Bispecific binding proteins and their uses

Abstract

To provide bispecific molecules or proteins that bind to two epitopes and are bivalent for binding to each of the two epitopes as cancer immunotherapeutics.SOLUTION: Provided is a protein comprising: a first binding domain (BD1) that binds to a first epitope; a second binding domain (BD2) that binds to a second epitope; and an Fc region comprising CH2 and CH3 domains. The Fc region comprises the BD2 at a solvent exposed loop in the CH2 domain, the CH3 domain, or at an interface of the CH2 and CH3 domains. The protein is bivalent for binding to each of the first and second epitopes. Also provided is a method for inducing an immune response in a subject, as well as a method for treating or preventing cancer in a subject by administering the proteins, nucleic acid molecules and / or compositions to the subject.SELECTED DRAWING: Figure 1-1

Description

[Technical Field]

[0001] Cross-reference of related applications This application claims priority and benefit to U.S. Provisional Patent Application No. 62 / 332,788, filed on 6 May 2016, the disclosures of which are incorporated herein by reference in their entirety.

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

[0003] Technical field This invention relates to a bispecific binding protein and its use. [Background technology]

[0004] Cancer remains a major global health burden. Despite advances in cancer treatment, there remains an unmet need for more effective and less toxic therapies, particularly for patients with advanced disease or cancer that is resistant to existing treatments.

[0005] The role of the immune system, particularly T-cell-mediated cytotoxicity, in tumor control is well recognized. There is growing evidence that T cells control tumor growth and patient survival in cancer patients at both the early and late stages of the disease. However, enhancing and sustaining tumor-specific T-cell responses in cancer patients remains challenging. The continued advancements and successes of cancer immunotherapies that stimulate or enhance the innate immune response to cancer are making these therapies an attractive treatment option compared to therapies using nonspecific chemotherapy and / or radiation.

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

[0007] Therefore, there remains a need to identify candidate targets for IO therapies, develop novel therapies against existing targets, and develop therapeutic strategies that avoid the shortcomings of currently used immunotherapies (such as lack of patient responsiveness and increased risk of side effects associated with combination therapy). IO therapies with bispecificity for combinations of target molecules (e.g., binding proteins), particularly those exhibiting superior binding affinity to target molecules compared to binding affinity to combinations of individual monospecific binding proteins, represent a particularly desirable class of molecules in terms of therapeutic efficacy. [Overview of the project] [Means for solving the problem]

[0008] The present invention provides a bispecific molecule or protein that binds to two epitopes (e.g., a first and a second epitope) and is divalent to bind to each of the first and second epitopes. The present invention also provides a method for inducing an immune response in a subject, as well as a method for treating or preventing cancer in a subject (e.g., a human subject) by administering a protein, nucleic acid molecule and / or composition to the subject.

[0009] In one embodiment, the present invention comprises a first binding domain (BD1) that binds to a first epitope, a second binding domain (BD2) that binds to a second epitope, and C H 2 domains and C H A protein comprising an Fc region having 3 domains, wherein the Fc region is C H 2 domains, C H 3 domains, or C H 2 and C H The protein provides a three-domain solvent exposure loop containing BD2, and the protein is divalent to bind to each of the first and second epitopes.

[0010] In another embodiment, the present invention provides a composition comprising a protein or antibody according to any embodiment described herein and a pharmaceutically acceptable carrier.

[0011] In another embodiment, the present invention provides a method for treating or preventing cancer in a subject, the method comprising the step of administering a protein or antibody to a subject (e.g., a human subject) according to any embodiment described herein. 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.

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

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

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

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

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

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

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

[0019] In one embodiment, the present 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.

[0020] In one embodiment, the present 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.

[0021] In another embodiment, the present 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.

[0022] In another embodiment, the present 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.

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

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

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

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

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

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

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

[0030] In other aspects, the present 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, a method of treating or preventing cancer in a subject by administering a bispecific binding protein, and a method of enhancing an immune response in a subject by administering a bispecific binding protein.

[0031] In various embodiments of any aspect described herein, the Fc region is C H 2 domain, C H 3 domain, or C H 2 and CH The solvent exposure loop in the amino acid sequence at the interface of the three domains includes BD2.

[0032] In various embodiments of any of the aspects described herein, the solvent exposure loop is C H It contains amino acid sequences derived from two domains. In certain embodiments, the solvent exposure loop contains the amino acid sequence ISRTP (SEQ ID NO: 39).

[0033] In various embodiments of any of the aspects described herein, the solvent exposure loop is C H It contains amino acid sequences derived from three domains. In certain embodiments, the solvent exposure loop includes the amino acid sequence SNG.

[0034] In various embodiments of any of the aspects described herein, the solvent exposure loop is C H 2 domains and C H It includes an amino acid sequence derived from the interface of the three domains. In a particular embodiment, the solvent exposure loop is the protein according to claim 7, comprising the amino acid sequence AKGQP (SEQ ID NO: 40).

[0035] In various embodiments of any of the aspects described herein, BD2 is or includes a single-stranded variable fragment (scFv).

[0036] In various embodiments of any aspect described herein, BD1 is a binding domain which is one or more of a Fab domain, scFv, a single-domain antibody, and an antibody variable domain. In certain embodiments, BD1 comprises a Fab domain.

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

[0038] In various embodiments of any aspect described herein, the protein further comprises a protein linker L1 between BD2 and the Fc region. In various embodiments of any aspect described herein, the protein further comprises a first protein linker L1 and a second protein linker L2 between BD2 and the Fc region. In various embodiments of any aspect described herein, BD2 is linked to the Fc region via protein linker L1. In various embodiments of any aspect described herein, BD2 is linked to the Fc region via two protein linkers L1 and L2. In some 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).

[0039] In various embodiments of any aspect described herein, the protein has the following polypeptide domain:V from the N-terminus to the C-terminus. H 1-C H 1-C H 2(N-terminus)-BD2-C H 2(C-terminus)-C H It contains a chimeric heavy chain having 3, and BD1 contains a Fab domain, V H 1 includes the heavy chain variable domain of the Fab domain, and C H Domain 1 contains Fab's heavy chain constant domain 1.

[0040] In various embodiments of any aspect described herein, the protein has the following polypeptide domain:V from the N-terminus to the C-terminus. H 1-C H 1-C H 2-BD2-C H It contains a chimeric heavy chain having 3, and BD1 contains a Fab domain, V H 1 includes the heavy chain variable domain of the Fab domain, and C H Domain 1 contains Fab's heavy chain constant domain 1.

[0041] In various embodiments of any aspect described herein, the protein has the following polypeptide domain:V from the N-terminus to the C-terminus. H 1-C H 1-C H 2-C H 3(N-terminus)-BD2-C H It contains a chimeric heavy chain having 3 (C-terminus), and BD1 contains a Fab domain, V H 1 includes the heavy chain variable domain of the Fab domain, and C H Domain 1 contains Fab's heavy chain constant domain 1.

[0042] In various embodiments of any aspect described herein, BD2 is or includes scFv. In certain embodiments, scFv has a V from the N-terminus to the C-terminus. H 2-Polypeptide Linker-V L 2 or V L 2-Poteptide Linker-V H Including 2, V H 2 contains the heavy chain variable domain of scFv, and V L 2 contains the light chain variable domain of scFv.

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

[0044] In various embodiments of any of the features described herein, BD2 is connected to the C region of Fc via the linker (L1). H 2 domains, C H 2 domains, or C H 2 and C H It is bound to the interface of the three domains.

[0045] In various embodiments of any of the features described herein, BD2 is connected to the C region of the Fc region via two protein linkers L1 and L2. H 2 domains, C H 3 domains, or C H 2 and C H It binds to the interface of the three domains. In various embodiments, L1 and L2 are independently selected from protein linkers having a length of 1 to 25 amino acids. In specific embodiments, L1 and L2 are independently selected from (G4S)2 (SEQ ID NO: 41), (G4S)3 (SEQ ID NO: 42), and (G4S)4 (SEQ ID NO: 43).

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

[0047] In describing this disclosure, certain aspects of the disclosure are illustrated in the figures. However, this disclosure is not limited to the exact arrangement and means of the aspects shown in the figures. [Brief explanation of the drawing]

[0048] [Figure 1-1]General schematic diagrams of some exemplary proteins described herein are shown. The CH2 and CH3 regions are shown in Figures 1A-1C using PyMOL, revealing the solvent-exposed surface loop regions as spheres. Figure 1A depicts the loop within the CH2 region; Figure 1B depicts the loop within the CH2-CH3 interface; and Figure 1C depicts the loop within the CH3 region. Exemplary constructs are illustrated in Figures 1D-1F, which include the BD1 and BD2 domains, representative as Fab and scFv domains, respectively. Figure 1D depicts BD1 bound at the hinge region and BD2 bound at the solvent-exposed loop within the CH2 region. Figure 1E depicts BD1 bound at the hinge region and BD2 bound at the solvent-exposed loop within the CH2-CH3 interface. Figure 1F depicts BD1 bound at the hinge region and BD2 bound at the solvent-exposed loop within the CH3 region. [Figure 1-2] This is a continuation of Figure 1-1. [Figure 2-1] As described herein, enlarged views of solvent-contactable loop sequences in CH2, the CH2-CH3 interface, and CH3 are provided. Examples of constructs incorporating BD2(scFv) are included in Figures 2A-2C, respectively. Figure 2A illustrates a representative loop sequence ISRTP (SEQ ID NO: 39) identified within the CH2 loop upstream of the CH2-CH3 interface. Figure 2B illustrates a representative loop sequence AKGQP (SEQ ID NO: 40) within the CH2-CH3 interface. Figure 2C illustrates a representative loop sequence SNG within the CH3 region downstream of the CH2-CH3 interface. [Figure 2-2] This is a continuation of Figure 2-1. [Figure 2-3] This is a continuation of Figure 2-2. [Figure 3] This demonstrates the simultaneous binding of BiS2, BiS3, and BiS5 constructs targeting PD-1 and CTLA-4. Trace A9 shows BiS2 PD-1 / CTLA-4, trace B9 shows BiS3 PD-1 / CTLA-4, and trace C9 shows BiS5 PD-1 / CTLA-4. [Figure 4] A schematic diagram of the proposed mechanism for PD-1 / CTLA-4 blockade is shown. [Figure 5]The results of an octet binding assay demonstrating the simultaneous binding of DuetMab constructs targeting PD-1 and CTLA-4 are shown. [Figure 6-1] This shows the PD-1 / CTLA-4 bispecific binding protein that inhibits the PD-1 and CTLA-4 pathways in the reporter gene assay. Figure 6A illustrates T cell activation by PD-1 blockade. Figure 6B illustrates 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. [Figure 6-2] This is a continuation of Figure 6-1. [Figure 7] The results of a Staphylococcus enterotoxin B (SEB) assay demonstrate that PD-1 / CTLA-4 DuetMab and BiS5Ab have equivalent activity in the SEB assay. [Figure 8] In the SEB assay, the activity of PD-1 / CTLA-4 DuetMab is shown compared to the isotype and parental mAb control. [Figure 9] In the SEB assay, PD-1 / CTLA-4 BiS5Ab exhibits activity compared to PD-1 / CTLA-4 DuetMab. [Figure 10-1] In a mixed leukocyte reaction assay (MLR), PD-1 / CTLA-4 DuetMab and BiS5Ab exhibited equivalent activity. Figure 10A is a schematic diagram of the assay (n=4 donors; 2 independent experiments). [Figure 10-2] This is a continuation of Figure 10-1. [Figure 11-1] In the MLR assay, PD-1 / CTLA-4 DuetMab demonstrated activity compared to isotype controls (n=2 donors; 1 experiment). [Figure 11-2] This is a continuation of Figure 11-1. [Figure 12-1] In the MLR assay, PD-1 / CTLA-4 DuetMab showed activity compared to the parent mAb control (n=2 donors; 1 experiment). [Figure 12-2] This is a continuation of Figure 12-1. [Figure 13-1] In the MLR assay, PD-1 / CTLA-4 DuetMab demonstrated activity compared to a competing antibody (n=2 donors; 1 experiment). [Figure 13-2] This is a continuation of Figure 13-1. [Figure 14] This document describes the study design for a single-dose pharmacokinetic / pharmacological (PK / PD) study in cynomolgus monkeys. [Figure 15] This study demonstrates that PD-1 / CTLA-4 DuetMab exhibits clear pharmacological effects (PD) in cynomolgus monkeys. [Figure 16] This shows the T cell-dependent antibody response (TDAR) in cynomolgus monkeys treated with PD-1 / CTLA-4 DuetMab (MEDI5752) and PD-1 / CTLA-4 BiS5Ab (MEDI8500). [Figure 17] This model system demonstrates how stable CHO cells express diverse levels of human PD-1 and / or CTLA-4 to test PD-1 / CTLA-4 bispecific molecules. [Figure 18] We demonstrate that PD-1 / CTLA-4 DuetMab simultaneously binds to PD-1 and CTLA-4 on the surface of the same cell. [Figure 19] This paper describes an experiment to determine whether cooperative binding differs between anti-PD-1 and CTLA-4 antibody combinations in cells expressing excessive PD-1, under CTLA-4 saturation conditions. [Figure 20] We demonstrate that PD-1 and CTLA-4 parental monoclonal antibodies bind to and occupy their target receptors without exerting any measurable effect on non-target receptors. [Figure 21] This study demonstrates that PD-1 / CTLA-4 DuetMab saturates CTLA-4 on CHO cells expressing excessive PD-1 at approximately 250 times lower concentrations compared to monoclonal antibody combinations. [Figure 22]This study demonstrates that PD-1 / CTLA-4 DuetMab saturates CTLA-4 on CHO cells expressing excessive PD-1 at a concentration approximately 500 times lower than that of cells expressing CTLA-4 alone. [Figure 23-1] We demonstrate that the PD-1 / CTLA-4 DuetMab preferentially cis-binds with PD-1 and CTLA-4 on the surface of the same cell. Treme in Figure 23B is a CTLA-4 mAb. [Figure 23-2] This is a continuation of Figure 23-1. [Figure 24] This demonstrates the binding and internalization of PD-1 / CTLA-4 DuetMab and the parental monoclonal antibody to cultured T cells. PD-1 / CTLA-4 DuetMab possesses the internalization properties of tremelimumab. [Figure 25] Figure 25A: Schematic of the internalization assay. Figure 25B: Demonstrates that PD-1 / CTLA-4 DuetMab exhibits the internalization properties of tremelimumab in stable CHO cells expressing 10-fold excess PD-1. [Figure 26] This study reveals the simultaneous binding of BiS2, BiS3, and BiS5 constructs targeting PD-L1 and CTLA-4. Trace A11 shows BiS2 PD-L1 CTLA-4, trace B11 shows BiS3 PD-L1 CTLA-4, and trace C11 shows BiS5 PD-L1 CTLA-4. [Figure 27A] This study clarifies the simultaneous binding of the BiS3 construct to PD-1 and TIM3 (clone 62, wild type). [Figure 27B] We will elucidate the simultaneous binding of the DuetMab construct to PD-1 and TIM3 (clone 62, wild type). [Figure 28-1]This report provides an overview of the cytotoxic activity of single-specific TIM3 and bispecific PD-1 / PD-1 constructs in cell killing assays. Figure 28A shows bright-field images of co-cultured wells at 18 hours. Combinations of anti-TIM-3 + anti-PD1 or TIM-3 / PD-1 bispecific formats enhance tumor cell death and increase T cell activation, as assessed by a decrease in adherent cells and an increase in T cell blasting (aggregation). Figure 28B shows the assessment of vital stain uptake by tumor cells after 18 hours of co-culture with melanoma-specific CD8+ T cells. Figure 28C shows IFNγ secretion after 18 hours of co-culture. The bispecific constructs generally exhibit superior cytotoxic activity, with the DuetMab format showing the most robust cytotoxic activity. [Figure 28-2] This is a continuation of Figure 28-1. [Figure 29] We will elucidate the simultaneous binding of the PD-1 / TIM3 DuetMab construct, which possesses the TIM3 arm sequence and is an affinity-matured mutant of clone 62. [Figure 30] This shows the binding of PD-1 / TIM3 bispecific antibodies such as BiS3, BiS5, and DuetMab to CHO cells overexpressing human TIM3 or human PD1. PD-1 and TIM3 expression data are shown in the inset. [Figure 31] The following shows TIM3 and PD-1 expression data on activated T cell clones (DMF4). [Figure 32] The results from the CMV antigen recall assay are shown, demonstrating that PD-1 / TIM3 bispecific antibodies, including BiS3, BiS5, and DuetMab, exhibit enhanced activity compared to isotyped treatment (3 donors (1-2 replicates per treatment / 1 donor), 1 experiment). [Figure 33] This study demonstrates that PD-1 / TIM3 bispecific antibodies, including BiS3, BiS5, and DuetMab, enhanced interferon (IFNγ) to concentrations greater than 8 nM in mixed lymphocyte reaction (MLR) assays (treatment / 1 donor pair / 2 independent experiments, with 2–4 replicates per run). [Figure 34] The results of a PD-1 reporter assay (dual cell system) using PD-1 / TIM3 bispecific antibodies including BiS3, BiS5, and DuetMab are shown. All bispecific formats showed similar activity to the parent LO115IgG1 (collection of 3-5 independent experiments per treatment / 3 biological replicates). [Figure 35] The results of octet assays using BiS2 and BiS3 OX40 / PD-L1 bispecific molecules are shown. [Figure 36-1] The results of SEB assays using BiS2 and BiS3 OX40 / PD-L1 bispecific molecules are shown. [Figure 36-2] This is a continuation of Figure 36-1. [Figure 37] The results of a PD-L1 reporter assay using BiS2 and BiS3 OX40 / PD-L1 bispecific molecules are shown. [Figure 38] The results of a CMV Ag recall assay using BiS2 and BiS3 OX40 / PD-L1 bispecific molecules are shown. [Figure 39] The results of an octet binding assay demonstrating the simultaneous binding of the OX40(SLR) / PD-L1 BiS5 construct, which targets PD-L1 and OX40, are shown. [Figure 40-1] This demonstrates the binding of a bispecific construct targeting PD-L1 and OX40 to human or cynomolgus monkey CHO cells expressing OX40 and PD-L1 / B7H1. [Figure 40-2] This is a continuation of Figure 40-1. [Figure 41] Flow cytometry (HyperCyt) demonstrates the binding of a bispecific construct targeting PD-L1 and OX40 to Jurkat OX40 reporter cells, NCI H358, CHOK1 B7H1(PD-L1) / OKT3 cells, and HEK CD32a cells. [Figure 42] The results of a PD-L1 reporter assay using the OX40 / PD-L1 bispecific molecule are shown. [Figure 43]The results of an OX40 reporter assay in HEK CD32a cells using the OX40 / PD-L1 bispecific molecule are shown. [Figure 44] The results of an OX40 reporter assay in CHOK PD-L1 overexpressing cells using the OX40 / PD-L1 bispecific molecule are shown. [Figure 45] This study demonstrates PD-L1-mediated OX40 agonism in tumor cells using an OX40 / PD-L1 bispecific molecule. [Figure 46] The results show that no agonism was detected in NCI H358PD-L1 KO cells using the OX40 / PD-L1 bispecific molecule in the control group of the OX40 / tumor cell assay. [Figure 47-1] The results of the SEB assay using the OX40 / PD-L1 bispecific molecule are shown. [Figure 47-2] This is a continuation of Figure 47-1. [Figure 48-1] Figure 48A: Schematic diagram of a Treg suppression assay experiment to test the OX40 / PD-L1 bispecific molecule. Figures 48B-C: Show Treg suppression based on the binding of the bispecific molecule. [Figure 48-2] This is a continuation of Figure 48-1. [Figure 49] The results of a Treg suppression assay using the OX40 / PD-L1 bispecific molecule are shown. [Figure 50] The results of a Treg suppression assay using the OX40 / PD-L1 bispecific molecule are shown. [Figure 51] This document describes the design of a mixed leukocyte reaction (MLR) assay experiment for testing the OX40 / PD-L1 bispecific molecule. [Figure 52-1] The results of an MLR assay using the OX40 / PD-L1 bispecific molecule are shown. [Figure 52-2] This is a continuation of Figure 52-1. [Figure 53]We demonstrate that the BiS2 and BiS5 OX40 / PD-L1 bispecific molecules mediate antibody-dependent cell-mediated cytotoxicity (ADCC) by natural killer (NK) cells against PD-L1 or OX40-expressing CHO cells. [Figure 54] We demonstrate that the BiS2 and BiS5 OX40 / PD-L1 bispecific molecules mediate antibody-dependent cell-mediated cytotoxicity (ADCC) by NK cells against PD-L1 and OX40-expressing CHO cells. [Figure 55] We demonstrate that the BiS2 and BiS5 OX40 / PD-L1 bispecific molecules increase CD107a recruitment of NK cells to PD-L1 and OX40-expressing CHO cells in antibody-dependent cell-mediated cytotoxicity (ADCC). [Figure 56] We demonstrate that the BiS2 and BiS5 OX40 / PD-L1 bispecific molecules mediate antibody-dependent cell-mediated cytotoxicity (ADCC) by NK cells against activated allogeneic T cells. [Figure 57] This study demonstrates that BiS5 OX40 / PD-L1 increases the recruitment of CD107a from NK cells derived from two different donors to activated allogeneic T cells in antibody-dependent cell-mediated cytotoxicity (ADCC). [Figure 58] This document describes a test design for comparing the PK / PD of the OX40 / PD-L1 bispecific molecule. [Figure 59] This paper compares the serum concentration-time profiles of the PD-L1 / OX40 bispecific molecule in cynomolgus monkeys. [Figure 60] This shows the depletion of soluble PD-L1 in serum by the PD-L1 / OX40 bispecific molecule. [Figure 61-1] This report provides an overview of pharmacodynamic data for the PD-L1 / OX40 bispecific molecule. Baseline is defined as the mean values ​​from pre-administration to day 5 and day 0. [Figure 61-2] This is a continuation of Figure 61-1. [Figure 62] A schematic diagram of the PD-1 / OX40 BiS2 IgG4P monoclonal antibody (mAb) is drawn. [Figure 63] A schematic diagram of the potential mechanism of action of PD-1 / OX40 BiS2 mAb is drawn. [Figure 64] The simultaneous binding activity of two different lots of PD-1(LO115) / OX40 BiS2 mAb to PD1-His and human OX40-Fc is depicted. [Figure 65] Figure 65A: A schematic diagram of the OX40 reporter assay is shown. Figure 65B: Results of the OX40 reporter assay using PD1 LO115 mAb, OX40 mAb, control mAb, and PD1(LO115) / OX40 BiS2 mAb are shown. [Figure 66] Figure 66A: A schematic diagram of the PD-1 / PD-L1 reporter assay is shown. Figure 66B: Results of the PD1 / PD-L1 reporter assay using PD1 LO115 mAb, OX40 mAb, control mAb, and PD1(LO115) / OX40 BiS2 mAb are shown. [Figure 67] The results of the SEB assay using BiS2 variants and controls of the PD-1(LO115) / OX40 bispecific molecule are shown. [Figure 68] The results of a CMV antigen recall assay using the BiS2 variant of the PD-1(LO115) / OX40 bispecific molecule and a control are shown. [Figure 69] The results of a CMV antigen recall assay using BiS2 and BiS3 variants of the PD-1(AMP514) / OX40 bispecific molecule, as well as a control, are shown. [Figure 70] This shows the serum concentration-time profile of PD1(LO115) / OX40 BiS2mAb after a single IV dose in cynomolgus monkeys. Data represent the mean ± standard deviation of 3 male monkeys / group. LLOQ (5 ng / mL) is indicated by the dotted line. PKC = pharmacokinetic curve; LLOQ = limit of quantification. [Figure 71]This shows the percentage of Ki67-positive CD4+ and CD8+ memory T cells after a single IV administration of PD-1(LO115) / OX40 BiS2mAb in cynomolgus monkeys. Data represent the mean ± standard deviation of 3 males / group. The left panel A shows CD4+ memory T cells, and the right panel shows CD8+ memory T cells. IV = intravenous. [Figure 72] This shows a typical standard curve for the quantitative determination of PD-1 / OX40 in cynomolgus monkey serum. [Figure 73-1] Exemplary DSC thermograms of the bispecific binding protein ("BiS5") disclosed herein, compared to different BiS formats ("BiS4"), are provided at various pH values. Figure 73A shows the effect of pH value on the thermal stability of BiS4. Figure 73B shows the effect of pH value on the thermal stability of BiS5. Figure 73C plots a representative curve-fitted DSC thermogram of BiS4, including Tonset, Tm1, Tm2, ​​and Tm3. Figure 73D plots a representative curve-fitted DSC thermogram of BiS5, including Tonset, Tm1, Tm2, ​​and Tm3. Figure 73E plots the effect of pH on Tonset, Tm1, Tm2, ​​and Tm3 for BiS4 and BiS5 formats. [Figure 73-2] This is a continuation of Figure 73-1. [Figure 74-1] The HP-SEC analysis of samples at pH 7.5 before and after storage at 40°C for 4 weeks is depicted. Figure 74A shows overlaid plots of SEC chromatograms of BiS4 and BiS5 before and after heat stress, with the solid line corresponding to the BiS4 and BiS5 samples at time zero (unstressed) and the dotted line corresponding to the BiS4 and BiS5 samples incubated at 40°C for 4 weeks (stressed). Figure 74B provides a bar graph showing the effect of pH 7.5 on various species (monomers, fragments, and aggregates) of BiS4 and BiS5 measured on day zero and after 4 weeks at 40°C. [Figure 74-2] This is a continuation of Figure 74-1. [Figure 75]We provide dynamic plots showing the effect of pH 7.5 on accelerated stability at 40°C and short-term storage stability for BiS4 (triangular trace) and BiS5 (circular trace). Figure 75A shows the percentage of residual monomer measured by HP-SEC over 4 weeks. Figure 75B shows the percentage of fragmentation measured by HP-SEC over 4 weeks. Figure 75C shows the percentage of aggregation measured by HP-SEC over 4 weeks. The data shown are from single-vial analysis. [Figure 76] The pH-rate profile plots for BiS4 (triangles) and BiS5 (circles) are plotted. Figure 76A shows the effect of various pH conditions on the rate of monomer loss at 40°C. Figure 76B shows the effect of various pH conditions on the rate of fragmentation at 40°C. Figure 76C shows the effect of various pH conditions on the rate of aggregation at 40°C. [Figure 77-1] Further analysis of BiS4 and BiS5 fragmentation is depicted. Figure 77A shows that neither molecule exhibits obvious fragmentation at pH 7.5 and 40°C (time = 0). Figure 77B shows that, under the same conditions as Figure 77A, after 2 weeks of storage at 40°C, obvious fragmentation is observed for BiS4 and slight fragmentation for BiS5. [Figure 77-2] This is a continuation of Figure 77-1. [Figure 78] The analysis of BiS4 and BiS5 fragmentation (left panel) and aggregation (right panel) as a function of pH is plotted. Both formats reduced fragmentation and aggregation at low (5.5) pH, but BiS5 performed better for both fragmentation and aggregation at all pH values. [Figure 79]The DSC thermograms of several BiS5 bispecific binding proteins disclosed herein for constructs A, B, C, and D (left panel) and E, F, G, and H (right panel) are shown. Constructs A and E contain scFv in IS-RTP; B and F contain scFv in AK-GQP; C and G contain scFv in S-NG; and D and H contain scFv in SN-G. The scFv in constructs A, B, C, and D is 2F4 (IgG is LC10), and the scFv in E, F, G, and H is LC10 (IgG is 2F4). Various TM values ​​are associated with the following domains: TM1=CH2 / scFv; TM2=Fab; TM3=CH3. [Figure 80] This specification presents representative datasets regarding FcRn binding with the bispecific binding proteins (constructs D and H). The scFv position in the CH2-CH3 domain (i.e., within the ISTRP loop) may have an effect on FcRn binding activity. [Figure 81] Representative datasets regarding FcγR binding with the bispecific binding proteins disclosed herein (constructs E, G, and H, including FcγRIIIa-158V) are shown. The inset figures represent the same data as the main figure, re-normalized upon FcγRIIIa-158V injection. All constructs were able to bind to FcγR, but there were observable differences in affinity based on the position of the scFv domain. [Figure 82] We demonstrate that an intact ISRTP loop is crucial for FcRn binding in exemplary bispecific binding constructs (e.g., A and E). The introduction of an N3 loop does not compensate for the disruption of the ISRTP loop (BiS5E+N3). The introduction of scFv into the N3 loop inserted into IgG1 Fc makes IgG unable to bind to FcRn. Time on the x-axis is measured in seconds. [Figure 83] The general schematic structural formats for each of the BiS1, BiS2, BiS3, BiS4, and BiS5 constructs are depicted. The designations "k1" and "k2" indicate the fragmentation patterns used in the kinetic analysis discussed in Example 3. [Figure 84] The fragmentation rates of BiS1, BiS2, BiS3, BiS4, and BiS5 as functions of pH are shown. [Figure 85] The aggregation rates of BiS1, BiS2, BiS3, BiS4, and BiS5 as functions of pH are shown. [Figure 86] The monomer loss rates of BiS1, BiS2, BiS3, BiS4, and BiS5 are shown as functions of pH. [Figure 87] Figures illustrating the fragmentation patterns for BiS1, BiS2, BiS3, BiS4, and BiS5, and their correspondence with the peaks in the HPSEC chromatograms, are plotted. [Figure 88] This section illustrates a typical analysis of fragmentation patterns under reducing conditions. [Figure 89] The structural configurations of BiS5 under reducing and non-reducing conditions are depicted. [Figure 90] Draw the SEB assay format. [Modes for carrying out the invention]

[0049] Before proceeding to describe this disclosure in further detail, it should be understood that this disclosure is not limited to any particular composition or process step, and may therefore vary. It should be noted that, as used herein and in the appended claims, the singular forms "a," "an," and "the" have plural meanings unless the context clearly indicates otherwise.

[0050] Unless otherwise stated, all technical and scientific terms used herein have the same meaning as those generally understood by those skilled in the art in the relevant fields. For example, *Concise Dictionary of Biomedicine and Molecular Biology*, Juo, Pei-Show, 2nd ed., 2002, CRC Press; *The Dictionary of Cell and Molecular Biology*, 3rd ed., 1999, Academic Press; and *Oxford Dictionary of Biochemistry and Molecular Biology*, Revised, 2000, Oxford University Press are common dictionaries for many of the terms used in this invention for those skilled in the art.

[0051] In this specification, amino acids may be denoted by their commonly known three-letter symbols or by the single-letter symbols recommended by the IUPAC-IUB Biochemical Nomenclature Commission. Similarly, nucleotides may be denoted by their commonly accepted single-letter symbols.

[0052] The amino acid numbering of the antibody's variable domain, complementarity-determining region (CDR), and framework region (FR) follows Kabat's definition as described in Kabat et al. Sequences of Proteins of Immunological Interest, 5th Ed. Public Health Service, National Institutes of Health, Bethesda, MD. (1991), unless otherwise noted. Using this numbering system, the actual linear amino acid sequence may contain fewer or additional amino acids in response to shortening or insertion into the FR or CDR of the variable domain. For example, the heavy chain variable domain may include an insertion of one amino acid after H2 residue 52 (residue 52a according to Kabat) and residues inserted after heavy chain FR residue 82 (e.g., residues 82a, 82b, and 82c according to Kabat). Kabat's numbering of residues can be determined for each antibody by aligning the antibody sequence with a homology region of a sequence with "standard" Kabat numbering. The optimal alignment of framework residues often requires the insertion of "spacer" residues in this numbering system, which are used in the Fv region. Furthermore, the type of specific residues at any given Kabat site number may vary between antibody chains due to interspecies or allele differences.

[0053] As used herein, the term “antibody” is also called immunoglobulin and includes 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., International Publication 2009 / 018386, International Application PCT / US Patent Application Publication 2012 / 045229, which is incorporated herein in its entirety), BiSAb, human antibodies, humanized antibodies, camelized antibodies, single-chain Fv(scFv), single-chain antibodies, single-domain antibodies, domain antibodies, Fab fragments, F(ab')2 fragments, antibody fragments exhibiting desired biological activity (e.g., antigen-binding moieties), disulfide-bound Fv(dsFv), and anti-idiotype (anti-Id) antibodies (e.g., anti-Id antibodies against the antibodies of the present invention), intrabodies, and any of the epitope-binding fragments described above. In particular, antibodies include immunoglobulin molecules and immunologically active fragments of immunoglobulin molecules, i.e., molecules containing at least one antigen-binding site. Antibodies also include peptide fusions with the antibody or a portion thereof, such as fusion proteins with an Fc domain. Immunoglobulin molecules may be any isotype (e.g., IgG, IgE, IgM, IgD, IgA, and IgY), subisotype (e.g., IgG1, IgG2, IgG3, IgG4, IgA1, and IgA2), or allotype (e.g., Gm, e.g., G1m(f, z, a, or x), G2m(n), G3m(g, b, or c), Am, Em, and Km(1, 2, or 3)). Antibodies may be derived from any mammal, including but not limited to humans, monkeys, pigs, horses, rabbits, dogs, cats, mice, or other animals, such as birds (e.g., chickens).

[0054] CTLA-4 Cytotoxic T lymphocyte-associated protein 4 (CTLA-4) is expressed on activated T cells and acts as a co-inhibitor that blocks the T cell response after CD28-mediated T cell activation. CTLA-4 is thought to be part of a central inhibitory pathway that influences both anti-tumor and autoimmune responses, while also regulating the magnitude of early activation of naive and memory T cells after TCR engagement. CTLA-4 is expressed only on T cells, and the expression of its ligands, CD80 (B7.1) and CD86 (B7.2), is primarily limited to antigen-presenting cells, T cells, and other immune-mediated cells. Antagonist anti-CTLA-4 antibodies that block the CTLA-4 signaling pathway have been reported to enhance T cell activation. One such antibody, ipilimumab, was approved by the FDA in 2011 for the treatment of metastatic melanoma. The use of anti-CTLA-4 antibodies to treat infections and tumors and to upmodulate adoptive immune responses 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. Patent Publication No. 20150104409; these are incorporated herein by reference in their entirety).

[0055] PD-L1 Programmed death ligand 1 (PD-L1) is also part of a receptor-ligand complex system involved in regulating T cell activation. In normal tissues, PD-L1 is expressed on T cells, B cells, dendritic cells, macrophages, mesenchymal stem cells, myeloid mast cells, and 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, acting in many sites to help tumors evade detection and elimination by the host immune system. PD-L1 is frequently expressed in a variety of cancers. In some cancers, PD-L1 expression is associated with low survival rates and poor prognosis. Antibodies that block the interaction between PD-L1 and its receptors can reduce PD-L1-dependent immunosuppressive effects and enhance the cytotoxic activity of antitumor T cells in vitro. Durvalmab is a human monoclonal antibody against human PD-L1 that can block the binding of PD-L1 to both the PD-1 and CD80 receptors. The use of anti-PD-L1 antibodies to treat infections and tumors and to enhance adoptive immune responses has been reported (see U.S. Patent Nos. 8,779,108 and 9,493,565, which are incorporated herein by reference in their entirety).

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

[0057] PD-1 is expressed in 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 that, after activation by binding to PD-L1 or PD-L2, causes downregulation of the immune system (Martin-Orozco, N. et al. (2007) "Inhibitory Costimulation and Anti-Tumor Immunity," Semin. Cancer Biol. 17(4):288-298), and acts as a cell death inducer (Ishida, Y. et al. (1992) "Induced Expression of PD-1, A Novel Member of The Immunoglobulin Gene Superfamily, Upon Programmed Cell Death," EMBO J. 11:3887-3895; Subudhi, S. et al. (2005) "The Balance of Immune Responses: Costimulation Verse Coinhibition," J. Molec. Med. 83:193-202). This process is utilized in many tumors via PD-L1 overexpression, leading to suppression of the immune response.

[0058] PD-1 is a well-validated target for immuno-mediated therapy in oncology, yielding promising results from clinical trials, particularly in the treatment of melanoma and non-small cell lung cancer (NSCLC). Antagonistic inhibition of the PD-1 / PD-L-1 interaction enhances T cell activation and strengthens the host immune system's recognition and elimination of tumor cells. The use of anti-PD-1 antibodies to treat infections and tumors and to increase adoptive immune responses has been proposed (see U.S. Patent Nos. 7,521,051; 7,563,869; and 7,595,048).

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

[0060] Immunohistochemical studies and early flow cytometry analyses have demonstrated that OX40 is expressed in T cells infiltrating various 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). While we do not intend to be bound by any particular theory, OX40 expression in tumor-infiltrating lymphocytes has been correlated with longer-term survival in several human cancers, suggesting that OX40 signaling may play a role in establishing an anti-tumor immune response (Ladanyi et al., 2004, Clin Cancer Res. 10:521-30; Petty et al., 2002, Am J Surg. 183:512-8).

[0061] In various nonclinical mouse tumor models, the use of OX40 agonists, including antibodies and OX40 ligand fusion proteins, has been successful and yielded promising results (Kjaergaard et al., 2000, Cancer Res. 60:5514-21; Ndhlovu et al., 2001, J Immunol. 167:2991-9; Weinberg et al., 2000, J Immunol. 164:2160-9). Co-stimulation of T cells with OX40 sometimes promotes sustained antitumor activity, resulting in long-lasting protection against subsequent tumor attacks (Weinberg et al., 2000, J Immunol. 164:2160-9). Co-stimulation of Treg cells and effector T cells has been shown to be necessary for OX40 agonist inhibition of tumor growth (Piconese et al., 2008, J Exp Med. 205:825-39). To enhance the antitumor effects of OX40 agonist therapy, numerous strategies and techniques have been explored, including combination therapy 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 upmodulate adoptive immune responses has been proposed (see U.S. Patent Application Publication No. 20160137740, incorporated herein by reference in its entirety).

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

[0063] A. Bispecific binding protein Adding multiple binding sites to a molecule that has specificity for a single binding domain can significantly enhance the molecule's capabilities (e.g., therapeutic or diagnostic capabilities). For example, a bispecific antibody can bind to two or more regions of the same target biomolecule, giving it greater specificity than a monospecific polypeptide that binds to only one epitope of the target. Alternatively, a bispecific antibody can bind to multiple target biomolecules, such as targets present in a complex, or targets where blockage and / or clustering are desirable. In a third scenario, the same bispecific antibody can exert different functions depending on the localization and / or expression of its target molecule.

[0064] This specification describes novel binding proteins. One such structure of these novel binding proteins is referred to as "DuetMab". DuetMab has the following basic structure: a modified heavy chain (where the CH1 region of the modified heavy chain has substitution of a native cysteine ​​for a non-cysteine ​​amino acid and substitution of a native non-cysteine ​​amino acid for a cysteine ​​amino acid); a corresponding modified light chain (where the CL region of the modified light chain also has substitution of a native cysteine ​​for a non-cysteine ​​amino acid and substitution of a native non-cysteine ​​amino acid for a cysteine ​​amino acid); a second Fc region having a second heavy chain; and a corresponding Fc region having a second modified light chain (where the modified heavy chain is directly linked to the corresponding modified light chain, and on separate target binding arms, the second heavy chain is directly linked to the second corresponding light chain, and the substituted cysteine ​​of the modified heavy chain obtained by substitution of a native non-cysteine ​​amino acid for a cysteine ​​amino acid, and the substituted cysteine ​​of the corresponding modified light chain obtained by substitution of a native non-cysteine ​​amino acid for a cysteine ​​amino acid, can form a disulfide bond). Disclosures relating to DuetMab can be found, for example, in U.S. Patent No. 9,527,927, which is incorporated herein by reference in whole.

[0065] Another exemplary structure of these novel binding proteins is called “BiSAb” or “BiSAbs”. Schematic diagrams of exemplary BiSAbs, as well as specific examples of particular BiSAbs, are provided herein. More generally, a BiSAb is a polypeptide containing two binding units, each of which binds to an epitope (for example, binding unit 1 binds to a first epitope, and binding unit 2 binds to a second epitope). A basic BiSAb is bivalent to bind to each of the two epitopes (for example, the polypeptide contains two binding units 1 ("BD1" or "BU1") and two binding units 2 ("BD2" or "BU2")). Thus, when binding units 1 and 2 bind to different epitopes, the BiSAb has the multispecificity of a conventional bispecific antibody and the bivalence of a conventional antibody molecule. In embodiments where binding units 1 and 2 bind to the same epitope, the BiSAb has the monospecificity of a normal antibody but is tetravalent. In addition to the binding unit, BiSAb also includes a linker polypeptide and an Fc moiety. This disclosure relates to a broad set of bispecific binding proteins, such as BiSAb and proteins containing the BiSAb core, which target molecules that modulate immune responses. Generally, the novel binding protein platforms and exemplary bispecific binding proteins (BiSAb) described herein include a binding unit / domain, a linker polypeptide, and an Fc moiety. This disclosure also provides nucleic acid molecules encoding such BiSAb, as well as vectors and host cells containing such nucleic acids that can be used in methods for generating and using such BiSAb. BiSAb, binding proteins containing the BiSAb core, and various parts of BiSAb are described in further detail herein.

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

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

[0068] The BiSAb disclosed herein has an overall structure similar to that of a typical antibody, but can be distinguished by the presence of additional binding units that bind at locations within the Fab domain and at locations away from the Fab domain and within hinge or Fc regions (e.g., at the interfaces of such regions, such as the CH2, CH3, or CH4 region or the CH2-CH3 interface). Thus, unlike typical antibodies which are bivalent to bind to a single epitope, BiSAb is bivalent to bind to two epitopes. However, as described herein, BiSAb can still retain many desirable properties of a typical antibody, such as the ability to bind to C1q and FcRn, as well as the ability to bind to the Fcγ receptor (e.g., exhibiting the ability to mediate antibody and complement-dependent cytotoxicity).

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

[0070] Unlike conventional antibodies, BiSAbs are bivalent (or tetravalent to one epitope), for example, for binding to two different epitopes, although much of a BiSAb is derived from or similar to a conventional antibody moiety. Any mAb domain and / or fragment known in the art may be used in the BiSAbs described herein. In particular, a BiSAb may include a Fab fragment and / or an scFv fragment or a variant thereof. Exemplary and non-limiting variants of scFv include, but are not limited to, tandemdi-scFv, tandemtri-scFv, diabody, and tribody or triabody.

[0071] This disclosure, in general, exemplifies novel binding proteins, with BiSAb being an example. Other examples include binding proteins containing a BiSAb core in addition to one or more additional binding units, and / or binding proteins containing an elongated BiSAb core. When BiSAb or the features of BiSAb are described herein, it should be understood that such descriptions apply generally to the novel binding proteins of this disclosure, regardless of whether such binding proteins contain two or three or more binding units. Accordingly, the term BiSAb is illustrative of the binding proteins described herein, and any such reference to BiSAb, where contextually permissible, may be used to describe binding proteins containing a BiSAb core.

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

[0073] Accordingly, one aspect of the present 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 both divalent with respect to each epitope. Each heavy-light chain pair comprises a binding domain (BD) (binding unit 1) which may include a Fab domain that binds to a first epitope, a second binding domain (BD2) (or binding unit 2 which may be, for example, scFv) that binds to a second epitope, and an Fc region. In some embodiments, the second binding domain may be bound to the Fab domain. In some embodiments, the Fc region of the BiSAb may be bound to the second binding domain (BD2) (binding unit 2; depicted as scFv in Figures 1D-1F) that binds to a second epitope.

[0074] In some embodiments, the Disclosure provides a BiSAb having a general 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 linker region, and an Fc region comprising a CH2 domain and a CH3 domain, where optionally, a second binding domain (BD2) flanked on one or both sides by a polypeptide linker (L1 and / or L2) is bound to the solvent exposure loop in the Fc region in the order of (i) the CH2 region, (ii) the CH2 and CH3 regions, or (iii) the CH3 region. The BiSAb in 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), forming part of the first binding domain (BD1). The binding domain (BD2) of a particular BiSAb shown in Figures 1D-1F is scFv.

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

[0076] Similarly, the polypeptide chain comprises a heavy chain having a VH1 domain, a CH1 domain, a hinge / linker, a CH2 domain, an optional linker (referred to herein as L1 or the first polypeptide linker), binding unit 2 (such as VL2 and VH2 of scFv), another optional linker (e.g., L2 or the second polypeptide linker), and a CH3 domain, as shown in Figure 1E. In this embodiment, BiSAb comprises a second binding domain, shown as scFv, bound to Fc in sequence at the interface of the CH2 and CH3 regions.

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

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

[0079] While a general format has been described for aspects relating to specific structural platforms of some of the BiSAb molecules disclosed herein, the various parts of the disclosed BiSAb and their exemplary functional properties are described in further detail below. In other embodiments, this disclosure considers and provides other BiSAb-binding proteins, including alternative structural formats and configurations, as briefly described herein and in other disclosures incorporated herein by reference (see, for example, U.S. Patent Application Publication No. 20090155275 and U.S. Patent No. 9,580,509).

[0080] 1. Connecting Unit The BiSAb of this disclosure comprises at least two binding units or binding domains (binding unit / domain 1 and binding unit / domain 2). In some embodiments, each binding unit binds to a different epitope, whether different epitopes on the same target molecule or epitopes on different targets. Since the binding units of BiSAb exist in pairs (there are two binding unit 1 and two binding unit 2), BiSAb exhibits divalent binding to each epitope. From the teachings herein, it will be understood that if each binding unit binds to the same epitope, BiSAb exhibits tetravalent binding to the epitope.

[0081] In one embodiment, the first binding unit is a Fab fragment, such as a Fab fragment of a conventional monoclonal antibody, or an antigen-binding fragment prepared by recombinant technology, 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 linked to each other via one or more disulfide bonds, such as a suitable antibody hinge region. The Fab binds to the first epitope.

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

[0083] In some embodiments, the BiSAb of this disclosure comprises a binding unit 2, which comprises a binding domain that binds to a second epitope. The binding unit 2 (or binding domain 2 (BD2)) may bind to BiSAb using any preferred strategy. When used herein, BD2 “bound” to BiSAb (for example, within an Fc region in some embodiments, and within a Fab region in other embodiments) means that the two molecules have interactions with each other such that they maintain orientation for target binding and binding to the Fc or Fab portion of the BiSAb structure. Examples of such interactions include covalent bonding via amino acid linkers, covalent bonding by recombinant expression of BD2 in the Fab region, hinge region, or Fc region at the CH2, CH3, or CH2-CH3 interface, or in the CH4 region, and non-covalent interactions such as van der Waals forces and hydrogen bonding interactions within these same regions. Non-limiting examples of binding domains (or “BD” or “binding unit”) included in the scope of this disclosure include antibody variable regions, antibody fragments, scFv, single-stranded diabodies, or other binding domains known in the art. Binding domains also include bispecific single-stranded diabodies, or single-stranded diabodies designed to bind two different epitopes. In one embodiment, epitope-binding domains useful for constructing the multispecific epitope-binding domains of this disclosure are exemplified in U.S. Patent Application Publications No. 20100298541 and No. 20130079280, which are incorporated herein by reference for all purposes.

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

[0085] In some embodiments, the BD may include a ligand-binding domain of the receptor or a receptor-binding domain of the ligand. In some embodiments, the BD includes a sequence having binding affinity to 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 to a target selected from the group consisting of CTLA-4, PD-1, PD-L1, OX40, and TIM3. The BiSAbs disclosed herein may include any combination of binding domains having binding affinity or specific binding activity to the molecular targets disclosed herein. For example, the BiSAbs disclosed herein may include combinations of binding domains that enable bispecific binding to targets including CTLA-4 and PD-1; CTLA-4 and PD-L1; CTLA-4 and TIM3; PD-1 and PD-L1; PD-L1 and OX40; PD-1 and TIM3; PD-L1 and TIM3; and TIM3. BiSAbs containing binding domains that bind to specific target combinations include, but are not limited to, PD-1 / CTLA-4;PD-L1 / CTLA-4;PD-1 / OX40;PD-L1 / OX40; and PD-1 / TIM3, as illustrated in the examples.

[0086] In some further embodiments, BiSAb exhibits greater binding activity (e.g., binding affinity and / or binding specificity) to at least one of the target molecules than the binding activity of the parent monospecific binding sequence used to generate the BiSAb. In similar embodiments, BiSAb may exhibit greater binding activity (e.g., binding affinity and / or binding specificity) to both target molecules than the binding activity of both parent monospecific binding sequences used to generate the BiSAb. In yet another embodiment, BiSAb may exhibit greater binding activity (e.g., binding affinity and / or binding specificity) to both target molecules than the binding activity of the combination of parent monospecific binding sequences used to generate the BiSAb. Whether used alone or in combination, the enhancement of BiSAb's binding properties to parent monospecific binding sequences provides unexpected advantages compared to the use of monospecific therapeutics targeting the same molecule, even when used in combination.

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

[0088] In some embodiments, the bispecific binding proteins (BiSAbs) disclosed herein can be generated using all or a domain containing the antigen-binding region of the aforementioned “parent” antibody. Non-limiting embodiments illustrated in the examples provide an explanation of how antibody sequences can be identified and combined to generate BiSAbs that exhibit bispecific binding to a combination of molecular targets.

[0089] Several methods, either alone or in combination, may be used to enhance the stability of BiSAb containing scFv molecules. One method that can be used alone or in combination with one or more of the other methods described herein is to manipulate the length and / or composition of the linker that ligates the scFv domains to stabilize the scFv portion.

[0090] Another possible approach is to introduce at least two amino acid substitutions (also called modifications or mutations) into the VH and / or VL domains of scFv to promote the formation of disulfide bonds (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 may be used alone or in combination with one or more of the other methods described herein.

[0091] In one embodiment, one or more mutations can be introduced into the VH and VL domains of scFv to promote the formation of interchain disulfide bonds between the VH and VL domains when BiSAb containing scFv is expressed. In another embodiment, two mutations are introduced into the same domain of the chain. In one embodiment, two mutations are introduced into different chains. In one embodiment, multiple complementary mutations are introduced to promote the formation of multiple disulfide bonds or other stabilizing interactions. In one embodiment, cysteine ​​is introduced to promote the formation of disulfide bonds. Exemplary amino acids that may be mutated with 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 aforementioned numbering is based on Kabat's numbering that identifies the positions of scFv only for VH2 and VL2 (and not for the full-length sequence of BiSAb or the amino acid positions of the sequence numbers provided herein). Exemplary combinations of amino acid positions that may be mutated with cysteine ​​residues include VH44-VL100, VH105-VL43, VH105-VL42, VH44-VL101, VH106-VL43, VH104-VL43, VH44-VL99, VH45-VL98, VH46-VL98, VH103-VL43, VH103-VL44, and VH103-VL45. In some embodiments, amino acid 44 of VH and amino acid 100 of VL are mutated with cysteine.

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

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

[0094] Another method, which may be used on its own or in combination with other methods, involves mutating one or more residues in the VH domain and / or VL domain of scFv to introduce one or more amino acid substitutions to match the most frequent residue at the aforementioned specific position in the consensus sequence of the VH domain and / or VL domain of a known antibody. In one embodiment, substitutions are introduced in one or both of the VH domain and / or VL domain of scFv at positions 1, 2, 3, 4, 5, 6, or more than 6. The same number of changes may be made in each domain, or different numbers of changes may be made in each domain. In one embodiment, one or more changes in the sequence to match a certain consensus sequence are conserved amino acid substitutions of residues present in the unmodified VH and / or VL sequence. In other embodiments, one or more changes are non-conserved amino acid substitutions of residues present in the unmodified VH and / or VL sequence. When multiple substitutions are made in one or both of the VH domain or VL domain of scFv, each substitution is independently either a conserved or non-conserved substitution. In some embodiments, all substitutions are conservative substitutions. In some embodiments, all substitutions are non-conservative substitutions. In some embodiments, at least one of the substitutions is conservative. In some embodiments, at least one or all of the substitutions are non-conservative.

[0095] It should be noted that any modification described as useful for modifying or stabilizing the scFv portion can also be applied to modifying the Fab portion. For example, modifying the variable domain of the Fab portion of BiSAb can improve stability, antigen binding, and homogeneity. Furthermore, modifying either the Fab or scFv portion can reduce immunogenicity.

[0096] In one embodiment, binding unit 2 (BD) is an scFv derived from a conventional monoclonal antibody containing 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 heavy chains of the scFv may be further interconnected via one or more disulfide bonds and may contain one or more mutations or modifications as described above. The scFv binds to a second epitope. In one embodiment, the second epitope is different from the first epitope to which binding unit 1 binds. In another embodiment, the second epitope is the same as the first epitope to which binding unit 1 binds. In one embodiment, the scFv is derived from or based on the sequence of a conventional monoclonal antibody, such as a conventional mouse antibody, a humanized antibody, or a human antibody. In some embodiments, a BiSAb containing scFv derived from or based on the sequence of a conventional monoclonal antibody retains one or more functional activities of the conventional antibody (for example, retaining at least 80% or more of the functional activity (80%, 85%, 90%, 95%, 97%, 98%, 99%, or 100%)). For example, in some embodiments, such a BiSAb containing scFv retains one or more of the affinity, inhibitory activity, or cytotoxic activity of the conventional antibody toward an antigen.

[0097] In some embodiments, BiSAb comprises any combination of binding unit 1 and binding unit 2 as described herein. For example, in some embodiments, the Disclosure provides polypeptides comprising Fabs that bind to a specific target (e.g., an epitope on a specific target), for example, Fabs comprising a specific amino acid sequence or encoded by a specific nucleotide sequence, and / or scFvs that bind to a specific target (e.g., an epitope on a specific target), for example, scFvs comprising a specific amino acid sequence or encoded by a specific nucleotide sequence.

[0098] As detailed above, binding units 1 and 2 may be linked to BiSAb by covalent bonds via linker polypeptide 1 (L1, L2). Generally, this bond is linked to heavy chain C H 2 domains, heavy chain C H Is it via 3 domains, or heavy chain C? H 2 domains and C H The linkage is via a chimeric heavy chain of BiSAb, either at the interface of the three domains or, in some embodiments, within the hinge region or Fab domain. L1 and L2 may differ in length and sequence independently of each other, and exemplary structures are described herein. This disclosure intends for BiSAb to include any combination of a linker polypeptide and a specific binding unit that binds to a desired target, such as any combination of a specific binding unit and the particular L1 and L2 polypeptide linkers described herein.

[0099] 2.Fc area As used herein, “Fc region” encompasses a domain derived from the constant region of an immunoglobulin, preferably a human immunoglobulin, including fragments, analogs, variants, mutants, or derivatives of the constant region. Suitable immunoglobulins include IgG1, IgG2, IgG3, IgG4, and other classes, such as IgA, IgD, IgE, and IgM. The Fc region may be a naturally occurring Fc region or a modified Fc region. The Fc region of an immunoglobulin generally consists of two constant domains, C H 2 domains and C H Includes 3 domains, optionally C H It includes four domains. The BiSAb of this disclosure includes an Fc domain of the same class as the hinge portion of either L1 or L2, or both.

[0100] a. Modified Fc region Modified Fc regions (also referred to herein as “mutant Fc regions”) may be used to alter the effector function and / or half-life of the BiSAb of this disclosure. One or more modifications may be made to the Fc region to alter the functional and / or pharmacokinetic properties of the molecule. As a result of such modifications, IgG C1q binding and complement-dependent cell-mediated cytotoxicity (CDC), or FcγR binding and antibody-dependent cell-mediated cytotoxicity (ADCC), or antibody-dependent phagocytosis (ADCP) may be reduced or enhanced. This disclosure encompasses BiSAbs modified to fine-tune effector function by enhancing or reducing function or providing desired effector function. Accordingly, in one aspect of this disclosure, the BiSAb includes a mutant Fc region (i.e., a modified Fc region as discussed below). In this specification, a BiSAb containing a mutant Fc region is also referred to as “Fc mutant BiSAb”. As used herein, “natural” refers to the unmodified parent sequence, and in this specification, a BiSAb containing a natural Fc region is referred to as “natural Fc BiSAb”. Fc-mutated BiSAbs can be prepared by many methods well known to those skilled in the art. Non-limiting examples include isolation of the antibody-coding region (e.g., from a hybridoma) and implementation of one or more desired substitutions in the Fc region. Alternatively, the antigen-binding portion of BiSAb (e.g., the variable region) may be subcloned into a vector encoding the mutant Fc region. In one embodiment, the mutant Fc region induces effector function at a level equivalent to that of the native Fc region. In another embodiment, the mutant Fc region induces effector function more highly than that of the native Fc region. In yet another embodiment, the mutant Fc region induces effector function less highly than that of the native Fc region. Several specific embodiments of mutant Fc regions are described below. Methods for measuring effector function are well known in the art.

[0101] In general, the effector function is modified by altering the Fc region, including but not limited to amino acid substitution, amino acid addition, amino acid deletion, and post-translational modifications of the Fc amino acid (e.g., glycosylation). The effector function of the BiSAb disclosed herein and the ratio of Fc region binding to FcR (e.g., affinity and specificity) may be fine-tuned using the methods described below to obtain a BiSAb with desired properties.

[0102] As used herein, the Fc region will be understood to include the polypeptide comprising the constant region of the antibody molecule, excluding the immunoglobulin domain of the first constant region. Therefore, Fc refers to the immunoglobulin domains of the last two constant regions of IgA, IgD, and IgG, and the immunoglobulin domains of the last three constant regions of IgE and IgM, and optionally, all or part of the N-terminus of the flexible hinge of these domains. In the case of IgA and IgM, Fc may include the J chain. In the case of IgG, Fc includes the immunoglobulin domains C-gamma 2 and C-gamma 3 (Cγ2 and Cγ3), and optionally the lower hinge portion between C-gamma 1 (Cγ1) and C-gamma 2 (Cγ2). While the boundaries of the Fc region may differ, as used herein, the Fc region of the human IgG heavy chain includes residue A231 to its carboxyl terminus, and the numbering follows the EU index as described in Kabat. Fc may refer to this isolated region, or to this region in relation to an antibody, antibody fragment, or Fc fusion protein. Polymorphisms have been observed at many different Fc positions, including, but are not limited to, positions 270, 272, 312, 315, 356, and 358 of IgG1 as numbered by the EU index, and therefore some differences may exist between this sequence and the sequences of prior art.

[0103] In one embodiment, the disclosure includes Fc mutant BiSAbs in which binding affinity to Fc ligands (e.g., Fc receptors, C1q) is modified compared to natural Fc BiSAbs. Examples of binding affinity include binding specificity and equilibrium dissociation constant (K). d), dissociation rate and binding rate (each k off and k on ), binding affinity and / or avidity are, but are not limited to, these. Equilibrium dissociation constant (K d ) is k off / k on It is publicly known in the art that it is defined as low K. d BiSAb containing an Fc mutation region having high K d It may be preferable to BiSAb having [a certain characteristic]. However, in some cases, k on or k off The value of K d This value may be more important than the value of . A person skilled in the art can determine which kinetic parameter is most important for each application. For example, modifications that reduce binding to one or more positive regulators (e.g., FcγRIIIA) and / or enhance binding to inhibitory Fc receptors (e.g., FcγRIIB) are considered suitable for reducing ADCC activity. Therefore, the ratio of binding affinity to various receptors (e.g., equilibrium dissociation constant (K)) may be important. d The ratio of () may indicate whether the ADCC activity of the Fc mutant BiSAb of this disclosure is enhanced or reduced. In addition, modifications that reduce binding to C1q are also considered suitable for reducing or removing CDC activity.

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

[0105] In one embodiment, the Fc mutant BiSAb has increased affinity for the Fc ligand. In another embodiment, the Fc mutant BiSAb has decreased affinity for the Fc ligand compared to the natural Fc BiSAb.

[0106] In certain embodiments, the Fc mutant BiSAb has enhanced binding to the Fc receptor FcγRIIIA. In other specific embodiments, the Fc mutant BiSAb has enhanced binding to the Fc receptor FcγRIIB. In yet another specific embodiment, the Fc mutant BiSAb has enhanced binding to both the Fc receptors FcγRIIIA and FcγRIIB. In some embodiments, the Fc mutant BiSAb with enhanced binding to FcγRIIIA does not simultaneously increase binding to the FcγRIIB receptor compared to the native Fc BiSAb. In certain embodiments, the Fc mutant BiSAb has decreased binding to the Fc receptor FcγRIIIA. In yet another specific embodiment, the Fc mutant BiSAb has decreased binding to the Fc receptor FcγRIIB. In yet another specific embodiment, the Fc mutant BiSAb has enhanced binding to the Fc receptor FcRn. In another specific embodiment, Fc mutant BiSAb with modified affinity for FcγRIIIA and / or FcγRIIB has enhanced binding to the Fc receptor FcRn. In yet another specific embodiment, Fc mutant BiSAb with modified affinity for FcγRIIIA and / or FcγRIIB has modified binding to C1q compared to natural Fc BiSAb.

[0107] In another embodiment, the Fc mutant BiSAb increases or decreases its affinity for C1q compared to the native Fc BiSAb. In yet another specific embodiment, the Fc mutant BiSAb with modified affinity for C1q has enhanced binding to the Fc receptor FcRn. In yet another specific embodiment, the Fc mutant BiSAb with modified affinity for C1q has modified binding to FcγRIIIA and / or FcγRIIB compared to the native Fc BiSAb.

[0108] It is recognized that antibodies can induce attack and destruction of target antigens through a series of processes collectively referred to in this technology as antibody effector function. One of these processes is called "antibody-dependent cell-mediated cytotoxicity" or "ADCC," which refers to a form of cytotoxicity in which, when secreted Ig antibodies bind to Fcγ receptors (FcγRs) present on certain cytotoxic cells (e.g., natural killer (NK) cells, neutrophils, and macrophages), these cytotoxic effector cells specifically bind to target cells containing the antigen, and subsequently kill the target cells with cytotoxicity. High-affinity IgG antibodies specific to the surface of target cells "arm" the cytotoxic cells and are essential for this killing. Lysis of target cells occurs extracellularly and requires direct cell-cell contact, but does not require complement. Another process encompassed by the term effector function is complement-dependent cell-mediated cytotoxicity (hereinafter referred to as "CDC"), which refers to a biochemical phenomenon in which target cells are destroyed by the complement system via antibodies. The complement system is a complex system of proteins found in normal plasma that works in conjunction with antibodies to destroy pathogens and other foreign cells. Another process encompassed by the term effector function is antibody-dependent phagocytosis (ADCP), which refers to a cellular response in which nonspecific cytotoxic cells expressing one or more effector ligands recognize antibodies bound to target cells and subsequently trigger phagocytosis of the target cells.

[0109] Fc mutant BiSAb is intended to be characterized by in vitro functional assays relating to the function of one or more FcγR-mediated effector cells. In some embodiments, Fc mutant BiSAb exhibits similar binding affinity and effector cell function (such as those described and disclosed herein) in in vivo models as in in vitro assays. However, this disclosure does not exclude Fc mutant BiSAb that do not exhibit the desired phenotype in in vitro assays but do exhibit the desired phenotype in vivo.

[0110] The serum half-life of proteins containing an Fc region can be extended by increasing the binding affinity of the Fc region to FcRn. The term "antibody half-life," as used herein, refers to the pharmacokinetic property of an antibody, which is an indicator of the mean survival time of an antibody molecule after administration. Antibody half-life can be expressed as the time required to remove 50 percent of a known amount of immunoglobulin from a patient's body (or other mammal) or a particular compartment of that body, for example, in serum (i.e., blood half-life) or in other tissues. Half-life may vary depending on the immunoglobulin or class of immunoglobulin. Generally, extending the antibody (or BiSAb) half-life increases the mean residence time (MRT) of the administered BiSAb in the circulating blood.

[0111] Extending the half-life can reduce the amount of drug administered to a patient, as well as potentially decrease the frequency of administration. To extend the serum half-life of BiSAb, salvage receptor-binding epitopes, such as those described in U.S. Patent No. 5,739,277, may be incorporated into BiSAb (particularly the antibody fragment). As used herein, the term “salvage receptor-binding epitope” refers to an epitope in the Fc region of an IgG molecule (e.g., IgG1, IgG2, IgG3, or IgG4) that is involved in the in vivo serum half-life of the IgG molecule. Alternatively, the BiSAb of this disclosure with an extended half-life may be prepared by modifying amino acid residues that have been identified as being involved in the interaction between Fc and the FcRn receptor (see, for example, U.S. Patents No. 6,821,505 and No. 7,083,784). Furthermore, the half-life of the BiSAb of this disclosure may be extended by conjugation to PEG or albumin using techniques widely available in the art.

[0112] It is considered that both insertion of another binding domain into the Fc region as described herein and / or subsequent binding by an antigen may affect Fc activity. For example, a binding antigen may increase or decrease the binding affinity and activity to FcgR, Clq, and FcRn. This may form an antigen-dependent switch for modulating various antibody-dependent processes. In one embodiment, antigen binding may reduce interaction with FcRn, allowing free BiSAb to interact with FcRn and have a normal half-life, while enabling rapid clearance / intercellularization of the BiSAb-Ag complex. Furthermore, this may allow BD2-antigen-mediated interactions to affect the clearance of antigens bound to BD1. In another embodiment, BiSAb may contain an Fc region directly inserted into BD2 (Fc-BD2).

[0113] In one embodiment, the disclosure relates to the Fc regions numbered by the EU index listed in Kabat: 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 The present invention provides Fc variants comprising modifications (e.g., amino acid substitutions, amino acid insertions, amino acid deletions) at one or more positions selected from the group consisting of 308, 313, 316, 318, 320, 322, 325, 326, 327, 328, 329, 330, 331, 332, 333, 334, 339, 341, 343, 370, 373, 378, 392, 416, 419, 421, 428, 433, 434, 435, 436, 440, and 443. Optionally, the Fc area may include modifications in additional and / or alternative positions known to those skilled in the art (see, for example, U.S. Patent Nos. 5,624,821; 6,277,375; 6,737,056; 7,083,784; 7,317,091; 7,217,797; 7,276,585; and 7,355,008). Furthermore, useful amino acid positions and specific substitutions are illustrated in Tables 2 and 6-10 of U.S. Patent No. 6,737,056; the table shown in Figure 41 of U.S. Patent Application Publication No. 2006 / 024298; the tables shown in Figures 5, 12, and 15 of U.S. Patent Application Publication No. 2006 / 235208; the tables shown in Figures 8, 9, and 10 of U.S. Patent Application Publication No. 2006 / 0173170 and the tables shown in Figures 8-10, 13, and 14 of International Publication No. 09 / 058492.

[0114] In certain embodiments, this disclosure refers to the Fc region numbered by the EU index listed in Kabat as 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, 23 9T, 239H, 239Y, 240I, 240A, 240T, 240M, 241W, 241L, 241Y, 241E, 241R, 243W , 243L, 243Y, 243R, 243Q, 244H, 245A, 247L, 247V, 247G, 250E, 250Q, 251F, 2 52L, 252Y, 254S, 254T, 255L, 256E, 256F, 256M, 257C, 257M, 257N, 262I, 262 A, 262T, 262E, 263I, 263A, 263T, 263M, 264L, 264I, 264W, 264T, 264R, 264F, 2 64M, 264Y, 264E, 265A, 265G, 265N, 265Q, 265Y, 265F, 265V, 265I, 265L, 265 H, 265T, 266I, 266A, 266T, 266M, 267Q, 267L, 268E, 269H, 269Y, 269F, 269R, 270E, 280A, 284M, 292P, 292L, 296E, 296Q, 296D, 296N, 296S, 296T, 296L, 29 6I, 296H, 296G, 297S, 297D, 297E, 298A, 298H, 298I, 298T, 298F, 299I, 299L, 299A, 299S, 299V, 299H, 299F, 299E, 305I, 308F, 313F, 316D, 318A, 318S, 32 0A, 320S, 322A, 322S, 325Q, 325L, 325I, 325D, 325E, 325A, 325T, 325V, 325H, 326A, 326D, 326E, 326G, 326M, 326V, 327G, 327W, 327N, 327L, 328S, 328M, 32 8D, 328E, 328N, 328Q, 328F, 328I, 328V, 328T, 328H, 328A, 329F, 329H, 329Q,330K, 330G, 330T, 330C, 330L, 330Y, 330V, 330I, 330F, 330R, 330H, 331G, 331A, 331L, 331M, 331F, 331W, 331K, 331Q, 331E, 331 S, 331V, 331I, 331C, 331Y, 331H, 331R, 331N, 331D, 331T, 332D, 332S, 332W, 332F, 332E, 332N, 332Q, 332T, 332H, 332Y, 332A, 3 The present invention provides Fc variants comprising at least one substitution selected from the group consisting of 33A, 333D, 333G, 333Q, 333S, 333V, 334A, 334E, 334H, 334L, 334M, 334Q, 334V, 334Y, 339T, 370E, 370N, 378D, 392T, 396L, 416G, 419H, 421K, 428L, 428F, 433K, 433L, 434A, 424F, 434W, 434Y, 436H, 440Y, and 443W. Optionally, the Fc region may include, but is not limited to, additional and / or alternative amino acid substitutions known to those skilled in the art, all of which are illustrated by reference herein by reference in Tables 2 and 6-10 of U.S. Patent No. 6,737,056; the table shown in Figure 41 of U.S. Patent Application Publication No. 2006 / 024298; the tables shown in Figures 5, 12, and 15 of U.S. Patent Application Publication No. 2006 / 235208; the tables shown in Figures 8, 9, and 10 of U.S. Patent Application Publication No. 2006 / 0173170 and the tables shown in Figures 8, 9, and 10 of U.S. Patent Application Publication No. 20090041770.

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

[0116] In another particular embodiment, the disclosure provides an Fc mutant BiSAb in which the Fc region is the Fc region of IgG4 and comprises at least one modification at one or more positions selected from the group consisting of 228 and 235, numbered by the EU index described in Kabat. In yet another particular embodiment, the Fc region is the Fc region of IgG4 and the non-natural amino acid is selected from the group consisting of 228P, 235E and 235Y, numbered by the EU index described in Kabat.

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

[0118] In certain embodiments, the disclosure provides an Fc mutant BiSAb in which the Fc region comprises at least one non-natural amino acid at one or more positions selected from the group consisting of 252, 254, and 256, numbered by the EU index described in Kabat. In one embodiment, the modification is at least one substitution selected from the group consisting of 252Y, 254T, and 256E, numbered by the EU index described in Kabat. See U.S. Patent No. 7,083,784, which is incorporated herein by reference in its entirety.

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

[0120] In some embodiments, effector function induced by IgG antibodies is strongly dependent on the sugar portion bound to the Fc region of the protein (Claudia Ferrara et al., 2006, Biotechnology and Bioengineering 93:851-861). Therefore, effector function can be enhanced or reduced by glycosylation modification of the Fc region (e.g., Umana et al, 1999, Nat. Biotechnol 17:176-180; Davies et al., 2001, Biotechnol Bioeng 74:288-294; Shields et al, 2002, J Biol Chem 277:26733-26740; Shinkawa et al., 2003, J Biol Chem See 278:3466-3473; U.S. Patent No. 6,602,684; U.S. Patent No. 6,946,292; U.S. Patent No. 7,064,191; U.S. Patent No. 7,214,775; U.S. Patent No. 7,393,683; U.S. Patent No. 7,425,446; U.S. Patent No. 7,504,256; POTELLIGENT™ technology (Biowa, Inc., Princeton, NJ); GLYCOMAB™ glycosylation technology (GLYCART biotechnology AG, Zurich, Switzerland). Accordingly, in one embodiment, the Fc region of the BiSAb of this disclosure includes glycosylation of altered amino acid residues. In another embodiment, alteration of the glycosylation of amino acid residues results in reduced effector function. In yet another embodiment, alteration of the glycosylation of amino acid residues results in enhanced effector function. In certain embodiments, the Fc region suppresses fucosylation. In other embodiments, the Fc region is made low-fucose (see, for example, U.S. Patent Application Publication 2005 / 0226867).In one embodiment, such BiSAbs with enhanced effector function, particularly ADCC, are expressed in host cells (e.g., CHO cells, Lemna minor) that have been engineered to produce highly fucose-removed polypeptides with more than 100 times the ADCC compared to polypeptides produced by parental cells (Mori et al., 2004, Biotechnol Bioeng 88:901-908; Cox et al., 2006, Nat Biotechnol., 24:1591-7).

[0121] Adding sialic acid to oligosaccharides on IgG molecules can enhance their anti-inflammatory activity and alter their cytotoxicity (Keneko et al., Science, 2006, 313:670-673; Scallon et al., Mol.Immuno. 2007 Mar;44(7):1524-34). The above studies demonstrate that IgG molecules with increased sialylation possess anti-inflammatory properties, while IgG molecules with decreased sialylation enhance immunostimulatory activity (e.g., enhance ADCC activity). Therefore, BiSAb may be modified with a sialylation profile appropriate for specific applications (US Patent Application Publication No. 2009 / 0004179 and International Publication No. 2007 / 005786).

[0122] In one embodiment, the Fc region of the BiSAb of this disclosure includes a modified sialylation profile compared to the natural Fc region. In another embodiment, the Fc region of the BiSAb of this disclosure includes an increased sialylation profile compared to the natural Fc region. In yet another embodiment, the Fc region of the BiSAb of this disclosure includes a decreased sialylation profile compared to the natural Fc region.

[0123] In one embodiment, the Fc variant of this disclosure is compared with other known Fc variants, for example: Ghetie et al., 1997, Nat Biotech. 15:637-40; Duncan et al, 1988, Nature 332:563-564; Lund et al., 1991, J. Immunol 147:2657-2662; Lund et al, 1992, Mol Immunol 29:53-59; Alegre et al, 1994, Transplantation 57:1537-1543; Hutchins et al., 1995, Proc Natl. Acad Sci USA 92:11980-11984; Jefferis et al, 1995, Immunol Lett. 44:111-117; Lund et al., 1995, Faceb J 9:115-119;Jefferis et al,1996,Immunol Lett 54:101-104;Lund et al,1996,J Immunol 157:4963-4969;Armour et al.,1999,Eur J Immunol 29:2613-2624;Idusogie et al,2000,J Immunol 164:4178-4184;Reddy et al,2000,J Immunol 164:1925-1933;Xu et al.,2000,Cell Immunol 200:16-26;Idusogie et al,2001,J Immunol 166:2571-2575;Shields et al. al.,2001,J Biol Chem 276:6591-6604;Jefferis et al,2002,Immunol Lett 82:57-65;Presta et al.,2002, Biochem Soc Trans 30:487-490); US Patent No. 5,624,821; US ​​Patent No. 5,885,573; US Patent No. 5,677,425; US Patent No. 6,165,745; US Patent No. 6,277,375; US Patent No. 5,869,046; US Patent No. 6,121,022; US Patent No. 5,624,821; US ​​Patent No. 5,648,260; US Patent No. 6,528,624; US Patent No. 6,194,5 This may be combined with the specifications disclosed in Specification No. 51; Specification No. 6,737,056; Specification No. 7,122,637; Specification No. 7,183,387; Specification No. 7,332,581; Specification No. 7,335,742; Specification No. 7,371,826; Specification No. 6,821,505; Specification No. 6,180,377; Specification No. 7,317,091; and Specification No. 7,355,008. Other modifications and / or substitutions and / or additions and / or deletions of the Fc domain will also be readily apparent to those skilled in the art.

[0124] It should be noted that polypeptides presented in BiSAb format containing natural Fc retain the ability to bind to FcRn and C1q and mediate ADCC, as shown in the example. Therefore, in some embodiments, BiSAb retains the ability to bind to FcRn and / or C1q and / or one or more Fcγ receptors (FcγR). For example, in some embodiments, 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 FcγR compared to a conventional antibody that binds to one of the epitopes to which BiSAb binds. In some embodiments, BiSAb is prepared from the binding domains of one or two conventional antibodies, and its activity is compared with one or both of these conventional antibodies.

[0125] Alternatively, a heavy-chain heterodimer may be constructed using the modified Fc region to obtain a BiSAb containing two different heavy-chain-light-chain pairs. To facilitate heterodimer formation, the interface between the pair of Fc regions is manipulated to maximize the proportion of heterodimers recovered from recombinant cell culture. In one embodiment, the interface includes at least a portion of the CH3 domain. In this method, a "projection" is created by substituting one or more smaller amino acid side chains on the interface of the first antibody molecule with a larger side chain (e.g., tyrosine or tryptophan). By substituting the larger amino acid side chain with a smaller amino acid side chain (e.g., alanine or threonine), a "cavity" of the same or similar size as the larger side chain is formed on the interface of the second antibody molecule as compensation. This is the mechanism by which the yield of heterodimers is increased compared to other undesirable end products such as homodimers. Examples of CH3 modifications include Y407V / T366S / L368A on one heavy chain and T366W on the other; and S354C / T366W on one heavy chain and Y349C / Y407V / T366S / L368A on the other. Further modifications that give a projection on one chain and a cavity on the other are described in U.S. Patent No. 7,183,076; U.S. Patent Application Publication No. 2014 / 0348839; and Merchant et al., 1998, Nat. Biotech 16:677-681. Some non-limiting examples of modifications that can result in a projection-cavity configuration are shown in Table 1a. Other modifications that can be used to form heterodimers include, but are not limited to, modifications that alter the charge polarity at both ends of the Fc dimer interface so that heterodimerization occurs when electrostatically matched Fc regions are co-expressed. Modifications that alter charge polarity include, but are not limited to, those shown in Table 1b below (see also International Publication No. 20090182127; Gunasekaran et al., 2010, JBC 285:19637-46).Furthermore, Davis et al. (2010, Prot.Eng.Design&Selection 23:195-202) describe a heterodimer Fc platform using the strand-exchanged engineered domain (SEED) CH3 region, which is a derivative of the human IgG domain and the IgA CH3 domain (see also International Publication No. 2007 / 110205).

[0126] [Table 1]

[0127] [Table 2]

[0128] Those skilled in the art will understand that, in some embodiments, Fc fusion proteins may form dimers due to the homodimerization of molecules containing the Fc region. In some embodiments, different operations may be performed on the Fc region of the binding protein (e.g., BiSAb) to promote and / or maintain heterodimerization (e.g., different operations using chimeric mutations, complementary mutations, dock-and-lock mutations, knob-into-hole mutations, strand exchange domain (SEED) mutations, 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; International Publication No. 2007 / 110205; International Publication No. 2007 / 147901; Gunasekaran et al. (2010) JBC See 285:19637-46. All of these are incorporated herein by reference. Thus, binding proteins can be manipulated to form heterodimers comprising, for example, a BiSAb fused with a first binding protein, binding domain, or first Fc region or fragment thereof, and a BiSAb fused with a second (i.e., different) binding protein, binding domain, or second Fc region or fragment thereof, where the first and second Fc regions or fragments are manipulated to heterodimerize.

[0129] 3. Glycosylation Glycosylation can not only alter the effector function of a polypeptide, but glycosylation modification of the variable region can also alter the affinity of the antibody (or BiSAb) to a target antigen. In one embodiment, the glycosylation pattern of the variable region of the BiSAb is modified. For example, a non-glycosylated BiSAb may be created (i.e., the BiSAb lacks glycosylation). Glycosylation can be modified, for example, to increase the affinity of the BiSAb to a target antigen. Such glycosylation can be achieved, for example, by altering one or more glycosylation sites in the BiSAb sequence. For example, glycosylation sites in one or more variable region frameworks may be removed, and one or more amino acid substitutions may be made to remove glycosylation at those sites. Such nonglycosylation can increase the affinity of the BiSAb to the antigen. These approaches are described in more detail in U.S. Patent No. 5,714,350 and U.S. Patent No. 6,350,861. Furthermore, one or more amino acid substitutions may be performed to remove the glycosylation site present in the Fc region (for example, asparagine 297 in IgG). In addition, non-glycosylated BiSAb may be prepared using bacterial cells lacking the necessary glycosylation mechanism.

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

[0131] An exemplary and non-limiting example of a linker is a polypeptide chain containing at least four residues. Some of these linkers may be flexible, hydrophilic, and may have little to no secondary structure of their own (linker portion or flexible linker portion). Linkers of at least four amino acids may be used to link neighboring domains and / or regions after the molecules have assembled. Longer or shorter linkers may also be used. For this reason, 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 parts of a molecule together, the linkers may be identical or different (for example, identical or different in length and / or amino acid sequence).

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

[0133] Linkers facilitate the formation of the desired structure. Linkers are (Gly-Ser) n The linker may contain residues, and to enhance solubility, Glu or Lys residues may be dispersed to some extent throughout. Alternatively, or in addition to this, the linker may contain no serine residues at all, which may be preferable when the linker undergoes O-linked glycosylation. In some embodiments, for example, when linker dimerization is used to make the domain of BiSAb a properly folded structure, the linker may contain cysteine ​​residues. In some embodiments, BiSAb contains at least two polypeptide linkers that ligate the polypeptide domains. In other embodiments, BiSAb contains at least three polypeptide linkers. In other embodiments, BiSAb contains four or more polypeptide linkers.

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

[0135] The polypeptide linker may contain all or part of a hinge region naturally containing three cysteine ​​residues. In some embodiments, a selected hinge region is cleaved, or otherwise modified or substituted, such that only one or two cysteine ​​residues remain in the complete and / or natural hinge region. Similarly, in certain other embodiments, the polypeptide linker may contain a mutant or otherwise modified portion of a hinge region in which the number of cysteine ​​residues is reduced by amino acid substitution or deletion, for example, a mutant or otherwise modified hinge region containing 0, 1, or 2 cysteine ​​residues as described herein.

[0136] Accordingly, the mutant hinge domain or other modified hinge domain may be derived from or constructed using a wild-type immunoglobulin hinge domain containing one or more cysteine ​​residues. In some embodiments, the mutant or other modified portion of the hinge region may not contain cysteine ​​residues, or may contain only one cysteine ​​residue, and the mutant or other modified hinge region is or is derived from a wild-type immunoglobulin hinge region containing one or more cysteine ​​residues or two or more cysteine ​​residues, respectively. In the mutant or other modified portion of the hinge region, the cysteine ​​residues of the wild-type immunoglobulin hinge region are preferably deleted or substituted with amino acids that cannot form disulfide bonds. In some embodiments, the mutant or other modified portion of the hinge region is or is derived from a human IgG wild-type hinge region which may contain any of the four human IgG isotype subclasses IgG1, IgG2, IgG3, or IgG4.

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

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

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

[0140] In some embodiments, a linker (e.g., L1 and / or L2 and / or L3) includes both a hinge portion and a linker portion, such as a gly-ser linker. In other embodiments, L1 and / or L2 includes only a hinge portion or only a linker portion, such as a gly-ser linker. In other embodiments, L1 and L2 include gly-ser linker portions. In some embodiments, the gly-ser linkers in a BiSAb are of the same length, while in other embodiments, the gly-ser linker portions in a BiSAb (e.g., L1 and L2) are of different lengths. If a BiSAb includes an scFv, the heavy and light chains of the scFv may be connected to a BiSAb (e.g., BD1, Fab, Fc, etc.) by a flexible linker. This flexible linker generally does not include a hinge portion and is rather a gly-ser linker or other flexible linker. The length and amino acid sequence of the flexible linkers that interconnect the domains of scFv can be easily selected and optimized (for example, (Gly4Ser)n, (SEQ ID NO: 48), where n is 2, 3, or 4 or greater).

[0141] Regardless of the various binding units and domains (e.g., binding domains / units (e.g., Fab-scFv) or polypeptide linkers used to link binding domains / units to Fc (e.g., scFv via L1 and L2)), BiSAb may optionally include additional polypeptide linkers. The length and sequence of such additional polypeptide linkers are independently selected. For example, BiSAb may further include flexible polypeptide linkers that link the variable heavy and light chains of scFv. These flexible polypeptide linkers may include gly-ser linkers. Generally, these linkers do not include hinge portions.

[0142] Here, it is considered that changes in the length of the linker flanking to BD2 can affect the orientation of the BD2 antigen-binding site and its spacing from the rest of the BiSAb molecule. For example, a short N-terminal linker and a long C-terminal linker may produce an orientation in which the binding site is conformed in one direction, while a long N-terminal linker and a short C-terminal linker may affect the opposite conformational orientation. Therefore, the length of the linker can modulate the orientation of the BD2 antigen-binding site and may have a significant impact on the generation or avoidance of steric effects between BD1 and BD2 and / or between BD2 and other entities that bind to antibody molecules in the Fc or other domains.

[0143] 5. Specific structure of BiSAb As previously stated, one aspect of this disclosure relates to a BiSAb structural configuration (platform) comprising two heavy-chain-light-chain pairs (shown in Figures 1A-1F). In some embodiments of this aspect, the polypeptide sequence of the BiSAb chimeric heavy chain may include 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 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 remaining Fc domain. In some embodiments, the Fc domain is C H 2 domains and C H It contains 3 domains. Therefore, in one embodiment, the orientation from the N-terminus to the C-terminus is as follows: VH1-C H 1-C H 2(N-terminus)-L1-BD2-L2-C H 2(C-terminus)-C H 3; VH1-C H 1-C H 2-L1-BD2-L2-C H 3; and VH1-C H 1-C H 2-C H 3(N-terminus)-L1-BD2-L2-C H3 provides a BiSAb chimeric heavy chain which may contain a polypeptide sequence at the C-terminus. The polypeptide sequence of the BiSAb light chain may contain a light chain variable domain (VL1) and a light chain constant domain (CL). Thus, the BiSAb light chain may contain a polypeptide sequence from the N-terminus to the C-terminus in the following orientation: VL1-CL. Note that VH1, VL1, and CL are used to indicate the "binding unit 1" (BD1) portion which binds to a first epitope. BD2 is used to indicate the "binding unit 2" portion which binds to a second epitope.

[0144] In embodiments where the binding domain is scFv, the BiSAb chimeric heavy chain may include a polypeptide sequence containing an antibody heavy chain variable domain (VH1), a polypeptide sequence containing an antibody heavy chain constant domain 1 (CH1), a polypeptide sequence containing a first polypeptide linker (L1), a polypeptide sequence containing an antibody light chain variable domain (VL2), a polypeptide sequence containing a flexible linker, a polypeptide sequence containing an antibody heavy chain variable domain (VH2), a polypeptide sequence containing a second polypeptide linker (L2), and a polypeptide sequence containing an antibody Fc domain. Therefore, the chimeric heavy chain of BiSAb containing scFv as BD2 has the following orientation from the N-terminus to the C-terminus: VH1-CH1-CH2(N-terminus)-L1-VL2-L3-VH2-L2-CH2(C-terminus)-CH3;VH1-CH1-CH2-L1-VL2-L3-VH2-L2-CH3;VH1-CH1-CH2-CH3(N-terminus)-L1-VL2-L3-VH The polypeptide sequence may also include 2-L2-CH3(C-terminus);VH1-CH1-CH2(N-terminus)-L1-VH2-L3-VL2-L2-CH2(C-terminus)-CH3;VH1-CH1-CH2-L1-VH2-L3-VL2-L2-CH3; and VH1-CH1-CH2-CH3(N-terminus)-L1-VH2-L3-VL2-L2-CH3(C-terminus).

[0145] A chimeric heavy chain is a polypeptide chain containing an amino acid sequence (for example, the amino acid sequence of each polypeptide domain). Note that VH1, VL1, and CL are used to indicate parts of binding unit 1, with VH1 and VL1 indicating the portion that binds to its first epitope. VH2 and VL2 are used to indicate parts of binding unit 2 that bind to a second epitope. In some embodiments, additional scFv binding domains are present at the N-terminus and / or C-terminus of the polypeptide constituting the BiSAb core (where BiSAb further includes binding units (BD) 3 and / or 4 and / or 5). In some embodiments, two or more scFv binding domains are present within the BiSAb core. The additional scFvs include antibody heavy chain variable regions indicated by VH3, VH4, and VH5, and corresponding antibody light chain variable regions indicated by VL3, VL4, and VL5, respectively.

[0146] 6. Signs, conjugates and parts In certain features, drugs and other molecules may target BiSAb by site-specific conjugation. For example, BiSAb may include a cysteine ​​manipulation domain (containing cysteine ​​introduced into the binding unit and / or Fc domain) from which a free thiol group for the conjugation reaction is obtained. In some embodiments, BiSAb may be manipulated to include a specific conjugation site. In some embodiments, the disclosure provides an Fc mutant BiSAb in which the Fc region contains amino acid substitutions at one or more of the EU indexed positions 239, 282, 289, 297, 312, 324, 330, 335, 337, 339, 356, 359, 361, 383, 384, 398, 400, 440, 422, and 442. In some embodiments, the Fc region includes substitutions in one or more of the following groups: a) positions 289 and 440; b) positions 330 and 440; c) positions 339 and 440; d) positions 359 and 440; e) positions 289 and 359; f) positions 330 and 359; g) positions 339 and 359; h) positions 289 and 339; i) positions 330 and 339; j) positions 289 and 330; k) positions 339 and 442; l) positions 289, 339 and 442; m) positions 289, 330 and 339; n) positions 330, 339 and 442; and o) positions 289, 330 and 442. In other embodiments, the Disclosure provides a BiSAb in which the CH1 domain of the Fab arm includes a substitution at one or more of the EU indexed positions 131, 132, 134, 135, 136, and 139. In one embodiment, the substitution includes an amino acid substitution selected from cysteine, lysine, tyrosine, histidine, selenocysteine, and selenomethionine. In a particular embodiment, the substitution is cysteine. Methods for producing stable cysteine-manipulated antibodies are described in U.S. Patent No. 7,855,275, U.S. Patent Application Publication No. 20110033378, and U.S. Patent Application Publication No. 20120213705, which are incorporated herein by reference in their entirety.

[0147] 7. Exemplary Target While the various aspects and embodiments relating to the DuetMab and BiSAb platforms described herein may be formulated to bind to any one or more desired targets, the BiSAbs disclosed herein are preferably designed to target specific target molecule pairs (for example, binding unit 1 binds to one target and binding unit 2 binds to the other target). As described above and illustrated in the following exemplary examples, the antibodies, DuetMab, and BiSAb disclosed herein target recipient subjects or molecules that modulate the immune response in immune cells in culture. In some embodiments, the binding domains exhibit specific binding activity to targets selected from the group consisting of CTLA-4, PD-1, PD-L1, OX40, and TIM3. DuetMab and BiSAb may include different combinations of binding domains in various orders and orientations, these domains having binding affinity to or specifically binding to the targets disclosed herein. For example, the DuetMab and BiSAb disclosed herein may include combinations of binding domains that enable bispecific binding to targets such as CTLA-4 and PD-1; CTLA-4 and PD-L1; and CTLA-4; CTLA-4 and TIM3; PD-1 and PD-L1; PD-L1 and OX40; PD-1 and TIM3; and PD-L1 and TIM3. DuetMab and BiSAb containing binding domains that bind to specific target combinations are exemplified in the examples and include non-limiting combinations of PD-1 / CTLA-4; PD-L1 / CTLA-4; PD-1 / TIM3; and PD-L1 / OX40. In some embodiments, the BiSAb has enhanced binding properties compared to the binding properties of individual single-specific binding protein combinations used to construct the BiSAb.

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

[0149] In some embodiments, due to the multimeric nature of the DuetMab or BiSAb of the present disclosure, a labeling or therapeutic agent can target a specific cell type or molecular target. For example, one functional domain of the DuetMab or BiSAb can bind to a cell surface target, while at the same time another functional domain of the same DuetMab or BiSAb can bind to a hapten or labeling agent useful for detection. Similarly, one functional domain can bind to a cell target, while at the same time a second functional domain can bind to a toxin. Since both binding reactions are mediated by a single molecule, the toxin can be placed in the vicinity of the cell target where the cytotoxic function affects the cell.

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

[0151] C. Vectors and host cells for producing BiSAb and subsequent purification This disclosure relates to a method for producing BiSAb. In some embodiments, a recombinant nucleic acid molecule encoding all or part of the BiSAb disclosed herein may be operably ligated to one or more regulatory nucleotide sequences of an expression construct. The nucleic acid sequences encoding the light chain and chimeric heavy chain of BiSAb may be cloned into the same expression vector in any orientation (e.g., light chain before or after heavy chain) or into two different vectors. When expression is performed using one vector, the two coding genes may have their own gene elements (e.g., promoter, RBS, leader, stop, polyA, etc.) or may be cloned as a set of gene elements, ligated to a cistron element. The regulatory nucleotide sequences are generally appropriate for the host cell used for expression. Many types of expression vectors and suitable regulatory sequences appropriate for various host cells are known in the art. Typically, the one or more regulatory nucleotide sequences include, but are not limited to, promoter sequences, leader or signal sequences, ribosome binding sites, transcription start and termination sequences, translation start and termination sequences, and enhancer or activator sequences. This disclosure intends constitutive or inducible promoters known in the art. The promoter may be a native promoter or a hybrid promoter combining elements of two or more promoters. The expression construct may reside on the episome of a cell, for example, on a plasmid, or the expression construct may be inserted into a chromosome.

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

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

[0154] Recombinant nucleic acids may be produced by ligating a cloned gene or a portion thereof into a vector suitable for expression in prokaryotic cells, eukaryotic cells (yeast, birds, insects, or mammals), or both. Expression media that produce recombinant polypeptides include plasmids and other vectors. For example, suitable vectors include the following types of plasmids for expression in prokaryotic cells, such as E. coli: plasmids derived from pBR322, pEMBL, pEX, pBTac, and pUC. In some embodiments, a mammalian expression vector includes both a prokaryotic sequence that facilitates vector propagation in bacteria and one or more eukaryotic transcription units that are expressed in eukaryotic cells. Vectors derived from pcDNAI / amp, pcDNAI / neo, pRc / CMV, pSV2gpt, pSV2neo, pSV2-dhfr, pTk2, pRSVneo, pMSG, pSVT7, pko-neo, and pHyg are examples of mammalian expression vectors suitable for eukaryotic cell transfection. Some of these vectors are modified with bacterial plasmids, such as sequences derived from pBR322, to facilitate replication and drug resistance selection in both prokaryotic and eukaryotic cells. Alternatively, for transient protein expression in eukaryotic cells, derivatives of viruses, such as bovine papillomavirus (BPV-1) or Epstein-Barr virus (pHEBo, pREP, and p205), may be used. Various methods used for plasmid preparation and host organism transformation are known in the art. For other expression systems suitable for both prokaryotic and eukaryotic cells, as well as general recombination 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 preferable to express recombinant polypeptides using a baculovirus expression system.Examples of such baculovirus expression systems include vectors derived from pVL (such as pVL1392, pVL1393, and pVL941), vectors derived from pAcUW (such as pAcUW1), and vectors derived from pBlueBac (such as pBlueBac III containing β-gal).

[0155] Once the molecules are produced, they may be purified by any method known in the art for purifying proteins, immunoglobulin molecules, or other polymer molecules, for example, by chromatography (e.g., ion exchange chromatography, affinity chromatography (particularly by affinity for specific antigen protein A or protein G), and size column chromatography), centrifugation, solubility difference, or any other polymer molecule technique for protein purification. Furthermore, the molecules disclosed herein may be fused with heterologous polypeptide sequences (e.g., affinity tags) that are routinely used to facilitate purification.

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

[0157] D. Pharmaceutical preparations In some embodiments, the present disclosure provides a pharmaceutical composition. Such a pharmaceutical composition may comprise a nucleic acid molecule encoding BiSAb. Alternatively, such a pharmaceutical composition may comprise DuetMab, BiSAb, a combination of DuetMab, or a combination of BiSAb with a pharmaceutically acceptable excipient. In some embodiments, the pharmaceutical compositions of the present disclosure are used as drugs.

[0158] E. Use As described herein, DuetMab and BiSAb can be conjugated to targets associated with cancer and cell proliferation disorders or disorders, which may be responsive to immunotherapy, for example, by inhibiting immunosuppressive activity and / or inducing an immune response associated with the target molecule. For example, abnormal signaling and / or inhibited immune responses may contribute to undesirable cell proliferation and cancer. Accordingly, the DuetMab, BiSAb and antibodies disclosed herein can be used to target cancer, treat undesirable cell proliferation and / or cancer associated with inhibited, reduced or inadequate immune responses. In particular, the tumor growth curve of tumors and / or tumor volume may be reduced by administration of DuetMab to BiSAb that induces and / or stimulates an immune response in subjects such as human patients with cancer.

[0159] Thus, the Disclosure also relates to various methods, including the administration of the binding proteins disclosed herein to subjects in need thereof. In one embodiment, the Disclosure relates to a method for inducing an immune response in subjects having or at risk of developing cancer, which includes administering the binding proteins disclosed herein to the subjects. In some embodiments, the method activates an immune response against cancer in the subjects. In some embodiments, the method enhances an immune response against cancer in the subjects. In some embodiments, the method activates an inhibited immune response pathway in the subjects, and this activation increases an immune response targeting cancer in the subjects. In some embodiments, the method enhances an immune response pathway targeting cancer in the subjects.

[0160] In another embodiment, the Disclosure relates to a method for treating cancer in a subject that requires such treatment, which includes administering a binding protein disclosed herein to the subject. In one embodiment, the method for treating cancer includes stopping or slowing the growth of the cancer in the subject. In one embodiment, the method for treating cancer includes stopping or slowing the metastasis of the cancer to other sites in the subject. In one embodiment, the method for treating cancer includes killing cancer cells in the subject. In one embodiment, the method for treating cancer includes stopping the proliferation and / or spread of cancer cells in the subject.

[0161] In various embodiments of the aforementioned aspects, the method relates to treating a patient for a tumor disease and / or cancerous disease. In some embodiments, cancer is selected from a group of cancers that are sensitive to an immune response induced in the subject. In some embodiments, 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, cancer is gastrointestinal or gastrointestinal cancer (e.g., anal cancer; bile duct cancer; extrahepatic bile duct cancer; appendiceal cancer; carcinoid tumors, colon cancer; colorectal cancer including pediatric colorectal cancer; esophageal cancer including pediatric esophageal cancer; bladder cancer; gastric (stomach) cancer including pediatric gastric cancer; hepatocellular carcinoma including adult (primary) hepatocellular carcinoma and pediatric hepatocellular carcinoma (e.g., hepatocellular carcinoma); pancreatic cancer including pediatric pancreatic cancer; sarcoma, rhabdomyosarcoma; islet cell carcinoma; rectal cancer; and small intestine cancer); lung cancer (and For example, small cell lung cancer (NSCLC) and small cell lung cancer (SCLC); head and neck cancers (e.g., lip and oral cancer; oral cancer including pediatric oral cancer; hypopharyngeal cancer; pharyngeal cancer including pediatric pharyngeal cancer; metastatic squamous cell cervical cancer of unknown primary origin; oral cancer; nasal and paranasal sinus cancer; nasopharyngeal cancer including pediatric nasopharyngeal cancer; oropharyngeal cancer; parathyroid cancer; pharyngeal cancer; salivary gland cancer including pediatric salivary gland cancer; throat cancer; and thyroid cancer); ovarian cancer and breast cancer are selected.

[0162] In the method described above, the amount of binding protein administered to the target is effective in inducing an immune response, increasing the immune response, stopping or slowing cancer growth, stopping or slowing cancer metastasis, killing cancer cells, and / or stopping or slowing the proliferation and / or spread of cancer cells.

[0163] In embodiments of the aforementioned methods, the binding protein includes Duetmab and BiSAb as disclosed herein. In some embodiments of the aforementioned methods, the binding protein includes an antibody or its antigen-binding fragment, as disclosed herein.

[0164] As used herein, the term “Subject” is intended to include humans and non-human animals, in particular mammals. Examples of subjects include human patients with disorders, such as cancer, as described herein, or human patients with normal conditions. “Non-human animals” include all vertebrates, such as non-mammals (e.g., chickens, amphibians, reptiles, etc.) as well as mammals such as non-human primates, livestock and / or agricultural animals (e.g., sheep, dogs, cats, cattle, pigs, etc.) and rodents (e.g., mice, rats, hamsters, guinea pigs, etc.). In certain embodiments, the subject is a human patient.

[0165] "Treatment" or "to treat" refers to both therapeutic treatment and preventive or protective measures. Those requiring treatment include individuals who already have a disability, those prone to disability, or those who should be prevented from developing a disability. When used in reference to a disease or subject requiring treatment, the term therefore includes, but is not limited to, stopping or slowing the progression of the disease, achieving remission of the disease, preventing symptoms, reducing the severity of the disease and / or symptoms, or shortening the duration of the disease compared to an untreated subject. In some embodiments, a treatment method may mitigate one or more clinical indications of a particular disease to be treated.

[0166] The examples described below are provided to illustrate specific aspects and embodiments of the present disclosure described above and should not be construed as limiting the scope of the description or the claimed subject matter of the appended patent.

Example

[0167] Materials and Methods Immune Response Modulation Assay To evaluate the potential immune responses induced by some of the immunotherapeutic molecules described herein, a cytomegalovirus (CMV) antigen recall assay was used. The reagents for the assay included: - CMV-reactive cryopreserved 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 (Miltenyi Biotec, cat#130-093-438, 50 μg / ml); - Ovalbumin (Thermo scientific cat#: 77,120, 1 mg / ml); - Coater, 96-well plate non-TC treated (Corning, cat#3788); and - Immunotherapeutic molecule.

[0168] The general assay protocol: The day before performing the assay, the cryopreserved PBMC were thawed in warm AIM V medium. The cells were washed twice in the Coater round well plate. The cell concentration was adjusted to 1×10 6 cells / mL.

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

[0170] The following day, 100 μL of AIMV medium containing a 2× PepTivator CMV peptide pool (0.1 μg / ml to 0.05 μg / ml final) and 2× immunotherapy drug molecules was added to the wells.

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

[0172] Staphylococcus enterotoxin A / B (SEA / SEB) assay protocol To determine the effect of DuetMab or BiSAb on the IL-2 immune response, the reagents used in either the SEB or SEA assay protocol include: - Leukocyte pyramid (NHSBT code NC24; from Addenbrookes Hospital); - 50ml Falcon tube (BD352070); - 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); - A 1 mg / ml stock solution of Staphylococcus aureus (Staphylococcus) enterotoxin A (SEA; Sigma, S-9399) or Staphylococcus aureus (Staphylococcus) enterotoxin B (SEB; Sigma, S-4881), stored at -20°C: - Culture media (all from Life Technologies): RPMI1640 containing 10% v / v thermally inactivated FCS (90005M) and glutamax (61870) supplemented with 100 U / mL penicillin + 100 U / mL streptomycin (15140-122); - V-shaped bottom plate (Greiner BioOne 651201); - 96-well flat-bottom plate (Corning Costar 7107).

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

[0174] General assay protocols PBMCs were isolated from human blood leukocyte pyramids (NHS Blood and Transplant Service code NC24) using density gradient centrifugation (Ficoll-Paque PLUS; GE Healthcare), followed by lysing of red blood cells in ammonium chloride solution (Stemcell Technologies). Anti-human CD3 (clonal OKT3 at 0.5 ug / ml in PBS; eBioscience) was coated in a flat-bottomed 96-well plate (Corning Costar 7107) at 37°C for 2 hours. Next, PBMCs were coated in medium (RPMI1640-Glutamax supplemented with 10% v / v heat-inactivated bovine serum and 100 U / 100 ug / ml streptomycin / penicillin (respectively) (Life Technologies)) at a rate of 0.2 × 10⁶ per well. 6 Cells were added. PBMCs were further stimulated by the addition of Staphylococcal Enterotoxin A or B (SEB; Sigma Aldrich) in the range of 0.0088 to 0.1 ug / ml, after which candidate DuetMab or BiSAb was added up to the final test concentration. After 3 days of incubation at 37°C and 5% CO2, the supernatant was removed from the cells, and IL-2 secretion was determined using a commercially available ELISA according to the manufacturer's instructions (R&D Systems Duoset product code DY202). See Figure 90.

[0175] Mixed leukocyte reaction (MLR) assay protocol (fresh blood) Furthermore, an MLR cell assay was also used to obtain in vitro correlations of T cell function in response to DuetMab and BiSAb as disclosed herein. The reagents used to perform the MLR assay from fresh blood samples include: - 8 mL CPT heparin tube; - AIM-V Medium (serum-free) Gibco #12055-091, no additives; - 50ml conical tube; - 2ml cryopreservation container; - ACK lysis buffer (Gibco #A10492-01); - 96-well tissue culture treated round-bottom plate BDfalcon #3077; - PHA (Roche) 1 mg / ml (final concentration 10 ug / mL) as a positive control.

[0176] General assay protocols PBMCs were prepared from blood samples introduced into CTP heparin tubes. The tubes were centrifuged at 2700 rpm for 20 minutes at 25°C without using a brake. The upper layer of serum was aspirated. The remaining material was gently collected with a pipette, and all of it above the CPT tube plug was collected and introduced into a 50 ml conical tube. The cells were washed with AIM-V medium (three times for 5 minutes at 1500 rpm, 25°C, with the brake on). All remaining red blood cells were lysed with erythrocyte lysis buffer (e.g., approximately 3 ml of buffer for approximately 5 minutes). The remaining cells were washed twice with AIM-V medium (5 minutes at 1500 rpm, 25°C, with the brake on). If necessary, the pellet was solidified in a single tube, resuspended in AIM-V medium, and cell counting was performed.

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

[0178] Human TH1 / TH2MSD10-plex protocol This assay was used to determine the amount of cytokines present in the culture supernatant in response to administration of DuetMab and BiSAb as disclosed herein. To perform this assay, an inhibitor was prepared by dissolving 200 mg of inhibitor B in 20 ml of PBS per plate. 150 µl of the dissolved inhibitor was added to each well. The plate was sealed and shaken at room temperature for 2 hours or overnight at 4°C. The wells were washed three times with PBST buffer. A calibrator was prepared by diluting 10 µl of frozen calibrator blend with 1 ml of diluent, and this was further diluted fourfold sequentially. 25 µl of calibrator (standard) and 25 µl of sample were added to separate the wells. The wells were incubated at room temperature for 2 hours with shaking. After incubation, the wells were washed three times with PBST.

[0179] The detection antibody was prepared, diluted to the required concentration, and added to each well. After incubation at room temperature for 2 hours with shaking, the wells were washed with PBST (3 times). Reading buffer was added to each well before reading with the MSD instrument.

[0180] Tumor-specific killing assay protocol Human gp100 209-217Human CD8+ T cell line (JR6C12) reactive to peptides was generously provided by Dr. Steven Rosenberg (National Cancer Institute, Bethesda, MD). JR6C12 cells were co-cultured with CFSE (CellTrace CFSE proliferation kit, ThermoFisher)-labeled human melanoma cell line (Mel624) in a 1:1 ratio (20,000 JR6C12 + 20,000 Mel624) in a 96-well flat-bottom plate at 37°C for 18 hours. Candidate molecules were added at a concentration of 69 nM at time 0 of co-culture. After 18 hours, the wells were visualized by bright-field microscopy. The supernatant was collected for MSD analysis, adherent cells were trypsinized, washed (2×) with PBS, and then stained in vivo (Zombie UV Fixable Viability kit, Biolegend). The uptake of vital stains by CFSE-labeled cells was evaluated by flow cytometry using LSRFortessa (BD).

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

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

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

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

[0185] [Table 3]

[0186] [Table 4]

[0187] [Table 5]

[0188] [Table 6]

[0189] [Table 7]

[0190] [Table 8]

[0191] [Table 9]

[0192] Example 2(a) PD-1 / CTLA-4 bispecific binding protein While not bound by any particular theory, the combination of PD-1 and CTLA-4 blockade has strong clinical and preclinical rationale. Therefore, it would be desirable to maximize the risk / benefit ratio of the PD-1 and CTLA-4 combination (Figure 4).

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

[0194] [Table 10]

[0195] [Table 11]

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

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

[0198] Using the parental sequences identified above in Table 2, we constructed bispecific binding proteins in DuetMab format that bind to PD-1 and CTLA-4. PD-1 / CTLA-4 DuetMab was constructed using the sequences shown in Table 4 below and evaluated, including a comparison with PD-1 / CTLA-4 BiS5, as described later.

[0199] [Table 12]

[0200] [Table 13]

[0201] [Table 14]

[0202] [Table 15]

[0203] Octet binding assay (DuetMab) A simultaneous binding assay for two individual antigens was performed by Octet analysis. After loading biotinylated human PD-1 onto streptavidin sensors, sequential interactions were first performed with DuetMab PD-1 / CTLA-4 and then with soluble CTLA-4 antigen. Streptavidin (SA) biosensors (ForteBio) were used to capture 5 μg / mL biotinylated human PD-1 in PBS pH 7.2, 3 mg / mL BSA, 0.05% (v / v) Tween® 20 (assay buffer). After the wash step, the loaded biosensors were subjected to sequential association and dissociation interactions with a sample well containing 133 nM DuetMab PD-1 / CTLA-4 bispecific antibody and then with a well containing 200 nM human CTLA-4 antigen. The results of the binding are shown in Fig. 5.

[0204] The intrinsic kinetics of the PD-1 / CTLA-4 DuetMab bispecific antibody were also evaluated by BiaCore. Binding experiments were performed using a BIAcore T200 instrument (BIAcore). Mouse anti-huIgG-Fab was immobilized on a CM5 chip to a target response of 2000 RU to capture the antibody. To achieve a captured antibody of approximately 100 response units, 100 nM DuetMab or mAb was flowed at 20 μL / min for 5 min. Next, the antigen was continuously injected for 5 min at a flow rate of 50 μl / min. Kinetic parameters (k on and k off ) as well as the dissociation constant (KD) were calculated from a non-linear fit using BIAevaluation 4.1 software. The binding results are shown in Table 5.

[0205]

Table 16

[0206] Reporter gene assay Results from reporter gene assays indicate that the PD-1 / CTLA-4 bispecific binding protein inhibited both the PD-1 and CTLA-4 pathways (Figures 6A-D). The BiS5 binding protein maintained PD-1 titer compared to the parent, but its titer was approximately 3-fold lower compared to anti-CTLA-4 IgG. The DuetMab antibody showed approximately 9-fold reduction in PD-1 titer and approximately 16-fold reduction in CTLA-4 (compared to IgG4P). In this field, there is a need for molecules that maintain functional activity (e.g., as shown in the SEB assay) but have low CTLA-4 targeting. Thus, the PD-1 / CTLA-4 bispecific binding protein has the potential to provide safety benefits to patients.

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

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

[0209] Pharmacokinetic and pharmacodynamic (PK / PD) studies A study was conducted to investigate the 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 a robust PD response was observed for both molecules (Figures 15A-B). Thus, the viability of PD-1 / CTLA-4 bispecific binding proteins in vivo was confirmed.

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

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

[0212] CHO cells expressing various levels of human PD-1 and / or CTLA-4 To test the PD-1 / CTLA-4 bispecific molecule, a model system was created using stable CHO cells expressing various levels of human PD-1 and / or CTLA-4 (Figure 17). Free 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 revealed that PD-1 / CTLA-4 DuetMab simultaneously binds to PD-1 and CTLA-4 on the surface of the same cell (Figure 18A-C).

[0213] CTLA-4 is constantly taken up within clathrin-coated pits, becoming only a small fraction of the receptor expressed on the cell surface at any given time. Reuse of cell surface CTLA-4 is rapid, with less than 80% of surface CTLA-4 being internalized within 5 minutes. Therefore, experiments were conducted to investigate whether cooperative binding to anti-PD-1 and anti-CTLA-4 antibody combinations recognizes PD-1 / CTLA-4 DuetMab in a CTLA-4 saturated state on cells expressing excessive PD-1 (Figure 19A-C). Receptor occupancy of each target antigen was independently determined using targeted anti-PD-1 and anti-CTLA-4 mAbs.

[0214] The parental monoclonal antibody was found to bind to and occupy target receptors without exerting a measurable effect on non-target receptors (Figures 20A-D). PD-1 / CTLA-4 DuetMab saturated CTLA-4 on CHO cells expressing over-PD-1 at approximately 250-fold lower concentrations compared to monoclonal antibody combinations (Figures 21A-D). PD-1 / CTLA-4 DuetMab saturated CTLA-4 on CHO cells expressing over-PD-1 at approximately 500-fold lower concentrations compared to cells expressing CTLA-4 alone (Figures 22A-F). As determined by quantification of doublet formation within a pre-mixed whole CHO population, PD-1 / CTLA-4 DuetMab preferentially cis-bound to PD-1 and CTLA-4 on the surface of the same cell (Figures 23A-B). However, PD-1 / CTLA-4 DuetMab can also trans-bound to single-expression cells. PD-1 / CTLA-4 DuetMab exhibited the internalization properties of the parental anti-CTLA-4 antibody, tremelimumab (Figures 24A-D). While not bound by any specific theory, the effects of this molecule suggest the possibility of inducing 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).

[0215] Example 2(b) PD-L1 / CTLA-4 bispecific binding protein Using the parental sequences identified above in Table 2, the following bispecific binding proteins that bind to PD-1 and CTLA-4 were constructed. The proteins identified as BiS2, BiS3, and BiS5 were constructed using the sequences shown in Table 6 below, and the simultaneous antigen-binding activity of the identified sequences was evaluated using the Octet binding assay as described above in Section 2(a) (Figure 26).

[0216] [Table 17]

[0217] [Table 18]

[0218] Example 2(c) PD-1 / TIM3 bispecific binding protein Using the parental sequences identified above in Table 2, the following bispecific binding proteins (Table 7) that bind to PD-1 and TIM3 were constructed. The proteins identified as BiS3, BiS5, and DuetMab were constructed using the sequences identified below, and their co-binding assays were evaluated by Octet analysis. Briefly, the biotinylated human TIM3-IgV domain was captured using a biosensor (ForteBio) containing 2 μg / ml streptavidin (SA) in PBS pH 7.2, 3 mg / ml BSA, and 0.05% (v / v) Tween® 20 (assay buffer). After the washing step, the loaded biosensor was subjected to sequential binding and dissociation interactions, first with a sample well containing 200 nM bispecific antibody, and then with a well containing 200 nM PD-1 antigen. After loading the biotinylated human TIM3-IgV domain into a streptavidin sensor, it was subjected to sequential interactions first with a bispecific molecule and then with the PD-1 antigen. The binding results are shown in Figures 27A-27B.

[0219] [Table 19]

[0220] [Table 20]

[0221] [Table 21]

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

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

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

[0225] In addition to clone 62, another bispecific binding protein in DuetMab format, binding to PD-1 and TIM3, was constructed using the parental sequences identified above in Table 2. The PD-1 / TIM3 DuetMab was constructed using the sequences shown in Table 8 below. The TIM3 arm sequence was obtained from O13-1, an affinity-matured mutant of clone 62, and the anti-PD-1 arm sequence was obtained from LO115, which is identical to the PD-1 arm used in the aforementioned PD-1 / CTLA-4 DuetMab bispecific antibody. The PD-1(LO115) / TIM3(O13-1) bispecific antibody was evaluated, including comparisons with PD-1 / TIM BiS3 and BiS5, as discussed below.

[0226] [Table 22]

[0227] [Table 23]

[0228] [Table 24]

[0229] [Table 25]

[0230] Octet binding assay (DuetMab, TIM3 arm affinity mature mutant) Simultaneous binding assays for two separate antigens, PD-1 and TIM3, were performed using the Octet assay. Biotinylated human TIM3 was loaded onto a streptavidin sensor and then sequentially interacted with PD-1 / TIM3 DuetMab first, and then with soluble PD-1 antigen. Biotinylated human TIM3 was captured using a ForteBio biosensor containing 5 μg / ml streptavidin (SA) in PBS pH 7.2, 3 mg / ml BSA, and 0.05% (v / v) Tween® 20 (assay buffer). After the washing step, the loaded biosensor was subjected to sequential association and dissociation interactions, first with a sample well containing 200 nM DuetMab PD-1 / CTLA-4 bispecific antibody, which has an affinity-matured variant of clone 62 TIM antibody (O13-1), and then with a well containing 200 nM human PD-1 antigen. The results of the combination are shown in Figure 29.

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

[0232] [Table 26]

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

[0234] [Table 27]

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

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

[0237] In summary, three bispecificity formats (DuetMab, BiS3, and BiS5) were developed for PD-1 / TIM3. All bispecificity formats exhibited equivalent or superior in vitro functionality to anti-PD-1, suggesting that these molecules may offer significant advantages over existing cancer immunotherapy strategies.

[0238] Example 2(d) OX40 / PD-L1 bispecific binding protein Using the parental sequences identified above in Table 2, the following bispecific binding proteins that bind to PD-1 and OX40 were constructed. The proteins identified as BiS2, BiS3, and BiS5 were constructed using the sequences shown in Table 10 below, and their co-binding was evaluated using the Octet binding assay described later.

[0239] [Table 28]

[0240] [Table 29]

[0241] Octet binding assay To evaluate the binding of the bispecific binding molecules disclosed herein, an Octet QK equipped with a Ni-NTA biosensor chip and 10× kinetic buffer were 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.

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

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

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

[0245] method For a two-cell bioactivity assay for PD-1 inhibition neutralization, PD-L1-expressing CHO scFv OKT3 cells were trypsin-treated, neutralized in warm RPMI complete medium, and collected in 50 mL conical tubes. The cells were pelleted at 380 g for 5 minutes in RT, then suspended in fresh RPMI complete medium and counted on a Vi cell counter. The PD-L1 expressed by the CHO scFv OKT3 cells was adjusted to 0.4e6 / mL, and 25 μL (10,000 cells) per well was cultured on plates as shown in the plate layout. Cells were allowed to adhere to the plates for 3 hours. Then, 50 μL of RPMI containing the test reagent (2 × final concentration) was equally added to the CHO cells and incubated for a further 1 hour. This incubation provides time for the test reagent to bind to PD-L1 on the surface of the CHO cells. After 1 hour, PD-1-expressing Jurkat NFAT luciferase reporter cells were collected in 50 mL conical tubes, pelleted at 380 g for 5 minutes in RT, and resuspended in fresh, warm RPMI complete medium. The cells were adjusted to 1.2 e6 / mL, and 25 μL (30,000) of cells were cultured in wells containing PD-L1-expressing CHO scFv OKT3 cells and test samples.

[0246] Cells and test reagents were incubated for an additional 18 hours to activate PD-1 Jurkat reporter cells. SteadyGlo luciferase reagent was then prepared, and 100 μL was added to each well in equal portions. Complete lysis was achieved by gentle shaking for 15 minutes at RT (200 rpm orbital shaker). After lysis, luciferase activity was measured using the US96 luminescence protocol with an Envision Multilabel Plate Reader. Luciferase RLU was plotted against log [test reagent] using Graphpad Prism software, and PD-L1 antagonism EC was analyzed using nonlinear regression analysis and 4-parameter fitting of S-shaped dose-response curves. 50 The value was determined.

[0247] result: Using a five-point dose titration starting with 100 nM (PD-L1), OX40 / PD-L1 BiS2 / 3 were tested against the PD-L1 / PD-1 parent and NIP228 (G4P) control. All OX40-PD-L1 BisAbs were found to be active and exhibited stronger agonism than the PD-L1(4736) parent (Figure 37). The BiS2 and BiS3 formats functioned similarly.

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

[0249] All of the aforementioned binding and immune response assays provide illustrative data demonstrating that the bispecific binding molecules disclosed herein exhibit specific binding to both target molecules, and in some cases exhibit higher activity than combinations of individual monospecific parent binding molecules (antibodies), thus enabling them to induce or enhance immune responses. Furthermore, they are also shown to possess cytotoxic activity against cancer cell lines. Thus, the data clearly indicate that these molecules and bispecific platform structures are excellent candidates for cancer immunotherapy agents.

[0250] Octet binding assay (OX40(SLR) / PD-L1 BiS5) To evaluate the binding of the bispecific binding molecules disclosed herein, Octet QK equipped with a Ni-NTA biosensor chip and 10× kinetic buffer were 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 revealed that the BiS5 OX40Ab / PD-L1 molecule binds to both PD-L1-His and hOX40-Fc (Figure 39).

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

[0252] PD-L1 and OX40 receptor assay In the PD-L1 receptor assay (using the protocol described above), all PD-L1scFv-containing bispecific molecules and positive control IgG showed activity (Figure 42A-B). Single-arm OX40 control and isotype control did not show activity in this assay. EC 50 The hill slope is consistent with values ​​obtained in previous assays for anti-PD-L1 parental controls and PD-L1 Bis2, Bis3, and Bis5 constructs.

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

[0254] In the OX40 reporter gene assay using CHOK1 PD-L1 on expressing cells, the OX40 / PD-L1 bispecific molecule shows equivalent agonism (Figures 44A - B). OX40 / PD-L1 Bis5 N434A IgG1 had an EC 50 activity equivalent to other Fc mutants of the OX40 / PD-L1 Bis5 bispecific Mab tested. No agonism by OX40 IgG or PD-L1 IgG was detected. Thus, PD-L1-mediated OX40 agonism was demonstrated.

[0255] PD-L1-mediated OX40 agonism with tumor cells was detected using the OX40 / PD-L1 bispecific molecule (Figures 45A - B). The OX40 / PD-L1 bispecific molecule showed an equal agonism - standard curve in this assay. Since no agonism by OX40 IgG was observed, the benefit of using the bispecific molecule over the combination of OX40 IgG and PD-L1 IgG was demonstrated. No agonism was detected with NCI H358 PD-L1 KO cells (Figures 46A - D), which indicates that the NCI H358 agonism observed in the cells is PD-L1 specific.

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

[0257] Treg suppression assay The OX40 / PD-L1 bispecific molecule was tested by performing a Treg suppression assay (Figures 4A - D). The OX40 / PD-L1 bispecific molecule only mediated CD4+T effIt was active above (Figures 49 and 50). Although not bound by any particular theory, this indicated crosslinking of OX40 in trans. The OX40 / PD-L1 bispecific molecule is T reg The inhibitory effect was suppressed, but this was limited to cases where crosslinking occurred through binding to plate-immobilized PD-L1.

[0258] Mixed leukocyte reaction (MLR) assay MLR assays were performed to test the OX40 / PD-L1 bispecific molecule (Figures 51A-B). The OX40 / PD-L1 bispecific molecule exhibited higher activity than combinations of individual antibodies against OX40 and PD-L1 (Figures 52A-E).

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

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

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

[0262] [Table 30]

[0263] The OX40 / PD-L1 bispecific molecule reduced serum soluble PD-L1 concentration to a lower LLOQ than the assay (Figure 60). The N434A mutation improved the pharmacokinetics of Bis5 OX40 / PD-L1-G1. In particular, CL was reduced by almost half; correspondingly, T1 / 2 and AUCinf were doubled; and Cmax and Vss were unaffected. This is consistent with the effects of this mutation on monoclonal antibody PKs reported to date. Thus, progress towards mAb-like PK of Bis5 OX40 / PD-L1-G1 IO BisAb has been achieved. Serum concentrations of Bis5 OX40 / PD-L1-G1 BisAb were below the limit of quantification (BLOQ) at 2 weeks, which is thought to be related to ADA.

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

[0265] Example 2(e). OX40 / PD-1 bispecific binding protein Using the parental sequences identified above in Table 2, the following bispecific binding proteins that bind to PD-1 and OX40 were constructed. The proteins identified as BiS2 and BiS3 were constructed using the sequences shown in Table 24 below, and evaluated for simultaneous antigen binding using the Octet binding assay described later.

[0266] [Table 31]

[0267] The PD-1 / OX40 BiS2 monoclonal antibody (mAb) is a bispecific antibody engineered to simultaneously bind to human and cynomolgus monkey PD-1 and human and cynomolgus monkey OX40 (Figure 62; PD-1 binding proteins are shown in gray, and OX40 binding proteins in light gray). While not bound by any particular theory, the proposed mechanism of action suggests dual signaling to T cells after cis-binding to both OX40 and PD-1, agonism of the T cell co-stimulation surface receptor OX40, and blockade of immunosuppressive PD-1 (Figure 63).

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

[0269] OX40 Reporter Assay PD-1(LO115) / OX40 BiS2 mAb showed comparable activity to other OX40 agonists (Figure 65A-B). Proteins were stored at 4°C, used immediately, frozen / thawed three times, stored at 4°C for 7 days, and then stored at 40°C for 7 days. Activity was recorded as relative luminescence with respect to mAb concentration. Day 0, EC of PD1(LO115) / OX40 BiS2 mAb at 4°C 50 It was approximately 2nM.

[0270] PD-1 / PD-L1 Reporter Assay PD-1(LO115) / OX40 BiS2 mAb showed activity comparable to other PD-1 agonists (Figure 66A-B). Proteins were stored at 4°C, used immediately, frozen / thawed three times, stored at 4°C for 7 days, and then stored at 40°C for 7 days. Activity was recorded as relative luminescence with respect to mAb concentration. On day 0, EC was measured for PD1(LO115) / OX40 BiS2 mAb at 4°C. 50The concentration was approximately 1 nM. Two primary human in vitro efficacy assays were performed: an antigen recall T cell assay and simultaneous T cell stimulation using Staphylococcus enterotoxin B (SEB).

[0271] Staphylococcal Enterotoxin B (SEB) Assay 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 of culture (Figure 67). Therefore, PD-1 / OX40 BiS2 mAb can simultaneously bind to its human target antigen and simultaneously stimulate T cells in vitro.

[0272] In antigen recall assays, PD-1 / OX40 BiS2 mAbs drove increased levels of interferon (IFN)γ compared to parental mAbs and parental mAb combinations (Figures 68 and 69).

[0273] CMV Ag Recall Assay Results from the CMV Ag recall assay (using the protocol described above) showed that the BiS2 and BiS3 molecules did not exhibit equal activity compared to the combination (Figure 70). The data indicate that PD-1 / OX40 BiS2 IgG4P mAb is active both in vitro and in vivo. PD-1 / OX40 BiS3, which has a different structure from PD-1 / OX40 BiS2, was not detected to a detectable level. Therefore, PD-1 / OX40 BiS3 (not active) is different from BiS2 (active).

[0274] Pharmacokinetic and pharmacodynamic (PK / PD) studies Cynomolgus monkeys were considered a pharmacologically appropriate nonclinical 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 non-GLP (Good Laboratory Practices) studies in cynomolgus monkeys. PD-1(LO115) / OX40 BiS2 mAb PK and PD (%) of Ki67-positive CD4+ and CD8+ total memory T cells were evaluated in cynomolgus monkeys (n=3; male) after a single intravenous (IV) administration over a dose range of 0.1 mg / kg to 30 mg / kg. PBMCs were collected before administration and at 1, 8, 11, and 15 days after administration, cryopreserved, thawed, and analyzed by flow cytometry. In summary, PD-1(LO115) / OX40 BiS2 mAb showed a nearly linear pharmacokinetic profile with a short half-life of 0.6–1.7 days (Figure 70; Table 12).

[0275] [Table 32]

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

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

[0278] [Table 33]

[0279] Example 3.1 Further comparisons were made between the BiS format disclosed herein ("BiS5") and another BiS format identified as "BiS4," which includes two binding domains (scFv domains) linked at a hinge region (e.g., between the Fc and Fab regions). BiS4 and BiS5 proteins were expressed in Chinese hamster ovaries (CHO) and purified by conventional chromatography. As noted above, these two formats have similar Fab and scFv sequences, and their main difference is the location of the scFv domain (for BiS4, as described herein, the scFv is located within the hinge region; for this particular BiS5, the scFv is located in C H (Present within the SNG loop in 3 domains). Purified BiS molecules were added to PBS buffer and prepared at 1.54 M -1 cm -1 Using the extinction coefficient, the protein concentration was determined using NanoDrop ND-1000 (Thermo Scientific, Wilmington, Delaware).

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

[0281] Differential Scanning Calorimetry (DSC) Differential scanning calorimetry (DSC) thermograms were acquired for zero-time samples using a VP-Capilary DSC connected to a temperature-controlled autosampler (Malvern Instruments Ltd., Westborough, Massachusetts). A protein concentration of 1 mg / mL was used with a scanning rate of 90°C / hour over a temperature range of 20°C to 100°C to acquire the thermograms. After subtracting buffer from the BiS4 and BiS5 thermograms under different pH conditions ranging from 5.0 to 7.5, baseline correction was performed. Data analysis was performed using a DSC plugin for the Origin 7 SR4 software package. The experimental results were fitted to a multi-state model with three transitions to determine the melting temperature (T m The value was calculated. Heat capacity of the first temperature transition (C pThe value is 500 cal mol. -1 ℃ -1 The point at which the temperature reaches the starting temperature (T onset ) was considered to be the case.

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

[0283] Thermal stability of BiS4 and BiS5 The effect of pH on the thermal stability of BiS4 and BiS5 was evaluated by analyzing thermograms obtained using a capillary-type DSC and prepared under six different pH conditions. Figures 74A and 74B show superimposed DSC thermograms of BiS4 and BiS5 from pH 5.0 to 7.5, respectively. As shown in Figure 73, each thermogram represents the transition temperature T m 1. T m 2, and T mThree thermal denaturation events due to 3 are shown. The first transition (T m 1) is C H This is thought to be related to the simultaneous mutation of the 2 and scFv domains, and the second (T m 2) and the third (T m 3) The transition is C H 3 and F ab It is associated with domain denaturation. For all formats, with increasing pH up to 6.5, T onset , T m 1. T m 2, and T m An increase of 3 was observed (Figures 73A, 73B, 73E and Table 14 below). For BiS4 and BiS5, T was observed under all pH conditions. onset , T m 2, and T m No difference was observed in 3 (Figure 73E and Table 15), but this is because the hinge area or C H The presence of scFv in any of the three domains is C H 3 and F ab This shows that it does not affect the thermal stability. Interestingly, under all pH conditions, T m A slight increase of 1 was observed for BiS5, which means that scFv is C H If located within 3 domains, scFv, C H This indicates an increase in thermal stability of either or both of the two factors.

[0284] [Table 34]

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

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

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

[0288] Further analysis of the physical and chemical stability, monthly monomer loss, fragmentation, and aggregation rates of the BiS4 and BiS5 formats was performed by plotting the above values ​​against six pH conditions (Figures 76A–76C). Across all six pH conditions ranging from pH 5.0 to 7.5, the monomer loss rate was lower for the BiS5 format than for the BiS4 format (Figure 76A), suggesting that the BiS5 format disclosed herein possesses unexpectedly superior physical and chemical stability compared to other bispecific protein formats. In the case of BiS4, the majority of monomer degradation was due to fragmentation at even lower pH conditions (Figure 76B). BiS5 exhibited lower fragmentation rates than BiS4 at all pH conditions tested. Surprisingly, the fragmentation rate of BiS5 appears to be flat over a wide pH range and lower than that of BiS4. While not bound by any particular theory, the relatively low fragmentation rate observed in BiS5 may be attributable to the G4S linker at either end of the scFv that links it to Fc. Fragmentation at one G4S linker linking to Fc would not release scFv because it may still be linked to Fc via the other G4S linker. In BiS4 and BiS5, the aggregation rates appeared to be similar across all pH conditions tested (Figure 76C), which suggests that the position of scFv has minimal effect on aggregation dynamics, and that this is consistent with the T between the two formats at all pH conditions, as measured using capillary DSC. onsetThis is further supported by the fact that no changes were observed (Figure 73E and Table 14 above). At pH 7.5 and 40°C (time = 0), neither molecule showed any obvious fragmentation (Figure 77A), but under the same conditions after storage at 40°C for two weeks, obvious fragmentation was observed for BiS4 and slight fragmentation was observed for BiS5 (Figure 77B). Both BiS4 and BiS5 showed reduced fragmentation and aggregation at low (5.5) pH, but BiS5 exhibited superior performance in both fragmentation and aggregation at all pH values ​​(Figure 78). This series of experiments demonstrates that BiS5 disclosed herein has superior chemical stability compared to BiS4 and similar physical stability to BiS4.

[0289] Example 3.2 Further tests were conducted to evaluate the physical and chemical stability of various embodiments of the bispecific binding proteins disclosed herein and identified as constructs A–H (e.g., Table 13 and related examples above). These constructs were analyzed using DSC, accelerated storage stability, and FcRn and FcgR binding assays, as described below.

[0290] Differential scanning calorimetry analysis DSC experiments for this dataset were performed using a Microcal VP-DSC scanning microcalorimeter (Microcal). All solutions and samples used for DSC were filtered through a 0.22 μm filter and degassed before loading into the calorimeter. Antibodies used for DSC testing were determined to be >98% monomer by analytical SEC. All samples were thoroughly dialyzed in 25 mM histidine-HCl (pH 6.0) before DSC analysis (at least 3 buffer exchanges). The buffer from this dialyze was used as the standard buffer for subsequent DSC experiments. Baseline measurements (buffer vs. buffer) were subtracted from sample measurements before sample measurement. Dialyzed samples (concentration 1 mg / ml) were added to the sample wells, and DSC measurements were performed at a scanning rate of 1°C / min. Data analysis and deconvolution were performed using Origin® DSC software provided by Microcal. Deconvolution analysis was performed using a non-dual-state model, and the best fit was obtained using 100 repeat cycles. onset T is defined as the qualitative temperature at which it becomes clear that the thermogram has a non-zero slope. m This is defined as the temperature at which half of the molecules in a pair are denatured, and it is calculated as the temperature value corresponding to the maximum value of each peak in the thermogram.

[0291] The results for different constructs are shown in Figure 79. In general, constructs A, C, and D, which include 2F4 as scFv, have lower T values ​​compared to constructs E, G, and H, which include LC10 as scFv. M It has a value of 1. Although not bound by any particular theory, the difference in TM1 values ​​is thought to be due to the inherently superior thermal stability of the LC10 scFv domain compared to the 2F4 variable domain. These data suggest that constructs A-D, which have 2F4 as the scFv, will have lower thermal stability than constructs E-H, which have LC10 as the scFv.

[0292] Accelerated storage stability analysis The construct concentration was normalized to 1 mg / mL. 1 mL of each bispecific construct or IgG control was divided equally into 1.5 mL Eppendorf tubes and introduced. Samples were incubated in a static incubator at 45°C for 2 weeks. Samples were analyzed at 3, 7, and 14 days to assess stability. Visual inspection was performed at each time point to record any increase in turbidity or precipitate. Samples were filtered using a 0.2 μm spin column, and 120 μm of sample was divided equally and introduced into HPLC, confirming that there were no bubbles at the bottom of the vial. Next, TSK-GEL G3000SW containing 0.1 M sodium phosphate and 0.1 M sodium sulfate (pH 6.8) as electrophoresis buffer was used. XL Samples were tested using an Agilent 1100 series HPLC-SEC to check for aggregation and degradation using a (300 × 7.8 mm) Tosoh Bioscience column. 60 μL of sample was injected and flowed at a flow rate of 1 mL / min. Monomer retention time (min), total peak area, % monomer, % aggregates, % fragments, and % monomer loss were obtained and used for analytical SEC analysis. The results are summarized in Table 15.

[0293] [Table 35]

[0294] As described herein, the locations of the scFv domains in the aforementioned construct are as follows ("-" indicates scFv): A and E are IS-RTP; B and F are AK-GQP; C and G are S-NG; D and H are SN-G. M The value is associated with the following domain: T M 1 = CH2 / scFv; T M 2=Fab;T M3=CH3. The data show a tendency for constructs A and C, which have 2F4 scFv inserted into the ISRTP(A) and SNG(C) loops, to aggregate more readily than constructs E and G, which have LC10 scfv inserted at the same positions. This observation suggests that the sequence identity and behavior of the scFv domain may influence the stability of bispecific binding protein constructs. Furthermore, from the above, it can be predicted that construct D, which contains 2F4 scFv, will have similarly low stability to A and C, but inserting 2F4 scFv into the SNG loop appears to stabilize the molecule and reduce its tendency to form aggregates. Taken together, this accelerated stability test shows that the scFv sequence and position within the Fc region may play a significant role in the stability of BiSAb constructs.

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

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

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

[0298] [Table 36]

[0299] Compared to other constructs (C, D, G, and H) with scFv inserted into the SNG loop downstream of the CH2-CH3 interface, the difference observed in FcγR binding between constructs (A and E) with scFv inserted into the ISRTP loop upstream of the CH2-CH3 interface is that they consistently exhibit lower FcγR binding.

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

[0301] Example 3.3 In addition to comparing BiS4 with the bispecific binding construct (BiS5) disclosed herein, tests were conducted to evaluate three other BiS structural motif platforms, which were identified as BiS1, BiS2, and BiS3 (see Figure 83). As can be understood by referring to Figure 83, these platforms vary with respect to the position of one binding domain (shown as the scFv domain). Of the five motifs, only BiS4 and BiS5 contain two linker moieties as binding sites to larger proteins, while the others (BiS1, BiS2, and BiS3) are bound by a single linker.

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

[0303] Furthermore, SEC data was used to map various peaks to corresponding fragments of the BiS molecule (Figure 87). The mapping was based on estimations such as the occurrence of fragmentation in the molecular hinge and linker regions, fragment size, theoretical fragmentation, and how expected fragment species are adjusted in relation to the fragment species observed in other formats. While there was good degradation between low molecular weight fragments (LMWFs) in each format, degradation between monomers and high molecular weight fragments (HMWFs) was poor or nonexistent in all formats. Based on the information in Table 17, we concluded that the HP-SEC technique significantly underestimates fragmentation in the BiS format compared to monoclonal antibodies. Alternative analyses were developed as described below.

[0304] [Table 37]

[0305] An alternative analysis was developed to calculate the fragmentation rate of HMWF using a molar extinction coefficient based on the assumption that (i) if small fragments are detected during degradation, corresponding larger fragments should also be present; (ii) no significant secondary fragmentation (fragmentation of fragments) occurred during the stability test period; and (iii) fragmentation occurs in the linker region and / or hinge region. The fragmentation rate was determined based on the following relationship.

number

[0306] [Table 38]

[0307] Furthermore, to determine whether the disulfide bond influenced fragmentation and stability, the fragmentation rate was analyzed using constructs under reducing conditions. Representative data from this assay are shown in Figure 86. Under reducing conditions, higher fragmentation rates were observed for all BiS formats except BiS1 (Table 19). The higher fragmentation rate under reducing conditions was concluded to support the idea that the scFv portion in the BiS5 construct is tethered to the CH3 region (Figure 89).

[0308] [Table 39]

[0309] [Table 40]

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

[0311] [Table 41]

[0312] [Table 42]

[0313] Based on all the data above, the bispecificity binding protein format disclosed herein appears to be the most stable of all the formats tested. Furthermore, BiSAb5 is found to be the most stable in that it minimizes both fragmentation and aggregation at the low pH values ​​tested (e.g., 5.0, 5.5, and 6.0). Thus, the unexpectedly remarkable stability of BisAb disclosed herein provides further advantages over other structural platforms and formats used to create bispecificity binding molecules.

[0314] Reference All publications and patents referenced herein are incorporated herein by reference in whole, as if each individual publication or patent were specifically shown to be incorporated by reference.

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

[0316] [Sequence List] TIFF0007898564000044.tif199141TIFF0007898564000045.tif199141TIFF0007898564000046.tif199141TIFF0007898564000047.tif199141TIFF0007898564000048.tif199141TIFF0007898564000049.tif199141TIFF0007898564000050.tif199141TIFF0007898564000051.tif199141TIFF0007898564000052.tif199141TIFF0007898564000053.tif199141TIFF0007898564000054.tif199141TIFF0007898564000055.tif199141TIFF0007898564000056.tif199141TIFF0007898564000057.tif199141TIFF0007898564000058.tif199141TIFF0007898564000059.tif199141TIFF0007898564000060.tif199141TIFF0007898564000061.tif199141TIFF0007898564000062.tif199141TIFF0007898564000063.tif199141TIFF0007898564000064.tif199141TIFF0007898564000065.tif199141TIFF0007898564000066.tif199141TIFF0007898564000067.tif199141TIFF0007898564000068.tif199141TIFF0007898564000069.tif199141TIFF0007898564000070.tif199141TIFF0007898564000071.tif199141TIFF0007898564000072.tif199141TIFF0007898564000073.tif199141TIFF0007898564000074.tif199141TIFF0007898564000075.tif199141TIFF0007898564000076.tif199141TIFF0007898564000077.tif199141TIFF0007898564000078.tif199141TIFF0007898564000079.tif199141TIFF0007898564000080.tif199141TIFF0007898564000081.tif199141TIFF0007898564000082.tif199141TIFF0007898564000083.tif199141TIFF0007898564000084.tif199141TIFF0007898564000085.tif199141TIFF0007898564000086.tif199141TIFF0007898564000087.tif199141TIFF0007898564000088.tif199141TIFF0007898564000089.tif199141TIFF0007898564000090.tif199141TIFF0007898564000091.tif199141TIFF0007898564000092.tif199141TIFF0007898564000093.tif199141TIFF0007898564000094.tif199141TIFF0007898564000095.tif199141TIFF0007898564000096.tif199141TIFF0007898564000097.tif199141TIFF0007898564000098.tif199141TIFF0007898564000099.tif199141TIFF0007898564000100.tif199141TIFF0007898564000101.tif199141TIFF0007898564000102.tif199141TIFF0007898564000103.tif199141TIFF0007898564000104.tif199141TIFF0007898564000105.tif199141TIFF0007898564000106.tif199141TIFF0007898564000107.tif199141TIFF0007898564000108.tif199141TIFF0007898564000109.tif199141TIFF0007898564000110.tif199141TIFF0007898564000111.tif199141TIFF0007898564000112.tif199141TIFF0007898564000113.tif199141TIFF0007898564000114.tif199141TIFF0007898564000115.tif199141TIFF0007898564000116.tif199141TIFF0007898564000117.tif199141TIFF0007898564000118.tif199141TIFF0007898564000119.tif199141TIFF0007898564000120.tif199141TIFF0007898564000121.tif199141TIFF0007898564000122.tif199141TIFF0007898564000123.tif199141TIFF0007898564000124.tif199141TIFF0007898564000125.tif199141TIFF0007898564000126.tif199141TIFF0007898564000127.tif199141TIFF0007898564000128.tif199141TIFF0007898564000129.tif199141TIFF0007898564000130.tif199141TIFF0007898564000131.tif199141TIFF0007898564000132.tif199141TIFF0007898564000133.tif199141TIFF0007898564000134.tif199141TIFF0007898564000135.tif199141TIFF0007898564000136.tif199141TIFF0007898564000137.tif199141TIFF0007898564000138.tif199141TIFF0007898564000139.tif199141TIFF0007898564000140.tif199141TIFF0007898564000141.tif199141TIFF0007898564000142.tif199141TIFF0007898564000143.tif199141TIFF0007898564000144.tif199141TIFF0007898564000145.tif199141TIFF0007898564000146.tif199141TIFF0007898564000147.tif199141TIFF0007898564000148.tif199141TIFF0007898564000149.tif199141TIFF0007898564000150.tif199141TIFF0007898564000151.tif199141TIFF0007898564000152.tif199141TIFF0007898564000153.tif199141TIFF0007898564000154.tif199141TIFF0007898564000155.tif199141TIFF0007898564000156.tif199141TIFF0007898564000157.tif199141TIFF0007898564000158.tif199141TIFF0007898564000159.tif199141TIFF0007898564000160.tif199141TIFF0007898564000161.tif199141TIFF0007898564000162.tif199141TIFF0007898564000163.tif199141TIFF0007898564000164.tif199141TIFF0007898564000165.tif199141TIFF0007898564000166.tif199141TIFF0007898564000167.tif199141TIFF0007898564000168.tif199141TIFF0007898564000169.tif199141TIFF0007898564000170.tif199141TIFF0007898564000171.tif199141TIFF0007898564000172.tif199141TIFF0007898564000173.tif199141TIFF0007898564000174.tif199141TIFF0007898564000175.tif199141TIFF0007898564000176.tif199141TIFF0007898564000177.tif199141TIFF0007898564000178.tif199141TIFF0007898564000179.tif199141TIFF0007898564000180.tif199141TIFF0007898564000181.tif199141TIFF0007898564000182.tif199141TIFF0007898564000183.tif199141.

Claims

[Claim 1] A bispecific binding protein that binds to OX40 and PD-L1, having a first peptide with the amino acid sequence of SEQ ID NO: 35 and a second peptide with the amino acid sequence of SEQ ID NO: 32.

Images