-L1 bispecific molecules.

[0082] Figures 36A-B show the results of a SEB assay using BiS2 and BiS3 OX40/PD-L1 bispecific molecules.

[0083] Figure 37 shows the results of a PD-L1 assay using BiS2 and BiS3 OX40/PD-L1 bispecific molecules.

[0084] Figure 38 shows the results of a CMV Ag recovery assay using BiS2 and BiS3 OX40/PD-L1 bispecific molecules.

[0085] Figure 39 shows the results of an octet binding assay demonstrating concurrent binding of the OX40(SLR)/PD-L1 BiS5 construct targeting PD-L1 and OX40.

[0086] Figures 40A-F show the binding of bispecific constructs targeting PD-L1 and OX40 to CHO cells expressing human or cynomolgus OX40 and PD-L1/B7H1.

[0087] Figure 41 shows the binding of bispecific constructs targeting PD-L1 and OX40 to Jurkat OX40 reporter cells, NCI H358 cells, CHOK1 B7H1(PD-L1)/OKT3, and HEK CD32A cells, measured by flow cytometry (HyperCyt).

[0088] Figures 42A-B show the results of a PD-L1 reporter assay using OX40/PD-L1 bispecific molecules.

[0089] Figures 43A-B show the results of an OX40 reporter assay in HEK CD32a cells using OX40/PD-L1 bispecific molecules.

[0090] Figures 44A-B show the results of an OX40 reporter assay in PDL1 overexpressing CHOK cells using OX40/PD-L1 bispecific molecules.

[0091] Figure 45 shows PD-L1-mediated OX40 agonism with tumor cells using OX40/PD-L1 bispecific molecules.

[0092] Figures 46A-D show results indicating that no agonism was detected with NCI H358 PD-L1 KO cells using OX40/PD-L1 bispecific molecules in controls for the OX40/tumor cell assay.

[0093] Figures 47A-D show the results of a SEB assay using OX40/PD-L1 bispecific molecules.

[0094] Figure 48A shows a schematic of the Treg suppression assay experiment to test OX40/PD-L1 bispecific molecules. Figures 48B-C show Treg suppression based on bispecific molecule binding.

[0095] Figure 49 depicts the results of the Treg suppression assay using the OX40/PD-L1 bispecific molecules.

[0096] Figure 50 shows the results of the Treg suppression assay using the OX40/PD-L1 bispecific molecules.

[0097] Figures 51A-B show the design of a mixed leukocyte reaction (MLR) assay experiment to test OX40/PD-L1 bispecific molecules.

[0098] Figures 52A-E show the results of the MLR assay using the OX40/PD-L1 bispecific molecules.

[0099] Figure 53 shows that OX40/PD-L1 bispecific molecules BiS2 and BiS5 mediate antibody-dependent cell-mediated cytotoxicity (ADCC) of natural killer (NK) cells against CHO cells expressing PD-L1 or OX40.

[00100] Figure 54 shows that OX40/PD-L1 bispecific molecules BiS2 and BiS5 mediate antibody-dependent cell-mediated cytotoxicity (ADCC) of NK cells against CHO cells expressing PD-L1 and OX40.

[00101] Figure 55 shows that the BiS2 and BiS5 OX40/PD-L1 bispecific molecules enhanced CD107a mobilization of NK cells against CHO cells expressing PD-L1 and OX40 in antibody-dependent cell-mediated cytotoxicity (ADCC).

[00102] Figure 56 shows that bispecific molecules BiS2 and BiS5 OX40/PD-L1 mediate antibody-dependent cell-mediated cytotoxicity (ADCC) of NK cells against activated allogeneic T cells.

[00103] Figures 57A-B show that BiS5 OX40/PD-L1 enhanced CD107a mobilization of NK cells from two different donors against activated allogeneic T cells in antibody-dependent cell-mediated cytotoxicity (ADCC).

[00104] Figure 58 shows a study design to compare PK/PD of OX40/PD-L1 bispecific molecules.

[00105] Figure 59 shows a comparison of serum concentration time profiles for PD-L1/OX40 bispecific molecules in cynomolgus monkeys.

[00106] Figure 60 shows the depletion of serum soluble PD-L1 by PD-L1/OX40 bispecific molecules.

[00107] Figures 61A-F provide a summary of the pharmacodynamic data for the PD-L1/OX40 bispecific molecules. Baseline defined as the mean of Day-5 and Day-0 pre-dose.

[00108] Figure 62 depicts a schematic of a PD-1/OX40 BiS2 IgG4P monoclonal antibody (mAb).

[00109] Figure 63 depicts a potential mechanism of action of the PD-1/OX40 BiS2 mAb.

[00110] Figure 64 depicts the concurrent binding activity for two different lots of PD-1(LO115)/OX40 BiS2 mAb to human PD1-His and OX40-Fc.

[00111] Figure 65A depicts a schematic of an OX40 reporter assay. Figure 65B shows the results of the OX40 reporter assay using PD1 LO115 mAb, OX40 mAb, a control mAb, and PD1(LO115)/OX40 BiS2 mAb.

[00112] Figure 66A depicts a schematic of a PD-1/PD-L1 reporter assay. Figure 66B shows the results of the PD1/PD-L1 reporter assay using PD1 LO115 mAb, OX40 mAb, a control mAb, and PD1(LO115)/OX40 BiS2 mAb.

[00113] Figure 67 shows the results of a SEB assay using the BiS2 variant of the bispecific molecule PD-1(LO115)/OX40 and controls.

[00114] Figure 68 shows the results of a CMV antigen recall assay using the BiS2 variant of the PD-1(LO115)/OX40 bispecific molecule and controls.

[00115] Figure 69 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.

[00116] Figure 70 shows the serum concentration-time profiles of PD1(LO115)/OX40 BiS2 mAb following a single IV dose in cynomolgus monkeys. Data represent the mean ± standard deviation of 3 males/group. The LLOQ (5 ng/mL) is shown by the dotted line. PKC = pharmacokinetic concentration; LLOQ = lower limit of quantification.

[00117] Figure 71 shows the percentage of Ki67-positive CD4+ and CD8+ memory T cells after single IV administration of PD-1(LO115)/OX40 BiS2 mAb in cynomolgus monkeys. Data represent mean ± standard deviation of 3 males/group. Left panel A represents CD4+ memory T cells and right panel shows CD8+ memory T cells. IV = intravenous.

[00118] Figure 72 shows a representative standard curve for quantification of PD-1/OX40 in cynomolgus monkey serum.

[00119] Figures 73A-73E provide illustrative DSC thermograms of the bispecific binding protein disclosed herein ("BiS5") relative to a different format of BiS ("BiS4") at different pH values. Figure 73A illustrates the effect of pH on the thermal stability of BiS4. Figure 73B illustrates the effect of pH on the thermal stability of BiS5. Figure 73C is a representative curve-fitted DSC thermogram for Bis4 with Tinitial, Tm1, Tm2, and Tm3. Figure 73D is a representative curve-fitted DSC thermogram for Bis5 with Tinitial, Tm1, Tm2, and Tm3. Figure 73E is a graph representing the effect of pH on Tinitial, Tm1, Tm2, and Tm3 for BiS4 and BiS5 formats.

[00120] Figures 74A and 74B depict HP-SEC analysis of samples at pH 7.5 before and after storage at 40°C for 4 weeks. Figure 74A provides an overlay plot of SEC chromatograms of BiS4 and BiS5 before and after heat stress, the solid lines correspond to BiS4 and BiS5 samples at time zero (no stress) and the dotted lines correspond to BiS4 and BiS5 samples incubated at 40°C for 4 weeks (with stress). Figure 74B provides a bar graph depicting the effect of pH 7.5 on different species (monomer, fragments, and aggregates) of BiS4 and BiS5 as measured at day zero and after 4 weeks at 40°C.

[00121] Figures 75A-75C provide kinetic graphs showing the effect of pH 7.5 on short-term and accelerated storage stability at 40°C for BiS4 (triangular trace) and BiS5 (circular trace). Figure 75A shows the percentage of monomer remaining, 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. The data presented are from single vial analysis.

[00122] Figures 76A-76C depict pH rate profile plots for BiS4 (triangles) and BiS5 (circles). Figure 76A shows the effect of different pH conditions on the rate of monomer loss at 40°C. Figure 76B shows the effect of different pH conditions on the rate of fragmentation at 40°C. Figure 76C shows the effect of different pH conditions on the rate of aggregation at 40°C.

[00123] Figures 77A and 77B represent a further analysis of the fragmentation of BiS4 and BiS5. Figure 77A shows that at pH 7.5 and 40°C (at time = 0), neither molecule exhibits appreciable fragmentation. Figure 77B shows that under the same conditions as Figure 77A, but after 2 weeks of storage at 40°C, appreciable fragmentation is observed for BiS4 and minimal fragmentation for BiS5.

[00124] Figure 78 depicts the analysis of BiS4 and BiS5 fragmentation (left panel) and aggregation (right panel) as a function of pH. Both formats have lower fragmentation and aggregation at lower pH (5.5), while BiS5 has superior performance at both pH values for both fragmentation and aggregation.

[00125] Figure 79 depicts DSC thermograms of various BiS5 bispecific binding proteins disclosed herein for constructs (left panel) A, B, C, and D, as well as (right panel) E, F, G, and H. Constructs A and E include the scFv in IS-RTP; B and F include the scFv in AK-GQP; C and G include the scFv in S-NG; and D and H include the scFv in SN-G. The scFv for constructs A, B, C, and D is 2F4 (IgG is LC10), while the scFv for E, F, G, and H is LC10 (IgG is 2F4). The various TM values are associated with the following domains, TM1 = CH2/scFv; TM2 = Fab; TM3 = CH3.

[00126] Figure 80 depicts a representative data set for FcRn binding to the bispecific binding protein disclosed herein (constructs D and H). The location of the scFv in the CH2-CH3 domain (i.e., within the ISTRP loop) may have an effect on FcRn binding activity.

[00127] Figure 81 depicts a representative data set for FcyR binding to the bispecific binding protein disclosed herein (constructs E, G, and H with FcyRIIIa-158V). The inset reflects the same data as the main figure, re-normalized upon injection of FcyRIIIa-158V. All constructs were able to bind to FcyRs with an observable difference in affinities based on the location of the scFv domains.

[00128] Figure 82 shows that an intact ISRTP loop is important for FcRn binding of the illustrative bispecific binding constructs (e.g., A and E). Introduction of the N3 loop does not compensate for disruption of the ISRTP loop (BiS5E+N3). An scFv introduced into an N3 loop inserted into an IgG1 Fc renders the IgG unable to bind FcRn. Time on the x-axis is measured in seconds.

[00129] Figure 83 depicts the general schematic structural format of each construct BiS1, BiS2, BiS3, BiS4, and BiS5. The denotations of "k1" and "k2" indicate the fragmentation patterns used in the kinetic analysis discussed in Example 3.

[00130] Figure 84 depicts the fragmentation rate of each BiS1, BiS2, BiS3, BiS4, and BiS5 as a function of pH.

[00131] Figure 85 depicts the aggregation rate of each BiS1, BiS2, BiS3, BiS4, and BiS5 as a function of pH.

[00132] Figure 86 depicts the monomer loss rate of each of BiS1, BiS2, BiS3, BiS4, and BiS5 as a function of pH.

[00133] Figure 87 depicts a representation of the fragmentation pattern and the correspondence with the peaks in the HPSEC chromatograms for each of BiS1, BiS2, BiS3, BiS4 and BiS5.

[00134] Figure 88 depicts a representative analysis of the fragmentation pattern under reducing conditions.

[00135] Figure 89 represents the structural arrangement of BiS5 under reducing and non-reducing conditions.

[00136] Figure 90 represents the format of the SEB test.

DETAILED DESCRIPTION

[00137] Before continuing to describe the present disclosure in additional detail, it is to be understood that this disclosure is not limited to specific compositions or process steps, which as such 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 references unless the context clearly dictates otherwise.

[00138] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of skill in the art to which this invention relates. For example, the 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 the Oxford Dictionary Of Biochemistry And Molecular Biology, Revised, 2000, Oxford University Press, provide one skilled in the art with a general dictionary of many of the terms used in this invention.

[00139] 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 Commission on Biochemical Nomenclature. Nucleotides may similarly be referred to by their commonly accepted one-letter codes.

[00140] The numbering of amino acids in the variable domain, complementarity determining regions (CDRs), and framework regions (FR), of an antibody follows, unless otherwise indicated, 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). Using this numbering system, the actual linear amino acid sequence may contain fewer or more amino acids corresponding to a shortening of, or insertion of, a FR or CDR in the variable domain. For example, a heavy chain variable domain may include a single amino acid insertion (residue 52a according to Kabat) after residue 52 of H2 and inserted residues (e.g., residues 82a, 82b, and 82c, etc. according to Kabat) after residue 82 of the heavy chain FR. Kabat residue numbering can be determined for a given antibody by aligning the homology regions of the antibody sequence with a numbered "standard" Kabat sequence. Maximal alignment of framework residues often requires the insertion of "spacer" residues into the numbering system to be used for the Fv region. In addition, the identity of individual residues at any given Kabat site number may vary from antibody chain to antibody chain due to species or allelic divergence.

[00141] As used herein, the terms "antibody" and "antibodies", also known as immunoglobulins, encompass 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., PCT Publication WO2009018386, PCT Application No. PCT/US2012/045229, incorporated herein in their entirety by reference), BiSAbs, human antibodies, humanized antibodies, camelized antibodies, single-chain Fvs (scFv), single-chain antibodies, single-domain antibodies, domain antibodies, Fab fragments, F(ab')2 fragments, antibody fragments that exhibit the desired biological activity (e.g., antigen-binding moiety), disulfide-linked Fvs (dsFv), and antibodies anti-idiotypic (anti-Id) antibodies (including, for example, anti-Id antibodies to the antibodies of the invention), intrabodies, and epitope-binding fragments of any of the foregoing. 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 a protein fused to an Fc domain. Immunoglobulin molecules can 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 can be derived from any mammal, including, but not limited to, humans, monkeys, pigs, horses, rabbits, dogs, cats, mice, etc., or other animals such as birds (e.g., chickens). CTLA-4

[00142] Cytotoxic T lymphocyte-associated protein 4 (CTLA-4) is expressed on activated T cells and serves as a co-inhibitor to keep T cell responses in check following CD28-mediated T cell activation. CTLA-4 is believed to regulate the extent of early activation of naive and memory T cells following TCR engagement and to be part of a central inhibitory pathway that affects 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 largely restricted to antigen-presenting cells, T cells, and other immune mediator cells. Antagonistic 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 infections and tumors and to upmodulate an adaptive immune response has been proposed (see, U.S. Patent 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. Publication No. 20150104409, incorporated herein in their entirety by reference).

[00143] PD-L1

[00144] Programmed death ligand 1 (PD-L1) is also part of a complex system of receptors and ligands that are involved in the control of T cell activation. In normal tissue, PD-L1 is expressed on T cells, B cells, dendritic cells, macrophages, mesenchymal stem cells, bone marrow-derived mast cells, as well as various non-hematopoietic cells. Its normal function is to regulate the balance between T cell activation and tolerance through 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 multiple sites to help tumors escape detection and elimination by the host immune system. PD-L1 is expressed on a wide range of cancers at high frequency. In some cancers, PD-L1 expression has been associated with reduced survival and poor prognosis. Antibodies that block the interaction between PD-L1 and its receptors are capable of alleviating PD-L1-dependent immunosuppressive effects and enhancing antitumor T cell cytotoxic activity in vitro. Durvalumab is a human monoclonal antibody directed against human PD-L1 that is capable of blocking 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 enhance an adaptive immune response has been proposed (see, U.S. Patent Nos. 8,779,108 and 9,493,565, incorporated herein in their entirety by reference). PD-1

[00145] 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," EM-BO J. 11 :3887-3895.

[00146] 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 the negative regulation of the immune system following activation by PDL-1 or PDL-2 binding (Martin-Orozco, N. et al. (2007) "Inhibitory Costimulation and Anti-Tumor Immunity," Semin. Cancer Biol. 17(4):288-298) and functions as an inducer of cell death (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, S.K. et al. (2005) "The Balance of Immune Responses: Costimulation Verse Coinhibition," J. Molec. Med. 83: 193-202). This process is exploited in many tumors through the overexpression of PD-L1, leading to a suppressed immune response.

[00147] PD-1 is a well-validated target for immune-mediated therapy in oncology, with positive results from clinical trials in the treatment of melanoma and non-small cell lung cancer (NSCLC), among others. Antagonistic inhibition of the PD-1/PD-L-1 interaction increases T-cell activation, increasing 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 enhance an adaptive immune response has been proposed (see, U.S. Patent Nos. 7,521,051; 7,563,869; 7,595,048). OX40

[00148] OX40 (CD134; TNFRSF4) is a tumor necrosis factor receptor found primarily on activated CD4+ and CD8+ T cells, regulatory T cells (Treg), and natural killer (NK) cells (Croft et al., 2009, Immunol Rev. 229: 173-91). OX40 has a known endogenous ligand, OX40 ligand (OX40L; CD152; TNFSF4), which exists in a trimeric form and can cluster OX40, resulting in potent cell signaling events within T cells. Id. Signaling through OX40 in activated CD4+ and CD8+ T cells leads to enhanced cytokine production, granzyme and perforin release, and expansion of effector and memory T cell clusters (Jensen et al., 2010, Semin Oncol. 37:524-32). Furthermore, OX40 signaling in Treg cells inhibits Treg expansion, turns off Treg induction, and blocks Treg suppressive function (Voo et al., 2013, J Immunol. 191:3641-50; Vu et al., 2007, Blood. 110:2501-10).

[00149] Immunohistochemical studies and initial flow cytometric analyses have shown that OX40 is expressed on T cells infiltrating a wide range 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). Although not wishing to be bound by theory, OX40 expression on tumor-infiltrating lymphocytes correlates with longer survival in several human cancers, suggesting that OX40 signals may play a role in establishing an antitumor immune response (Ladanyi et al., 2004, Clin Cancer Res. 10:521-30; Petty et al., 2002, Am J Surg. 183: 512-8).

[00150] In a variety of nonclinical mouse tumor models, OX40 agonists, including antibodies and OX40 ligand fusion proteins, have been used successfully, with promising results (Kjaergaard et al., 2000, Cancer Res. 60:5514-21; Ndhlo- vu et al., 2001, J Immunol. 167:2991-9; Weinberg et al., 2000, J Immunol. 164:2160-9). Costimulation of T cells by OX40 promoted antitumor activity that in some cases was durable, providing long-lasting protection against subsequent tumor challenge (Weinberg et al., 2000, J Immunol. 164:2160-9). Inhibition of Treg cells and costimulation of effector T cells have been shown to be necessary for tumor growth inhibition of OX40 agonists (Piconese et al., 2008, J Exp Med. 205:825–39). Many strategies and technologies have been explored to enhance the antitumor effect of OX40 agonist therapy through combinations with vaccines, chemotherapy, radiotherapy, 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 to treat infections and tumors and to positively shape an adaptive immune response has been proposed (see, U.S. Publication No. 20160137740, incorporated herein in its entirety by reference). TIM3

[00151] The T-cell inhibitory receptor Tim-3 (T-cell immunoglobulin and mucin-containing domain-3) plays a role in the regulation of anti-tumor immunity, as it is expressed on CD4+ helper 1 (Th1) cells and CD8+ cytotoxic T 1 (Tc1) cells. It was initially identified as an inhibitory T-cell receptor, acting as an immune checkpoint receptor that functions specifically to limit the duration and magnitude of Th1 and Tc1 T cell responses. Further research has identified that the Tim-3 pathway can cooperate with the PD-1 pathway to promote the development of a severe dysfunctional phenotype in CD8+ T cells in cancer. It has also been expressed on regulatory T (Treg) cells in certain cancers. Given the involvement of the TIM3 pathway in key immune cell populations that are immunosuppressed in some cancers, it represents an attractive candidate for immuno-oncology therapy. See, Anderson, A.C., Cancer Immunol Res., (2014) 2:393-398; and Ferris, R.L., et al., J Immunol. (2014) 193:1525-1530.

Bispecific Binding Proteins

[00152] The addition of multiple binding sites to a molecule having specificity for a single binding domain can greatly enhance the capabilities (e.g., therapeutic, diagnostic, etc.) of the molecule. For example, a bispecific antibody can bind to more than one region of the same target biomolecule, conferring greater specificity than a monospecific polypeptide that binds to only one epitope on a target. Alternatively, a bispecific antibody can bind to multiple target biomolecules, such as targets that are present in a complex, or targets for which sequestration and/or clustering is desired. In a third scenario, the same bispecific antibody can perform distinct functions at any given time, depending on the location and/or expression of its target molecules.

[00153] Novel binding proteins are described herein. One such embodiment of these novel binding proteins is referred to as "DuetMab". DuetMab has the following basic structure: an Fc region having a modified heavy chain, wherein 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, wherein 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 having a second heavy chain; and the corresponding modified second light chain, wherein the modified heavy chain is directly linked to the corresponding modified light chain, and in a separate target binding arm, the second heavy chain is directly linked to the corresponding second light chain, and wherein the substituted cysteine of the modified heavy chain resulting from the substitution of the native non-cysteine amino acid for the cysteine amino acid and the substituted cysteine of the corresponding modified light chain resulting from the substitution of the native cysteine amino acid for the cysteine amino acid can form a disulfide bond. Disclosure relating to DuetMab can be found, for example, in U.S. Patent No. 9,527,927, which is incorporated in its entirety herein by reference.

[00154] Other additional configurations of these novel binding proteins are referred to as "BiSAb" or "BiSAbs". Schematic representations of exemplary BiSAbs, as well as specific examples of particular BiSAbs are provided herein. More generally, a BiSAb is a polypeptide containing two binding moieties, each of which binds to an epitope (e.g., binding moiety 1 binds to a first epitope and binding moiety 2 binds to a second epitope). The basic BiSAb is bivalent in binding to each of the two epitopes (e.g., the polypeptide comprises two binding moiety 1's ("BD1" or "BU1") and two binding moiety 2's ("BD2" or "BU2")). Thus, when binding moiety 1 and 2 bind to different epitopes, the BiSAb possesses the multispecificity of a conventional bispecific antibody and the bivalency of a conventional antibody molecule. In embodiments where binding moieties 1 and 2 bind to the same epitope, the BiSAb has the monospecificity of a conventional antibody, but is tetravalent. In addition to the binding moieties, the BiSAbs also include linker polypeptides and an Fc portion. The disclosure relates to a broad set of bispecific binding proteins, such as the BiSAb and proteins comprising a BiSAb core, that target molecules that modulate the immune response. Generally, the novel binding protein platforms and exemplary bispecific binding proteins (BiSAbs) described herein comprise binding moieties/domains, linker polypeptides, and an Fc portion. The disclosure also provides nucleic acid molecules encoding such BiSAbs, as well as vectors and host cells that include such nucleic acids and that can be used in methods of producing and using such BiSAbs. BiSAbs, binding proteins comprising a BiSAb core, and the various portions of BiSAbs are described herein in greater detail.

[00155] In some aspects, a BiSAb may comprise two pairs of heavy light chains derived from a specific binding protein (i.e., antibody), wherein the heavy and light chains each comprise a variable region (e.g., VL and VH), which together form a first binding unit, and wherein the heavy chains each comprise a second binding unit (e.g., an Fc-linked scFv or Fab domain). When the first and second binding units bind to different epitopes, each heavy light chain pair is bispecific and the two pairs together are bivalent for each epitope. When the first and second binding units bind to the same epitope, each heavy light chain pair is monospecific and the two pairs together are tetravalent for the epitope. In some aspects, the two heavy light chain pairs are identical. In some aspects, the two heavy light chain pairs are not identical.

[00156] In specific embodiments, the domains of the BiSAbs can be based on known immunoglobulin domains. Immunoglobulin molecules such as monoclonal antibodies (mAbs) are widely used as diagnostic and therapeutic agents, and methods for engineering binding fragments of mAbs are well known in the art. Monoclonal antibodies, like all immunoglobulin molecules, are comprised of heavy chain and light chain peptide subunits, each of which includes variable and constant domains that confer binding specificity (variable domain) and isotype (constant domain).

[00157] The BiSAbs disclosed herein may have an overall structure similar to a conventional antibody, but are distinguishable by the presence of an additional binding moiety that is linked to a location within the Fab domain, linked to a location remote from the Fab domain, and within the hinge or Fc regions (e.g., within 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 may still retain numerous desirable properties of conventional antibodies, such as the ability to bind to FcRn and the ability to bind to C1q and Fcy receptors (e.g., indicative of the ability to mediate antibody- and complement-dependent cytotoxicity).

[00158] The binding domains described herein may comprise antigen-binding fragments containing only portions of a mAb molecule, such as Fab, F(ab')2, Fab', scFv, di-scFv, sdAb fragments, as such fragments have found use as diagnostic or therapeutic agents. Additionally, specific residues in the variable domains may be altered in order to improve the binding specificity and/or stability of antibodies and antibody fragments. Other residues not directly involved in antigen binding may be substituted in order to "humanize" regions of non-human antibodies and reduce the immunogenicity of the mAb.

[00159] Although BiSAbs differ from conventional antibodies in that, for example, they are bivalent for binding to two different epitopes (or tetravalent for binding to a single epitope), many of the portions of BiSAbs are derived from or analogous to portions of conventional antibodies. Any mAb domains and/or fragments known in the art may be used in the BiSAbs described herein. In particular, the BiSAb may comprise Fab and/or scFv fragments, or variants thereof. Examples of non-limiting scFv variants include, but are not limited to, tandem di-scFvs, tandem tri-scFvs, diabodies, and tri(a)bodies.

[00160] The disclosure generally relates to novel binding proteins, of which BiSAbs are an illustrative example. Additional examples are binding proteins comprising a BiSAb core as well as one or more additional binding units and/or binding proteins comprising an extended BiSAb core. It should be understood that whenever BiSAbs or features of BiSAbs are described herein, such description generally applies to the novel binding proteins of the disclosure, regardless of whether such binding proteins include two binding units or more than two binding units. Accordingly, the term BiSAb is exemplary of binding proteins described herein and, where the context permits, any such reference to BiSAb may also be used to describe binding proteins comprising a BiSAb core.

New BiSAb structural platform.

[00161] In one aspect, the disclosure provides BiSAb binding proteins having a structural platform comprising domains that are generally illustrated by the schematic diagrams in Figures 1A-1F. These diagrams are illustrative and thus insertion between additional residues is 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 an IgG1 that was modeled using PyMOL, and illustrate several exemplary BiSAbs of the disclosure. Three surface-exposed loops were identified at or near the CH2-CH3 interface that were likely capable of supporting insertion of a second binding moiety (e.g., a 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 ISRTP loop (SEQ ID NO:39) identified in the CH2 region near the CH2-CH3 interface. Figure 1D also shows a representative IS-scFv-RTP construct, in which a scFv is inserted between S and R of the ISRTP loop. Figure 1B is a schematic diagram of the representative AKGQP (SEQ ID NO:40) loop identified at the CH2-CH3 interface. Figure 1E shows a representative AK-scfv-GQP construct, in which a scFv is inserted between K and G of the AKGQP loop. Figure 1C is a schematic diagram of the representative SNG loop identified in the CH3 region downstream of the CH2-CH3 interface. Figure 1F also shows the representative S-scfv-NG construct, in which a scFv is inserted between S and N of the SNG loop. Examples herein provide illustration of constructs oriented as SN-scFv-G, in which a scFv is inserted between N and G of the SNG loop.

[00162] Thus, one aspect of the disclosure relates to a BiSAb comprising two identical heavy-light chain pairs, wherein each heavy-light chain pair is bispecific and the two identical pairs are in bivalent sets for each epitope. Each heavy-light chain pair comprises a binding domain (BD) which may comprise a Fab domain that binds to a first epitope (binding unit 1), a second binding domain (BD2) that binds to a second epitope (or binding unit 2 which may be, for example, a scFv), and an Fc region. In some embodiments, the second binding domain may be associated with the Fab domain. In some embodiments, the Fc region of the BiSAb may be associated with the second binding domain (BD2) that binds to a second epitope (binding unit 2; depicted as a scFv in Figures 1D-1F).

[00163] In some embodiments, the disclosure provides a BiSAb having an overall platform structure comprising two chimeric heavy chains, each comprising a heavy chain variable region (VH1), a heavy chain constant region (CH1), a hinge or polypeptide binding region, an Fc region comprising a CH2 domain, and a CH3 domain, wherein a second binding domain (BD2), optionally flanked on one or both sides by a polypeptide linker (L1 and/or L2) is associated with solvent-exposed loops in the Fc region following (i) the CH2 region, (ii) the interface of the CH2 and CH3 region, or (iii) the CH3 region. The BiSAb of this aspect of the disclosure also comprises two conventional antibody light chains, each comprising a light chain variable region (VL1) and a light chain constant region (CL), which is part of the first binding domain (BD1). The binding domain (BD2) of the particular BiSAb illustrated in Figures 1D-1F is a scFv.

[00164] Figures 1D-1F provide a useful schematic representation of a BiSAb, which may also be referred to herein as a BiSAb "core." The polypeptide chain as shown in Figure 1D 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 a first polypeptide linker), linker moiety 2 (such as VL2 and VH2 of an scFv), another optional linker (e.g., L2 or a second polypeptide linker), the remaining C-terminal CH2 domain, and a CH3 domain. Since such a heavy chain may include BD2 having alternative binding proteins and/or traditional light chain regions, it is referred to herein as a chimeric heavy chain. A BiSAb comprises two such chimeric heavy chains, and these may be the same or different. Note that the variable heavy chain (VH) domain for binding unit 1 is referred to as VH1. In certain aspects, this is a variable heavy chain of a Fab that binds to a first epitope. Likewise, the variable light chain (VL) domain for binding unit 1 is referred to as VL1. In certain aspects, this is a variable light chain of a Fab that binds to a first epitope. In contrast, the domains for binding unit two are denoted with the number "2", such as VH2 and VL2 for aspects where binding unit 2 is a scFv that binds to a second epitope.

[00165] Likewise, the polypeptide chain as shown in Figure 1E 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 a first polypeptide linker), linker moiety 2 (such as VL2 and VH2 of a scFv), another optional linker (e.g., L2 or a second polypeptide linker), and a CH3 domain. In this embodiment, the BiSAb comprises a second binding domain, illustrated as scFv associated with the Fc in sequence at the interface of the CH2 and CH3 regions.

[00166] The polypeptide chain as shown in Figure 1F 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), linker unit 2 (such as VL2 and VH2 of an scFv), another optional linker (e.g., L2 or a second polypeptide linker), the CH3 domain.

[00167] In these embodiments, the BiSAbs typically include an antibody hinge region typical or modified in the chimeric heavy chain sequences. Non-limiting examples of amino acid sequences that contain a hinge region include: EPKSCDKTHTCPPCP (SEQ ID NO:44); EPKSCDKT (SEQ ID NO:45); EPKSCGKT (SEQ ID NO:46); EPKSC (SEQ ID NO:47).

[00168] Having described the general format for aspects relating to the particular structural platform for certain BiSAb molecules disclosed herein, various exemplary functional moieties and properties of the disclosed BiSAbs are described in greater detail below. In other embodiments, the disclosure contemplates and provides other BiSAb binding proteins comprising alternative structural shapes and arrangements that are briefly described herein as well as in other disclosures (see, e.g., U.S. Publication No. 20090155275 and U.S. Patent No. 9,580,509) which are incorporated herein by reference.

Connection Units

[00169] The BiSAbs of the disclosure comprise at least two binding units or binding domains (binding unit/domain 1 and binding unit/domain 2). In certain aspects, each binding unit binds to a different epitope, either to different epitopes located on the same target molecule or to epitopes on different targets. Since each binding unit of a BiSAb is present as a pair (there are two binding units 1s and two binding units 2s), the BiSAbs exhibit bivalent binding to each epitope. It will be understood from the teachings herein that where each binding unit binds to the same epitope, a BiSAb will exhibit tetravalent binding to the epitope.

[00170] In certain aspects, the first binding moiety is a Fab fragment, e.g., a Fab fragment of a conventional monoclonal antibody or a recombinantly produced antigen-binding fragment comprising 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 interconnected through one or more disulfide bonds, such as, for example, through a suitable antibody hinge region. The Fab binds to a first epitope.

[00171] In certain aspects, 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 aspects, the BiSAb containing the Fab derived from or based on the sequence of a conventional monoclonal antibody retains one or more functional activities of the conventional antibody (e.g., retains at least 80% or more (80%, 85%, 90%, 95%, 97%, 98%, 99%, or 100%) of a functional activity). For example, in certain aspects, the BiSAb containing such a Fab retains one or more of the antigen affinity, inhibitory activity, immune system modulation activity, activation or induction of an immune response, and/or cell (e.g., cancer cell) killing activity of the conventional antibody.

[00172] In certain aspects, the BiSAbs of the disclosure comprise binding unit 2, and binding unit 2 comprises a binding domain that binds to a second epitope. Binding unit 2 (or binding domain 2 (BD2)) may be associated with the BiSAb using any suitable strategy. As used herein, a BD2 that is "associated" with the BiSAb (e.g., within the Fc region in some embodiments, within the Fab region in other embodiments) means that the two molecules have an interaction with each other such that the BD2 retains the orientation for binding to the target and association with the Fc portion or the Fab portion of the BiSAb structure. Examples of such interactions include covalent binding via amino acid linkers, covalent binding via recombinant expression of BD2 within the Fab region, within the hinge region, or within the Fc region at CH2, CH3, or at the interface of CH2 and CH3, or at the CH4 region, and non-covalent interactions such as van der Waals interactions and hydrogen bonding within these same regions. Non-limiting examples of binding domains (or "BD" or "binding moieties") encompassed by the scope of the 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 to two distinct epitopes. In one aspect, epitope-binding domains useful in constructing multispecific epitope-binding domains of the disclosure are exemplified in US20100298541 and US20130079280 which are incorporated herein by reference for all purposes.

[00173] In certain aspects, the BiSAb may comprise a binding domain that includes a scFv. Thus, in certain aspects, binding unit 2 comprises a scFv. It should be understood that a scFv encompasses a polypeptide chain comprising a heavy chain variable domain (VH) linked to a light chain variable domain (VL) via a flexible polypeptide linker. Figures 1D-1F show a schematic of an exemplary BiSAb, wherein the BD (represented herein as binding unit 2) is a scFv having domains as described herein that may be designated as VL2 and VH2. In some aspects, 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 or different antibodies. In some aspects, a VH or VL of the scFv can comprise one or more CDRs that bind to a target of interest, while the remainder of the VH or VL domain is derived from a different antibody or is synthetic. In some aspects, the scFv comprises at least one CDR from an antibody, e.g., an antibody known in the art to bind to a target of interest. In some aspects, the scFv comprises at least two CDRs from a given antibody. In some aspects, the scFv comprises at least three CDRs from a given antibody. In some aspects, the scFv comprises at least four CDRs from a given antibody. In some aspects, the scFv comprises at least five CDRs from a given antibody. In some aspects, the scFv comprises at least six CDRs from a given antibody.

[00174] In certain 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 that has 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 that have binding affinity or specific binding activity for the molecular targets disclosed herein. For example, the BiSAbs disclosed herein may comprise 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 andTIM3; and TIM3. BiSAbs that include binding domains that bind to specific target combinations illustrated in the Examples and include the non-limiting combinations of PD-1/CTLA-4; PD-L1/CTLA-4; PD-1/OX40; PD-L1/OX40; and PD-1/TIM3.

[00175] In some additional embodiments, the BiSAbs exhibit a binding activity (e.g., binding affinity and/or binding specificity) for at least one of the target molecules that is greater than the binding activity of the parental monospecific binding sequence used to generate the BiSAb. In similar embodiments, the BiSAbs can exhibit a binding activity (e.g., binding affinity and/or binding specificity) for both target molecules that is greater than the binding activity of both of the parental monospecific binding sequences used to generate the BiSAbs. In yet a further embodiment, the BiSAbs can exhibit a binding activity (e.g., binding affinity and/or binding specificity) for both of the target molecules that is greater than the binding activity for the combination of the parental monospecific binding sequences used to generate the BiSAbs. The enhanced binding properties of BiSAbs relative to the parent monospecific binding sequences, whether alone or in combination, provide unexpected advantages over the use of monospecific therapeutic agents that target the same molecules, even when used in combination.

[00176] 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, can 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, can comprise a heavy chain region variable sequence (HCv) and a light chain region variable sequence (LCv), or a portion of an HCv and an LCv comprising the CDR1, CDR2, and CDR3 sequences of the heavy and light chain sequences. In still other embodiments, the antibody, or an antigen-binding fragment thereof, can comprise the CDR1, CDR2, and CDR3 sequences of the heavy and light chain sequences. In some embodiments, the antibody can be a chimeric, humanized, or human antibody. In some embodiments, the antibody can be a polyclonal or monoclonal antibody. In additional embodiments, the antibody is a monoclonal antibody.

[00177] In some embodiments, domains comprising all or a portion of an antigen-binding region of such "parent" antibodies as disclosed above can be used to generate the bispecific binding proteins (BiSAbs) that are disclosed herein. The non-limiting embodiments that are illustrated in the Examples provide a description related to how antibody sequences can be identified and combined so as to produce a BiSAb that exhibits bispecific binding to a combination of molecular targets.

[00178] Several methodologies may be used alone or in combination in order to improve the stability of a BiSAb comprising a scFv molecule. One potential methodology that may be used, alone or in combination with one or more of the other methodologies described herein, is the manipulation of the length and/or composition of the linker connecting the scFv domains in order to stabilize the scFv moiety.

[00179] Another potential methodology that can be used is to introduce at least two amino acid substitutions (also referred to as modifications or mutations) into the VH and/or VL domains of the scFv so as to promote disulfide bond formation (see, for example, 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 can be used alone or in combination with one or more of the other methodologies described in this document.

[00180] In certain aspects, one or more mutations can be introduced into each of the VH and VL domains of the scFv so as to promote interchain disulfide bond formation between the VH and VL domains upon expression of a BiSAb comprising an scFv. In another aspect, the two mutations are introduced into the same chain domain. In a certain aspect, the two mutations are introduced into different chains. In certain aspects, multiple complementary mutations are introduced so as to promote the formation of multiple disulfide bonds or other stabilizing interactions. In certain aspects, a cysteine is introduced so as to promote disulfide bond formation. Examples of amino acids that may be mutated to cysteine 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 foregoing numbering is based on Kabat numbering identifying the position relative only to VH2 and VL2 of the scFv (and not relative to the position of the amino acid in the full-length BiSAb sequence or to the SEQ ID NOs provided herein). Exemplary combinations of amino acid positions that can 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 aspects, amino acid 44 of VH and amino acid 100 of VL are mutated to cysteines.

[00181] Another method that may be used, alone or in combination with one or more of the other methods described herein, is selection of the order of the domains of the scFv. In certain aspects, the orientation of the VH domain relative to the VL domain is optimized for stability. In certain aspects, the scFv is in the VH-linker-VL orientation. In certain aspects, the scFv is in the VL-linker-VH orientation. In embodiments related to the novel BiSAb format disclosed herein, the orientation of the domains in the scFv may determine how the scFv associates with the Fc portion of the BiSAb. Although this is described in greater detail below in the context of polypeptide linkers. Briefly, however, given that the BD (e.g., a scFv) is linked to CH2, CH3, or at the interface of CH2 and CH3 by optional polypeptide linkers (L1) and (L2), the order of the domains determines which portion of the scFv is linked to L1 and which portion of the scFv is linked to L2.

[00182] An additional method that may be used, alone or in combination with the other methods described herein, is to introduce one or more stabilizing mutations by mutating one or more surface residues of the scFv. In some aspects, one, two, three, four, five, six, or more than six residues are mutated in one or both of the VH and/or VL domains of the scFv. In certain aspects, the changes are made only in the VH domain of the scFv. In certain aspects, the changes are made only in the VL domain of the scFv. In certain aspects, the changes are made in both the VH and 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 certain aspects, one or more of the changes is a conservative amino acid substitution from the residue present in the unmodified, parent scFv. In other aspects, one or more of the changes is a non-conservative amino acid substitution from a residue present in the unmodified parent scFv. When multiple substitutions are made in either one or both of the VH or VL domains of the scFv, each substitution is independently a conservative or non-conservative substitution. In certain aspects, all of the substitutions are conservative substitutions. In certain aspects, all of the substitutions are non-conservative. In certain aspects, at least one of the substitutions is conservative. In certain aspects, at least one of the substitutions is non-conservative.

[00183] Yet another method that may be used, alone or in combination with other methods, is to introduce one or more amino acid substitutions by mutating one or more residues present in the VH and/or VL domain of the scFv to match the most frequent residue at said particular position of a consensus sequence of the VH and/or VL domain of known antibodies. In certain aspects, the substitutions are introduced at one, two, three, four, five, six or more than six positions in one or both of the VH domain and/or the VL domain 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 certain aspects, one or more of the changes in the given consensus sequence is a conservative amino acid substitution from the residue present in the unmodified VH and/or VL sequence. In other aspects, one or more of the changes represents a non-conservative amino acid substitution from the residue present in the unmodified VH and/or VL sequence. When multiple substitutions are made in either one or both of the VH or VL domains of the scFv, each substitution is independently a conservative or non-conservative substitution. In certain respects, all of the substitutions are conservative substitutions. In certain respects, all of the substitutions are non-conservative substitutions. In certain respects, at least one of the substitutions is conservative. In certain respects, at least one of the substitutions is non-conservative.

[00184] It should be noted that any of the modifications described as useful for modifying or stabilizing the scFv portion may be applied to modify the Fab portion. For example, the variable domains of the Fab portion of a BiSAb may be modified so as to improve stability, antigen binding, and the like. Furthermore, both the Fab and the scFv portion may be modified to reduce immunogenicity.

[00185] In certain aspects, the binding unit 2 (BD) is a scFv, e.g., a scFv derived from a conventional monoclonal antibody comprising a variable light chain (VL2) and a variable heavy chain (VH2) interconnected by a flexible linker, such as a glycine-serine linker. Optionally, the variable light and variable heavy chains of the scFv may be further interconnected via one or more disulfide bonds and, as described above, may include one or more mutations or variations. The scFv binds to a second epitope. In certain aspects, the second epitope is different from the first epitope bound by binding unit 1. In other aspects, the second epitope is the same as the first epitope bound by binding unit 1. In certain aspects, the scFv is derived from or based on the sequence of a conventional monoclonal antibody, such as a conventional murine, humanized, or human antibody. In certain aspects, the BiSAb containing the scFv derived from or based on the sequence of a conventional monoclonal antibody retains one or more functional activities of the conventional antibody (e.g., retains at least 80% or more (80%, 85%, 90%, 95%, 97%, 98%, 99%, or 100%) of a functional activity). For example, in certain aspects, the BiSAb containing such a scFv retains one or more of the antigen affinity, inhibitory activity, or cell killing activity of the conventional antibody.

[00186] In certain aspects, a BiSAb comprises any binding unit 1s and binding unit 2s described herein, including any combination of a binding unit 1 and a binding unit 2. For example, in certain aspects, the disclosure provides a polypeptide comprising a Fab that binds to a particular target (e.g., that binds to an epitope on a particular target), such as a Fab comprising a particular amino acid sequence or encoded by a particular nucleotide sequence, and/or a scFv that binds to a particular target (e.g., that binds to an epitope on a particular target), such as a scFv comprising a particular amino acid sequence or encoded by a particular nucleotide sequence.

[00187] As described in detail above, linker moiety 1 and linker moiety 2 may be associated with the BiSAb via covalent linkage via a linker polypeptide 1 (L1, L2). Generally, the linkage is via the chimeric heavy chain of the BiSAb, such that the cross-linking is via the heavy chain CH2 domain, the heavy chain CH3 domain, or at the interface of the heavy chain CH2 and CH3 domains, or in some embodiments within the hinge region or Fab domain. L1 and L2 may vary in length and sequence independently of one another, and exemplary configurations are described herein. The disclosure contemplates BiSAbs comprising any combination of linker moieties and linker polypeptides, including any combination of the specific linker moieties that bind the desired target(s) and specific L1 and L2 polypeptide ligands described herein.

2. Region Fc

[00188] As used herein, "Fc region" encompasses domains derived from the constant region of an immunoglobulin, preferably a human immunoglobulin, including a fragment, analog, variant, mutant or derivative of the constant region. Suitable immunoglobulins include IgG1, IgG2, IgG3, IgG4 and other classes such as IgA, IgD, IgE and IgM. The Fc region can be a native sequence Fc region or an altered Fc region. The Fc region of an immunoglobulin generally comprises two constant domains, a CH2 domain and a CH3 domain, and optionally comprises a CH4 domain. The BiSAbs of the disclosure include an Fc region comprising a CH2 domain and a CH3 domain.

Altered Fc regions

[00189] Altered Fc regions (also referred to herein as "variant Fc regions") can be used to alter the effector function and/or half-life of a BiSAb of the disclosure. One or more alterations can be made to the Fc region in order to alter the functional and/or pharmacokinetic properties of the molecules. Such alterations can result in decreased or increased Clq binding and complement-dependent cytotoxicity (CDC) or FcyR binding, for IgG, and antibody-dependent cellular cytotoxicity (ADCC), or antibody-dependent cell-mediated phagocytosis (ADCP). The present disclosure encompasses BiSAbs in which alterations have been made to fine-tune the effector function, either by increasing or decreasing the function or by providing a desired effector function. Thus, in one aspect of the disclosure, the BiSAbs comprise a variant Fc region (i.e., Fc regions that have been altered as discussed below). BiSAbs comprising a variant Fc region are also referred to herein as "variant Fc BiSAbs." As used herein, "native" refers to the unmodified parent sequence and BiSAbs comprising a native Fc region are referred to herein as "native Fc BiSAbs." Variant Fc BiSAbs can be generated by various methods well known to one of skill in the art. Non-limiting examples include, isolating antibody coding regions (e.g., from hybridoma) and making one or more desired substitutions in the Fc region. Alternatively, the antigen-binding portion (e.g., variable regions) of a BiSAbs can be subcloned into a vector encoding a variant Fc region. In one aspect, the variant Fc region exhibits a similar level of induction of effector function compared to the native Fc region. In another aspect, the variant Fc region exhibits a higher induction of effector function compared to the native Fc. In another aspect, the variant Fc region exhibits a lower induction of effector function compared to the native Fc. Some specific aspects of variant Fc regions are detailed below. Methods for measuring effector function are well known in the art.

[00190] In general, effector function is modified through changes in the Fc region, including but not limited to amino acid substitutions, amino acid additions, amino acid deletions, and changes in post-translational modifications on Fc amino acids (e.g., glycosylation). The methods described below can be used to fine-tune the effector function of a BiSAb of the disclosure, a ratio of the binding properties of the Fc region to the FcR (e.g., affinity and specificity), resulting in a BiSAb with the desired properties.

[00191] It is understood that the Fc region, as used herein, includes the polypeptides comprising the constant region of an antibody molecule excluding the first immunoglobulin constant region 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 a portion of the N-terminal flexible hinge to these domains. For IgA and IgM, Fc may include the J chain. For IgG, Fc comprises the immunoglobulin domains Cgamma2 and Cgamma3 (CY2 and CY3) and optionally a lower hinge portion between Cgamma (CY1) and Cgamma2 (CY2). Although the boundaries of the Fc region may vary, as used herein, the human IgG heavy chain Fc region comprises residues A231 at its carboxyl terminus, where numbering is in accordance with the EU index as set forth in Kabat. Fc may refer to this region in isolation, or to this region 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 positions 270, 272, 312, 315, 356 and 358 of IgG1 as numbered by the EU index, and thus there may be slight differences between the sequence presented and sequences in the prior art.

[00192] In one aspect, the present disclosure encompasses Fc variant BiSAbs that have altered binding properties for an Fc ligand (e.g., an Fc receptor, C1q) relative to a native Fc Bi-SAb. Examples of binding properties include, but are not limited to, binding specificity, equilibrium dissociation constant (Kd), dissociation and association rates (koff and kon, respectively), binding affinity and/or avidity. It is known in the art that the equilibrium dissociation constant (Kd) is defined as koff/kon. In certain aspects, a BiSAb comprising an Fc variant region with a low Kd may be more desirable than a BiSAb with a high Kd. However, in some cases, the value of kon or koff may be more relevant than the value of Kd. One skilled in the art can determine which kinetic parameter is most important for a given application. For example, a modification that reduces binding to one or more positive regulators (e.g., FcyRIIIA) and/or enhances binding to an inhibitory Fc receptor (e.g., FCYRIIB) would be suitable to reduce ADCC activity. Thus, the ratio of binding affinities (e.g., the ratio of equilibrium dissociation constants (Kd)) for different receptors may indicate whether the ADCC activity of an Fc variant BiSAb of the disclosure is increased or decreased. Furthermore, a modification that reduces binding to C1q would be suitable to reduce or eliminate CDC activity.

[00193] In one aspect, Fc variant BiSAbs exhibit an altered binding affinity for one or more Fc receptors, including but not limited to FcRn, FCYRI (CD64), including the FCYRIA, FCYRIB, and FCYRIC isoforms; FCYRII (CD32 including the FCYRIIA, FCYRIIB, and FCYRIIC isoforms); and FCYRIII (CD16, including the FCYRIIIA and FcyRIIIB isoforms), compared to a native Fc BiSAb.

[00194] In certain aspects, an Fc variant BiSAb has a higher affinity for an Fc ligand. In other aspects, an Fc variant BiSAb has a lower affinity for an Fc ligand relative to the native Fc BiSAb.

[00195] In a specific aspect, an Fc variant BiSAb has improved binding to the Fc receptor FcyRIIIA. In another specific aspect, an Fc variant BiSAb has improved binding to the Fc receptor FcyRIIIB. In a further specific aspect, an Fc variant BiSAb has improved binding to both the Fc receptors FcyRIIIA and FcyRIIB. In certain aspects, Fc variant BiSAbs that have improved binding to FcyRIIIA do not have a concomitant increase in binding to the FcyRIIB receptor compared to an Fc BiSAb. In a specific aspect, an Fc variant BiSAb has reduced binding to the Fc receptor FcyRIIIA. In a further specific aspect, an Fc variant BiSAb has reduced binding to the Fc receptor FcyRIIB. In another specific aspect, an Fc variant BiSAb has improved binding to the Fc receptor FcRn. In yet another specific aspect, a variant Fc BiSAb exhibiting altered affinity for FCYRIIIA and/or FCYRIIB has improved binding to the Fc receptor FcRn. In yet another specific aspect, a variant Fc BiSAb exhibiting altered affinity for FCYRIIIA and/or FCYRIIB has altered binding to C1q relative to a native Fc BiSAb.

[00196] In another aspect, Fc variant BiSAbs exhibit improved or reduced affinities for C1q relative to a native Fc BiSAb. In yet another specific aspect, a Fc variant BiSAb exhibiting altered affinity for C1q has improved binding to the Fc receptor FcRn. In yet another specific aspect, a Fc variant BiSAb exhibiting altered affinity for C1q has altered binding to FcyRIIIA and/or FcyRIIB relative to a native Fc BiSAb.

[00197] It is recognized that antibodies are capable of directing attack and destroying targeted antigens through various processes collectively known in the art as antibody effector functions. One such process, known as "antibody-dependent cell-mediated cytotoxicity" or "ADCC," refers to a form of cytotoxicity in which secreted Ig binds to Fc gamma receptors (FcyRs) present on certain cytotoxic cells (e.g., Natural Killer (NK) cells, neutrophils, and macrophages), allowing these cytotoxic effector cells to specifically bind to an antigen-bearing target cell and subsequently kill the target cell with cytotoxins. Specific high-affinity IgG antibodies directed to the surface of target cells "arm" the cytotoxic cells and are necessary for such killing. Target cell lysis is extracellular, requires direct cell-to-cell contact, and does not involve complement. Another process encompassed by the term effector function is complement-dependent cytotoxicity (hereinafter referred to as "CDC"), which refers to a biochemical event of antibody-mediated target cell killing by the complement system. The complement system is a complex system of proteins found in normal blood plasma that combines with antibodies to destroy pathogenic bacteria and other foreign cells. Yet another process encompassed by the term effector function is antibody-dependent cell-mediated phagocytosis (ADCP), which refers to a cell-mediated reaction in which nonspecific cytotoxic cells expressing one or more effector ligands recognize antibody bound to a target cell and subsequently cause phagocytosis of the target cell.

[00198] It is contemplated that the Fc variant BiSAbs are characterized by in vitro functional assays to determine one or more FcyR-mediated effector cell functions. In certain aspects, the Fc variant BiSAbs have similar binding properties and effector cell functions in in vivo models (such as those described and disclosed herein) as compared to in vitro-based assays. However, the present disclosure does not exclude Fc variant BiSAbs that do not exhibit the desired phenotype in in vitro-based assays but exhibit the desired phenotype in vivo.

[00199] The serum half-life of proteins comprising Fc regions can be increased by increasing the binding affinity of the Fc region for FcRn. The term "antibody half-life" as used herein means a pharmacokinetic property of an antibody that is a measure of the mean survival time of antibody molecules following their administration. Antibody half-life can be expressed as the time required to eliminate 50 percent of a known quantity of immunoglobulin from the patient's (or other mammal's) body or from a specific compartment thereof, for example, as measured in serum, i.e., circulating half-life, or in other tissues. Half-life can vary from one immunoglobulin, or class of immunoglobulin, to another. In general, an increase in antibody (or BiSAb) half-life results in an increase in the mean residence time (MRT) in circulation for the administered BiSAb.

[00200] Increasing the half-life allows for a reduction in the amount of drug administered to a patient, as well as a reduction in the frequency of administration. To increase the half-life of the BiSAb, a salvage receptor-binding epitope (especially an antibody fragment) can be incorporated into the BiSAb as described in U.S. Patent No. 5,739,277, for example. As used herein, the term "salvage receptor-binding epitope" refers to an epitope of the Fc region of an IgG molecule (e.g., IgG1, IgG2, IgG3, or IgG4) that is responsible for increasing the in vivo serum half-life of the IgG molecule. Alternatively, BiSAbs of the disclosure with increased half-life can be generated by modification of amino acid residues identified as involved in the interaction between the Fc and the FcRn receptor (see, for examples, U.S. Patent Nos. 6,821,505 and 7,083,784). Furthermore, the half-life of the BiSAbs of the disclosure can be increased by conjugation with PEG or albumin through techniques widely used in the art.

[00201] It is contemplated that either the insertion of additional binding domains into the Fc region, as described herein, and/or subsequent binding by antigen, may affect Fc activity. For example, binding antigen may increase or decrease binding affinity and activity for FcgRs, C1q, and FcRn. This would create an antigen-dependent switch so as to modulate various antibody-dependent processes. In one aspect, antigen binding may decrease interaction with FcRn, allowing a free BiSAb to interact with FcRn and have a normal half-life, but allow rapid clearance/cellular internalization of BiSAb-Ag complexes. Furthermore, this could allow BD2 antigen-mediated interactions to have an effect on the clearance of BD1-bound antigens. In a further aspect, the BiSAb could comprise the Fc region directly inserted into BD2 (Fc-BD2).

[00202] In one aspect, the present disclosure provides Fc variants, wherein the Fc region comprises a modification (e.g., amino acid substitutions, amino acid insertions, amino acid deletions) at one or more positions 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, 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 as numbered by index EU as defined in Kabat. Optionally, the Fc region may comprise a modification at additional and/or alternative positions known to one of skill in the art (see, e.g., U.S. Patents 5,624,821; 6,277,375; 6,737,056; 7,083,784; 7,317,091; 7,217,797; 7,276,585; 7,355,008). In addition, useful amino acid positions and specific substitutions are exemplified in Tables 2 and 6-10 of U.S. Pat. No. 6,737,056; the tables set forth in Figure 41 of U.S. Pat. No. 2006/024298; the tables set forth in Figures 5, 12, and 15 of U.S. Pat. No. 2006/235208; the tables presented in Figures 8, 9 and 10 of US 2006/0173170 and the tables presented in Figures 8-10, 13 and 14 of WO 09/058492.

[00203] In a specific aspect, the present disclosure provides a variant Fc, wherein the Fc region comprises at least one substitution selected from the group consisting of 221K, 221Y, 225E, 225K, 225W, 228P, 234D, 234E, 234N, 234Q, 234T, 234H, 234Y, 234I, 234V, 234F, 235A, 235D, 235R, 235W, 235P, 235S, 235N, 235Q, 235T, 235H, 235Y, 235I, 235V, 235E, 235F, 236E, 237L, 237M, 237P, 239D, 239E, 239N, 239Q, 239F, 239T, 239H, 239Y, 240I, 240A, 240T, 240M, 241W, 241 L, 241Y, 241E, 241 R. 243W, 243L 243 Y, 243R, 243Q, 244H, 245A, 247L, 247V, 247G, 250E, 250Q, 251F, 252L, 252Y, 254S, 254T, 255L, 256E, 256F, 256M, 257C, 257M, 257N, 262I, 262A, 262T, 262E, 263I, 263A, 263T, 263M, 264L, 264I, 264W, 264T, 264R, 264F, 264M, 264Y, 264E, 265A, 265G, 26 5N, 265Q, 265Y, 265F, 265V, 265I, 265L, 265H, 265T, 266I, 266A, 266T, 266M, 267Q, 267L, 268E, 269H, 269Y, 269F, 269R, 280A, 284M, 292P, 292L, 296E, 296Q, 296D, 296N, 296S, 296T, 296L, 296I, 296H, 296G, 297S, 297D, 297E, 298A, 298H, 298I, 29 8T, 298F, 299I, 299L, 299A, 299S, 299V, 299H, 299F, 299E, 305I, 308F313F, 316D, 318A, 318S, 320A, 320S, 322A, 322S, 325Q, 325L, 325I, 325D, 325E, 325A, 325T, 325V, 325H, 326A, 326D, 326E, 326G, 326M, 326V, 327G, 327W, 327N, 327L, 328S, 328M, 8D, 328E, 328N, 328Q, 328F, 328I, 328V, 328T, 328H, 328A, 329F, 329H, 329Q, 330K, 330G, 330T, 330C, 330L, 330Y, 330V, 330F, 330R, 330H, 331G, 331A, 331L, 331M, 331F, 331W, 331K, 331Q, 331E, 331S, 331V, 331I, 331C, 331Y, 331H, 331R, 331N, 1D, 331T, 332D, 332S, 332W, 332F, 332E, 332N, 332Q, 332T, 332H, 332Y, 332A, 333A, 333D, 333G, 333Q, 333S, 333V, 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, as numbered by the EU index as defined in Kabat. Optionally, the Fc region may comprise additional amino acid substitutions and/or alternatives known to one of skill in the art, including but not limited to those exemplified in Tables 2 and 610 of US 6,737,056; the tables set forth in Figure 41 of US 2006/024298; the tables set forth in Figures 5, 12 and 15 of US 2006/235208; the tables set forth in Figures 8, 9 and 10 of US 2006/0173170 and the tables set forth in Figures 8, 9 and 10 of US20090041770, all of which are incorporated herein by reference.

[00204] In a specific aspect, the disclosure provides an Fc variant BiSAb, wherein the Fc region comprises at least one modification (e.g., amino acid substitutions, amino acid insertions, amino acid deletions) at one or more positions selected from the group consisting of 228, 234, 235, and 331 as numbered by the EU index as defined 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 defined in Kabat.

[00205] In another aspect, the present disclosure provides a variant Fc BiSAb, 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 defined in Kabat. In yet another specific aspect, the Fc region is an IgG4 Fc region and the non-naturally occurring amino acids are selected from the group consisting of 228P, 235E, and 235Y as numbered by the EU index as defined in Kabat.

[00206] In another specific aspect, the present disclosure provides an Fc variant BiSAb, wherein the Fc region comprises at least one non-naturally occurring amino acid at one or more positions selected from the group consisting of 239, 330, and 332 as numbered by the EU index as defined in Kabat. In one aspect, 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 defined in Kabat. See, U.S. Patent Number 7,317,091, incorporated in its entirety herein by reference.

[00207] In a specific aspect, the present disclosure provides an Fc variant BiSAb, wherein the Fc region comprises at least one non-naturally occurring amino acid at one or more positions selected from the group consisting of 252, 254, and 256 as numbered by the EU index as defined 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 defined in Kabat. See, U.S. Patent Number 7,083,784, incorporated in its entirety herein by reference.

[00208] In certain aspects, the present disclosure provides a variant Fc BiSAb, wherein the Fc region comprises a non-naturally occurring amino acid at position 428 as numbered by the EU index as defined in Kabat. In one aspect, the modification at position 428 is selected from the group consisting of 428T, 428L, 428F, and 428S as numbered by the EU index as defined in Kabat. See, U.S. Patent No. 7,670,600, incorporated in its entirety herein by reference. In another aspect, a variant Fc BiSAb can further include a non-naturally occurring amino acid at position 434 as numbered by the EU index as defined in Kabat. In one aspect, the modification at position 434 is selected from the group consisting of 434A, 434S, and 434F as numbered by the EU index as defined in Kabat. In other aspects, the present disclosure provides a variant Fc BiSAb, wherein the Fc region comprises a non-naturally occurring amino acid at positions 428 and 434 as numbered by the EU index as defined in Kabat. In a specific aspect, the Fc region comprises 428L, 434S. See, U.S. Patent Number 8,088,376.

[00209] In certain aspects, the effector functions elicited by IgG antibodies depend strongly on the carbohydrate moiety attached to the Fc region of the protein (Claudia Ferrara et al., 2006, Biotechnology and Bioengineering 93:851-861). Thus, glycosylation of the Fc region can be modified to enhance or diminish effector function (see for examples, 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 278:3466-3473; U.S. Pat. Nos. 6,602,684; 6,946,292; 7,064,191; 7,214,775; 7,393,683; 7,425,446; 7,504,256; POTELLI- GENT™ (Biowa, Inc. Princeton, N.J.); GLYCOMAB™ glycosylation engineering technology (GLYCART biotechnology AG, Zurich, Switzerland). Thus, in one aspect, the Fc regions of BiSAbs of the disclosure comprise altered glycosylation of amino acid residues. In another aspect, the altered glycosylation of the amino acid residues results in reduced effector function. In another aspect, the altered glycosylation of the amino acid residues results in increased effector function. In a specific aspect, the Fc region has reduced fucosylation. In another aspect, the Fc region is afucosylated (see for examples, U.S. Patent Application Publication No. 2005/0226867). In one aspect, these BiSAbs with enhanced effector function, specifically ADCC, are generated in host cells (e.g., CHO cells, Lemna minor) engineered to produce highly defucosylated polypeptide with over 100-fold higher ADCC compared to the polypeptide produced by the parental cells (Mori et al., 2004, Biotechnol Bioeng 88:901-908; Cox et al., 2006, Nat Biotechnol., 24:1591-7).

[00210] The addition of sialic acid to oligosaccharides in 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). Studies cited earlier in this document demonstrate that IgG molecules with increased sialylation have anti-inflammatory properties while IgG molecules with reduced sialylation have increased immunostimulatory properties (e.g., increased ADCC activity). Therefore, a BiSAb can be engineered with a sialylation profile appropriate for a particular therapeutic application (U.S. Publication No. 2009/0004179 and International Publication No. WO 2007/005786).

[00211] In one aspect, the Fc regions of BiSAbs of the disclosure comprise an altered sialylation profile compared to the native Fc region. In one aspect, the Fc regions of BiSAbs of the disclosure comprise an increased sialylation profile compared to the native Fc region. In another aspect, the Fc regions of BiSAbs of the disclosure comprise a reduced sialylation profile compared to the native Fc region.

[00212] In one aspect, the Fc variants of the present disclosure can be combined with other known Fc variants, such as those disclosed in 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 U S A 92:11980-11984; Jefferis et al, 1995, Immunol Lett. 44:111-117; Lund et al., 1995, Façab J 9:115-119; Jefferis et al, 1996, Immunol Lett 54:101-104; Lund et al, 1996, J Immunol 157:4963-4969; Armor et al., 1999, Eur J Immunol 29:2613-2624; Idusogie et al, 2000, J Immunol 164:4178-4184; Reddy et al, 2000, J Immunol 164:19251933; Xu et al., 2000, Cell Immunol 200:16-26; Idusogie et al, 2001, J Immunol 166:2571-2575; Shields et al., 2001, J Biol Chem 276:65916604; 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; 6,194,551; 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; 7,355,008. Other modifications and/or substitutions and/or additions and/or deletions of the Fc domain will be readily apparent to one skilled in the art.

[00213] It is notable that the polypeptides presented in the BiSAb format comprising a native Fc retain the ability to bind to FcRn and C1q and to mediate ADCC, as shown in the examples. Thus, in certain aspects, a BiSAb retains the ability to bind to FcRn and/or C1q and/or one or more Fcgamma receptors (FCYRS). For example, in certain aspects, a BiSAb retains at least 70%, 75%, 80%, 85%, 90%, 95%, 96%, 97%, 98%, 99%, or 100% of the ability to bind to FcRn and/or C1q and/or one or more FCYRS, as compared to a conventional antibody that binds to one of the epitopes to which the BiSAb binds. In certain aspects, a BiSAb is generated from the binding domains of one or two conventional antibodies, and comparison of activity is made to one or both of the conventional antibodies.

[00214] Altered Fc regions can also be used to generate heavy chain heterodimers, resulting in BiSAbs comprising two different heavy-light chain pairs. To facilitate heterodimer formation, the interface between a pair of Fc regions is engineered so as to maximize the percentage of heterodimers that are recovered from recombinant cell culture. In certain aspects, the interface comprises at least a portion of the CH3 domain. In such a method, a "protrusion" is generated by replacing one or more small amino acid side chains from the interface of the first antibody molecule with larger side chains (e.g., tyrosine or tryptophan). Compensatory "cavities" of identical or similar size to the larger side chain(s) are created at the interface of the second antibody molecule by replacing the larger amino acid side chains with smaller ones (e.g., alanine or threonine). This provides a mechanism to increase the yield of the heterodimer over other unwanted end products such as homodimers. CH3 modifications include, for example, Y407V/T366S/L368A on one heavy chain and T366W on the other heavy chain; S354C/T366W on one heavy chain and Y349C/Y407V/T366S/L368A on the other heavy chain. Additional modifications resulting in a protrusion on one chain and a cavity on the other are described in US 7,183,076; US 2014/0348839; and Merchant et al., 1998, Nat. Biotech 16:677-681. Some non-limiting examples of modifications that can result in a cavity-protrusion arrangement are shown in Table 1a. Other modifications that can be used to generate heterodimers include, but are not limited to, those that alter the charge polarity across the Fc dimer interface such that co-expression of electrostatically coincident Fc regions results in heterodimerization. Modifications that alter charge polarity include, but are not limited to, those presented in Table 1b below (see also US20090182127; 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 chain-exchanged engineered domain (SEED) CH3 regions that are derived from human IgG and IgA CH3 domains (also, see WO 2007/110205). Table 1a. CH3 modifications for heterodimerization (cavity-protrusion) Table 1b. CH3 modifications for heterodimerization

[00215] One of skill in the art will understand that, in some aspects, an Fc Fusion protein can form dimers due to the homodimeric nature of the molecules comprising an Fc region. In some aspects, the Fc regions of a binding protein (e.g., BiSAb) can be differentially engineered with mutations to: promote and/or maintain heterodimerization (e.g., chimeric mutations, complementary mutations, "dock and lock" mutations, "knobs and holes" mutations, strand exchange engineered domain (SEED) mutations, etc., see, e.g., U.S. Patent 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 incorporated herein by reference). Accordingly, a binding protein can be engineered to form a heterodimer 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, wherein the first and second Fc regions, or fragments thereof, have been engineered to heterodimerize.

3. Glycosylation

[00216] In addition to the ability of glycosylation to alter the effector function of polypeptides, modified glycosylation in the variable region can alter the affinity of the antibody (or BiSAb) for the target antigen. In one aspect, the glycosylation pattern in the variable region of the present BiSAbs is modified. For example, an aglycosylated BiSAb can be manufactured (i.e., the BiSAb does not undergo glycosylation). The glycosylation can be altered to, for example, increase the affinity of the BiSAb for a target antigen. Such carbohydrate modifications can be accomplished by, for example, altering one or more glycosylation sites within the BiSAb sequence. For example, one or more amino acid substitutions can be made that result in the elimination of one or more glycosylation sites from the variable framework region to thereby eliminate glycosylation at that site. Such aglycosylation can increase the affinity of the BiSAb for the antigen. This approach is described in greater detail in U.S. Pat. 5,714,350 and 6,350,861. One or more amino acid substitutions can also be made, which result in the elimination of a glycosylation site present in the Fc region (e.g., Asparagine 297 of IgG). In addition, aglycosylated BiSAbs can be produced in bacterial cells that lack the necessary glycosylation machinery.

4. Polypeptide Ligands

[00217] Linkers can be used to join chimeric heavy chain domains/regions of BiSAb into a contiguous molecule. As described herein, a BiSAb can include one, two, or more polypeptide linkers (e.g., L1 and L2). Additionally, a BiSAb can include additional linkers, such as a flexible linker that interconnects the heavy and light variable chains of an scFv. Additionally, a BiSAb can include additional linkers, such as a flexible linker that interconnects the heavy and light variable chains of an scFv and other linkers that attach other binding moieties to the core structure of the BiSAb.

[00218] A non-limiting example of a linker is a polypeptide chain comprising at least 4 residues. Portions of such linkers may be flexible, hydrophilic, and have little or no secondary structure of their own (linker portions or flexible linker portions). Linkers of at least 4 amino acids may be used to join domains and/or regions that are positioned close to each other after the molecule has been assembled. Longer or shorter linkers may also be used. Thus, linkers may be approximately 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 approximately 50 residues in length. When multiple linkers are used to link portions of the molecule, the linkers may be the same or different (e.g., the same or different length and/or amino acid sequence).

[00219] Linkers may be cleavable linkers, which contain at least one bond that can be selectively cleaved by a cleavage reagent. Cleavable linkers may be used to facilitate the removal of all or a portion of the linker sequence. Linkers may be engineered to contain protease cleavage sites such that cleavage occurs in the middle of the linker or at least one end of the linker. For example, thrombin sites may be engineered at each of the two flanking ends of a linker. Depending on the type of linker used, cleavage may also be mediated by agents such as TCEP, TFA, and DTT. Linkers may be designed such that cleavage reagents remove all linker residues from the cleavage product. Other exemplary non-limiting linkers include prodrug linkers whose bonds can 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 in cells of the body. When BiSAbs contain more than one polypeptide linker, each of the linkers may be different, or at least one of the linkers may be different from the others. In some aspects, a BiSAb comprises a cleavable linker. In a specific aspect, the BiSAb comprises a scFv, wherein the scFv comprises a cleavable linker between VH2 and VL2.

[00220] The linker(s) facilitate formation of the desired structure. The linkers may comprise (Gly-Ser)n residues, with some Glu or Lys residues interspersed to increase solubility. Alternatively, or additionally, the linkers may not comprise any Serine residues; such linkers may be preferable where the linker is subject to O-linked glycosylation. In some aspects, the linkers may contain cysteine residues, for example, if linker dimerization is used to bring the domains of the BiSAb into their correctly folded configuration. In some aspects, the BiSAb comprises at least two polypeptide linkers that join domains of the polypeptide. In other aspects, the BiSAb comprises at least three polypeptide linkers. In other aspects, the BiSAb comprises four or more polypeptide linkers.

[00221] In some aspects, the polypeptide linker comprises a portion of an Fc portion. For example, in some aspects, the polypeptide linker can comprise a portion of the immunoglobulin hinge domain of an IgG1, IgG2, IgG3, and/or IgG4 antibody. In some aspects, the polypeptide linker comprises a portion of the mutated immunoglobulin hinge domain of an IgG1, IgG2, IgG3, and/or IgG4 antibody. In some aspects, 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 aspects, the polypeptide linker comprises at least 5, 7, 8, or 15 amino acid residues of the modified immunoglobulin hinge region/domain of an IgG1, IgG2, IgG3, and/or IgG4 antibody.

[00222] The polypeptide linker may comprise all or a portion of a hinge region naturally comprising three cysteines. In certain aspects, the selected hinge region is truncated or otherwise altered, or substituted relative to the full-length and/or naturally occurring hinge region such that only one or two of the cysteine residues remain. Similarly, in certain other aspects, the polypeptide linker may comprise a mutated or otherwise altered portion of a hinge region in which the number of cysteine residues is reduced by amino acid substitution or deletion, e.g., a mutated or otherwise altered hinge region containing zero, one, or two cysteine residues as described herein.

[00223] A mutated or otherwise altered hinge domain may thus be derived or constructed from (or using) a wild-type immunoglobulin hinge domain containing one or more cysteine residues. In certain aspects, a mutated or otherwise altered portion of a hinge region may contain zero or only one cysteine residue, wherein the mutated or otherwise altered hinge region is or was derived from a wild-type immunoglobulin hinge region containing, respectively, one or more or two or more cysteine residues. In the mutated or otherwise altered portion of a hinge region, the cysteine residues of the wild-type immunoglobulin hinge region are preferably deleted or replaced with amino acids that are incapable of forming a disulfide bond. In some aspects, a mutated or otherwise altered portion of a hinge region is or was derived from a wild-type hinge region of human IgG, which may include any of the four human IgG isotype subclasses, IgG1, IgG2, IgG3, or IgG4.

[00224] In some aspects, the polypeptide linker comprises a portion of a hinge region comprising the cysteine residue that forms a disulfide bond with an immunoglobulin light chain (EU residue 220). In some aspects, the polypeptide linker comprises an altered portion of a hinge region comprising an amino acid substitution at EU residue C220. In some aspects, the polypeptide linker comprises the amino acid substitution C220V.

[00225] In some aspects, the polypeptide linker comprises an amino acid substitution that prevents spontaneous hinge-related self-cleavage. In some aspects, the polypeptide linker comprises an amino acid substitution at the EU position D221. In some aspects, the polypeptide linker comprises the amino acid substitution D221G. In some aspects, the polypeptide linker lacks the amino acid D221.

[00226] As discussed above, some embodiments include one or more polypeptide linkers comprising or consisting of a glycerol linker. As used herein, the term "Gly-Ser linker" refers to a peptide consisting of glycine and serine residues. An exemplary Gly-Ser linker comprises an amino acid sequence of the formula (Gly4Ser)n, wherein n encompasses a positive integer (e.g., 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10). Some preferred and non-limiting examples of a Gly-Ser linker include (Gly4Ser)2, (SEQ ID NO:41) and (Gly4Ser)4, (SEQ ID NO:42) as well as (Gly4Ser)3 (SEQ ID NO:43). In still other aspects, two or more gli-ser linkers are incorporated in series into a polypeptide linker. In some aspects, the polypeptide linker comprises at least a portion of a hinge region (e.g., derived from an IgG1, IgG2, IgG3, or IgG4 molecule) and a series of gli-ser amino acid residues (e.g., a gli-ser linker such as (Gli4Ser)n, wherein n is 2, 3, or 4).

[00227] In certain aspects, the linkers (e.g., L1 and/or L2 and/or L3, etc.) include both a hinge portion and a linker portion, such as a linker portion comprising a gli-ser linker. In other aspects, L1 and/or L2 include only a hinge portion or only a linker portion, such as a gli-ser linker. In other aspects, both L1 and L2 include a gli-ser linker portion. In certain aspects, the gli-ser linker within a BiSAb is the same length, while in other aspects, the gli-ser linker portion within a BiSAb (e.g., L1 and L2) are of different lengths. When a 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. Such a flexible linker generally does not include a hinge moiety, but is instead a gli-ser linker or other flexible linker. The length and amino acid sequence of a flexible linker that interconnects the domains of a scFv can be readily selected and optimized (e.g., (Gli4Ser)n, (SEQ ID NO:48) where n is 2, 3, or 4 or more).

[00228] Regardless of the polypeptide linker used to interconnect multiple binding units and domains (e.g., between binding domains/units (e.g., Fab-scFv) or Fc-binding domain/unit (e.g., scFv via L1 and L2), the BiSAb may optionally comprise additional polypeptide linkers. The lengths and sequence of such additional polypeptide linkers are independently selected. For example, the BiSAb may further comprise a flexible polypeptide linker interconnecting the heavy and light variable chains of an scFv. Such a flexible polypeptide linker may comprise a gli-ser linker. In general, such a linker does not include a hinge moiety.

[00229] It is contemplated herein that varying the length of the linkers flanking BD2 may impact the orientation of the BD2 antigen-binding site and spacing relative to the rest of the BiSAb molecule. For example, a short N-terminal linker and a long C-terminal linker may create an orientation where the binding site is conformed in one direction, while a long N-terminal and short C-terminal linker may impart an opposite conformational orientation. Thus, the length of the linker may be modulated so as to orient the BD2 antigen-binding site and have important impacts on creating or avoiding steric effects between BD1 and BD2 and/or BD2 and other entities that bind to the antibody molecule in the Fc domain or other domains.

5. Specific configuration of BiSAbs

[00230] As discussed above, one aspect of the disclosure relates to a structural arrangement (platform) of the BiSAb comprising two pairs of heavy light chains (illustrated in Figures 1A-1F). In some embodiments of this aspect, the chimeric BiSAb heavy chain polypeptide sequence can 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 the 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 remainder of the Fc domain. In some aspects, the Fc domain comprises a CH2 domain and a CH3 domain. Thus, certain embodiments provide a chimeric BiSAb heavy chain that can comprise polypeptide sequences in the following orientation from the N-terminus to the C-terminus: VH1-CH1-CH2(N-term)-L1-BD2-L2-CH2(C-term)-CH3; VH1-CH1-CH2-L1-BD2-L2-CH3; and VH1-CH1-CH2-CH3(N-term)-L1-BD2-L2-CH3(C-term). The BiSAb light chain polypeptide sequence can comprise a light chain variable domain (VL1) and a light chain constant domain (CL). Thus, a chimeric BiSAb heavy chain comprises polypeptide sequence in the following orientation from the N-terminus to the C-terminus: VL1-CL. Note that VH1, VL1, and CL are used to denote portions of "binding unit 1" (BD1) that binds to a first epitope. BD2 is used to denote portions of "binding unit 2" that binds to a second epitope.

[00231] In aspects wherein the binding domain is a scFv, the chimeric heavy chain BiSAb can 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. Thus, the chimeric heavy chain of a BiSAb comprising a scFv such as BD2 can comprise polypeptide sequences in the following orientation from the N-terminus to the 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-VH2-L2-CH3(C-term); VH1-CH1-CH2(N-term)-L1-VH2-L3-VL2-L2-CH2(C-term)-CH3; VH1- CH1-CH2-L1-VH2-L3-VL2-L2-CH3; and VH1-CH1-CH2-CH3(N-term)-L1-VH2-L3-VL2-L2-CH3(C-term).

[00232] The chimeric heavy chain is a polypeptide chain comprising an amino acid sequence (e.g., the amino acid sequence of each of the polypeptide domains). The chimeric heavy chain is a polypeptide chain comprising an amino acid sequence (e.g., the amino acid sequence of each of the polypeptide domains). Note that VH1, VL1, and CL are used to denote portions of binding unit 1, with VH1 and VL1 indicating the portion that binds the first epitope. VH2 and VL2 are used to denote portions of binding unit 2 that binds the second epitope. In certain aspects, additional scFv binding domains are present at the N-terminal and/or C-terminal ends of the polypeptides that constitute the BiSAb core (wherein the BiSAb core additionally comprises binding unit (BD) 3 and/or 4 and/or 5). In certain aspects, more than one scFv binding domain is present within the BiSAb core. Each additional scFv comprises an antibody heavy chain variable region denoted by VH3, VH4, VH5 and a corresponding antibody light chain variable region denoted by VL3, VL4, VL5.

6. Markers, conjugates and portions

[00233] In certain embodiments, drugs and other molecules can be targeted to the BiSAb via site-specific conjugation. For example, BiSAbs can comprise cysteine-engineered domains (including engineered cysteine(s) in a binding moiety and/or Fc domain), which result in free thiol groups for conjugation reactions. In certain embodiments, a BiSAb is engineered to incorporate specific conjugation sites. In some aspects, the present disclosure provides a variant Fc BiSAb, wherein the Fc region comprises an amino acid substitution 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 aspects, the Fc region comprises substitutions at one or more of the following groups of 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 other aspects, the present disclosure provides a BiSAb, wherein 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 for an amino acid chosen from cysteine, lysine, tyrosine, histidine, selenocysteine, and selenomethionine. In a specific aspect, the substitution is a cysteine. Methods for generating stable cysteine engineered antibodies are described in U.S. Pat. No. 7,855,275 , U.S. Pat. No. 20110033378 , and U.S. Pat. No. 20120213705 , the contents of which are incorporated in their entirety herein by reference.

7. Exemplary Targets

[00234] Although aspects and embodiments related to various DuetMab and BiSAb platforms described herein can be generated to bind to any desired target or targets, the BiSAbs disclosed herein preferentially target specific pairs of target molecules (e.g., binding moiety 1 binds one of the targets and binding moiety 2 binds the other target). As discussed above and as exemplified in the Illustrative Examples below, the antibodies, DuetMabs, and BiSAbs disclosed herein are targeted to a molecule that modulates an immune response in a recipient subject, or in immune cells in culture. 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. DuetMabs and BiSAbs may comprise a combination of different binding domains in various orders and orientations, wherein the domains 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 enable bispecific binding to targets including; 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 include binding domains that bind to specific target combinations illustrated in the Examples and include 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 relative to the binding properties of the combined individual monospecific binding proteins that are used to generate the BiSAbs.

[00235] In certain aspects, a DuetMab or BiSAb of the disclosure binds two different epitopes on the same target (e.g., binding moiety 1 binds a first epitope on a target and binding moiety 2 binds a second epitope on the same target).

[00236] In some aspects, the multimeric nature of the DuetMabs or BiSAbs of the disclosure confers the ability to direct markers or therapeutics to a specific cell type or molecular target. For example, one functional domain on a DuetMab or BiSAb may bind a target on the surface of a cell, while another functional domain on the same DuetMab or BiSAb binds a hapten or labeling agent useful for detection. Similarly, one functional domain may bind to a cellular target while a second functional domain binds to a toxin. Since both binding reactions are mediated through a single molecule, the toxin may be brought into proximity of the cellular target, where it exerts a cytotoxic function.

8. Nucleic Acid Molecules Encoding BiSAbs

[00237] The present disclosure provides nucleic acid molecules encoding BiSAbs. One aspect of the disclosure provides nucleic acid molecules encoding any of the BiSAbs of the disclosure. A nucleic acid molecule may encode a heavy chain and/or light chain of any of the BiSAb molecules that are disclosed herein, as well as any of the individual binding domains (e.g., scFvs) that are disclosed herein. One of skill in the art will appreciate that such polynucleotide molecules may vary in nucleotide sequence given the degeneracy of nucleic acid codons as well as the codon frequency for particular organisms, as is generally known in the art.

9. Vectors and host cells for BiSAbs production and subsequent purification

[00238] The disclosure relates to methods for producing BiSAbs. In certain aspects, recombinant nucleic acid molecules encoding all or a portion of the BiSAbs disclosed herein can be operably linked to one or more regulatory nucleotide sequences in an expression construct. The nucleic acid sequences encoding the chimeric BiSAb light and heavy chains can be cloned into the same expression vector in either orientation (e.g., light chain in front of the heavy chain or vice versa) or can be cloned into two different vectors. If expression is carried out using one vector, the two encoding genes can have their own genetic elements (e.g., promoter, RBS, leader, stop, polyA, etc.) or can be cloned with a single set of genetic elements but connected with a cistron element. Regulatory nucleotide sequences will generally be appropriate for a host cell used for expression. Numerous types of appropriate expression vectors and suitable regulatory sequences are known in the art for a variety of host cells. Typically, said one or more regulatory nucleotide sequences may include, but are not limited to, promoter sequences, leader or signal sequences, ribosomal binding sites, transcriptional initiation and termination sequences, translation initiation and termination sequences, and enhancer or activator sequences. Constitutive or inducible promoters as known in the art are contemplated by the disclosure. Promoters may be either naturally occurring promoters or hybrid promoters that combine elements of more than one promoter. An expression construct may be present in a cell in an episome, such as a plasmid, or the expression construct may be inserted into a chromosome.

[00239] In certain aspects, the expression vector contains a selectable marker gene to allow selection of transformed host cells. Selectable marker genes are well known in the art and will vary with the host cell used. In certain aspects, this disclosure relates to an expression vector comprising a nucleotide sequence encoding a polypeptide and operably linked to at least one regulatory sequence. Regulatory sequences are recognized in the art and are selected to direct expression of the encoded polypeptide. Accordingly, 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 depend on factors such as the choice of host cell to be transformed and/or the type of protein to be expressed. In addition, the copy number of the particular vector, the ability to control that copy number, and the expression of any other proteins encoded by the vector, such as antibiotic markers, must also be considered.

[00240] The disclosure further relates to methods of producing a BiSAb of the disclosure. For example, a host cell transfected with one or more than one expression vector encoding a BiSAb (e.g., a single vector encoding the chimeric heavy and light chain or two vectors, one encoding the chimeric heavy chain and the other encoding the light chain) can be cultured under appropriate conditions to allow expression of the polypeptide to occur. The BiSAb can be secreted and isolated from a mixture of cells and medium containing the polypeptide. Alternatively, the BiSAb can be retained in the cytoplasm or in a membrane fraction and the cells harvested, lysed and the protein isolated. A cell culture includes host cells, media and other by-products. Suitable media for cell culture are well known in the art. BiSAbs can be isolated from cell culture medium, host cells, or both using techniques known in the art for protein purification, including ion exchange chromatography, gel filtration chromatography, ultrafiltration, electrophoresis, and immunoaffinity purification. In certain aspects, the BiSAb is made as a fusion protein containing a domain that facilitates its purification.

[00241] A recombinant nucleic acid can be produced by ligating the cloned gene, or a portion thereof, into a vector suitable for expression in either prokaryotic cells, eukaryotic cells (yeast, avian, insect or mammalian), or both. Expression vehicles for production of a recombinant polypeptide include plasmids and other vectors. For example, suitable vectors include plasmids of the types: pBR322-derived plasmids, pEMBL-derived plasmids, pEX-derived plasmids, pBTac-derived plasmids and pUC-derived plasmids for expression in prokaryotic cells such as E. coli. In certain aspects, mammalian expression vectors contain both prokaryotic sequences to facilitate propagation of the vector in bacteria and one or more eukaryotic transcription units that are expressed in eukaryotic cells. Vectors derived from pcD-NAI/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 from bacterial plasmids, such as pBR322, to facilitate replication and selection for drug resistance in prokaryotic and eukaryotic cells. Alternatively, derivatives of viruses such as bovine papillomavirus (BPV-1) or Epstein-Barr virus (pHEBo, derived from pREP and p205) can be used for transient expression of proteins in eukaryotic cells. The various methods employed in preparing the plasmids and transforming host organisms are known in the art. For other expression systems suitable 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 the recombinant polypeptide 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).

[00242] Once a molecule has been produced, it can be purified by any method known in the art for purifying a protein, immunoglobulin molecule, or other multimeric molecules using techniques such as, for example, chromatography (e.g., ion exchange, affinity, particularly affinity for Protein A or Protein G specific antigens, and size exclusion column chromatography), centrifugation, differential solubility, or by any other standard technique for protein purification. In addition, the molecules disclosed herein can be fused to heterologous polypeptide sequences (e.g., affinity tags), as are routinely employed to facilitate purification.

[00243] Regardless of how a BiSAb is generated and purified, binding assays, e.g., dual ELISA assays, can be performed (before and/or after purification) to confirm the functional binding activity of the BiSAb. Such binding assays are generally known in the art.

10. Pharmaceutical Formulations

[00244] In certain aspects, the disclosure provides pharmaceutical compositions. Such pharmaceutical compositions can be compositions comprising a nucleic acid molecule encoding a BiSAb. Such pharmaceutical compositions can also be compositions comprising a DuetMabs, a BiSAb, a combination of DuetMabs, or a combination of BiSAbs, and a pharmaceutically acceptable excipient. In certain aspects, the pharmaceutical compositions of the disclosure are used as a medicament.

11. Uses

[00245] As discussed herein, DuetMabs and BiSAbs can be used to bind targets associated with cancer and cell proliferative diseases or disorders that may be responsive to an immunotherapy, e.g., by inhibiting an immunosuppressive activity and/or inducing an immune response that is associated with the target molecule(s). For example, aberrant signaling and/or inhibited immune response can contribute to unwanted cell proliferation and cancer. Accordingly, DuetMabs, BiSAbs and the antibodies disclosed herein can be used to treat unwanted cell proliferation and/or cancer associated with an inhibited, reduced or insufficient immune response against cancer. In particular, the tumor growth curve of a tumor and/or the volume of a tumor can be reduced by administering a DuetMab or BiSAb that induces and/or stimulates an immune response in a subject, such as, for example, a human patient suffering from cancer.

[00246] Thus, the disclosure also relates to various methods comprising administering the binding proteins disclosed herein to a subject in need thereof. In one aspect, the disclosure is for 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 immune response pathway that is inhibited in the subject, wherein the activation enhances an immune response that targets cancer in the subject. In some embodiments, the method enhances an immune response pathway that targets cancer in the subject.

[00247] In another aspect, the 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 areas 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.

[00248] In various embodiments of the above aspects, the methods relate to treating a subject for a tumor disease and/or a cancerous disease. In 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, colorectal 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 selected from digestive or gastrointestinal cancers (e.g., anal cancer; bile duct cancer; extrahepatic bile duct cancer; appendix cancer; carcinoid tumor, gastrointestinal cancer; colon cancer; colorectal cancer including childhood colorectal cancer; esophageal cancer including childhood esophageal cancer; gallbladder cancer; gastric (stomach) cancer including childhood gastric cancer; hepatocellular cancer (e.g., hepatocellular carcinoma) including adult (primary) hepatocellular cancer and childhood hepatocellular cancer; pancreatic cancer including childhood pancreatic cancer; sarcoma, rhabdomosarcoma; pancreatic islet cell cancer; rectal cancer; and small bowel cancer); lung cancer (e.g., non-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 childhood oral cancer; hypopharyngeal cancer; laryngeal cancer including childhood laryngeal cancer; metastatic squamous cell neck cancer with an occult primary; mouth cancer; nasal cavity and paranasal sinus cancer; nasopharyngeal cancer including childhood nasopharyngeal cancer; oropharyngeal cancer; parathyroid cancer; pharyngeal cancer; salivary gland cancer including childhood salivary gland cancer; throat cancer; and thyroid cancer); ovarian and breast cancer.

[00249] In the above methods, the amount of binding protein that is administered to the subject is effective to induce an immune response, enhance an immune response, stop or slow the growth of cancer, stop or slow the metastasis of cancer, kill cancer cells, and/or slow or stop the proliferation and/or spread of cancer cells in the subject.

[00250] In embodiments of the above methods, the binding protein comprises a DuetmAb or BiSAb as disclosed herein. In some embodiments of the above methods, the binding protein comprises an antibody, or an antigen-binding fragment thereof, as disclosed herein.

[00251] As used herein, the term "subject" is intended to include both human and non-human animals, particularly mammals. Examples of subjects include human subjects, e.g., a human patient having a disorder, e.g., a disorder described herein, such as cancer, or a normal individual. A "non-human animal" includes all vertebrates, e.g., non-mammals (such as chickens, amphibians, reptiles) and mammals, such as non-human primates, domesticated and/or agriculturally useful animals (such as sheep, dogs, cats, cows, pigs, etc.), and rodents (such as mice, rats, hamsters, guinea pigs, etc.). In particular embodiments, the subject is a human patient.

[00252] "Treatment" or "treating" relates to both therapeutic treatment and prophylactic or preventative measures. Those individuals in need of treatment include those already with the disorder as well as those likely to have the disorder or those in whom the disorder is to be prevented. When used with reference to a disease or an individual in need of treatment, the terms accordingly include, but are not limited to, stopping or slowing the progression of the disease, remission of the disease, prophylaxis of symptoms, reduction of the disease and/or severity of symptoms, or reduction in the duration of the disease compared to an untreated individual. In embodiments, the methods of treatment may abate one or more clinical indications of the particular disease being treated.

[00253] The following Examples are provided to illustrate particular aspects and embodiments of the disclosure provided above and are not to be construed as limiting the scope of the description or the subject matter claimed herein. Examples Materials and Methods Immune Response Modulation Assay

[00254] A cytomegalovirus (CMV) antigen recall assay was used to evaluate the potential immune response induced by some of the immunotherapeutic molecules described herein. Reagents for the assay include: - Frozen CMV peripheral blood mononuclear cells (PBMC); - AIM V® medium (Life Technologies, cat# 12055-091); - Phosphate buffered saline (PBS, Life technologies, cat# 20012-043); - PepTivator®, CMVpp65 peptide pool, (Mil- tenyi Biotec, cat#130-093-438, 50μg/mL); - Ovalbumin, (Thermo scientific cat# 77120, 1 mg/mL); - Costar, TC-untreated 96-well plate (Corning, cat # 3788); and - an immunotherapeutic molecule. General Assay Protocol:

[00255] One day before the assay was performed, frozen PBMCs were thawed in warm AIM V medium. The cells were washed twice in a 96-well Costar round plate. The cell concentration was adjusted to a concentration of 1x106 cells/mL.

[00256] Aliquots (100 μL) of cells were distributed into individual wells, leaving the outer columns and rows of the plate empty. The cells were allowed to settle overnight.

[00257] The next day, 100 μL of AIMV media containing 2X PepTivator CMV peptide pool (0.1 μg/mL-0.05 μg/mL final) and 2X immunotherapeutic molecule was added to the wells.

[00258] After 72 hours, 25 μL of supernatant from each well was transferred to pre-blocked and washed MSD plates (anti-human IFN gamma). After addition of standards, the plates were incubated for 2 hours at room temperature. After the incubation period, the MSD plates were washed three times. Following washing, 25 μL of SULFO-TAG detection antibody was added and allowed to react for 1 hour at room temperature. The plates were washed again and 150 μL of 2X MSD reading buffer was added before readings were taken. Staphylococcal Enterotoxin A/B (SEA/SEB) Assay Protocol

[00259] Reagents used in either the SEB or SEA assay protocol to determine the effect of DuetMabs or BiSAbs on the IL-2 immune response include: - White blood cell cones (NHSBT code NC24; from Addenbrookes Hospital); - 50 mL Falcon tubes (BD 352070); - Ficoll-Paque PLUS (GE Healthcare 17-1440-02); - Anti-CD3 (clone OKT3; 1 mg/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 solutions at 1 mg/mL stored at -20°C; - Culture media (all from Life Technologies): RPMI1640 with glutamax (61870) supplemented with 10% v/v heat-inactivated FCS (90005M) and 100U/mL penicillin + 100ug/ml streptomycin (15140-122); - V-bottom plate (Greiner BioOne 651201); - Flat-bottom 96-well plates (Corning Costar 7107).

[00260] Reagents for the DELFIA IL-2 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® enhancement solution (Perkin-Elmer 4001-0010); at RT prior to 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, sterile filtered; - Powdered milk (Marvel; Premier Foods); - Sample Diluent (RPMI1640 + 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 (EL406) with automatic plate loader (Biostack).

General Rehearsal Protocol

[00261] PBMCs were isolated from human blood leukocyte cones (NHS Blood and Transplant Service code NC24) using density gradient centrifugation (Ficoll-Paque PLUS; GE Healthcare), then red blood cells were lysed in ammonium chloride solution (Stemcell Technologies). Anti-human CD3 (OKT3 clone at 0.5 ug/mL in PBS; eBioscience) was coated onto flat-bottomed 96-well plates (Corning Costar 7107) for 2 hours at 37°C. Then, 0.2 × 106 cells were added per well of PBMC in culture medium (RPMI1640-Glutamax supplemented with 10% v/v heat-inactivated bovine serum and 100 U/100 ug per mL Steptomycin/Penicillin (respectively) (Life Technologies). PBMC were further stimulated by the addition of Staphylococcal Enterotoxin A or B (SEB; Sigma Aldrich) within a range of 0.0088–0.1 ug/mL and candidate DuetMabs or BiSAbs were added to the final concentrations tested. After 3 days, the culture 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.

Mixed Leukocyte Reaction (MLR) Assay Protocol (Fresh Blood)

[00262] The MLR cell-based assay was also used to provide in vitro correlation of T cell function in response to the DuetMabs and BiSAbs disclosed herein. Reagents used in performing the MLR assay from fresh blood samples include: - 8 mL CPT heparin tubes; - AIM-V (serum-free) medium Gibco #12055-091, without additives; - 50 mL conical tubes; - 2 mL cryopreservation vials; - ACK lysis buffer (Gibco #A10492-01); - BD falcon #3077 tissue culture-treated U-bottom 96-well plates; - PHA (Roche) 1mg/mL (final concentration 10ug/mL), as a positive control;

General Rehearsal Protocol

[00263] PBMCs were prepared from blood samples collected in heparin CPT tubes. The tubes were centrifuged for 20 minutes at 2700 rpm without stopping at 25°C. The top layer of serum was aspirated. The remaining material was gently pipetted, and everything above the CPT tube was collected and placed in 50 mL conical tubes. AIM-V medium was added to the cells to wash the cells (3 times at 1500 rpm, with the stop on, at 25°C for 5 minutes). Any remaining red blood cells were lysed using red blood cell lysis buffer (e.g., approximately 5 min with approximately 3 mL of buffer). The remaining cells were washed twice with AIM-V medium (1500 rpm, stopped, at 25°C for 5 minutes). If necessary, pellets were consolidated in a single tube and resuspended in AIM-V medium, and a cell count was performed.

[00264] To perform the MLR assay, cells were plated in 96-well plates at 200,000 cells per donor per well in AIM-V medium, 50 μL per donor (total of 400,000/100 uL). Candidate molecules were added (4x) 50 uL per well diluted in serum-free AIM-V medium. After 72 hours, plates were visualized and 30 uL of supernatant was removed for human TH1/TH2 cytokine assay (MSD).

Human TH1/TH2 MSD 10-plex protocol

[00265] This assay was used to determine the amounts of cytokines present in culture supernatants in response to administration of the DuetMabs and BiSAbs disclosed herein. To perform the assay, the blocking agent was prepared by dissolving 200 μL of blocker B in 20 mL of PBS per plate. 150 μL of dissolved blocker was added to each well. The plate was sealed and shaken for 2 hours at room temperature or overnight at 4°C. The wells were washed 3x with PBST buffer. A calibrator was prepared by diluting 10 μL of the frozen calibrator mixture in 1 mL of diluent, and was further diluted 4-fold serially. To separate the wells, 25 μL of calibrator (standard) and 25 μL of samples were added. The wells were incubated for 2 hours at room temperature with shaking. After incubation, wells were washed 3x with PBST.

[00266] The detection antibody is prepared and diluted to the required concentration, and was added to each well. After 2 hours of incubation at room temperature with shaking, the wells were washed (3x) in PBST. Before reading on the MSD machine, reading buffer was added to each well.

Tumor Specific Killing Assay Protocol

[00267] The human CD8+ T cell line (JR6C12) with reactivity against the human gp100209-217 peptide was kindly provided by Dr. Steven Rosenberg (National Cancer Institute, Bethesda, MD). JR6C12 cells were cocultured with a CFSE-labeled human melanoma line (Mel624) (CellTrace CFSE Proliferation Kit, ThermoFisher) for 18 h at 37 °C in a 1:1 ratio (20,000 JR6C12 + 20,000 Mel624) in a 96-well flat-bottom plate. Candidate molecules were added at time 0 of coculture at a concentration of 69 nM. After 18 h, the wells were visualized by brightfield microscopy. Supernatants were collected for MSD analysis and adherent cells were trypsinized and washed (2x) in PBS prior to viability staining (Zombie UV Fixed Viability kit, Biolegend). Viability dye uptake of CFSE-labeled cells was assessed by flow cytometry on a LSRFortessa (BD).

Example 1. Identification of candidate Fc sites for binding to the binding domain

[00268] Using the open source PyMOL molecular visualization system, the structure of the antibody was investigated at the CH2 and CH3 regions, as well as at or near the CH2-CH3 interface, in order to identify candidate regions, such as surface-exposed loops, for binding to the binding domain. These regions would accommodate the insertion of a second binding domain (e.g., an scFv) without compromising the structural integrity or stability of the IgG or the second binding domain itself. From the analysis, three regions were identified (represented as spheres in Figures 1A-1C). Figures 1D, 1E, and 1F depict embodiments of the binding proteins, showing the binding of a second binding domain (with an scFv for illustration purposes) at each of the loops identified in Figures 1A, 1B, and 1C, respectively.

[00269] Figure 2A provides a more detailed schematic diagram of the amino acid sequence of one of the representative loops identified in the CH2 region proximal to the CH2-CH3 interface and comprising the sequence ISRTP (SEQ ID NO:39). A binding domain may be inserted within such an amino acid sequence to generate any number of representative constructs such as, for example, insertion of scFv domains 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 the "-scFv-" identifies the point in the native loop sequence to which the binding domain may be associated. Figure 2B is a similar schematic diagram that is representative of the loop identified at the CH2-CH3 interface and comprising the amino acid sequence AKGQP (SEQ ID NO:40). Representative constructs described herein may include a binding domain (such as, for example, a scFv domain) linked to such a loop sequence as described herein, including A-scFv-KGQP, AK-scFv-GQP, AKG- scFv-QP, AKGQ-scFv-P, scFV-AKGQ, AKGQ-scFV, where the "-scFv-" identifies the point in the native loop sequence to which the binding domain can be associated. Figure 2C provides a schematic diagram of the representative loop identified downstream of the CH2-CH3 interface, within the CH3 region and comprising the amino acid sequence SNG. Representative constructs of this loop sequence, discussed in terms of the illustrative embodiments for the other two loop regions above, include scFV-SNG, S-scFv-NG SN- scFv-G, and SNG-scFV.

[00270] Example 2. Generation and characterization of a series of parental antibodies and bispecific binding proteins, including the combination of binding moieties

[00271] A series of monoclonal antibodies have been developed and characterized. Using combinations of antigen-binding sequences (e.g., CDRs, HCv, LCv, HC, LC) derived from these "parent" antibodies, a series of bispecific binding proteins have been generated and shown to have bispecific binding activity for the combined target antigens. The bispecific binding proteins have been designed to have the particular structural platform motif that is disclosed herein (i.e., "BiS5").

[00272] The parental antibody sequences are described in the following tables: Table 2a. PD-1 Parental Antibody Sequences

Example 2 (a) PD-1/CTLA-4 bispecific binding proteins

[00273] Without being bound by theory, there is a strong clinical and preclinical rationale for the combination of PD-1 and CTLA-4 blockade. Thus, it would be desirable to maximize the risk/benefit ratio of the combination of PD-1 and CTLA-4 (Figure 4).

[00274] The following bispecific binding proteins that bind to PD-1 and CTLA-4 were created using the parental sequences identified above in Table 2. Proteins identified as Bis2, Bis3, and Bis5 were generated with the sequences identified below and were evaluated for competitor antigen binding activity using the Octet binding assay as discussed below. Table 3. BiS Constructs for PD-1/CTLA-4

Octet binding assay (BiS2, BiS3 and BiS5)

[00275] To assess binding of the bispecific binding molecules disclosed herein, an Octet QK equipped with Ni-NTA biosensor tips and 10X kinetic 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.

[00276] Sample plates were shaken at 1000 rpm prior to analysis. Ni-NTA biosensor tips were pre-wetted for 5 min in 1X kinetic buffer. 1X kinetic buffer also served as processing buffer for baseline determination and as dilution buffer for antigens and bispecific antibodies. Ni-NTA biosensor tips were dipped in 100 nM His-tagged PD-L1-Fc (see, (b), below) or His-tagged PD-1-Fc for antigen capture for ~1 min. Antigen-coated biosensor tips were each dipped in 10 μg/mL bispecific antibodies for ~5 min and then moved to a column of wells containing 100 nM CTLA-4-Fc antigen for 2 min. The connection results are shown in Figure 3.

[00277] A bispecific binding protein in the DuetMab format that binds to both PD-1 and CTLA-4 was created using the parental sequences identified above in Table 2. The PD-1/CTLA-4 DuetMab was generated with the sequences in Table 4 below and was evaluated as discussed below, including in comparison to PD-1/CTLA-4 Bis5. Table 4. DuetMab Constructs for PD-1/CTLA-4

Octet Binding Assay (DuetMab)

[00278] Concurrent binding studies to two distinct antigens were performed by Octet analysis. Biotinylated human PD-1 was loaded onto Streptavidin sensors, followed by sequential interactions first with DuetMab PD-1/CTLA-4 and then with soluble CTLA-4 antigen. Streptavidin (SA) biosensors (ForteBio) were used to capture biotinylated human PD-1 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 biosensors were subjected to successive association and dissociation interactions, first with sample wells containing the bispecific antibody Du-etMab PD-1/CTLA-4 at 133 nM and then with wells carrying the human CTLA-4 antigen at 200 nM. The binding results are shown in Figure 5.

[00279] The intrinsic kinetics of the PD-1/CTLA-4 bispecific antibody DuetMab were also assessed via BiaCore. Binding experiments were performed using a BIAcore T200 instrument (BIAcore). To capture the antibody, mouse anti-huIgG-Fab was immobilized on a CM5 chip to a target response of 2000 RU. 100 nM of the DuetMab or mAbs was flowed at 20 μL/min for 5 min to achieve approximately 100 units of captured antibody response. Antigen was then serially injected at a flow rate of 50 μL/min for 5 min. Kinetic parameters (kon and koff) and dissociation constant (KD) were calculated from a nonlinear fit using BIAevaluation 4.1 software. Binding results are shown in Table 5. Table 5. BiaCOre data for PD-1/CTLA-4

Reporter Gene Assays

[00280] Results from the reporter gene assays show that the PD-1/CTLA-4 bispecific binding proteins inhibited both the PD-1 and CTLA-4 pathways (Figure 6A-D). The BiS5 binding protein retained PD-1 potency compared to the parent, but had a ~3-fold reduced potency compared to anti-CTLA-4 IgG. The DuetMab antibody had a ~9-fold reduction in PD-1 potency and a ~16-fold reduction for CTLA-4 (compared to IgG4P). There is a need in the art for a molecule that has reduced CTLA-4 potency but functional activity (e.g., as shown below in the SEB assay). PD-1/CTLA-4 bispecific binding proteins therefore have the potential to provide a safety benefit to patients.

Staphylococal Enterotoxin B (SEB) Assay

[00281] Results from a SEB assay show that DuetMab and BiS5Ab had equivalent activity in the SEB primary immune cell assays (Figure 7), and DuetMab showed increased activity compared to the DummyDuet control arms (Figure 8A). DuetMab showed approximately equivalent activity to the parental molecule combination and increased activity compared to LO115 or the CTLA-4 antibody MEDI1123 (tremelimumab) (Figure 8B). Finally, BiS5 and DuetMab showed increased activity compared to a combination of novel isotype controls (Figures 9A-B). Data were obtained from four donors across two independent experiments, and these particular assays required the use of IFNY.

Mixed Leukocyte Reaction (MLR) Assay

[00282] MLR assays were performed to test PD-1/CTLA-4 bispecific molecules. PD-1/CTLA-4 DuetMab and BiS5Ab have equivalent activity in the mixed lymphocyte reaction (MLR) assay (Figures 10A-C). PD-1/CTLA-4 DuetMab had greater activity compared to the combination of DummyDuet/isotype control arms (Figures 11A-D). PD-1/CTLA-4 DuetMab had approximately equivalent activity compared to the combination of parental antibody controls (Figures 12A-D). Finally, PD-1/CTLA-4 DuetMab showed equivalent activity compared to the PD-1/CTLA-4 competitor and had greater activity than anti-PD-1 alone (e.g., pembrolizumab and nivolumab) (Figures 13A-D).

Pharmacokinetic and Pharmacodynamic Studies (PK/PD)

[00283] A study was performed to examine 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, and robust PD responses were observed for both molecules (Figures 15A-B). Thus confirming the viability of the PD-1/CTLA-4 bispecific binding protein in an in vivo setting.

T cell-dependent antibody response (TDAR)

[00284] Cynomolgus monkeys were dosed intravenously (saphenous or cephalic vein) with the indicated dose (0.5, 5, 50 mg/kg) of DuetMab or BiS5 bispecific molecules. Keyhole limpet hemocyanin (KLH) protein was reconstituted with the appropriate amount of sterile water for injection under sterile conditions. The low dose of KLH solution was administered subcutaneously into the back of each animal on two occasions (day 1 and day 29). Blood samples for subsequent analysis were obtained from all animals. Assessment of KLH-specific IgM and IgG antibody titers was performed. Anti-KLH antibodies in the monkeys' serum were detected using ELISA.

[00285] T cell-dependent antibody response (TDAR) was seen in cynomolgus monkeys dosed with PD-1/CTLA-4 DuetMab (Figure 16A) and PD-1/CTLA-4 BiS5Ab (Figure 16B).

CHO cells expressing varying levels of human PD-1 and/or CTLA-4

[00286] A model system was developed for the study of PD-1/CTLA-4 bispecific molecules using stable CHO cells expressing varying levels of human PD-1 and/or CTLA-4 (Figure 17). Missing antigen-binding arms on cell-bound DuetMab were detected by flow cytometry using fluorescently labeled soluble PD-1 and CTLA-4 proteins. The results of this assay show that PD-1/CTLA-4 DuetMab concurrently binds PD-1 and CTLA-4 on the surface of the same cell (Figures 18A-C).

[00287] CTLA-4 is continuously endocytosed into clathrin-coated pits, resulting in only a small fraction of the receptor expressed on the cell surface at any given time. CTLA-4 recycling to the cell surface is rapid, with over 80% of surface CTLA-4 being internalized within 5 minutes. Therefore, an experiment was performed to determine whether cooperative binding differentiates PD-1/CTLA-4 DuetMab relative to a combination of anti-PD-1 and anti-CTLA-4 antibodies at CTLA-4 saturation in cells overexpressing PD-1 (Figures 19A-C). Receptor occupancy of each target antigen was determined independently using labeled anti-PD-1 and anti-CTLA-4 mAbs.

[00288] The parent monoclonal antibodies were found to bind and occupy their target receptor without a measurable effect on the unlabeled receptor (Figures 20A-D). PD-1/CTLA-4 DuetMab saturated CTLA-4 on CHO cells overexpressing PD-1 at ~250-fold lower concentrations compared to a combination of the monoclonal antibodies (Figure 21A-D). PD-1/CTLA-4 DuetMab saturated CTLA-4 on CHO cells overexpressing PD-1 at ~500-fold lower concentrations compared to cells expressing CTLA-4 alone (Figures 22A-F). PD-1/CTLA-4 DuetMab preferentially bound in cis to both PD-1 and CTLA-4 on the surface of the same cell, as determined by quantification of doublet formation within the total premixed CHO population (Figures 23A-B). However, PD-1/CTLA-4 DuetMab could also bind in trans to single-expressing cells. PD-1/CTLA-4 DuetMab took on internalization properties from the parental anti-CTLA-4 antibody, tremelimumab (Figures 24A-D). Without being limited by theory, the effect shown by this molecule has the potential to induce PD-1 downregulation. The internalization properties of PD-1/CTLA-4 DuetMab were also observed in stable CHO cells expressing 10-fold excess PD-1 (Figure 25B). Example 2 (b) PD-L1/CTLA-4 bispecific binding proteins The following bispecific binding proteins that bind to both PD-L1 and CTLA-4 were created using the parental sequences identified above in Table 2. Proteins identified as Bis2, Bis3, and Bis5 were generated with the sequences in Table 6 below, and the sequences identified below were assessed for competitor antigen binding activity using the Octet binding assay as described below in section 2(a) (Figure 26) Table 6. BiS constructs for PD-L1/CTLA-4 Example 2 (c) PD-1/TIM3 bispecific binding proteins

[00289] The following bispecific binding proteins (Table 7) that bind to PD-1 and TIM3 were created using the parental sequences identified above in Table 2. Proteins identified as Bis3, Bis5, and DuetMab were generated with the sequences identified below and were evaluated for concurrent binding activity studies via Octet analysis. Briefly, streptavidin (SA) biosensors (ForteBio) were used to capture biotinylated human TIM3-IgV domain 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 biosensors were subjected to successive association and dissociation interactions, first with sample wells containing 200 nM bispecific antibodies and then with wells carrying 200 nM PD-1 antigen. The biotinylated human TIM3-IgV domain was loaded onto Streptavidin sensors, followed by sequential interactions first with bispecific molecules and then with PD-1 antigen. The binding results are shown in Figures 27A–27B. Table 7. BiS constructs for PD-1/TIM3 Tumor-specific killing activity assay Rosenberg clone-melanoma killing assay.

[00290] The general cell killing activity of the TIM3/PD-1 bispecific binding molecules and the parental TIM3 antibody was tested using the Rosenberg Clone: JR6C12 and Melanoma: Mel324 cell lines.

[00291] General Test Protocol

[00292] JR6C12 cells functioned as effectors, and are a human CD8+ T cell line expanded from a melanoma patient and specific for the gp100-melanoma antigen. To assess therapeutic potential, Mel624 tumor cells were fluorescently labeled and added to effectors (JR6C12) and candidate antibody that binds to TIM3 and/or PD-1. The cells were co-cultured for 16 hours. The multiple panels in Figure 28A provide a visual representation that the addition of TIM3 62 either in combination with anti-PD1 or PD-1/TIM3 bispecific molecules (as described in Table 7) enhances T cell activation and tumor killing.

[00293] Furthermore, as shown in Figures 28B-28C, PD-1/TIM3 bispecific molecules demonstrate greater tumor killing potency relative to anti-TIM3, anti-PD-1, or isotype control monotherapy, as assessed by (b) tumor cell viability uptake and (c) IFN-gamma secretion.

[00294] In addition to clone 62, other bispecific binding proteins that bind to both the PD-1 and TIM3 format on DuetMab were created using the parental sequences identified above in Table 2. The PD-1/TIM3 DuetMab was generated with the sequences in Table 8 below. The TIM3 arm sequence was obtained from O13-1, which is an affinity mature variant of clone 62, and the anti-PD-1 arm sequence was obtained from LO115, which is identical to the PD-1 arm used for 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 in comparison to PD-1/TIM3 BiS3 and BiS5. Table 8. DuetMab Constructs for PD-1/TIM3 Octet binding assay (DuetMab, mature TIM3 arm affinity variant)

[00295] Concurrent binding studies to two distinct antigens, PD-1 and TIM3, were performed by Octet analysis. Biotinylated human TIM3 was loaded onto Streptavidin sensors, followed by sequential interactions first with PD-1/TIM3 DuetMab and then with soluble PD-1 antigen. Streptavidin (SA) biosensors (ForteBio) were used to capture biotinylated human TIM3 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 biosensors were subjected to successive association and dissociation interactions, first with sample wells containing Duet-Mab PD-1/CTLA-4 bispecific antibody bearing the TIM3 arm (O13-1), which is the affinity mature variant of TIM3 antibody clone 62, loaded at 200 nM and then with wells carrying human PD-1 antigen at 200 nM. The binding results are shown in Figure 29.

[00296] The intrinsic kinetics of the PD-1/TIM3 bispecific antibody DuetMab were also assessed via BiaCore. Binding experiments were performed using a BIAcore T200 instrument (BIAcore). To capture the antibody, mouse anti-huIgG-Fab was immobilized on a CM5 chip to a target response of 2000 RU. 100 nM of the DuetMab or mAbs was flowed at 20 μL/min for 5 min to achieve approximately 100 units of captured antibody response. Antigen was then serially injected at a flow rate of 50 μL/min for 5 min. Kinetic parameters (kon and koff) and dissociation constant (KD) were calculated from a nonlinear fit using BIAevaluation 4.1 software. Binding results are shown in Table 9. Table 9. BiaCore data for PD-1/TIM3

[00297] PD-1/TIM3 bispecific antibodies, including BiS3, BiS5, and DuetMab, bound to CHO cells overexpressing human TIM3 or human PD-1 (Figure 30 and Table 25), PD-1 and TIM3 (DMF4) expression are shown in Figure 31.

CMV Ag recovery assay

[00298] PD-1/TIM3 bispecific antibodies, including BiS3, BiS5, and DuetMab, enhance CD8+ T cell proliferation in a CMV antigen retrieval assay compared to isotype treatment (Figures 32A-C).

Mixed Leukocyte Reaction (MLR) Assay

[00299] PD-1/TIM3 bispecific antibodies, including BiS3, BiS5, and DuetMab, enhanced interferon (IFNY) secretion in the mixed lymphocyte response (MLR) assay, with activity above that of monotherapy and combination therapy (Figures 33A-D). PD-1/TIM3 bispecific antibodies, including BiS3, BiS5, and DuetMab, demonstrated similar activity to parental LO115 IgG1 in a Jurkat NFKB reporter line that predominantly expresses PD-1 (87% single PD-1 positive) (Figures 34A-C).

[00300] In summary, three bispecific formats (DuetMab, Bis3 and Bis5) were generated for PD-1/TIM-3. All bispecific formats exhibit in vitro functionality equivalent to or better than anti-PD-1, suggesting that these molecules may provide a superior advantage to current immuno-oncology strategies.

Example 2 (d) OX40/PD-L1 bispecific binding proteins

[00301] The following bispecific binding proteins that bind to PD-L1 and OX40 were created using the parental sequences identified above in Table 2. Proteins identified as Bis2, Bis3, and Bis5 were generated with the sequences in Table 10 below and were evaluated for competitor antigen binding activity using the Octet binding assay as discussed below. Table 10. BiS Constructs for OX40/PD-L1

Octet binding assay

[00302] To assess the binding of the bispecific binding molecules disclosed herein, an Octet QK equipped with Ni-NTA biosensor tips and 10X kinetic 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 proteins) were purchased from R&D Systems (Minneapolis, MN). All binding assays were performed at 25°C.

[00303] Sample plates were shaken at 1000 rpm prior to analysis. Ni-NTA biosensor tips were pre-wetted for 5 min in 1X kinetic buffer. 1X kinetic buffer also served as processing buffer for baseline determination and as dilution buffer for antigens and bispecific antibodies. Ni-NTA biosensor tips were dipped in 100 nM His-tagged PD-L1-Fc (see, (b), below) or His-tagged PD-1-Fc for antigen capture for ~1 min. Antigen-coated biosensor tips were each dipped in 10 μg/mL bispecific antibodies for ~5 min and then moved to a 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 PD-1/OX40 was used as a control (Figure 35).

Staphylococal Enterotoxin B (SEB) Assay

[00304] A SEB assay using the protocol described above showed that the OX40/PD-L1 bispecific molecule was active in both the BiS2 and BiS3 formats (Figures 36A-B). PD-L1 reporter assay Materials: - Cell lines and culture conditions: - Jurkat NFAT luciferase Human PD-1 clone 2 reporter - PD-L1 expressing CHO scFv OKT3 (UBC) (All cells were maintained in RPMI 1640 medium plus 10% FBS and 1X pen/strep antibiotic (complete RPMI medium) at 37°C in a humidified tissue culture incubator). - RPMI-1640, LifeTechnologies cat# A1049101 - Heat-Inactivated Newborn Calf Serum (FBS), LifeTechnologies cat#26010074 - Complete RPMI Medium: RPMI-1640 + 10% FBS - 100X Penicillin/Streptomycin, LifeTechnologies cat# 15140-122 - TC-Treated Flat Bottom Culture 96-Well Plates, Costar 3903, VWR cat# 29444-010 - SteadyGlo Luciferase Assay System, Promega, cat# E2510 - Test Antibodies - EnVision Multilabel Plate Reader, Perkin Elmer Methods:

[00305] For 2-cell bioactivity assay for neutralization of PD-L1 inhibition, PD-L1-expressing scFv OKT3 CHO cells were trypsinized, neutralized with warm RPMI complete medium, and harvested in a 50 mL conical tube. Cells were pelleted at 380 g for 5 min at RT, and then suspended in fresh RPMI complete medium and counted in a Vi cell counter. PD-L1-expressing scFv OKT3 cells were adjusted to 0.4e6/mL and 25 μL (10,000 cells) per well were plated as shown in the plate layout. Cells were allowed to adhere to the plates for 3 hours. Subsequently, 50 μL of RPMI containing test reagents (2X final conc) was aliquoted onto the CHO cells and incubated for an additional 1 hour. This incubation allows time for the test reagent to bind to PD-L1 on the surface of CHO cells. At 1 h, PD-1-expressing Jurkat NFAT luciferase reporter cells were collected in a 50 mL conical tube, pelleted at 380 g for 5 min at RT, and resuspended in fresh, warm RPMI complete medium. Cells were adjusted to 1.2e6/mL, and 25 μL (30,000) of cells were plated into wells with PD-L1-expressing OKT3 scFv CHT cells and test articles.

[00306] Cells and test reagent were further incubated for 18 h for activation of the Jurkat PD-1 reporter cell. Subsequently, SteadyGlo luciferase reagent was prepared and 100 μL aliquoted into each well. Complete lysis was achieved by gentle shaking at RT (200 rpm orbital shaker) for 15 min. After lysis, luciferase activity was measured on an Envision Multilabel plate reader using the US96 luminescence protocol. Luciferase RLU was plotted against log [test reagent] in Graphpad Prism software, and EC50 values for PD-L1 antagonism were determined using nonlinear regression analysis, fitting a 4-parameter sigmoidal dose response curve.

Results:

[00307] OX40/PD-L1 BiS2/3 were tested against parental PD-L1/PD-1 and NIP228(G4P) controls using a five-point dose titration with a starting point of 100nM (PD-L1). The OX40-PD-L1 BisAbs were both shown to be active and to have stronger agonism than the parental PD-L1 (4736) (Figure 37). The BiS2 and BiS3 formats performed similarly.

CMV Ag recovery assay

[00308] In the CMV Ag recovery assay (using the protocol described above), the BiS2 and BiS3 molecules demonstrated equal activity towards the combination (Figure 38).

[00309] All of the binding and immune response assays discussed above provide illustrative data that the bispecific binding molecules disclosed herein exhibit specific binding to both target molecules - in some cases greater binding activity than the combination of the individual monospecific parent binding molecules (antibodies), and can induce or enhance an immune response. Furthermore, the molecules are demonstrated to have cell killing activity against a cancer cell line. As such, the data show that these molecules and bispecific platform construct(s) represent excellent candidates for immuno-oncology therapeutics.

Octet binding assay (OX40(SLR)/PD-L1 BiS5)

[00310] To assess the binding of the bispecific binding molecules disclosed herein, an Octet QK equipped with Ni-NTA biosensor tips and 10X kinetic 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 proteins) were purchased from R&D Systems (Minneapolis, MN). All binding assays were performed at 25°C. The binding results show that the BiS5 OX40Ab/PD-L1 molecules bind to both PD-L1-His and hOX40-Fc (Figure 39).

[00311] PD-L1/OX40 BiS5 binds to CHO cells expressing human OX40 and PD-L1/B7H1 or cynomolgus (Figures 40A-F). Binding of PD-L1/OX40 BiS5 constructs was also measured by flow cytometry (HyperCyt) (Figure 42). OX40 IgG4P and OX40/PD-L1 bispecifics bind to Jurkat OX40 reporter cells. PD-L1 IgG and OX40/PD-L1 bispecifics bind to NCI H358 and CHOK1 B7H1(PD-L1)/OKT3 cells. All IgG and bispecifics bind to HEK CD32a cells.

PD-L1 and OX40 reporter assay

[00312] In the PD-L1 reporter assay (using the protocol described above), all PD-L1 scFv containing bispecifics and positive control IgG exhibited activity (Figures 42A-B). Single-arm OX40 controls and isotype controls did not exhibit any activity in this assay. The EC50 values and slopes are consistent with values obtained in previous assays for the PD-L1 parental controls and the PD-L1 Bis2, Bis3, and Bis5 constructs.

[00313] In an OX40 reporter gene assay using HEK CD32a cells, the bispecific constructs had equal activity to each other, and Fc-mediated agonism was observed (Figures 43A-B). OX40/PD-L1 Bis5 N434A IgG1 had equivalent EC50 activity to OX40 IgG4P and MEDI0562 (OX40 IgG1).

[00314] In an OX40 reporter gene assay using cells overexpressing CHOK1 PD-L1, OX40/PD-L1 bispecifics show equal agonism (Figures 44A-B). OX40/PD-L1 Bis5 N434A IgG1 had equivalent EC50 activity to other Fc variants of OX40/PD-L1 Bis5 bispecific Mab tested. No agonism was detected with OX40 IgGs or PD-L1 IgG. Thus, PD-L1-mediated OX40 agonism was demonstrated.

[00315] PD-L1-mediated OX40 agonism with tumor cells using OX40/PD-L1 bispecific molecules was detected (Figures 45A-B). The OX40/PD-L1 bispecific molecules showed equal agonism in this assay - bell-shaped curves. No agonism was observed with OX40 IgGs, thus demonstrating a benefit of using bispecifics over the combination of OX40 IgG plus PD-L1 IgG. No agonism was observed with NCI H358 PD-L1 KO cells (Figures 46A-D) showing that the agonism seen with NCI H358 cells is PD-L1 specific.

Staphylococal Enterotoxin B (SEB) Assay

[00316] In the SEB assay, the OX40/PD-L1 bispecific molecules had greater activity than the combination of individual antibodies to OX40 and PD-L1 (Figures 47A-D). In particular, the G4P construct had greater activity than the G1 construct. A wild-type YTE-containing variant and the N434A variant had equivalent activity.

Treg suppression assay

[00317] A Treg suppression assay was performed to test the OX40/PD-L1 bispecific molecules (Figures 4A-D). The OX40/PD-L1 bispecific molecules were active on CD4+ Tef only in the presence of PD-L1 (Figures 49 and 50). Without being bound by theory, this indicated cross-linking of OX40 in trans. The OX40/PD-L1 bispecific molecules suppressed the inhibitory effects of Treg, but only when cross-linked by binding to plate-immobilized PD-L1.

[00318] Mixed Leukocyte Reaction (MLR) Assay

[00319] An MLR assay was performed to test the OX40/PD-L1 bispecific molecules (Figures 51A-B). The OX40/PD-L1 bispecific molecules had greater activity than the combination of individual antibodies to OX40 and PD-L1 (Figure 52A-E).

Antibody-dependent Cell-mediated Cytotoxicity (ADCC) Assay

[00320] An ADCC assay was performed to test OX40/PD-L1 bi-specific molecules. ADCC assay using fresh isolated NK cells as effector cells and CHOK1 overexpressing PD-L1 B7H1 and CHOK1 OX40 cells, respectively, as target cells at an effector to target (E:T) ratio of 20:1. Target cell lysis was analyzed using Europium release from the labeled target cells after 5 hours. In the ADCC assay, OX40/PD-L1 BiS2 and BiS5 mediated ADCC against CHO cells expressing PD-L1 or OX40 (Figures 53 and 54).

[00321] A CD107a mobilization assay was performed using freshly isolated NK cells as effector cells and CHO K1 cells overexpressing PD-L1 and OX40 as target cells at an E:T ratio of 10:1. CD107a mobilization on the cell surface of NK cells was analyzed by flow cytometry after 4 hours. The OX40/PD-L1 bispecific molecules BiS2 and BiS5 enhanced CD107a mobilization of NK cells against CHO cells expressing PD-L1 and OX40 in antibody-dependent cell-mediated cytotoxicity (ADCC) assays (Figure 55). The bispecific molecules BiS2 and BiS5 OX40/PD-L1 increased CD107a mobilization of NK cells against activated allogeneic T cells that upregulated OX40 and PD-L1 (Figure 56). BiS5 OX40/PD-L1 increased CD107a mobilization of NK cells from two different donors against activated allogeneic T cells (Figures 57A-B).

Pharmacokinetic and Pharmacodynamic Studies (PK/PD)

[00322] A study was designed to compare the PK/PD of OX40/PD-L1 bispecific molecules (Figure 58). Serum concentration-time profiles for PD-L1/OX40 bispecific molecules were compared in cynomolgus monkeys (Figure 59 and Table 11). The mean T1/2 for the Bis5 OX40/PD-L1 IgG1 N434A molecule was greater than that for the WT Bis5 molecule; the clearance rate was lower for the Bis5 OX40/PD-L1 IgG1 N434A molecule compared to the WT Bis5 molecule. Both molecules similarly reduced serum soluble PD-L1, and elicited a significant increase in the percentage of total Ki67+ memory CD4+ T cells, total memory CD8+ T cells, and NK cells. Table 11. Pharmacokinetic parameters of BiS5-OX40/PD-L1-G1

[00323] OX40/PD-L1 bispecific molecules reduced serum soluble PD-L1 concentrations below the assay LLOQ (Figure 60). The N434A mutation improved the pharmacokinetics of BiS5-OX40/PD-L1-G1. In particular, CL was reduced by approximately half; there was a corresponding 2-fold increase in T1/2 and AUCinf; and Cmax and Vss were not impacted. This was consistent with previously reported effects of this mutation on monoclonal antibody PK. Thus, progress was made toward mAb-like PK for BiS5-OX40/PD-L1-G1 IO BisAb. Serum concentrations of BiSA5-OX40/PD-L1-G1 BiSAbs were below the limit of quantification (BLOQ) at 2 weeks, and may have been related to ADA.

[00324] Substantial and statistically significant increases were observed in total memory CD4, total memory CD8, and NK cell proliferation (percentage of Ki67+ cells) comparing the PD-L1 OX40 Bis5 groups to the control Ab group (anti-PcrV-Psl control) (Figures 61A-F). A trend towards 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 were no statistically significant differences in proliferation between the PD-L1 OX40 Bis5 N434A (extended half-life) version and the G1 version. The PD-L1 OX40 Bis5 N434A and IgG1 versions are biologically active bispecific molecules. Example 2(e). OX40/PD-1 bispecific binding proteins

[00325] The following bispecific binding proteins that bind to PD-1 and OX40 were created using the parental sequences identified above in Table 2. Proteins identified as Bis2 and Bis3 were generated with the sequences in Table 24 below and were evaluated for competitor antigen binding activity using the Octet binding assay as discussed below. Table 24. BiS Constructs for OX40/PD-1

[00326] The PD-1/OX40 monoclonal antibody (mAb) BiS2 is a bispecific antibody (Figure 62; PD-1 binding proteins depicted in gray and OX40 binding proteins depicted in light gray) designed to concurrently bind to human and cynomolgus monkey PD-1 and human and cynomolgus monkey OX40. Without being bound by theory, a proposed mechanism of action suggests dual signaling effects on T cells following binding in cis to both OX40 and PD-1, agonism of the T cell costimulatory surface receptor OX40, and blockade of the immunosuppressive PD-1 (Figure 63).

Octet binding assay

[00327] Concurrent binding activity was shown for two different lots of the PD-1(LO115)/OX40 BiS2 mAb to human PD1-His and OX40-Fc (Figure 64).

OX40 reporter essay

[00328] PD-1(LO115)/OX40 BiS2 mAb showed comparable activity to other OX40 agonists (Figures 65A-B). Proteins were stored at 4°C and used immediately, freeze/thawed three times, stored at 4°C for 7 days, and stored at 40°C for 7 days. Activity was reported as relative light units versus mAb concentration. The EC50 was ~2 nM for PD1(LO115)/OX40 BiS2 mAb at 4°C on day 0.

PD-1/PD-L1 reporter assay

[00329] PD-1(LO115)/OX40 BiS2 mAb showed comparable activity to other PD-1 agonists (Figures 66A-B). Proteins were stored at 4°C and used immediately, freeze/thawed three times, stored at 4°C for 7 days, and stored at 40°C for 7 days. Activity was reported as relative light units versus mAb concentration. The EC50 was ~1 nM for PD1(LO115)/OX40 BiS2 mAb at 4°C on day 0. Two sets of primary human in vitro potency assays were performed; an antigen memory T cell assay and a T cell costimulation using staphylococcal enterotoxin B (SEB).

Staphylococal Enterotoxin B (SEB) Assay

[00330] In the SEB assay, PD-1(LO115)/OX40 BiS2 mAb induced an increase in IL-2 levels detected in the cell supernatant after 3 days in culture (Figure 67). Thus, PD-1/OX40 BiS2 mAb can concurrently bind to its human target antigens and can co-stimulate T cells in vitro.

[00331] In the antigen retrieval assay, PD-1/OX40 BiS2 mAb promoted an increase in interferon (IFN)-gamma levels compared to the parental mAbs and the combination of parental mAbs (Figures 68 and 69).

CMV Ag recovery assay

[00332] The results of the CMV Ag recovery assay (using the protocol described above), the BiS2 and BiS3 molecules did not demonstrate equal activity towards the combination (Figure 70). The data show that PD-1/OX40 BiS2 IgG4P mAb is active in vitro and in vivo. PD-1/OX40 BiS3, which differs in structure from PD1/OX40 BiS2, was not detectably active. Thus, PD-1/OX40 BiS3 (not active) is distinct from BiS2 (active).

Pharmacokinetic and Pharmacodynamic Studies (PK/PD)

[00333] The cynomolgus monkey was considered a pharmacologically relevant non-clinical species for testing 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. The PK and PD of PD-1(LO115)/OX40 BiS2 mAb (percentage of Ki67-positive total CD4+ and CD8+ memory T cells) were evaluated in cynomolgus monkeys (n = 3; males) following a single intravenous (IV) dose across the dose scale of 0.1 mg/kg to 30 mg/kg. PBMC were collected pre-dose and on days 1, 8, 11, and 15 post-dose, cryopreserved, and thawed prior to analysis by flow cytometry. In summary, PD1(LO115)/OX40 BiS2 mAb exhibited approximately linear PK with a short half-life of 0.61.7 days (Figure 70; Table 12). Table 12. Mean pharmacokinetic parameters of Values are presented as Mean (Standard Deviation). AUClast = area under the concentration time curve to the last measurable concentration; AUCINF = area under the concentration time curve to infinity; Cmax = C max = maximum observed concentration; CL: systemic clearance; T1/2 = half-life; Vss: terminal phase volume of distribution; Vss: volume of distribution at steady state.

[00334] Mean maximum concentrations (Cmax) increased approximately dose-proportionally from 2.0 μg/mL at 0.1 mg/kg to 607 μg/mL at 30 mg/kg. AUC~ increased approximately dose-proportionally 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 total CD4+ memory T cell proliferation (Ki67) and an increase in total CD8+ memory T cell proliferation (Ki67). A representative standard curve for quantification of PD-1/OX40 in cynomolgus monkey serum is shown (Figure 72). Example 3. Physical and chemical stability of BiSAb constructs

[00335] A series of experiments were performed in order to evaluate and verify the physical and chemical stability of the BiSAb constructs as described herein, relative to other bispecific binding protein structural strategies and platforms. In particular, the series of stability studies discussed below identified and analyzed the effect of various pH ranges on the stability of BiSAbs (e.g., hydrolysis, fragmentation, aggregation, thermal stability). As the data show, for the different illustrative embodiments of the various BiSAb formats, the BiSAb disclosed herein (identified as “BiS5” in the studies below, and in the D/H format as shown in Table 13) demonstrated unexpected and surprising physical and chemical stability relative to all other BiSAb structural motifs. Table 13. BiS5 Stability Study Constructs

Example 3.1

[00336] A further comparison was made between the BiS format disclosed herein ("BiS5") and another BiS format that includes two binding domains (scFv domains) linked at the hinge region (i.e., between the Fc and Fab regions), identified as "BiS4". The BiS4 and BiS5 proteins were expressed in Chinese hamster ovary (CHO) and purified by routine chromatographic methods. As noted above, these two formats have similar Fab and scFv sequences, with the main difference between them being the location of the scFv domain (for BiS4, the scFv is located at the hinge region; for this particular BiS5, an scFv is located at the SNG loop in the CH3 domain, as discussed herein). Purified BiS molecules were supplied in PBS buffer and the protein concentration was determined using NanoDrop ND-1000 (Thermo Scientific, Wilmington, Delaware) using an extinction coefficient of 1.54 M-1 cm-1.

Study of pH Assessment and Short-Term Stability

[00337] For pH evaluation studies, BiS4 and BiS5 antibodies were concentrated to ~12mg/mL and dialyzed against 6 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 cutoff (MWCO), Thermo-Fisher, Rockford, Illinois). After completion of dialysis, 0.02% polysorbate 80 was applied and the final protein concentration was adjusted to ~10mg/mL. The BiS4 and BiS5 formulations were sterilized using a 0.22 μm filter (Millipore, Billerica, Massachusetts) in a presanitized laminar flow dryer. One-milliliter aliquots were dispensed into 3-mL type I borosilicate glass vials (West Pharmaceutical Services, Exton, Pennsylvania). Samples were stored at 40°C and analyzed by SEC at time zero and after storage for 1, 2, 3, and 4 weeks.

Differential Scanning Calorimetry (DSC)

[00338] Differential scanning calorimetry thermograms for samples at time zero were obtained using a VP-Capillary DSC connected to a temperature-regulated autosampler (Malvern Instruments Ltd., Westborough, Massachusetts). A protein concentration of 1 mg/mL was used for thermogram acquisition along with a scan rate of 90°C/h over the temperature range of 20°C-100°C. Thermograms of BiS4 and BiS5 at different pH conditions ranging from 5.0 to 7.5 were buffer subtracted and baseline corrected. Data analysis was performed using the DSC plug-in for the Origin 7 SR4 software package. The experimental results were fitted to a multistate model with three transitions in order to calculate the melting temperature (Tm) values. The point where the heat capacity value (Cp) for the first thermal transition reached 500 cal mol-1°C-1 was considered as the initial temperature (Tinicial).

High Performance Size Exclusion Chromatography (HP-SEC)

[00339] To separate the aggregate and fragment species of the monomer based on size, stability samples were analyzed using an Agilent high performance liquid chromatography system with a photodiode array detector capable of recording UV absorbance spectra from 200-400 nm with 7.8 x 30 cm2, 5μm, 250Å, Tosho TSKgel G3000SWxl (TOSOH Biosciences, King of Prussia, Pennsylvania) and a corresponding guard column. To separate the species, a mobile phase containing 0.1 M sodium phosphate dibasic anhydrous, 0.1 M sodium sulfate, 0.01% sodium azide, pH 6.8 and a flow rate of 1 mL min min -1 were used. The amount of protein injected was approximately 250 μg. The separation of BiS4 and BiS5 was monitored using the absorbance spectrum at 280 nm. The peak areas for soluble aggregates (multimer and dimer), monomer, and fragments were quantified. Then, the percentage of each species was calculated and plotted against the incubation time to develop kinetic plots. The pH profile curves for the rate of monomer loss, fragmentation, and aggregation per month were developed by calculating the slope of each kinetic plot.

Thermal stability of BiS4 and BiS5

[00340] The effect of pH on the thermal stability of BiS4 and BiS5 was evaluated by analyzing thermograms obtained using capillary DSC and generated at six different pH conditions. Figures 74A and 74B show overlays of DSC thermograms of BiS4 and BiS5, respectively, from pH 5.0 to 7.5. As shown in Figures 74C and 74D, each thermogram shows three thermal unfolding events with transition temperatures Tm1, Tm2, and Tm3. The first transition (Tm1) is likely associated with the simultaneous unfolding of the CH2 and scFv domains, while the second (Tm2) and third (Tm3) transitions are associated with the unfolding of the CH3 and Fab domains. For both formats, an increase in Tonset, Tm1, Tm2, and Tm3 was observed with increasing pH up to 6.5 (Figures 73A, 73B, 73E and Table 14, below). For BiS4 and BiS5, no differences in Tonset, Tm2, and Tm3 were observed across all pH conditions (Figure 73E and Table 15), indicating that the presence of scFv in either the hinge region or the CH3 domain does not impact the thermal stability of CH3 and Fab. Interestingly, a slight increase in Tm1 was observed for BiS5 across all pH conditions, indicating an increase in the thermal stability of either scFv, CH2, or both when the scFv is located in the CH3 domain. Table 14: Effect of pH on thermal onset temperature (Tinicial) and thermal melting temperatures (Tm1, Tm2, and Tm3) for BiS4 and BiS5 as measured by capillary DSC.

Physical and Chemical Stability of BiS4 and BiS5

[00341] The physical and chemical stability of the BiS4 and BiS5 formats at different pH values (ranging from 5.0 to 7.5) was evaluated at 40°C for up to 4 weeks. The HP-SEC chromatograms at "time zero" were used to compare the total area, monomer, aggregate, and fragment content of the HP-SEC chromatograms with 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 lines), the majority of the sample is monomer, with no other notable differences except for a small difference in peak height between the two samples, likely due to a small difference in concentration (Figure 74A). The dashed lines show an overlay of HP-SEC chromatograms of both formats under the same pH conditions after storage for 4 weeks at 40 °C. Under accelerated temperature stress conditions, both formats show an additional peak, early elution peak (multimeric species), a decrease in monomer, and elevated levels of fragments (Figure 74A). Monomer loss due to fragmentation was more prominent in BiS4 compared to BiS5, indicating that BiS5 is chemically more stable. Based on its structure, possible fragmentation sites, and retention time, we speculate that the small fragment peak (TR - 10.8 min) is a Fab, and the large fragment peak (TR -9.8 min) and shoulder peak (TR -8.7 min) is a Fab with scFv and its corresponding higher molecular weight fragment (HMWF) with Fab, scFv, and Fc, respectively.

[00342] To further assess the effect of scFv location on the physical and chemical stability of BiS4 and BiS5, the percent total area for each species was plotted in a bar graph for time zero and 4 weeks at 40°C for pH 7.5 (Figure 74B). As shown in Figure 74B, at time zero, the monomer purity for 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. For BiS4 samples, 11.8%, 7.2%, and 3.5% of the shoulder peak (TR -8.7 min), large fragment (TR -9.8 min), and small fragment (TR -10.8 min) were formed, respectively (Figure 74B). Surprisingly, the BiS5 sample showed only 1.4% small fragment (RT -10.8 min), likely due to scFv tethering on both sides of the para Fc domain (Figure 89).

[00343] Figures 75A-75C show the kinetics of aggregation, fragmentation, and monomer loss for BiS4 and BiS5 in samples incubated at 40°C at pH 7.5. The BiS4 samples showed a faster rate of monomer loss compared to BiS5 (Figure 75A). The rate of monomer loss for BiS4 and BiS5 at pH 7.5 was 27.4%/month and 4.5%/month, respectively (Figure 75A). For BiS4, the majority of the monomer loss is due to fragmentation which was 23.9%/month and to a lesser extent due to aggregation which was 3.5%/month (Figures 75B and 75C). Interestingly, for BiS5, aggregation appeared to be at a slightly higher rate (2.8%/month) compared to the fragmentation rate (1.7%/month) (Figures 75B- 75C).

[00344] Further analysis of the effect of pH on the physical and chemical stability of the BiS4 and BiS5 formats, and the rate of monomer loss, fragmentation, and aggregation per month, was performed by plotting these values against the 6 pH conditions (Figures 76A-76C). Across all six pH conditions ranging from pH 5.0-7.5, the rate of monomer loss was lower for the BiS5 format compared to BiS4 (Figure 76A), suggesting that the BiS5 format disclosed herein has unexpectedly superior physical and chemical stability to other bispecific protein formats. For BiS4, the majority of monomer degradation was due to fragmentation, even at lower pH conditions (Figure 76B). BiS5 exhibited lower fragmentation rates when compared to BiS4 at all pH conditions tested. Surprisingly, the fragmentation rates on BiS5 appeared to be flat and lower over a wider pH range compared to BiS4. Without being bound by any particular theory, it may be that the lower fragmentation rates observed on BiS5 may arise from the G4S linkers at either end of scFv linking it to Fc. Fragmentation at one of the G4S linkers linking Fc may not release the scFv, as it may still be linked to Fc via the other G4S linker. On BiS4 and BiS5, the aggregation rates appeared to be similar across all pH conditions tested (Figure 76C), suggesting that the location of scFv has minimal effect on aggregation kinetics, which is also supported by the unobserved change in Tinitial between the two formats across all pH conditions as measured using capillary DSC (Figure 73E and Table 14, discussed above). At pH 7.5 and 40 °C (at time = 0), no molecules exhibited appreciable fragmentation (Figure 77A), but under the same conditions after 2 weeks of storage at 40 °C, appreciable fragmentation is observed for BiS4 and minimal fragmentation for BiS5 (Figure 77B). Both BiS4 and BiS5 have lower fragmentation and aggregation at lower pH (5.5), while BiS5 outperforms at both pH values for both fragmentation and aggregation (Figure 78). This series of experiments demonstrates that BiS5, disclosed herein, has superior chemical stability and similar physical stability to BiS4.

Example 3.2

[00345] Additional studies were performed in order to evaluate the physical and chemical stability of different embodiments of the bispecific binding proteins that are disclosed herein and identified as Constructs A-H (see, for example, Table 13 and related Examples, above). These constructs were analyzed using DSC, accelerated storage stability, and FcRn and FcgR binding assays, as follows below.

Differential scanning calorimetry analysis

[00346] DSC experiments for this dataset 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 prior to loading into the calorimeter. Antibodies used for DSC studies were >98% monomeric as determined by analytical SEC. Prior to DSC analysis, all samples were extensively dialyzed (at least 3 buffer changes) into 25 mM histidine-HCl (pH 6.0). The buffer from this dialysis was used as the reference buffer for subsequent DSC experiments. Prior to sample measurement, baseline measurements (buffer versus buffer) were subtracted from the sample measurement. Dialyzed samples (at a concentration of 1 mg/mL) were added to the sample well 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-2-state model and best fits were obtained using 100 iteration cycles. Tinitial is defined as the qualitative temperature at which the thermogram appears to have a non-zero slope, Tm is defined as the temperature at which half of the molecules in an ensemble are unfolded, and is calculated as the temperature value corresponding to each maximum peak in the thermogram.

[00347] The results for the different constructs are presented in Figure 79. Generally, constructs A, C, D, which include 2F4 as scFv have lower TM1 when compared to constructs E, G, H which include LC10 as scFv. Without being bound by theory, the difference in TM1 value may be due to an inherently better thermal stability of the scFv domain of LC10 relative to the 2F4 variable domain. The data suggest that constructs A-D, having 2F4 as scFv, would be less thermally stable than constructs E-H with LC10 as scFv.

Accelerated storage stability analysis

[00348] Construct concentrations were normalized to 1 mg/mL. 1 mL of each bispecific construct or IgG control was aliquoted into 1.5 mL Eppendorf tubes. Samples were incubated in a static incubator for 2 weeks at 45°C. Samples were analyzed at days 3, 7, and 14 and assessed for stability. At each time point, a visual inspection was performed to record any increased turbidity or precipitation. Samples were filtered using a 0.2 μm spin column and 120 μL of sample was aliquoted into an HPLC vial, ensuring that there were no air bubbles at the bottom of the vial. The samples were then tested on an Agilent 1100 series HPLC-SEC to check for aggregation and degradation using a Tosoh Bioscience TSK-GEL G3000SWXL (300 x 7.8 mm) column with 0.1 M sodium phosphate, 0.1 M sodium sulfate, pH 6.8 as the processing buffer. 60 μL of sample was injected and processed at a flow rate of 1 mL/min. The monomer retention time (mins), total peak area, % monomer, % aggregate, % fragment, % monomer loss were captured and used for analytical SEC analysis. The results are summarized in Table 15. Table 15: Accelerated stability studies.

[00349] As described herein, the location of the scFv domain in the above constructs is as follows (where the "-" indicates scFv): A and E are in IS-RTP; B and F are in AK-GQP; C and G are in S-NG; and D and H are in SN-G. The various TM values are associated with the following domains, M1 = CH2/scFv; TM2 = Fab; TM3 = CH3. The data tend to show 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 which have LC10 scfv inserted into the same locations. This observation suggests that the sequence identity and behavior of the scFv domain may have an effect on the stability of the bispecific binding protein constructs. Furthermore, from the above, it could be predicted that construct D, which contains the 2F4 scFv, would have a similar lower stability to A and C; however, it appears that the insertion of 2F4 scFv into the SNG loop stabilizes the molecule and reduces the tendency to form aggregates. Taken together, this accelerated stability study indicates that the sequence and location of scFv within the Fc region may play a measurable role in the stability of the BiSAb construct.

FcRn and FcyR binding analysis

[00350] Binding experiments were performed using a BIAcore 3000 instrument (BIAcore). To capture antibody, 1000 RU IsdH antigen (Fab) was immobilized on a CM5 chip. 100 nM of the BiSAb construct or mAb controls were flowed at 20 μL/min for 5 min to capture antibody. 5uM of huFcRn or FcyR I, IIa, IIb, IIIa-158V or IIIA158F was flowed at 5 μL/min for 20 min. FcRn binding was performed in PBS + 5 uM EDTA at pH 6.0 while FcRn binding was performed in PBS + 5 uM EDTA at pH 7.4.

[00351] Constructs A, C, D, E, G, and H were evaluated for FcRn binding. Representative data are shown in Figure 80 for each of the bispecific constructs E and H, and for the binding of 2F4 IgG to FcRn. Constructs having scFv downstream of the CH2-CH3 interface appear to retain FcRn binding (e.g., constructs D & H). Constructs having scFv located within the ISRTP loop upstream of the CH2-CH3 interface appear to abolish detectable FcRn binding (e.g., constructs A & E). The ISRTP loop is within the region of the half-life-extending YTE mutations in the Fc (M252Y/S254T/T256E) that are known to be important for FcRn binding.

[00352] Constructs A, C, D, E, G, and H were tested for binding to FcyRI, FcyRIIa, FcyRIIb, FcyRIIIa-158F, and FcyRIIIa-158V. Representative data are presented in Figure 81 for constructs E, G, and H that bind to FcyRIIIa-158V. All constructs tested retained binding to FcyRs, albeit with different affinities (Figure 81, inset). Table 16 shows the observed binding trends of the various constructs to FcyRs. Table 16: FcyR binding trends.

[00353] The observed differences in FcyR binding with the constructs having scFv inserted into the ISRTP loop upstream of the CH2-CH3 interface (A and E) consistently show reduced FcyR binding, when compared to the other constructs having scFv inserted into the SNG loop downstream of the CH2-CH3 interface (C, D, G and H).

[00354] Attempts to assess whether FcRn binding in constructs A and E could be improved or restored were made by introducing a half-life extending loop (N3) into the Fc region. Figure 82 is representative of the data and shows that for constructs E, neither the BiS5Ab E nor the E construct with the introduced N3 loop (BiS5Ab E + N3) were able to bind FcRn. Furthermore, inserting LC10 scFv into the N3 loop (N3 scFv) and keeping the ISRTP loop intact also decreased FcRn binding below detectable levels. These data indicate that, at least for construct E, if not for each construct disclosed herein, both the ISRTP loop and the N3 loop, if present, need to be intact and unmodified in order to retain FcRn binding.

Example 3.3

[00355] In addition to the comparison between BiS4 and the bispecific binding constructs disclosed herein (BiS5), a study was conducted to evaluate three other BiS structural motif platforms, identified as BiS1, BiS2 and BiS3 (see, Figure 83). As will be appreciated by reference to Figure 83, these platforms vary in terms of the location of one of the binding domains (illustrated as an scFv domain). Of the five motifs, only BiS4 and BiS5 include two linker moieties as binding sites to the larger protein, the others (BiS1, BiS2 and BiS3) are bound by a single linker.

[00356] Briefly, representative molecules of each construct were analyzed for stability using the techniques discussed above in Examples 3.1 and 3.2. Samples of each construct were added to pH 5.0, 5.5, 6.0, 6.5, 7.0, and 7.5 buffers and stored at 40°C for a period of two months. The samples were then analyzed for fragmentation rate (Figure 84)), aggregation rate (Figure 85), and monomer loss rate (Figure 86) using HP-SEC. Under these conditions, the analysis indicated that the bispecific binding protein format disclosed herein (“BiS5”; and the D/H format as indicated above in Table 13) had superior physical and chemical stability relative to all other formats under all pH conditions.

[00357] The SEC data were also used to map the various peaks in the corresponding fragments of the BiS molecules (Figure 87). The mapping was based on assumptions including that fragmentation occurs in the hinge and linker region of the molecules, the fragment size, the theoretical fragmentation, and how the expected fragment species align relative to the fragment species observed in other formats. Although there is good resolution between low molecular weight fragments (LMWF) in each format, there is also poor or no resolution between monomers and higher molecular weight fragments (HMWF) in all formats. Based on the information in Table 17, it was concluded that the HP-SEC technique underestimates fragmentation in BiS formats to a greater extent than for monoclonal antibodies. An alternative analysis was developed as discussed below. Table 17: Fragmentation Analysis - SEC Underestimates Fragmentation Rates of BiS Formats. *Retention Time 1Monomer loss includes loss due to fragmentation only and does not include aggregation. 2Rates may be underestimated due to coelution of HMWF with monomer. 3Rates for the shoulder peak may vary due to suspended integration.

[00358] An alternative analysis was developed in order to calculate HMWF fragmentation rates using a molar extinction coefficient based on the assumptions that (i) if a small fragment is detected during degradation, a corresponding large fragment must also be present; (ii) secondary fragmentation (fragment fragments) does not occur significantly during the duration of the stability study; and (iii) fragmentation occurs in the linker region and/or the hinge region. Fragmentation rates were determined based on the following relationships: Table 18: Inferred fragmentation analysis. 1Monomer loss includes loss due to fragmentation only and does not include aggregation. Fragmentation was observed for all BiS formats except for BiS1 (Table 19). It was concluded that the higher fragmentation rates under reducing conditions confirm that the scFv moiety in the BiS5 construct is bound to the CH3 region (Figure 89). Table 19: Fragmentation analysis-reducing conditions. Table 20: Overview of fragmentation analyses.

[00359] The characteristics of the stabilizing disulfide bonds 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 into the SN-G loop) formats with and without the stabilizing disulfide bond in the scFv. The accelerated stability study indicated that the BiS4 constructs without a 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 removal of the stabilizing disulfide bond in the scFv in the BiS5 construct did not have a significant effect on its stability.

[00360] Based on all of the above data, it appears that the bispecific binding protein format disclosed herein is the most stable of all formats tested. Furthermore, BiSAb5 appears to be most stable at the lower pH values tested (e.g., 5.0, 5.5, and 6.0) in terms of minimizing both fragmentation and aggregation. As such, the unexpected and surprising stability characteristics of the BisAbs disclosed herein provide an additional advantage over other platforms and structural formats that are used in the engineering of bispecific binding molecules.

Incorporation by Reference

[00361] All publications and patents referenced herein are hereby incorporated by reference in their entirety as if each individual publication or patent were specifically and individually indicated to be incorporated by reference.

[00362] Although specific aspects of the present disclosure have been discussed, the above specification is illustrative and not restrictive. Many variations of the disclosure will become apparent to those skilled in the art upon review of the present specification and claims below. The full scope of the disclosure should be determined by reference to the claims, together with their full scope of equivalents, and the specification, together with such variations.

Claims (19)

1. A bispecific binding protein that binds to PD-1 and TIM3, characterized in that it comprises a first heavy chain, a first light chain, a second heavy chain, and a second light chain, wherein the first heavy chain comprises: a first heavy chain CDR1 comprising amino acids having the sequence set forth in residues 24-35 of SEQ ID NO: 9, a first heavy chain CDR2 comprising amino acids having the sequence set forth in residues 50-66 of SEQ ID NO: 9, and a first heavy chain CDR3 comprising amino acids having the sequence set forth in residues 99-111 of SEQ ID NO: 9; wherein the first light chain comprises: a first light chain CDR1 comprising amino acids having the sequence set forth in residues 24-41 of SEQ ID NO: 7, a first light chain CDR2 comprising amino acids having the sequence set forth in residues 55-61 of SEQ ID NO: 7, and a first light chain CDR3 comprising amino acids having the sequence set forth in residues 94-100 of SEQ ID NO: 7; wherein the second heavy chain comprises: a second heavy chain CDR1 comprising SEQ ID NO: 79, a second heavy chain CDR2 comprises SEQ ID NO: 80, a second heavy chain CDR3 comprises SEQ ID NO: 81; and wherein the second light chain comprises: a second light chain CDR1 comprising SEQ ID NO: 82, a second light chain CDR2 comprising SEQ ID NO: 83, and a second light chain CDR3 comprising SEQ ID NO: 84. 2. The bispecific binding protein of claim 1, wherein the first heavy chain comprises the amino acid sequence of SEQ ID NO: 9, wherein the first light chain comprises the amino acid sequence of SEQ ID NO: 7, wherein the second heavy chain comprises the amino acid sequence of SEQ ID NO: 30, and wherein the second light chain comprises the amino acid sequence of SEQ ID NO: 28. 3. 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 CDR1 of the heavy chain comprises SEQ ID NO: 79, the CDR2 of the heavy chain comprises SEQ ID NO: 80, the CDR3 of the heavy chain comprises SEQ ID NO: 81, and the CDR1 of the light chain comprises SEQ ID NO: 82, the CDR2 of the light chain comprises SEQ ID NO: 83, the CDR3 of the light chain comprises SEQ ID NO: 84. 4. The antibody or antigen-binding fragment thereof of claim 3, 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. 5. The antibody or antigen-binding fragment thereof of claim 3, wherein the heavy chain comprises SEQ ID NO: 87 and the light chain comprises: SEQ ID NO: 88. 6. Composition characterized by the fact that it comprises the bispecific binding protein as defined in claim 2 and a pharmaceutically acceptable carrier. 7. A nucleic acid molecule comprising a nucleotide sequence encoding the bispecific binding protein as defined in claim 2. 8. Vector, characterized by the fact that it comprises the nucleic acid molecule as defined in claim 7. 9. Host cell, characterized by the fact that it comprises the vector as defined in claim 8. 10. Use of the bispecific binding protein as defined in claim 2, characterized in that it is for the preparation of a medicament for treating cancer in a subject. 11. Use according to claim 10, characterized in that the cancer is one or more of ovarian cancer, breast cancer, colorectal cancer, prostate cancer, cervical cancer, uterine cancer, testicular cancer, bladder cancer, head and neck cancer, melanoma, pancreatic cancer, renal cell carcinoma and lung cancer. 12. Use of the bispecific binding protein as defined in claim 2, characterized in that it is for the preparation of a medicament for increasing an immune response in a subject. 13. Composition, characterized by the fact that it comprises the bispecific binding protein as defined in claim 1 and a pharmaceutically acceptable carrier. 14. A nucleic acid molecule comprising a nucleotide sequence encoding the bispecific binding protein as defined in claim 1. 15. Vector characterized by the fact that it comprises the nucleic acid molecule as defined in claim 14. 16. Host cell, characterized by the fact that it comprises the vector as defined in claim 15. 17. Use of the bispecific binding protein as defined in claim 1, characterized in that it is for the preparation of a medicament for treating cancer in a subject. 18. Use according to claim 17, characterized in that the cancer is one or more of ovarian cancer, breast cancer, colorectal cancer, prostate cancer, cervical cancer, uterine cancer, testicular cancer, bladder cancer, head and neck cancer, melanoma, pancreatic cancer, renal cell carcinoma and lung cancer. 19. Use of the bispecific binding protein as defined in claim 1, characterized in that it is for the preparation of a medicament for increasing an immune response in a subject.