Bispecific pd-l1xcd28 antibodies and methods of use thereof

Abstract

The present disclosure provides bispecific antigen binding molecules (PD-L1 * CD28) that bind to PD-L1 and CD28. In certain embodiments, the present disclosure provides a bispecific PD-L1 * CD28 antigen binding molecule (or antibody), or antigen binding fragment thereof, comprising a first antigen binding domain that specifically binds CD28 and a second antigen binding domain that specifically binds PD-L1. In certain embodiments, the bispecific antigen-binding molecules of the present disclosure bind CD28 on T cells with a first antigen-binding domain and bind PD-L1 expressed on tumor cells or antigen presenting cells with a second antigen-binding domain. In certain embodiments, the bispecific antigen binding molecules are capable of inhibiting tumor growth. Bispecific antigen binding molecules of the present disclosure are useful for the treatment of cancer.

Description

Technical Field

[0001] This disclosure relates to antibodies that bind to PD-L1 and CD28 and methods of using them, for example, for the treatment or prevention of cancer.

[0002] sequence list

[0003] This application contains a sequence list that has been electronically submitted in XML format and is incorporated herein by reference in its entirety. The XML copy created on May 10, 2024, is named SeqList11374.XML and has a size of 103,464 bytes. This sequence list is part of the specification and is incorporated herein by reference in its entirety. The following sequences are below the minimum length allowed under the sequence list format ST.26: gctgcatcc (SEQ ID NO: 21), AAS (SEQ ID NO: 22), ggggcaagt (SEQ ID NO: 69), and GAS (SEQ ID NO: 70). Background Technology

[0004] The ability of T cells to recognize and kill their cellular targets, such as virus-infected cells or tumor cells, depends on a series of coordinated interactions. The most important of these is the recognition and binding of the T-cell receptor (TCR) complex (which comprises associated CD3 γ, δ, ε, and ζ chains) to target cells, and this interaction is known as “signal 1” for T cell activation. The TCR recognizes viral or tumor peptides presented on the grooves of MHC proteins expressed on the surface of target cells. Because such binding is typically low-affinity, successful triggering of “signal 1” requires the aggregation of numerous TCR complexes along the interface between the T cell and its target cell; this interface is called the “immune synapse.” T cell activation can be further promoted by additional interactions. For example, T cells have a molecule called CD28 on their surface, which can provide a co-stimulatory “signal 2” to enhance activation via the TCR complex. T cell activation is enhanced when T cells recognize their target cells via their TCR complex and subsequently engage “signal 2” by binding to their homologous ligands on the target cells via CD28; like “signal 1”, CD28-mediated “signal 2” is thought to occur through co-aggregation at the immune synapse.

[0005] Agonisttic anti-CD28 (monoclonal antibody) (mAb) can be used for sustained in vitro expansion of cultured T cells; however, the use of anti-CD28 antibodies is discouraged due to a series of acute and serious adverse events in a phase I clinical trial in which a super agonist anti-CD28 mAb was tested systemically (Hünig, Nat. Rev. Immunol. 2012; 12:317–318). Local or targeted use of anti-CD28 mAb can be used to promote anti-tumor immunity with low risk. Jung et al., Int J Cancer. 2001 Jan 15; 91(2):225-30.

[0006] Programmed death ligand 1 (PD-L1), the ligand for programmed death (PD-1 or PD1), is expressed on antigen-presenting cells (such as activated monocytes and dendritic cells) and in some cancer cells. Stimulation with PD-L1 is known to inhibit the activation of PD-1-expressing T lymphocytes (induction of cell proliferation and various cytokine production).

[0007] Monoclonal antibodies (mAbs) designed to enhance T-cell activation are in clinical development as anti-tumor therapeutics. However, most current treatments struggle to overcome the inhibitory nature of the tumor microenvironment, thus failing to generate effective tumor-specific T-cell activation and subsequent tumor cell killing. Several blocking mAbs targeting checkpoint inhibitors such as CTLA-4 (cytotoxic T-lymphocyte-associated protein) and programmed cell death 1 (PD-1) / programmed cell death ligand 1 (PD-L1) have been clinically approved for melanoma, renal cell carcinoma, non-small cell lung cancer, and advanced metastatic squamous cell carcinoma of the skin. Blocking PD-1 release breaks the inhibition of T-cell activation, but its potency as a single agent is often insufficient to achieve tumor clearance and a durable anti-tumor response. Therefore, additional cancer-specific therapies, such as improved immunotherapies, are needed. Summary of the Invention

[0008] This disclosure provides a bispecific antigen-binding molecule that binds to Cluster of Differentiation 28 (CD28) and programmed death-ligand 1 (PD-L1), including a bispecific antibody (“PD-L1xCD28” or “CD28xPD-L1”). The bispecific antigen-binding molecule of this disclosure provides a pan-tumor co-stimulation approach (tumor-agnostic targeting) in which the inhibitory checkpoint PD-L1, present in multiple tumor types, is converted into a co-stimulatory signal that activates CD28 on T cells. The PD-L1xCD28 bispecific antigen-binding molecule blocks the interaction between PD-L1 and PD-1, as well as the interaction between PD-L1 and CD80, thereby protecting CD80 from CTLA4 trogocytosis and making it available for binding to CD28, thereby providing a co-stimulatory signal to T cells. The PD-L1xCD28 bispecific antigen-binding molecule of this disclosure binds and conjugates to PD-L1-expressing tumor cells, APCs, and other tumor-infiltrating immune cells, thus providing a broad range of tumor-unaware targeting compared to bispecific targeting of CD28 and tumor-associated antigens (TAAs). In some embodiments, when used in combination with a CD3-based bispecific agent targeting this TAA, the bispecific antigen-binding molecule enhances T cell-mediated killing of tumor cells expressing both PD-L1 and TAA. The bispecific antigen-binding molecule binds to CD28 with low affinity, which in turn leads to reduced cytokine release. Therefore, combining such CD28-based bispecific agents with PD-1 inhibitory antibodies and / or bispecific TAAxCD3 antibodies can provide safe and effective antitumor therapy against many tumor types.

[0009] In one aspect, this disclosure provides an isolated bispecific antigen-binding molecule comprising a first antigen-binding domain and a second antigen-binding domain, wherein the first antigen-binding domain has a capacitance of less than about 3 × 10⁻⁶. -8 M of K D Specifically binds to human CD28, as measured by surface plasmon resonance at 25°C; the second antigen-binding domain has a binding frequency of less than approximately 2 × 10⁻⁶. -10 M of K D Specific binding to programmed death ligand 1 (PD-L1), as measured by surface plasmon resonance at 25 °C.

[0010] In some implementations, the bispecific antigen-binding molecule has a capacitance of less than about 2 × 10⁻⁶. -8 M's EC 50It binds to the surface of human T cells, as detected by in vitro FACS binding assays. In some embodiments, this bispecific antigen-binding molecule binds at a concentration of less than about 8 × 10⁻⁶. -8 M's EC 50 It binds to the surface of human T cells, as measured by in vitro FACS binding assays.

[0011] In some implementations, the bispecific antigen-binding molecule is at a concentration of less than about 3 × 10⁻⁶. -9 M's EC 50 It binds to the surface of cells expressing PD-L1, as measured by in vitro FACS binding assays.

[0012] In some embodiments, the bispecific antigen-binding molecule has an IC50 of less than about 1.3 nM. 50 Blocking the binding of PD-L1 to PD-1, as measured by an ELISA-based blocking assay.

[0013] In some embodiments, the isolated bispecific antigen-binding molecule exhibits a co-stimulatory effect when used in combination with an anti-mucin 16 (MUC16) x CD3 bispecific antibody and tested on tumor cells expressing PD-L1 and MUC16. In one embodiment, the co-stimulatory effect is manifested by one or more of the following: (a) the ability of human T cells to activate and direct the killing of PD-L1-expressing target cells; (b) the ability to upregulate PD-1 on T cells; (c) the ability to increase the release of cytokines IFNγ and TNF from PBMCs; (d) the ability to deplete tumor cells; or (f) the ability to enhance tumor clearance. In another embodiment, the co-stimulatory effect is also manifested by one or more of the following: (g) activation of NFκB activity in a T cell / APC luciferase-based reporting assay; or (h) the use of primary CD4 + T cell / APC function assay is a measure of IL-2 cytokine production.

[0014] In some embodiments, the bispecific antigen-binding molecule is combined with the bispecific MUC16xCD3 antibody at a concentration of less than about 10. -10 M's EC 50 Mediates in vitro T cell killing of OVCAR-3 cells expressing PD-L1.

[0015] For example, in one embodiment, a bispecific antigen-binding molecule (e.g., an antibody or its antigen-binding fragment) that binds PD-L1 and CD28 comprises: (1) a PD-L1 binding arm comprising: (a) its heavy chain variable region comprising an amino acid sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% amino acid sequence identity with the amino acid sequences shown in SEQ ID NO: 2 and 42, comprising the amino acid sequences shown in SEQ ID NO: 2 and 42, or consisting of the amino acid sequences shown in SEQ ID NO: 2 and 42, or variants thereof; and / or (b) its light chain variable region comprising an amino acid sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% amino acid sequence identity with the amino acid sequences shown in SEQ ID NO: 18 and 58, comprising the amino acid sequences shown in SEQ ID NO: 18 and 58, or consisting of the amino acid sequences shown in SEQ ID NO: 18 and 58, or variants thereof; and / or (b) its light chain variable region comprising an amino acid sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% amino acid sequence identity with the amino acid sequences shown in SEQ ID NO: 18 and 58, comprising the amino acid sequences shown in SEQ ID NO: 18 and 58, or consisting of the amino acid sequences shown in SEQ ID NO: 18 and 58, or variants thereof; and / or (b) its light chain variable region. The amino acid sequences shown in SEQ ID NO: 18 and 58, or variants thereof; and / or (2) comprising a CD28 binding arm comprising: (c) a heavy chain immunoglobulin or its variable region comprising an amino acid sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% amino acid sequence identity with the amino acid sequences shown in SEQ ID NO: 10, 32 and 50, comprising the amino acid sequences shown in SEQ ID NO: 10, 32 and 50, or comprising the amino acid sequences shown in SEQ ID NO: 10, 32 and 50, or variants thereof; and / or (d) a light chain immunoglobulin or its variable region comprising an amino acid sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% amino acid sequence identity with the amino acid sequences shown in SEQ ID NO: 18 and 58, comprising the amino acid sequences shown in SEQ ID NO: 18 and 58, or ... The amino acid sequences shown in 18 and 58 are composed of [the amino acid sequence].

[0016] In one embodiment, a bispecific antigen-binding protein (e.g., an antibody or its antigen-binding fragment) that binds to PD-L1 and CD28 comprises: (1) a PD-L1 binding arm and a CD28 binding arm, the PD-L1 binding arm comprising: (a) a heavy chain immunoglobulin or its variable region comprising HCDR1, HCDR2, and HCDR3 of the heavy chain variable region, the heavy chain variable region comprising the amino acid sequence shown in SEQ ID NO: 2 or 42 and having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% amino acid sequence identity with the amino acid sequence shown in SEQ ID NO: 2 or 42, respectively, comprising or consisting of the amino acid sequence shown in SEQ ID NO: 2 or 42; and / or (b) a light chain immunoglobulin or its variable region comprising LCDR1, LCDR2, and LCDR3 of the light chain variable region, the light chain variable region comprising the amino acid sequence shown in SEQ ID NO: 18 or 58 ..., and having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% amino acid sequence identity with the amino acid sequence shown in SEQ ID NO: 2 or 42, respectively, comprising or The amino acid sequence shown in SEQ ID NO: 18 or 58 has at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% amino acid sequence identity, comprising or consisting of the amino acid sequence shown in SEQ ID NO: 18 or 58; or (2) a CD28 binding arm and a PD-L1 binding arm, the CD28 binding arm comprising: (c) a heavy chain immunoglobulin or its variable region, comprising HCDR1, HCDR2, and HCDR3 of the heavy chain variable region, the heavy chain variable region comprising the amino acid sequence shown in SEQ ID NO: 10, 32, or 50, and having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% amino acid sequence identity with the amino acid sequence shown in SEQ ID NO: 10, 32, or 50, respectively, comprising or consisting of the amino acid sequence shown in SEQ ID NO: 10, 32, or 50. Composed of the amino acid sequence shown in SEQ ID NO: 10, 32 or 50; and / or (d) a light chain immunoglobulin or its variable region comprising the light chain variable regions LCDR1, LCDR2 and LCDR3, wherein the light chain variable regions comprise the amino acid sequence shown in SEQ ID NO: 18 or 58 and have at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% amino acid sequence identity with the amino acid sequence shown in SEQ ID NO: 18 or 58, respectively, comprising or composed of the amino acid sequence shown in SEQ ID NO: 18 or 58.

[0017] In some embodiments, the first antigen-binding domain comprises: (a) three heavy chain complementarity-determining regions (HCDR1, HCDR2, and HCDR3) contained within a heavy chain variable region (HCVR), the heavy chain variable region (HCVR) containing amino acid sequences selected from SEQ ID NO: 50, 32, and 10 or variants thereof; and (b) three light chain complementarity-determining regions (LCDR1, LCDR2, and LCDR3) contained within a light chain variable region (LCVR), the light chain variable region (LCVR) containing amino acid sequences selected from SEQ ID NO: 58 and 18 or variants thereof.

[0018] In some embodiments, the isolated bispecific antigen-binding molecule comprises: HCDR1, which contains an amino acid sequence selected from SEQ ID NO: 52, 34 and 12; HCDR2, which contains an amino acid sequence selected from SEQ ID NO: 54, 36 and 14; and HCDR3, which contains an amino acid sequence selected from SEQ ID NO: 56, 38 and 16.

[0019] In some embodiments, the isolated bispecific antigen-binding molecule comprises: LCDR1, which contains an amino acid sequence selected from SEQ ID NO: 60 and 20; LCDR2, which contains an amino acid sequence selected from SEQ ID NO: 62 and 22; and LCDR3, which contains an amino acid sequence selected from SEQ ID NO: 64 and 24.

[0020] In some embodiments, the first antigen-binding domain comprises HCVR and LCVR, wherein the HCVR contains the amino acid sequence of SEQ ID NO: 50 or a variant thereof, and the LCVR contains the amino acid sequence of SEQ ID NO: 58 or a variant thereof.

[0021] In some embodiments, the first antigen-binding domain comprises HCVR and LCVR, wherein the HCVR contains the amino acid sequence of SEQ ID NO: 32 or a variant thereof, and the LCVR contains the amino acid sequence of SEQ ID NO: 18 or a variant thereof.

[0022] In some embodiments, the first antigen-binding domain comprises HCVR and LCVR, wherein the HCVR contains the amino acid sequence of SEQ ID NO: 10 or a variant thereof, and the LCVR contains the amino acid sequence of SEQ ID NO: 18 or a variant thereof.

[0023] In some embodiments, the second antigen-binding domain comprises: (a) three heavy chain complementarity-determining regions (HCDR1, HCDR2, and HCDR3) contained within a heavy chain variable region (HCVR), the heavy chain variable region (HCVR) containing an amino acid sequence selected from SEQ ID NO: 42 and 2 or a variant thereof; and (b) three light chain complementarity-determining regions (LCDR1, LCDR2, and LCDR3) contained within a light chain variable region (LCVR), the light chain variable region (LCVR) containing an amino acid sequence selected from SEQ ID NO: 58 and 18 or a variant thereof.

[0024] In some embodiments, the second antigen-binding domain comprises: (a) HCDR1, which comprises the amino acid sequence of SEQ ID NO: 44 or SEQ ID NO: 4; (b) HCDR2, which comprises the amino acid sequence of SEQ ID NO: 46 or SEQ ID NO: 6; and (c) HCDR3, which comprises the amino acid sequence of SEQ ID NO: 48 or SEQ ID NO: 8.

[0025] In some embodiments, the second antigen-binding domain comprises: LCDR1 containing an amino acid sequence selected from SEQ ID NO: 60 and 20; LCDR2 containing an amino acid sequence selected from SEQ ID NO: 62 and 22; and LCDR3 containing an amino acid sequence selected from SEQ ID NO: 64 and 24.

[0026] In some embodiments, the second antigen-binding domain comprises: (a) HCDR1, HCDR2, and HCDR3 domains containing the amino acid sequences of SEQ ID NO: 44, 46, and 48, respectively; and LCDR1, LCDR2, and LCDR3 domains containing the amino acid sequences of SEQ ID NO: 60, 62, and 64, respectively; or (b) HCDR1, HCDR2, and HCDR3 domains containing the amino acid sequences of SEQ ID NO: 4, 6, and 8, respectively; and LCDR1, LCDR2, and LCDR3 domains containing the amino acid sequences of SEQ ID NO: 20, 22, and 24, respectively.

[0027] In some embodiments, the second antigen-binding domain comprises: (a) an HCVR containing the amino acid sequence of SEQ ID NO: 42 or a variant thereof, and an LCVR containing the amino acid sequence of SEQ ID NO: 58 or a variant thereof; or (b) an HCVR containing the amino acid sequence of SEQ ID NO: 2 or a variant thereof, and an LCVR containing the amino acid sequence of SEQ ID NO: 18 or a variant thereof.

[0028] In some aspects, this disclosure provides isolated bispecific antigen-binding molecules comprising: (a) a first antigen-binding domain that specifically binds to human CD28, wherein the first antigen-binding domain comprises: HCDR1, HCDR2, HCDR3 domains containing the amino acid sequences of SEQ ID NO: 52, 54, 56, respectively, and LCDR1, LCDR2, LCDR3 domains containing the amino acid sequences of SEQ ID NO: 60, 62, 64, respectively; and (b) a second antigen-binding domain that specifically binds to human PD-L1, wherein the second antigen-binding domain comprises: HCDR1, HCDR2, HCDR3 domains containing the amino acid sequences of SEQ ID NO: 44, 46, 48, respectively, and LCDR1, LCDR2, LCDR3 domains containing the amino acid sequences of SEQ ID NO: 60, 62, 64, respectively.

[0029] In some aspects, this disclosure provides isolated bispecific antigen-binding molecules comprising: (a) a first antigen-binding domain that specifically binds to human CD28, wherein the first antigen-binding domain comprises: HCDR1, HCDR2, HCDR3 domains containing the amino acid sequences of SEQ ID NO: 34, 36, 38, respectively, and LCDR1, LCDR2, LCDR3 domains containing the amino acid sequences of SEQ ID NO: 20, 22, 24, respectively; and (b) a second antigen-binding domain that specifically binds to human PD-L1, wherein the second antigen-binding domain comprises: HCDR1, HCDR2, HCDR3 domains containing the amino acid sequences of SEQ ID NO: 4, 6, 8, respectively, and LCDR1, LCDR2, LCDR3 domains containing the amino acid sequences of SEQ ID NO: 20, 22, 24, respectively.

[0030] In some aspects, this disclosure provides isolated bispecific antigen-binding molecules comprising: (a) a first antigen-binding domain that specifically binds to human CD28, wherein the first antigen-binding domain comprises: HCDR1, HCDR2, HCDR3 domains containing the amino acid sequences of SEQ ID NO: 12, 14, 16, respectively, and LCDR1, LCDR2, LCDR3 domains containing the amino acid sequences of SEQ ID NO: 20, 22, 24, respectively; and (b) a second antigen-binding domain that specifically binds to human PD-L1, wherein the second antigen-binding domain comprises: HCDR1, HCDR2, HCDR3 domains containing the amino acid sequences of SEQ ID NO: 4, 6, 8, respectively, and LCDR1, LCDR2, LCDR3 domains containing the amino acid sequences of SEQ ID NO: 20, 22, 24, respectively.

[0031] In some embodiments, the isolated bispecific antigen-binding molecule comprises: (a) a first antigen-binding domain comprising an HCVR containing the amino acid sequence of SEQ ID NO: 50 and an LCVR containing the amino acid sequence of SEQ ID NO: 58; and (b) a second antigen-binding domain comprising an HCVR containing the amino acid sequence of SEQ ID NO: 42 and an LCVR containing the amino acid sequence of SEQ ID NO: 58.

[0032] In some embodiments, the isolated bispecific antigen-binding molecule comprises: (a) a first antigen-binding domain comprising an HCVR containing the amino acid sequence of SEQ ID NO: 32 and an LCVR containing the amino acid sequence of SEQ ID NO: 18; and (b) a second antigen-binding domain comprising an HCVR containing the amino acid sequence of SEQ ID NO: 2 and an LCVR containing the amino acid sequence of SEQ ID NO: 18.

[0033] In some embodiments, the isolated bispecific antigen-binding molecule comprises: (a) a first antigen-binding domain comprising an HCVR containing the amino acid sequence of SEQ ID NO: 10 and an LCVR containing the amino acid sequence of SEQ ID NO: 18; and (b) a second antigen-binding domain comprising an HCVR containing the amino acid sequence of SEQ ID NO: 2 and an LCVR containing the amino acid sequence of SEQ ID NO: 18.

[0034] In some aspects, this disclosure provides isolated bispecific antigen-binding molecules that compete with a reference antibody for binding to PD-L1 or to the same epitope on PD-L1, wherein the reference antibody comprises a first antigen-binding domain and a second antigen-binding domain, the first antigen-binding domain comprising an HCVR / LCVR pair containing an amino acid sequence selected from SEQ ID NO: 50 / 58, 32 / 18, and 10 / 18, and the second antigen-binding domain comprising an HCVR / LCVR pair containing an amino acid sequence selected from SEQ ID NO: 42 / 58 and 2 / 18.

[0035] In some aspects, this disclosure provides isolated bispecific antigen-binding molecules that compete with a reference antibody for binding to human CD28 or the same epitope on human CD28, wherein the reference antibody comprises a first antigen-binding domain and a second antigen-binding domain, the first antigen-binding domain comprising an HCVR / LCVR pair containing an amino acid sequence selected from SEQ ID NO: 50 / 58, 32 / 18, and 10 / 18, and the second antigen-binding domain comprising an HCVR / LCVR pair containing an amino acid sequence selected from SEQ ID NO: 42 / 58 and 2 / 18.

[0036] In some implementations, the isolated bispecific antigen-binding molecule is a human bispecific antigen-binding molecule.

[0037] In some implementations, the isolated bispecific antigen-binding molecule is a bispecific antibody.

[0038] In some embodiments, the antibody comprises a human IgG heavy chain constant region linked to an HCVR of each of a first antigen-binding domain and a second antigen-binding domain. In some embodiments, the heavy chain constant region is allotype IgG1. In some embodiments, the heavy chain constant region is allotype IgG4.

[0039] In some embodiments, the heavy chain constant region linked to the HCVR of the first antigen-binding domain or the heavy chain constant region linked to the HCVR of the second antigen-binding domain contains amino acid modifications that reduce protein A binding, but not simultaneously, wherein the reduction in protein A binding is relative to the same isotype heavy chain without the modification.

[0040] In some implementations, the H435R substitution (EU number) in the heavy chain containing isotype IgG1 or IgG4 is modified.

[0041] In some implementations, the modification includes H435R substitution and Y436F substitution (EU number) in the isotype IgG1 or IgG4 heavy chain.

[0042] In some implementations, the bispecific antibody comprises a chimeric hinge that reduces the binding of the Fcγ receptor relative to the same isotype wild-type hinge.

[0043] In some embodiments, the antibody comprises a first heavy chain of HCVR containing a first antigen-binding domain and a second heavy chain of HCVR containing a second antigen-binding domain, wherein the first heavy chain comprises an amino acid sequence selected from SEQ ID NO: 68, 40 and 28; and the second heavy chain comprises an amino acid sequence selected from SEQ ID NO: 66 and 26.

[0044] In some embodiments, the antibody comprises a common light chain of LCVR containing a first antigen-binding domain and a second antigen-binding domain, wherein the common light chain comprises an amino acid sequence selected from SEQ ID NO: 70 and 30.

[0045] In some embodiments, the antibody comprises a first heavy chain of HCVR containing a first antigen-binding domain and a second heavy chain of HCVR containing a second antigen-binding domain, wherein the first heavy chain comprises the amino acid sequence of SEQ ID NO: 68 and the second heavy chain comprises the amino acid sequence of SEQ ID NO: 66.

[0046] In some embodiments, the antibody comprises a common light chain of LCVR containing a first antigen-binding domain and a second antigen-binding domain, wherein the common light chain comprises the amino acid sequence of SEQ ID NO: 70.

[0047] In some aspects, this disclosure provides a bispecific antibody comprising a first antigen-binding domain that specifically binds to human CD28 and a second antigen-binding domain that specifically binds to human PD-L1, wherein the bispecific antibody comprises a first heavy chain containing the amino acid sequence of SEQ ID NO: 68, which is paired with a common light chain containing the amino acid sequence of SEQ ID NO: 70; and a second heavy chain containing the amino acid sequence of SEQ ID NO: 66, which is paired with a common light chain containing the amino acid sequence of SEQ ID NO: 70.

[0048] In some aspects, this disclosure provides a bispecific antibody comprising a first antigen-binding domain that specifically binds to human CD28 and a second antigen-binding domain that specifically binds to human PD-L1, wherein the bispecific antibody comprises a first heavy chain containing the amino acid sequence of SEQ ID NO: 40, which is paired with a common light chain containing the amino acid sequence of SEQ ID NO: 30; and a second heavy chain containing the amino acid sequence of SEQ ID NO: 26, which is paired with a common light chain containing the amino acid sequence of SEQ ID NO: 30.

[0049] In some aspects, this disclosure provides a bispecific antibody comprising a first antigen-binding domain that specifically binds to human CD28 and a second antigen-binding domain that specifically binds to human PD-L1, wherein the bispecific antibody comprises a first heavy chain containing the amino acid sequence of SEQ ID NO: 28, which is paired with a common light chain containing the amino acid sequence of SEQ ID NO: 30; and a second heavy chain containing the amino acid sequence of SEQ ID NO: 26, which is paired with a common light chain containing the amino acid sequence of SEQ ID NO: 30.

[0050] In some implementations, the bispecific antibody is a human antibody.

[0051] In some aspects, this disclosure provides pharmaceutical compositions comprising a bispecific antigen-binding molecule disclosed herein and a pharmaceutically acceptable carrier or diluent. In other aspects, this disclosure provides pharmaceutical compositions comprising a bispecific antibody disclosed herein and a pharmaceutically acceptable carrier or diluent.

[0052] This disclosure also provides a method for preparing the bispecific antigen-binding protein described herein, comprising: (a) introducing one or more nucleic acid molecules containing a nucleic acid sequence encoding an immunoglobulin chain encoding the bispecific antigen-binding protein into a host cell (e.g., CHO cells); (b) culturing the host cell under conditions favorable to the expression of the nucleic acid molecule; and (c) optionally, isolating the antigen-binding protein or immunoglobulin chain from the host cell and / or the culture medium in which the host cell is cultured. Any antigen-binding protein or immunoglobulin chain that is a product of such a method is part of this disclosure.

[0053] In some implementations, the host cell is a Chinese hamster ovary (CHO) cell.

[0054] In some embodiments, the method further includes formulating the antigen-binding molecule into a pharmaceutical composition comprising an acceptable carrier.

[0055] Antigen-binding molecules or immunoglobulin chains that are products of the methods described herein are also part of this disclosure.

[0056] This disclosure also provides nucleic acid molecules comprising nucleotide sequences encoding the bispecific antigen-binding molecule described herein; or sets of nucleic acid molecules comprising nucleotide sequences encoding the following: an HCVR encoding a first antigen-binding domain specifically binding to human CD28, an HCVR encoding a second antigen-binding domain specifically binding to human PD-L1, and an LCVR encoding the isolated bispecific antigen-binding molecule described herein. Expression vectors containing nucleic acid molecules of this disclosure, or sets of expression vectors containing sets of nucleic acid molecules of this disclosure, and host cells (e.g., CHO cells) containing nucleic acid molecules, vectors, or antigen-binding proteins of this disclosure are also part of this disclosure.

[0057] This disclosure provides a method for generating a bispecific antigen-binding molecule that binds to PD-L1 and CD28, comprising: (a) culturing a host cell as described herein under conditions favorable for generating the bispecific antigen-binding molecule; and (b) optionally, isolating the antigen-binding molecule or immunoglobulin chain from the host cell and / or the culture medium in which the host cell is cultured. In some embodiments, the host cell is a CHO cell. In some embodiments, the method further includes formulating the antigen-binding molecule into a pharmaceutical composition comprising an acceptable carrier.

[0058] Antigen-binding molecules or immunoglobulin chains that are products of the methods disclosed herein are also part of this disclosure.

[0059] This disclosure also provides nucleic acid molecules comprising nucleotide sequences encoding bispecific antibodies as described herein; or sets of nucleic acid molecules comprising nucleotide sequences encoding a heavy chain of a first antigen-binding domain specifically binding to human CD28, a heavy chain of a second antigen-binding domain specifically binding to human PD-L1, and a light chain of a bispecific antibody as described herein. Expression vectors of nucleic acids containing this disclosure, as well as host cells (e.g., CHO cells) containing nucleic acid molecules, vectors, or antibodies of this disclosure, are also part of this disclosure.

[0060] This disclosure provides a method for generating bispecific antibodies that bind to PD-L1 and CD28, comprising: (a) culturing host cells as described herein under conditions favorable for generating the bispecific antigen-binding molecule; and (b) optionally, isolating the antigen-binding molecule or immunoglobulin chain from the host cells and / or the culture medium in which the host cells are cultured. In some embodiments, the host cells are CHO cells. In some embodiments, the method further includes formulating the bispecific antibody into a pharmaceutical composition comprising an acceptable carrier.

[0061] Antibodies that are products of the methods disclosed herein are also part of this disclosure.

[0062] This disclosure also provides methods for treating hyperproliferative diseases (e.g., cancer) in subjects (e.g., humans) in which such treatment is desired, comprising administering (e.g., subcutaneously, intravenously, or intramuscularly) an effective amount of a bispecific antigen-binding protein or composition or formulation. In one embodiment of this disclosure, the cancer is B-cell carcinoma, basal cell carcinoma, urothelial carcinoma of the bladder, brain cancer, breast cancer, cervical cancer, cervical squamous cell carcinoma, colorectal cancer, diffuse large B-cell lymphoma, endometrial adenocarcinoma, endometrial cancer, esophageal cancer, gastroesophageal adenocarcinoma, gastroesophageal cancer, glioblastoma multiforme, head and neck squamous cell carcinoma, hepatocellular carcinoma, lung adenocarcinoma, lung cancer, lung squamous cell carcinoma, melanoma, multiple myeloma, non-small cell lung cancer, ovarian cancer, pancreatic cancer, prostate cancer, rectal cancer, skin cancer, or T-cell cancer. In one embodiment, the cancer comprises cancer cells expressing PD-L1.

[0063] In some aspects, this disclosure provides methods for inhibiting tumor growth in a subject, including administering to the subject an isolated bispecific antigen-binding molecule, or a bispecific antibody, or a pharmaceutical composition as described herein.

[0064] In some implementation schemes, the tumor is B-cell carcinoma, basal cell carcinoma, bladder urothelial carcinoma, brain cancer, breast cancer, cervical cancer, cervical squamous cell carcinoma, colorectal cancer, diffuse large B-cell lymphoma, endometrial adenocarcinoma, endometrial cancer, esophageal cancer, gastroesophageal adenocarcinoma, gastroesophageal cancer, glioblastoma multiforme, head and neck squamous cell carcinoma, hepatocellular carcinoma, lung adenocarcinoma, lung cancer, lung squamous cell carcinoma, melanoma, multiple myeloma, non-small cell lung cancer, ovarian cancer, pancreatic cancer, prostate cancer, rectal cancer, skin cancer, or T-cell carcinoma.

[0065] In some implementations, the tumor expresses PD-L1.

[0066] In some implementations, the method also includes administering a second therapeutic agent or treatment regimen.

[0067] In some implementations, the second therapeutic agent or treatment regimen includes chemotherapy drugs, DNA alkylating agents, immunomodulators, proteasome inhibitors, histone deacetylase inhibitors, radiotherapy, surgery, stem cell transplantation, bispecific antibodies that interact with tumor-associated antigens (TAAs) and T-cell or immune cell antigens, antibody-drug conjugates, oncolytic viruses, bispecific antibodies conjugated to antitumor agents, VEGF inhibitors, checkpoint inhibitors, GITR agonists, CD27 agonists, 4-1BB activators, PD-1 inhibitors, CTLA-4 inhibitors, EGFR inhibitors, Ang2 inhibitors, MUC16 inhibitors, cancer vaccines, cytokines, modified IL2, modified IL12, IL4 inhibitors, IL6 inhibitors, corticosteroids, or combinations thereof.

[0068] In some implementations, the T cell or immune cell antigen is CD3.

[0069] In some implementations, TAA is selected from AFP, ALK, BAGE protein, BCMA, BIRC5 (survival protein), BIRC7, β-linkin, brc-abl, BRCA1, BORIS, CA9, carbonic anhydrase IX, caspase-8, CALR, CCR5, CD19, CD20 (MS4A1), CD22, CD30, CD40, CDK4, CEA, CTLA4, cyclin-B1, CYP1B1, EGFR, EGFRvIII, ErbB2 / Her2, ErbB3, ErbB4, ETV6-AML, EpCAM, EphA2, Fra-1, FOLR1, GAGE ​​protein (e.g., GAGE-1, -2), GD2, GD3, GloboH, phosphatidylinositol proteoglycan-3, GM3, gp100, Her2, HLA / B-raf, HLA / k-ras, HLA / MAGE-A3, hT ERT, LMP2, MAGE proteins (e.g., MAGE-1, -2, -3, -4, -6, and -12), MART-1, mesothelin, ML-IAP, Muc1, Muc2, Muc3, Muc4, Muc5, Muc16 (CA-125), MUM1, NA17, NY-BR1, NY-BR62, NY-BR85, NY-ESO1, OX40, p15, p53, PAP, PAX3, PAX5 PCTA-1, PLAC1, PRLR, PRAME, PSMA (FOLH1), RAGE protein, Ras, RGS5, Rho, SART-1, SART-3, STEAP1, STEAP2, TAG-72, TGF-β, TMPRSS2, Thompson-nouvelle antigen (Tn), TRP-1, TRP-2, tyrosinase, and uroplakin-3.

[0070] In other respects, this disclosure provides the use of bispecific antigen-binding molecules, or bispecific antibodies, or pharmaceutical compositions as described herein in the treatment of tumors.

[0071] In some implementation schemes, the tumor is B-cell carcinoma, basal cell carcinoma, bladder urothelial carcinoma, brain cancer, breast cancer, cervical cancer, cervical squamous cell carcinoma, colon cancer, colorectal cancer, diffuse large B-cell lymphoma, endometrial adenocarcinoma, endometrial cancer, esophageal cancer, gastroesophageal adenocarcinoma, gastroesophageal cancer, glioblastoma multiforme, head and neck squamous cell carcinoma, hepatocellular carcinoma, melanoma, multiple myeloma, leukemia, lung adenocarcinoma, lung cancer, lung squamous cell carcinoma, non-small cell lung cancer, ovarian cancer, pancreatic cancer, prostate cancer, rectal cancer, skin cancer, or T-cell carcinoma.

[0072] In some implementations, the tumor expresses PD-L1.

[0073] In some embodiments, the antigen-binding molecule or pharmaceutical composition is used in combination with a second therapeutic agent or treatment regimen, the second therapeutic agent or treatment regimen including chemotherapeutic agents, DNA alkylating agents, immunomodulators, proteasome inhibitors, histone deacetylase inhibitors, radiotherapy, surgery, stem cell transplantation, bispecific antibodies that interact with tumor-associated antigens (TAAs) and T-cell or immune cell antigens, antibody-drug conjugates, oncolytic viruses, bispecific antibodies conjugated to antitumor agents, VEGF inhibitors, checkpoint inhibitors, GITR agonists, CD27 agonists, 4-1BB activators, PD-1 inhibitors, CTLA-4 inhibitors, EGFR inhibitors, Ang2 inhibitors, MUC16 inhibitors, cancer vaccines, cytokines, modified IL2, modified IL12, IL4 inhibitors, IL6 inhibitors, corticosteroids, or combinations thereof. Attached Figure Description

[0074] Figures 1A to 1B This relates to Example 7. Figure 1A This is a graph showing the average tumor volume in mice that have been given the specified antibody. Figure 1B This is a graph showing the survival probability of mice administered the specified antibody.

[0075] Figure 2 This relates to Example 8, and is a graph showing the average tumor volume in mice administered the specified antibody.

[0076] Figures 3A to 3E Referring to Example 8, this is a graph showing tumor volume in individual mice administered the specific antibody. The percentage of tumor-free (TF) mice is shown. Figure 3A The tumor volume in individual mice treated with the same type of control antibody is shown. Figure 3B Tumor volumes in individual mice, batch 2, treated with REGN6192 are shown. Figure 3C Tumor volumes in individual mice, batch 2, treated with REGN6193 are shown. Figure 3D Tumor volume is shown in individual mice that were administered REGN6194. Figure 3E Tumor volumes in individual mice treated with REGN6194 and cimiprimab are shown.

[0077] Figure 4 This relates to Example 8, and is a graph showing the survival probability of mice administered the specified antibody.

[0078] Figures 5A to 5B Example 9 is involved. Figure 5A This is a graph showing the average tumor volume in mice that have been given the specified antibody. Figure 5B This is a graph showing the survival probability of mice administered the specified antibody.

[0079] Figures 6A to 6C This relates to Example 10 and is a graph showing the concentration of a specific cytokine in the blood of mice administered a specified antibody. Figure 6A The concentration of IL-2 is shown. Figure 6B The concentration of IL-5 is shown. Figure 6C The concentration of IL-4 is shown. TGN1412 is an anti-CD28 superagonist. Suntharalingam, N. Engl. J. Med., 355(10):1018-1028 (2006).

[0080] Figure 7 This relates to Example 11 and is a graph showing the average tumor volume in mice administered the specified antibody.

[0081] Figures 8A to 8E This relates to Example 12. Figures 8A to 8D This demonstrates the administration of a specified antibody and transplantation of parental M38 cells with a specific ratio of human PD-L1. + A graph showing the average tumor volume in mice with M38 tumor cells. Figure 8A The average tumor volume of mice transplanted at a ratio of 0:100 is shown. Figure 8B The average tumor volume of mice transplanted at a 50:50 ratio is shown. Figure 8C The average tumor volume of mice transplanted at a ratio of 90:10 is shown. Figure 8D The average tumor volume of mice transplanted at a ratio of 99:1 is shown. Figure 8E This demonstrates the administration of a specified antibody and transplantation of parental M38 cells with a specified ratio of human PD-L1. + A graph showing the survival probability of mice with M38 tumor cells.

[0082] Figures 9A to 9E Example 13 is relevant. Figure 9AThis is a graph showing the average tumor volume in mice treated with the specified antibody. Figures 9B to 9E This is a graph showing tumor volume in individual mice treated with specific antibodies. The percentage of mice without tumors (TF) is also shown. Figure 9B The tumor volume in individual mice treated with the same type of control antibody is shown. Figure 9C Tumor volume is shown in individual mice that were administered REGN6194. Figure 9D The tumor volume in individual mice treated with cimipril is shown. Figure 9E Tumor volumes in individual mice treated with REGN6194 and cimipril are shown. Detailed Implementation

[0083] Before describing this disclosure, it should be understood that this disclosure is not limited to the specific methods and experimental conditions described, as such methods and conditions can vary. It should also be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting, as the scope of this disclosure will be limited only by the appended claims.

[0084] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains. When used with respect to a specific numerical value, the term “about” as used herein means that the value may vary by no more than 1% from the stated value. For example, the expression “about 100” as used herein includes 99 and 101 and all values ​​in between (e.g., 99.1, 99.2, 99.3, 99.4, etc.).

[0085] Although any methods and materials similar to or equivalent to those described and materials herein may be used in the practice or testing of this disclosure, preferred methods and materials are now described.

[0086] The inhibitory release of checkpoint inhibitors from PD-L1-blocking mAbs is known to suppress T-cell activation, but their potency as a single agent is often insufficient to achieve tumor clearance and durable antitumor responses in many cancers. Several approaches to improving response rates to PD-1 / PD-L1 inhibition are currently being evaluated, such as identifying biomarkers predicting responsiveness to PD-L1 mAbs or combining them with chemotherapy or radiation therapy. However, the challenge lies in the fact that many of these combinations are often based on pre-existing drug availability and post-hoc rationale for combination therapy, rather than truly hypothesis-driven approaches, which in some cases leads to worse patient outcomes. Clearly, checkpoint inhibition and reactivation of the immune system offer the potential for long-term remission in many patients, thus requiring further methods to improve or enhance T-cell activity to promote more durable responses.

[0087] This disclosure illustrates the potential benefits of combining PD-L1 inhibition with CD28 activating (which provides "signal 2") to enhance antitumor activity. The PD-L1 binding arm of the bispecific antibody acts as a bridge anchoring the antibody's CD28 activation domain to the tumor, providing a pan-tumor co-stimulatory approach (tumor unaware) to convert the inhibitory checkpoint (PD-L1), which is widely present in a variety of tumor indications, into a co-stimulatory signal to activate CD28 on T cells.

[0088] The CAR-T approach also employs the artificial activation of chimeric receptors for both "Signal 1" and "Signal 2" to provide improved antitumor activity. The bispecific antigen-binding molecule of this disclosure offers practical benefits superior to CAR-T therapy: it eliminates the need for cumbersome cell therapy preparation tailored to each patient individually, and it avoids the need for patients to undergo prior "lymphocyte depletion" via toxic chemotherapy, which is often associated with adverse effects that preclude patients from receiving cell therapy. This bispecific approach offers the potential for enhanced efficacy and improved safety through the specificity of its action. In summary, the data presented here suggest that the bispecific PD-L1xCD28 antigen-binding molecule of this disclosure provides a well-tolerated biologic solution with significantly enhanced and synergistic antitumor activity, whether used as a monotherapy or in combination with PD-1 inhibitory antibodies (e.g., cimipril) and / or TAAxCD3 bispecific antibodies.

[0089] definition

[0090] As used herein, “PD-L1” and “PD-L1 fragment” refer to the human PD-L1 protein (also known as CD274, B7-H, B7H1, PDCD1L1, and PDCD1LG1) or fragments thereof, unless specified as originating from a non-human species (e.g., “mouse PD-L1”, “mouse PD-L1 fragment”, “monkey PD-L1”, “monkey PD-L1 fragment”, etc.). In one embodiment of this disclosure, human PD-L1 comprises the amino acid sequence shown in NCBI accession no. AAH69381.1. In one embodiment, a human PD-L1 fragment is shown having a C-terminal myc-myc-hexahistine tag (hPD-L1.mmH) (SEQ ID NO: 71) or having a human or mouse Fc (SEQ ID NO: 73 and 74).

[0091] As used herein, “CD28” refers to the human CD28 protein expressed on T cells as a co-stimulatory receptor, unless otherwise specified as originating from a non-human species. In one embodiment of this disclosure, human CD28 comprises the amino acid sequence shown in NCBI accession No. NP_006130.1. In one embodiment, human CD28 (hCD28.mFc, SEQ ID NO: 72) with a C-terminal mouse Fc tag is expressed.

[0092] "Isolated" antigen-binding proteins (e.g., antibodies or antigen-binding fragments thereof), polypeptides, polynucleotides, and carriers are at least partially free of other biomolecules from the cells or cell cultures in which they are produced. Such biomolecules include nucleic acids, proteins, other antibodies or antigen-binding fragments, lipids, carbohydrates, or other substances such as cell debris and growth media. Isolated antigen-binding proteins may also be at least partially free of expression system components, such as biomolecules from host cells or their growth media. Generally, the term "isolated" is not intended to mean: the complete absence of such biomolecules; or the absence of water, buffers, or salts; or a component of a pharmaceutical formulation containing an antigen-binding protein (e.g., an antibody or antigen-binding fragment).

[0093] The following references relate to the BLAST algorithm frequently used in sequence analysis: BLAST ALGORITHMS: Altschul et al., (2005) FEBS J. 272(20): 5101-5109; Altschul, SF et al., (1990) J.Mol. Biol. 215:403-410; Gish, W. et al., (1993) Nature Genet. 3:266-272; Madden, TL et al., (1996) Meth. Enzymol. 266:131-141; Altschul, SF et al., (1997) NucleicAcids Res. 25:3389-3402; Zhang, J. et al., (1997) Genome Res. 7:649-656; Wootton, JC et al., (1993) Comput. Chem. 17:149-163; Hancock, JM et al., (1994) Comput. Appl. Biosci. 10:67-70; ALIGNMENT SCORING SYSTEMS: Dayhoff, MO et al., "A model of evolutionary change in proteins." in Atlas of Protein Sequence and Structure, (1978) Vol. 5, Supplement 3. MO Dayhoff (ed.), pp. 345-352, Natl. Biomed. Res. Found., Washington, DC; Schwartz, RM et al., "Matrices for detecting distant relationships." in Atlas of Protein Sequence and Structure, (1978) Vol. 5, Supplement 3.'' MO Dayhoff (ed.), pp. 353-358, Natl. Biomed. Res. Found., Washington, DC; Altschul, SF, (1991) J. Mol. Biol. 219:555-565; States, DJ et al., (1991) Methods 3:66-70; Henikoff, S. et al., (1992) Proc. Natl. Acad. Sci.USA 89:10915-10919; Altschul, SF et al., (1993) J. Mol. Evol. 36:290-300; ALIGNMENTSTATISTICS: Karlin, S. et al., (1990) Proc. Natl. Acad. Sci. USA 87:2264-2268; Karlin, S. et al., (1993) Proc. Natl. Acad. Sci. USA 90:5873-5877; Dembo, A. et al., (1994) Ann. Prob. 22:2022-2039; and Altschul, SF "Evaluating the statistical significance of multiple distinct local alignments." in Theoretical and Computational Methods in Genome Research (S. Suhai, Editor), (1997) Pages 1-14, Plenum, NY. .

[0094] Antibodies are immunoglobulin molecules consisting of four polypeptide chains: two heavy chains (HC) and two light chains (LC) linked together by disulfide bonds. Each heavy chain (HC) contains a heavy chain variable region (abbreviated as HCVR or V in this article). H The heavy chain constant region (e.g., IgG, IgG1, or IgG4) contains three domains, C... H 1. C H 2 and C H 3. Each light chain (LC) contains a light chain variable region (abbreviated as LCVR or V in this document). L ) and light chain constant regions (e.g., λ or κ). The light chain constant region contains a structural domain (C L 1). V H and V L The region can be further subdivided into highly variable regions, called complementarity determining regions (CDRs), interspersed with more conservative regions, called framework regions (FRs). Each V... H and V LIt contains three CDRs and four FRs, arranged in the following order from the amino terminus to the carboxyl terminus: FR1, CDR1, FR2, CDR2, FR3, CDR3, FR4. The heavy chain CDR may be referred to as HCDR, and the light chain CDR may be referred to as LCDR. In different embodiments, the FRs of the antibody (or its antigen-binding portion) may be identical to the human germline sequence, or may be natural or artificially modified.

[0095] The antigen-binding arm of a Y-shaped IgG antibody (e.g., the CD28 or PD-L1 binding arm) refers to the structural portion of the antibody that confers specificity for binding to the antigen. For example, the antigen-binding arm of an IgG antibody has a heavy chain (HC) that associates with the light chain (LC).

[0096] For example, a bispecific antibody includes an arm (or domain) that binds to a first antigen and another arm (or domain) that binds to a second antigen. For example, a PD-L1xCD28 bispecific antibody includes an arm that binds to PD-L1 and another arm that binds to CD28.

[0097] A bispecific antigen-binding molecule (e.g., a bispecific antibody) may have an effector arm that binds to a first antigen and a targeting arm that binds to a second antigen. The effector arm may be a first antigen-binding domain (e.g., anti-CD28) that binds to an antigen on an effector cell (e.g., a tumor cell or an immune cell). The targeting arm may be a second antigen-binding domain (e.g., an anti-PD-L1 antibody) that binds to an antigen on a target cell (e.g., a tumor cell or an immune cell). In the context of this disclosure, the effector arm binds to CD28, while the targeting arm binds to the inhibitory checkpoint PD-L1.

[0098] The terms "antigen-binding moiety," "antigen-binding fragment," etc., used herein refer to any naturally occurring, enzymatically obtainable, synthetic, or genetically modified polypeptide or glycoprotein that specifically binds to an antigen to form a complex. Multispecific antigen-binding fragments of antibodies bind to multiple antigens (e.g., if the fragment is bispecific, it binds to two different antigens). Antigen-binding fragments of antibodies can be derived from, for example, intact antibody molecules using any suitable standard technique, such as proteolytic digestion or recombinant genetic engineering techniques involving the manipulation and expression of DNA encoding variable and optionally constant domains of the antibody. Some non-limiting examples of antigen-binding fragments include: (i) Fab fragments; (ii) F(ab')2 fragments; (iii) Fd fragments; (iv) Fv fragments; (v) single-stranded Fv (scFv) molecules; and (vi) dAb fragments.

[0099] In one embodiment, the antigen-binding fragment of the antibody will contain at least one variable domain. The variable domain can have any size or amino acid composition and will typically contain at least one CDR, which is adjacent to or co-framed with one or more frame sequences. L V of domain association H In the antigen-binding fragment of the domain, V H and V L Domains can be positioned relative to each other in any suitable arrangement. For example, variable regions can be dimers and contain V. H -V H V H -V L or V L -V L Dimer. Alternatively, the antigen-binding fragment of the antibody may contain monomer V. H or V L Structural domain.

[0100] In some embodiments, the antigen-binding fragment of the antibody may contain at least one variable domain covalently linked to at least one constant domain. Some non-limiting exemplary configurations of the variable and constant domains that may be present within the antigen-binding fragment of the antibody of this disclosure include: In any configuration of variable and constant domains (including any of the exemplary configurations listed above), the variable and constant domains may be directly connected to each other or connected via all or part of a hinge or connector region. The hinge region may consist of at least two amino acids (e.g., 5, 10, 15, 20, 40, 60, or more), which allows for flexible or semi-flexible connections between adjacent variable and / or constant domains in a single polypeptide molecule. Furthermore, the antigen-binding fragment of the antibody of this disclosure may comprise non-covalently associated fragments with each other and / or with one or more monomers V H or V L Homodimers or heterodimers (or other polymers) of any of the variable and constant domain configurations listed above, whose domains (e.g., via disulfide bonds) are non-covalently associated.

[0101] The term "recombinant" antigen-binding protein, such as an antibody or an antigen-binding fragment thereof, refers to a molecule produced, expressed, isolated, or obtained by techniques or methods known in the art as recombinant DNA technologies, including, for example, DNA splicing and transgenic expression. The term includes antibodies expressed in non-human mammals (including transgenic non-human mammals, such as transgenic mice) or host cells (e.g., Chinese hamster ovary (CHO) cells) or cell expression systems, or antibodies isolated from recombinant human antibody libraries. This disclosure includes recombinant antigen-binding proteins as illustrated herein.

[0102] The term "specific binding" refers to those bindings that target antigens (such as PD-L1 or CD28 proteins) using K-type receptors. D The binding affinity is less than about 10. -6 M (e.g., 10) -7 M, 10 -8 M, 10 -9 M, 10 -10 M, 10 -11 M or 10 -12 Antigen-binding proteins (e.g., antibodies or their antigen-binding fragments) of PD-L1, such as those measured by real-time, label-free biolayer interference, for example at 25°C or 37°C, such as the Octet® HTX biosensor, or by surface plasmon resonance, such as BIACORE™, or by solution affinity ELISA. "Anti-PD-L1" refers to an antigen-binding protein (or other molecule, such as an antigen-binding arm) that specifically binds to PD-L1, such as an antibody or its antigen-binding fragment; "Anti-CD28" refers to an antigen-binding protein (or other molecule, such as an antigen-binding arm) that specifically binds to CD28, such as an antibody or its antigen-binding fragment. "PD-L1xCD28" refers to an antigen-binding protein (or other molecule) that specifically binds to both PD-L1 and CD28 (and optionally, one or more other antigens), such as an antibody or its antigen-binding fragment.

[0103] This disclosure includes antigen-binding proteins, such as antibodies or antigen-binding fragments, that bind to the same PD-L1 and CD28 epitopes as the antigen-binding proteins of this disclosure.

[0104] The term "epitope" refers to a specific antigen-binding site that interacts with an antigen-binding protein, such as an antigenic determinant cluster interacting with a variable region called a complementary site in an antibody molecule (e.g., an antigenic determinant cluster on PD-L1 or CD28). A single antigen may have more than one epitope. Therefore, different antibodies can bind to different regions of an antigen and may have different biological effects. The term "epitope" may also refer to a site on an antigen to which B cells and / or T cells respond, and / or to an antigenic region bound by an antibody. Epitopes can be defined as structural or functional. Functional epitopes are typically a subset of structural epitopes and contain those residues with an affinity that directly contributes to the interaction. Epitopes can be linear or conformational, i.e., composed of non-linear amino acids. In some embodiments, an epitope may comprise a determinant cluster as a chemically active surface group of a molecule such as an amino acid, a sugar side chain, a phosphoryl group, or a sulfonyl group, and in some embodiments may have specific three-dimensional structural features and / or specific charge features.

[0105] Methods for identifying epitopes of antigen-binding proteins (e.g., antibodies, fragments, or peptides) include alanine scanning mutation analysis, peptide blotting (Reineke (2004) Methods Mol. Biol. 248: 443-63), peptide cleavage analysis, crystallographic studies, and NMR analysis. Alternatively, methods such as epitope excision, epitope extraction, and antigen chemical modification can be used (Tomer (2000) Prot. Sci. 9: 487-496). Another method for identifying amino acids in peptides that interact with antigen-binding proteins (e.g., antibodies, fragments, or peptides) is hydrogen / deuterium exchange detected by mass spectrometry. See, for example, Ehring (1999) Analytical Biochemistry 267: 252-259; Engen and Smith (2001) Anal. Chem. 73: 256A-265A.

[0106] This disclosure includes antigen-binding proteins that compete with the antigen-binding proteins of this disclosure for binding to CD28 and PD-L1. As used herein, the term "competition" means that an antigen-binding protein (e.g., an antibody or an antigen-binding fragment thereof) binds to an antigen and inhibits or blocks the binding of another antigen-binding protein (e.g., an antibody or an antigen-binding fragment thereof) to that antigen. Unless otherwise stated, the term also includes competition between two antigen-binding proteins (e.g., antibodies) in two directions, i.e., a first antibody binds to an antigen and blocks the binding of a second antibody, and vice versa. Therefore, in one embodiment of this disclosure, the competition occurs in one such direction. In some embodiments, the first antigen-binding protein (e.g., an antibody) and the second antigen-binding protein (e.g., an antibody) may bind to the same epitope. Alternatively, the first antigen-binding protein and the second antigen-binding protein (e.g., an antibody) may bind to different but, for example, overlapping or non-overlapping epitopes, wherein the binding of one inhibits or blocks the binding of the second antibody, for example, by steric hindrance. Competition between antigen-binding proteins (e.g., antibodies) can be measured by methods known in the art, such as by real-time, label-free biolayer interferometry. Furthermore, binding competition between antigen-binding proteins (e.g., monoclonal antibodies (mAbs)) can be determined using real-time, label-free biolayer interferometry on the Octet RED384 biosensor (Pall ForteBio Corp.).

[0107] Generally, the antibody or antigen-binding fragments of this disclosure modified in some way retain the ability to specifically bind to PD-L1 and CD28, for example, retaining at least 10% of their PD-L1 and CD28 binding activity (compared to the parent antibody) when the activity is expressed in molar terms. Preferably, the antibody or antigen-binding fragments of this disclosure retain at least 20%, 50%, 70%, 80%, 90%, 95%, or 100% or more of the PD-L1 and CD28 binding affinity as a parent antibody. The antibody or antigen-binding fragments of this disclosure are also intended to include conserved or non-conserved amino acid substitutions (referred to as “conserved variants” or “functionally conserved variants” of the antibody) that substantially do not alter its biological activity.

[0108] Polypeptides (such as immunoglobulin chains V containing the amino acid sequence specifically shown herein) H V L A “variant” (e.g., SEQ ID NO: 2, 4, 6; 8; 10; 12; 14; 16; 18; 20; 22; 24; 26; 28; 30; 32; 34; 36; 38; 40; 42; 44; 46; 48; 50; 52; 54; 56; 58; 60; 62; 64; 66;) is defined as containing a amino acid sequence identical to the reference amino acid sequence shown herein (e.g., SEQ ID NO: 2, 4, 6; 8; 10; 12; 14; 16; 18; 20; 22; 24; 26; 28; 30; 32; 34; 36; 38; 40; 42; 44; 46; 48; 50; 52; 54; 56; 58; 60; 62; 64; 66;). A polypeptide with at least approximately 70% to 99.9% (e.g., at least 70%, 72%, 74%, 75%, 76%, 79%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, 99.5%, or 99.9%) amino acid sequence identity or similarity, compared via a BLAST algorithm, where the parameters of the algorithm are selected to give the maximum match between the sequences over the entire length of each reference sequence (e.g., expected threshold: 10; word length: 3; maximum match within the query range: 0; BLOSUM 62 matrix; gap cost: 11 for existence, 1 for extension; conditional combination score matrix adjustment).

[0109] Furthermore, variants of the polypeptide may include polypeptides such as those with immunoglobulin chain V. H V LHC or LC or CDR: may include the amino acid sequence of a reference polypeptide, the amino acid sequence of which is specifically listed herein, but with one or more mutations (e.g., 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10), such as one or more missense mutations (e.g., conserved substitutions), nonsense mutations, deletions, or insertions. For example, this disclosure includes such a CD28xPD-L1 antigen-binding protein: which comprises the amino acid sequence shown in SEQ ID NO: 18, but with one or more such mutations in the PD-L1 binding arm immunoglobulin light chain (or V... L Variants and / or immunoglobulin heavy chains (or V) containing the amino acid sequence shown in SEQ ID NO: 2, but with one or more such mutations. H Variants. In one embodiment of this disclosure, the CD28xPD-L1 antigen-binding protein includes: immunoglobulin light chain variants comprising LCDR1, LCDR2, and LCDR3, wherein one or more (e.g., 1, 2, or 3) of such CDRs have one or more such mutations (e.g., conserved substitutions); and / or immunoglobulin heavy chain variants comprising HCDR1, HCDR2, and HCDR3, wherein one or more (e.g., 1, 2, or 3) of such CDRs have one or more such mutations (e.g., conserved substitutions).

[0110] "Conservatively modified variants" or "conservative substitutions," for example, the "conservatively modified variants" or "conservative substitutions" of immunoglobulin chains described herein, refer to variants in which one or more amino acids in a polypeptide are substituted with other amino acids having similar characteristics (e.g., charge, side chain size, hydrophobicity / hydrophilicity, ossicular structure, and rigidity). Such changes can often be made without significantly impairing the biological activity of the antibody or fragment. Those skilled in the art will recognize that, in general, a single amino acid substitution in a non-essential region of a polypeptide does not substantially alter the biological activity (see, for example, Watson et al., (1987) Molecular Biology of the Gene, The Benjamin / CummingsPub. Co., p. 224 (4th edition)). Furthermore, substitutions of structurally or functionally similar amino acids are unlikely to significantly impair biological activity. This disclosure includes PD-L1xCD28 antigen-binding proteins and / or binding arms comprising such conservedly modified variant immunoglobulin chains.

[0111] Examples of amino acid groups with side chains possessing similar chemical properties include: 1) aliphatic side chains: glycine, alanine, valine, leucine, and isoleucine; 2) aliphatic hydroxyl side chains: serine and threonine; 3) amide-containing side chains: asparagine and glutamine; 4) aromatic side chains: phenylalanine, tyrosine, and tryptophan; 5) basic side chains: lysine, arginine, and histidine; 6) acidic side chains: aspartic acid and glutamic acid; and 7) sulfur-containing side chains: cysteine ​​and methionine. Alternatively, a conservative substitution is any variation with a positive value in the PAM250 log-likelihood matrix disclosed in Gonnet et al., (1992) Science 256: 1443-45.

[0112] Unless the context clearly specifies otherwise, nouns without quantifiers as used herein include one / something and more / somethings. Unless otherwise stated, the terms “comprising,” “including,” “containing,” or “having,” and their variations, as used herein, mean to cover the items listed thereafter and their equivalents, as well as other subjects. The phrases “in one embodiment,” “in multiple embodiments,” “in some embodiments,” etc., are used repeatedly herein. Such phrases do not necessarily refer to the same embodiment, but they may refer to the same embodiment unless the context indicates otherwise. The terms “and / or” or “ / ” as used herein mean any one of the items associated with the term, any combination of the items, or all of the items.

[0113] PD-L1xCD28 antigen-binding molecule, antibody or its antigen-binding fragment

[0114] The antibodies (or antigen-binding molecules) of this disclosure can be bispecific or multispecific. Multispecific antibodies or antigen-binding molecules may be specific to different epitopes of a single target polypeptide, or may contain antigen-binding domains specific to more than one target polypeptide. See, for example, Tutt et al., 1991, J. Immunol. 147:60-69; Kufer et al., 2004, Trends Biotechnol. 22:238-244. The antibodies of this disclosure can be linked to or co-expressed with another functional molecule, such as another peptide or protein. For example, an antibody or fragment thereof can be functionally linked to one or more other molecular entities, such as another antibody or antibody fragment (e.g., through chemical coupling, genetic fusion, non-covalent association, or other means), to produce bispecific or multispecific antibodies with a second binding specificity.

[0115] The term "anti-CD28 antibody" as used herein is intended to include both monospecific anti-CD28 antibodies and multispecific (e.g., bispecific) antibodies or antigen-binding molecules comprising a CD28-binding arm and a second arm that binds to PD-L1. Therefore, this disclosure includes bispecific antibodies in which one arm of an immunoglobulin binds human CD28, and the other arm of the immunoglobulin is specific for PD-L1. The CD28-binding arm may comprise any HCVR / LCVR or CDR amino acid sequence as shown in Tables 3 and 8 herein. In some embodiments, the CD28-binding arm binds human CD28 and induces human T cell proliferation.

[0116] According to certain exemplary embodiments, this disclosure includes a bispecific antigen-binding molecule that specifically binds to CD28 and PD-L1. Such a molecule may be referred to herein as, for example, "anti-CD28 / anti-PD-L1", or "anti-CD28xPD-L1", or "CD28xPD-L1", or "PD-L1xCD28", or "anti-PD-L1 / anti-CD28", or "anti-PD-L1xCD28", or "PD-L1xCD28" bispecific molecule, or "anti-PD-L1 x anti-CD28" or "anti-CD28x anti-PD-L1", or other similar terms.

[0117] According to certain exemplary embodiments, a bispecific antigen-binding molecule (e.g., a bispecific antibody) may have an effector arm and a target arm. The effector arm may be a first antigen-binding domain that binds to CD28 on effector cells (e.g., T cells). The target arm may be a second antigen-binding domain that binds to an antigen (e.g., an immune checkpoint) on target cells (e.g., tumor cells or antigen-presenting cells). In the context of this disclosure, the effector arm binds to CD28, and the target arm binds to the inhibitory checkpoint ligand PD-L1. Bispecific anti-CD28 / PD-L1 can provide a pan-tumor co-stimulatory approach (tumor-unaware) to convert inhibitory checkpoints (PD-L1) widely present in various tumor indications and transform them into co-stimulatory signals to activate CD28 on T cells.

[0118] As used herein, the term "antigen-binding molecule" means a protein, polypeptide, or molecular complex that specifically binds to a particular antigen and comprises or consists of at least one complementary determinant region (CDR), alone or in combination with one or more additional CDRs and / or frame regions (FRs). In some embodiments, the antigen-binding molecule is an antibody or antibody fragment, as those terms are defined elsewhere herein.

[0119] As used herein, the term "bispecific antigen-binding molecule" refers to a protein, peptide, or molecular complex (e.g., an antibody or its antigen-binding fragment) comprising at least a first antigen-binding domain and a second antigen-binding domain. Each antigen-binding domain within a bispecific antigen-binding molecule contains at least one antigen-binding domain (CDR), which is alone or in combination with one or more additional CDRs and / or FRs, and binds specifically to a particular antigen. In the context of this disclosure, the first antigen-binding domain specifically binds to a first antigen (e.g., CD28), and the second antigen-binding domain specifically binds to a second, different antigen (e.g., PD-L1).

[0120] In some exemplary embodiments, the bispecific antigen-binding molecule is a bispecific antibody. Each antigen-binding domain of the bispecific antibody includes a heavy chain variable domain (HCVR) and a light chain variable domain (LCVR).

[0121] The first antigen-binding domain and the second antigen-binding domain may be directly or indirectly linked to each other to form the bispecific antigen-binding molecule of this disclosure. Alternatively, the first antigen-binding domain and the second antigen-binding domain may each be linked to a separate multimerizing domain. Association of one multimerizing domain with another multimerizing domain promotes association between the two antigen-binding domains, thereby forming a bispecific antigen-binding molecule. As used herein, "multimerizing domain" refers to any macromolecule, protein, polypeptide, peptide, or amino acid having the ability to associate with a second multimerizing domain of the same or similar structure or construction. For example, a multimerizing domain may be containing immunoglobulin C. H 3-domain polypeptides. A non-limiting example of a polymerized component is the Fc region of an immunoglobulin (containing C... H 2-C H 3. A domain, such as the Fc domain of IgG, wherein the IgG is selected from isotypes IgG1, IgG2, IgG3, and IgG4, and any allotypes within each isotype group. The Fc domain may contain wild-type or modified IgG isotypes.

[0122] The bispecific antigen-binding molecules disclosed herein typically comprise two multimerizing domains, such as two Fc domains, each independently forming part of a separate antibody heavy chain. The first and second multimerizing domains can be the same IgG isotype, such as IgG1 / IgG1, IgG2 / IgG2, or IgG4 / IgG4. Alternatively, the first and second multimerizing domains can be different IgG isotypes, such as IgG1 / IgG2, IgG1 / IgG4, or IgG2 / IgG4.

[0123] In some embodiments, the polymerizing domain is an amino acid sequence or Fc fragment of 1 to 200 amino acids containing at least one cysteine ​​residue. In other embodiments, the polymerizing domain is a cysteine ​​residue, or a short cysteine-containing peptide. Still other polymerizing domains include peptides or polypeptides comprising or composed of leucine zippers, helical-cyclic motifs, or coiled-coil motifs.

[0124] Any bispecific antibody form or technology can be used to prepare the bispecific antigen-binding molecules of this disclosure. For example, an antibody or antigen-binding fragment having a first antigen-binding specificity can be functionally linked with one or more other molecular entities, such as another antibody or antibody fragment having a second antigen-binding specificity (e.g., by chemical coupling, gene fusion, non-covalent association, or other means), to produce a bispecific antigen-binding molecule. Specific exemplary bispecific forms that can be used in the context of this disclosure include, but are not limited to, scFv-based forms or bispecific antibody forms, IgG-scFv fusions, dual variable domain (OVO)-Ig, quadromas, knocks-in-holes, common light chains (e.g., common light chains with bulge-in-hole structures, etc.), CrossMab, CrossFab, (SEEO) bodies, leucine zippers, Ouobody, IgG1 / IgG2, dual-action Fab (OAF)-IgG, and Mab. 2 Bispecific form (for a review of the aforementioned forms, see, for example, Klein et al., 2012, mAbs 4:6, 1-11 and references cited therein).

[0125] In the context of the bispecific antigen-binding molecules of this disclosure, compared to the wild-type, naturally occurring Fc domain form, the polymerized domain, such as the Fc domain, may contain one or more amino acid variations (e.g., insertions, deletions, or substitutions). For example, this disclosure includes bispecific antigen-binding molecules that contain one or more modifications in the Fc domain such that the modified Fc domain has a modified binding interaction (e.g., enhanced or reduced) between Fc and FcRn. In one embodiment, the bispecific antigen-binding molecule is located at C... H 2 or C HRegion 3 contains a modification that enhances the affinity of the Fc domain for FcRn in acidic environments (e.g., in endosomes with a pH range of about 5.5 to about 6.0). Some non-limiting examples of such Fc modifications include, for example, modifications at positions 250 (e.g., E or Q); 250 and 428 (e.g., L or F); 252 (e.g., LN / FIW or T), 254 (e.g., S or T), and 256 (e.g., S / R / Q / EID or T); or modifications at positions 428 and / or 433 (e.g., UR / S / P / Q or K) and / or 434 (e.g., H / F or V); or modifications at positions 250 and / or 428; or modifications at positions 307 or 308 (e.g., 308F, V308F) and 434. In one embodiment, the modifications include 428L (e.g., M428L) and 434S (e.g., N434S) modifications; 428L, 2591 (e.g., V2591) and 308F (e.g., V308F) modifications; 433K (e.g., H433K) and 434 (e.g., 434Y) modifications; 252, 254 and 256 (e.g., 252Y, 254T and 256E) modifications; 250Q and 428L modifications (e.g., T250Q and M428L); and 307 and / or 308 modifications (e.g., 308F or 308P).

[0126] This disclosure also includes the first C H 3 structural domains and second Ig C H A bispecific antigen-binding molecule with three domains, wherein the first and second IgC H The three domains differ from each other by at least one amino acid, and wherein at least one amino acid difference reduces the binding of the bispecific antibody to protein A compared to a bispecific antibody lacking amino acid difference. In one embodiment, the first IgC H 3-domain binding to protein A, and second IgC H Domain 3 contains mutations that reduce or eliminate protein A binding, such as H95R modification (IMGT exon number; H435R (EU number)). Second C H3 may also include Y96F modification (IMGT; Y436F (EU)). Other modifications that may exist in the second CH3 include: in the case of IgG1 antibody, D16E, L18M, N44S, K52N, V57M, and V821 (IMGT; D356E, L358M, N384S, K392N, V397M, and V4221 (EU)); in the case of IgG2 antibody, N44S, K52N, and V821 (IMGT; N384S, K392N, and V4221 (EU)); and in the case of IgG4 antibody, Q15R, N44S, K52N, V57M, R69K, E79Q, and V821 (IMGT; Q355R, N384S, K392N, V397M, R409K, E419Q, and V4221 (EU)).

[0127] In some implementations, the Fc domain can be chimeric, combining Fc sequences derived from more than one immunoglobulin isotype. For example, a chimeric Fc domain may contain partial or complete Fc sequences derived from human IgG1, human IgG2, or human IgG4C. H C in Zone 2 H 2 sequences, and C sequences derived partially or entirely from human IgG1, human IgG2, or human IgG4. H 3. Sequence. The chimeric Fc domain may also include a chimeric hinge region. For example, the chimeric hinge may include an "upper hinge" sequence derived from the hinge region of human IgG1, human IgG2, or human IgG4, combined with a "lower hinge" sequence derived from the hinge region of human IgG1, human IgG2, or human IgG4. A specific example of a chimeric Fc domain that may be included in any antigen-binding molecule described herein includes, from the N-terminus to the C-terminus: [IgG4C H 1]-[IgG4 upper hinge]-[IgG2 lower hinge]-[IgG4 CH2]-[IgG4 C H 3). Another example of a chimeric Fc domain that may be included in any antigen-binding molecule shown herein includes, from the N-terminus to the C-terminus: [IgG1 C H 1]-[IgG1 upper hinge]-[IgG2 lower hinge]-[IgG4 C H 2]-[IgG1 C H 3]. These and other examples of chimeric Fc domains that may be included in any antigen-binding molecule contained in this disclosure are described in WO2014 / 022540 A1. Chimeric Fc domains having these general structural arrangements and their variants may have altered Fc receptor binding, which in turn affects Fc effector function.

[0128] The antibody and antigen-binding fragments of this disclosure comprise immunoglobulin chains containing the amino acid sequences (and variants thereof) specifically shown herein, along with cellular and in vitro post-translational modifications to the antibody or fragment. For example, this disclosure includes antibodies and antigen-binding fragments that specifically bind to PD-L1 and CD28, comprising the heavy chain and / or light chain amino acid sequences shown herein; and antibodies and fragments with one or more asparagine, serine, and / or threonine residues glycosylated, one or more asparagine residues deamidated, one or more residues (e.g., Met, Trp, and / or His) oxidized, N-terminal glutamine being pyroglutamic acid (pyroE), and / or C-terminal lysine or other amino acid deletions.

[0129] The bispecific antigen-binding molecule of this disclosure comprises a first antigen-binding arm (“CD28-binding arm” or “CD28-binding domain”) that specifically binds to CD28. In some embodiments, the CD28-binding arm comprises HCVR and LCVR containing the amino acid sequences disclosed herein. The bispecific antigen-binding molecule also comprises a second antigen-binding arm (“PD-L1-binding arm” or “PD-L1-binding domain”) that specifically binds to PD-L1. In some embodiments, the PD-L1-binding arm comprises HCVR and LCVR containing the amino acid sequences disclosed herein. In some embodiments, the PD-L1-binding arm comprises a heavy chain immunoglobulin and a corresponding light chain immunoglobulin, said heavy chain immunoglobulin comprising a V-type immunoglobulin containing a combination of heavy chain CDRs (HCDR1, HCDR2, and HCDR3). H The light chain immunoglobulin comprises a V containing a combination of light chain CDRs (LCDR1, LCDR2, and LCDR3). L As described in this document or in International Patent Application Publication No. WO2014 / 004427.

[0130] Polynucleotides and their preparation methods

[0131] Isolated polynucleotide molecules or groups of polynucleotide molecules containing polynucleotide sequences that encode immunoglobulin chains of any PD-L1xCD28 multispecific antigen-binding protein described herein are part of this disclosure. This disclosure also includes vectors or vector groups containing said polynucleotide molecules and / or host cells (e.g., Chinese hamster ovary (CHO) cells) containing said polynucleotide molecules, vectors, or antigen-binding proteins described herein.

[0132] A polynucleotide molecule or sequence refers to DNA or RNA. This disclosure includes any polynucleotides of this disclosure, for example, immunoglobulin V encoding the PD-L1 binding arm and / or the CD28 binding arm. H VL CDR-H, CDR-L, HC, or LC, optionally, are operatively linked to a promoter or other expression control sequence. For example, this disclosure provides any polynucleotide molecule or group of polynucleotide molecules (e.g., DNA) comprising the nucleotide sequences shown in Table 2 and Table 4.

[0133] This disclosure includes polynucleotides comprising the nucleotide sequences shown in SEQ ID NO: 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 29, 31, 33, 35, 39, 41, 43, 45, 49, 51, 53, 55, 57, 59, 61, 63, 65, 67 and / or 69, optionally operatively linked to a promoter or other expression control sequence or other polynucleotide sequence.

[0134] Generally, a “promoter” or “promoter sequence” is a DNA regulatory region capable of binding to RNA polymerases in a cell (e.g., proteins or substances that bind directly or via other promoters) and initiating transcription of a coding sequence. Promoters can be operatively linked to other expression control sequences, including enhancer and repressor sequences and / or polynucleotides of this disclosure. Examples of promoters that can be used to control gene expression include, but are not limited to, the cytomegalovirus (CMV) promoter (US Patent Nos. 5,385,839 and 5,168,062), the early promoter region of SV40 (Benoist et al., (1981) Nature 290:304-310), the promoter contained in the 3' long terminal repeat of Rous sarcoma virus (Yamamoto et al., (1980) Cell 22:787-797), the herpes simplex kinase promoter (Wagner et al., (1981) Proc. Natl. Acad. Sci. USA 78:1441-1445), the regulatory sequence of the metallothionein gene (Brinster et al., (1982) Nature 296:39-42); and prokaryotic expression vectors such as the β-lactamase promoter (VIIIa-Komaroff et al., (1978) Proc. Natl. Acad. Sci. USA). 75:3727-3731) or tac promoter (DeBoer et al., (1983) Proc. Natl. Acad. Sci. USA 80:21-25); see also "Useful proteins from recombinant bacteria" in Scientific American (1980) 242:74-94; and promoter elements from yeast or other fungi, such as Gal4 promoter, ADC (alcohol dehydrogenase) promoter, PGK (phosphoglycerol kinase) promoter or alkaline phosphatase promoter.

[0135] The polynucleotide encoding the polypeptide is "operably linked" to a promoter or other expression control sequence: in a cell or other expression system, this sequence directs the RNA polymerase-mediated transcription of the coding sequence into RNA, preferably mRNA, which can then be spliced ​​(if it contains introns) and, optionally, translated into a protein encoded by the coding sequence.

[0136] This disclosure includes polynucleotides encoding immunoglobulin polypeptide chains that are variants of those nucleotide sequences specifically shown herein. A “variant” of a polynucleotide is a polynucleotide containing at least about 70% to 99.9% (e.g., 70%, 72%, 74%, 75%, 76%, 79%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, 99.5%, 99.9%) of the identity with the reference nucleotide sequences shown herein; this is done by comparison using a BLAST algorithm, where the parameters of the algorithm are selected to give the maximum match between the sequences over the entire length of the respective reference sequences (e.g., expected threshold: 10; word length: 28; maximum match within the query range: 0; match / non-match score: 1, -2; gap penalty: linear). In one embodiment of this disclosure, the variant of the nucleotide sequence specifically illustrated herein comprises one or more (e.g., 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, or 12) point mutations, insertions (e.g., in-frame insertions), or deletions (e.g., in-frame deletions) of one or more nucleotides. In one embodiment of this disclosure, such mutations may be missense or nonsense mutations. In one embodiment of this disclosure, such variant polynucleotides encode immunoglobulin polypeptide chains that can be incorporated into the PD-L1 binding arm and / or CD28 binding arm, i.e., such that the protein retains specific binding to PD-L1 and / or CD28.

[0137] Eukaryotic and prokaryotic host cells, including mammalian cells, can be used as hosts for expressing the PD-L1xCD28 antigen-binding protein (e.g., an antibody or its antigen-binding fragment) or its antigen-binding arm. Such host cells are well known in the art, and many are available from the American Type Culture Collection (ATCC). These host cells include, in particular, Chinese hamster ovary (CHO) cells, NSO, SP2 cells, HeLa cells, baby hamster kidney (BHK) cells, monkey kidney cells (COS) cells, human hepatocellular carcinoma cells (e.g., Hep G2), A549 cells, 3T3 cells, HEK-293 cells, and many other cell lines. Mammalian host cells include cells from humans, mice, rats, dogs, monkeys, pigs, goats, cattle, horses, and hamsters. Other cell lines that can be used include insect cell lines (such as fall armyworm (Spodoptera frugiperda) or white armyworm (Trichoplusia ni)), amphibian cells, bacterial cells, plant cells, and fungal cells.Fungal cells include cells of yeast and filamentous fungi, including, for example, *Pichia*, *Pichia pastoris*, *Pichia finlandica*, *Pichia trehalophila*, *Pichia koclamae*, *Pichia membranaefaciens*, *Pichia minuta* (Ogataea minuta, Pichia lindneri), *Pichia opuntiae*, *Pichia athermotolerans*, *Pichia salictaria*, *Pichia guercuum*, *Pichia pijperi*, *Pichia stiptis*, *Pichia methanolica*, *Pichia* sp., *Saccharomyces cerevisiae*, *Saccharomyces* sp., and *Hansenula polymorpha*. The fungi include *C. polymorpha*, *Kluyveromyces* sp., *Kluyveromyces lactis*, *Candida albicans*, *Aspergillus niduLans*, *Aspergillus niger*, *Aspergillus oryzae*, *Trichoderma reesei*, *Chrysosporium lucknowense*, *Fusarium* sp., *Fusarium gramineum*, *Fusarium venenatum*, *Physcomitrella patens*, and *Neurospora crassa*.This disclosure includes isolated host cells (e.g., CHO cells or any of the types of host cells described above) containing the anti-PD-L1 x anti-CD28 antigen-binding proteins of this disclosure, such as REGN6192, REGN6193, and REGN6194, as well as the anti-PD-L1 x anti-CD28 antigen-binding proteins shown in Table 9, or one or more polynucleotide molecules encoding the heavy and / or light chains of immunoglobulin (Ig); and / or one or more polynucleotides encoding the PD-L1 binding arm and CD28 binding arm of the multispecific antigen-binding protein of this disclosure.

[0138] This disclosure also includes cells expressing PD-L1 and / or CD28 or antigenic fragments or fusions thereof (e.g., His6, Fc, and / or myc), wherein the PD-L1 and / or CD28 or antigenic fragments or fusions thereof are bound by the PD-L1xCD28 antigen-binding protein of this disclosure (e.g., an antibody or an antigen-binding fragment thereof) (e.g., REGN6194) and any bispecific antibodies disclosed herein.

[0139] Several methods for generating recombinant antibodies are known in the art. One example of a method for generating recombinant antibodies is disclosed in US4816567. Transformation can be performed by any known method for introducing polynucleotides into host cells. Methods for introducing heterologous polynucleotides into mammalian cells are well known in the art and include dextran-mediated transfection, calcium phosphate precipitation, polybrene-mediated transfection, protoplast fusion, electroporation, encapsulation of polynucleotides in liposomes, biolistic injection, and direct microinjection of DNA into the cell nucleus. Additionally, nucleic acid molecules can be introduced into mammalian cells via viral vectors. Methods for transforming cells are well known in the art. See, for example, US Patent Nos. 4399216; 4912040; 4740461 and 4959455.

[0140] This disclosure includes a recombinant method for preparing the anti-PD-L1 x anti-CD28 (e.g., REGN6194) antigen-binding protein of this disclosure, such as the antibody or antigen-binding fragment thereof, or the immunoglobulin chain thereof, said recombinant method comprising:

[0141] (i) Introducing one or more polynucleotides encoding immunoglobulin light and heavy chains into a host cell, the immunoglobulin light and heavy chains encoding antigen-binding arms of PD-L1xCD28 antigen-binding protein, for example, wherein the polynucleotides are located in a vector; and / or integrating into the host cell chromosome and / or operatively linking to a promoter;

[0142] (ii) Culture host cells (e.g., CHO or Pichia pastoris or Pichia pastoris) under conditions favorable to the expression of the polynucleotide, and

[0143] (iii) Optionally, antigen-binding proteins (e.g., antibodies or antigen-binding fragments) or chains are isolated from host cells and / or the culture medium in which the host cells are cultured. This disclosure also includes PD-L1xCD28 antigen-binding proteins, such as antibodies and their antigen-binding fragments, as products of the methods described herein (and optionally, the purification methods described herein).

[0144] In one embodiment of this disclosure, a method for preparing a PD-L1xCD28 (e.g., REGN6194) antigen-binding protein, such as an antibody or an antigen-binding fragment thereof, includes methods for purifying the antigen-binding protein, such as by column chromatography, precipitation, and / or filtration. As discussed, the products of such methods also constitute part of this disclosure.

[0145] Sequence variants

[0146] Compared to corresponding germline sequences derived from a single antigen-binding domain, antibodies and bispecific antigen-binding molecules of this disclosure may contain one or more amino acid substitutions, insertions, and / or deletions in the frame regions and / or CDR regions of the heavy and light chain variable domains. Such mutations can be readily identified by comparing the amino acid sequences disclosed herein with germline sequences available from, for example, public antibody sequence databases. Antigen-binding molecules of this disclosure may contain antigen-binding fragments derived from any exemplary amino acid sequence disclosed herein, wherein one or more amino acids in one or more frame regions and / or CDR regions are mutated to corresponding residues of the germline sequence from which the antibody is derived, or mutated to corresponding residues of another human germline sequence, or mutated to conserved amino acid substitutions of the corresponding germline residues (such sequence changes are collectively referred to herein as “germline mutations”). Those skilled in the art can readily generate numerous antibody and antigen-binding fragments containing one or more single germline mutations or combinations thereof, starting from the heavy and light chain variable region sequences disclosed herein. In some embodiments, V H and / or V LAll frame and / or CDR residues within the domain mutate back to residues present in the original germline sequence from which the antigen-binding domain originally originated. In other embodiments, only certain residues mutate back to the original germline sequence, for example, mutated residues present only in the first 8 amino acids of FR1 or the last 8 amino acids of FR4, or mutated residues present only in CDR1, CDR2, or CDR3. In other embodiments, one or more frame and / or CDR residues mutate to corresponding residues of a different germline sequence (i.e., a germline sequence different from the original germline sequence from which the antigen-binding domain originally originated). Furthermore, the antigen-binding domain may contain any combination of two or more germline mutations within the frame and / or CDR regions, for example, where certain individual residues mutate to corresponding residues of a specific germline sequence, while certain other residues different from the original germline sequence are maintained or mutated to corresponding residues of a different germline sequence. Once obtained, one or more desired properties of the antigen-binding domain containing one or more germline mutations can be readily tested, such as improved binding specificity, increased binding affinity, improved or enhanced antagonistic or agonistic biological properties (as the case may be), reduced immunogenicity, etc. Bispecific antigen-binding molecules containing one or more antigen-binding domains obtained in this general manner are covered in this disclosure.

[0147] This disclosure also includes antigen-binding molecules in which one or both antigen-binding domains comprise variants of any of the HCVR, LCVR, and / or CDR amino acid sequences disclosed herein with one or more conserved substitutions. For example, this disclosure includes antigen-binding molecules comprising antigen-binding domains having HCVR, LCVR, and / or CDR amino acid sequences having, for example, 10 or fewer, 8 or fewer, 6 or fewer, 4 or fewer, etc., conserved amino acid substitutions relative to any of the HCVR, LCVR, and / or CDR amino acid sequences disclosed herein. A “conserved amino acid substitution” is an amino acid substitution in which an amino acid residue is replaced by another amino acid residue having a side chain (R group) with similar chemical properties (e.g., charge or hydrophobicity). Generally, conserved amino acid substitutions do not substantially alter the functional properties of the protein. Examples of amino acid groups with side chains having similar chemical properties include: (1) aliphatic side chains: glycine, alanine, valine, leucine, and isoleucine; (2) aliphatic hydroxyl side chains: serine and threonine; (3) amide-containing side chains: asparagine and glutamine; (4) aromatic side chains: phenylalanine, tyrosine, and tryptophan; (5) basic side chains: lysine, arginine, and histidine; (6) acidic side chains: aspartic acid and glutamic acid; and (7) sulfur-containing side chains: cysteine ​​and methionine. Preferred conservative amino acid substitution groups are: valine-leucine-isoleucine, phenylalanine-tyrosine, lysine-arginine, alanine-valine, glutamic acid-aspartic acid, and asparagine-glutamine. Alternatively, conservative substitutions are any changes with positive values ​​in the PAM250 log-likelihood matrix disclosed in Gonnet et al., (1992) Science 256: 1443-1445. A “moderately conservative” substitution is any change that has a non-negative value in the PAM250 log-likelihood matrix.

[0148] This disclosure also includes antigen-binding molecules comprising an antigen-binding domain that is substantially identical to any of the HCVR, LCVR, and / or CDR amino acid sequences disclosed herein. When referring to amino acid sequences, the terms “significant identity” or “substantially identical” mean that, when optimally aligned, for example, using the GAP or BESTFIT procedure with default vacancy weights, two amino acid sequences share at least 95% sequence identity, and more preferably at least 98% or 99% sequence identity. Preferably, dissimilar residue sites differ due to conserved amino acid substitutions. In cases where two or more amino acid sequences differ from each other due to conserved substitutions, the percentage of sequence identity or similarity can be upregulated to correct for the conservatism of the substitutions. Means for making such adjustments are well known to those skilled in the art. See, for example, Pearson (1994) Methods Mol. BioI. 24: 307-331.

[0149] Sequence similarity of peptides, also known as sequence identity, is typically determined using sequence analysis software. Protein analysis software uses similarity measures specified for various substitutions, deletions, and other modifications, including conserved amino acid substitutions, to match similar sequences. For example, GCG software includes programs such as Gap and Bestfit, which can use default parameters to determine sequence homology or sequence identity between closely related peptides, such as homologous peptides from different biological species, or between wild-type proteins and their mutant proteins. See, for example, GCG version 6.1. Peptide sequences can also be compared using FASTA (the program in GCG version 6.1) with default or recommended parameters. FASTA (e.g., FASTA2 and FASTA3) provides alignment of the best overlapping region between the query sequence and the retrieved sequence and a percentage of sequence identity (Pearson (2000) Methods Mol. BioI. 132: 185-219). When comparing sequences of this disclosure with databases containing large numbers of sequences from different organisms, another preferred algorithm is the computer program BLAST, particularly BLASTP or TBLASTN, using default parameters. See, for example, Altschul et al., (1990) J. Mol. BioI. 215:403-410; Altschul et al., (1997) Nucleic Acids Res. 25:3389-402.

[0150] Antibodies containing Fc variants

[0151] According to certain embodiments of this disclosure, an anti-PD-L1X anti-CD28 bispecific antigen-binding molecule comprising an Fc domain is provided, said Fc domain containing one or more mutations that enhance or weaken antibody binding to the FcRn receptor, for example, at acidic pH compared to neutral pH. For example, this disclosure includes CFc in the Fc domain. H 2 or C H Region 3 contains a mutated antibody and antigen-binding molecule, wherein the mutation increases the affinity of the Fc domain for FcRn in an acidic environment (e.g., in endosomes with a pH range of about 5.5 to about 6.0). When administered to animals, such a mutation can lead to a prolonged serum half-life of the antibody. Some non-limiting examples of such Fc modifications include, for example, modifications at the following positions:

[0152] 250 (e.g., E or Q);

[0153] 250 and 428 (e.g., L or F);

[0154] 252 (e.g., L / Y / F / W or T),

[0155] 254 (e.g., S or T), and / or

[0156] 256 (e.g., S / R / Q / E / D or T);

[0157] Or, as an embellishment in the following locations:

[0158] 428 and / or 433 (e.g., H / L / R / S / P / Q or K) and / or

[0159] 434 (e.g., H / F or Y);

[0160] Or, as an embellishment in the following locations:

[0161] 250 and / or 428;

[0162] Or, as an embellishment in the following locations:

[0163] 307 or 308 (e.g., 308F, V308F), and / or

[0164] 434.

[0165] In one implementation, the modification includes:

[0166] Modifications include 428L (e.g., M428L) and 434S (e.g., N434S);

[0167] Modifications include 428L, 259I (e.g., V259I) and 308F (e.g., V308F);

[0168] Modified with 433K (e.g., H433K) and 434 (e.g., 434Y);

[0169] Modifications 252, 254, and 256 (e.g., 252Y, 254T, and 256E);

[0170] 250Q and 428L modifications (e.g., T250Q and M428L); and / or

[0171] 307 and / or 308 modifications (e.g., 308F or 308P).

[0172] For example, this disclosure includes a PD-L1xCD28 bispecific antigen-binding molecule comprising an Fc domain, wherein the Fc domain comprises one or more pairs or one or more groups of mutations selected from:

[0173] 250Q and 248L (e.g., T250Q and M248L);

[0174] 252Y, 254T and 256E (e.g. M252Y, S254T and T256E);

[0175] 428L and 434S (e.g., M428L and N434S); and

[0176] 433K and 434F (e.g., H433K and N434F).

[0177] This disclosure also includes the first C H 3 structural domains and second Ig C H A bispecific antigen-binding molecule with three domains, wherein the first and second IgC HThe three domains differ from each other by at least one amino acid, and wherein at least one amino acid difference reduces the binding of the bispecific antibody to protein A compared to a bispecific antibody lacking amino acid difference. In one embodiment, the first IgC H 3-domain binding to protein A and second IgC H Domain 3 contains mutations that reduce or eliminate protein A binding, such as H95R modification (IMGT exon number; H435R (EU number)). Second C H 3 may also include Y96F modification (IMGT; Y436F (EU)). See, for example, U.S. Patent No. 8,586,713. Second C H Other modifications that may exist in 3 include: in the case of IgG1 antibody, D16E, L18M, N44S, K52N, V57M and V82I (IMGT; D356E, L358M, N384S, K392N, V397M and V422I (EU)); in the case of IgG2 antibody, N44S, K52N and V82I (IMGT; N384S, K392N and V422I (EU)); and in the case of IgG4 antibody, Q15R, N44S, K52N, V57M, R69K, E79Q and V82I (IMGT; Q355R, N384S, K392N, V397M, R409K, E419Q and V422I (EU)).

[0178] All possible combinations of the aforementioned Fc domain mutations and other mutations within the antibody variable domains disclosed herein are contemplated within the scope of this disclosure.

[0179] Biological characteristics of bispecific antibodies and antigen-binding molecules

[0180] This disclosure includes antibodies and antigen-binding fragments thereof that bind to human CD28 and PD-L1 with high affinity. This disclosure also includes antibodies and antigen-binding fragments thereof that bind to human CD28 and / or PD-L1 with intermediate or low affinity, depending on the therapeutic context and the specific targeting characteristics desired. For example, in the case of a bispecific antigen-binding molecule where one arm binds CD28 and the other arm binds to a target antigen (e.g., PD-L1), it is desirable for the target antigen-binding arm to bind to the target antigen with high affinity, while the anti-CD28 arm binds to CD28 with only intermediate or low affinity. In this way, the antigen-binding molecule can preferentially target cells expressing the target antigen, while avoiding general / untargeted CD28 binding and the associated adverse side effects.

[0181] According to certain implementations, this disclosure includes K with a density of less than about 200 nM. DAntibodies binding to human CD28 (e.g., at 25°C) and antigen-binding fragments of the antibodies are measured, for example, by surface plasmon resonance, using the assay format defined in Example 2 herein. In some embodiments, the antibodies or antigen-binding fragments of this disclosure are expressed in Kc values ​​of less than about 100 nM, less than about 90 nM, less than about 80 nM, less than about 60 nM, less than about 40 nM, less than about 30 nM, less than 20 nM, less than 10 nM, or less than 5 nM. D In conjunction with CD28, as measured by surface plasmon resonance, for example, using an assay format defined in Example 2 herein or a substantially similar assay. In some embodiments, the antibody or antigen-binding fragment of this disclosure is expressed at a Kc concentration of about 5 nM to about 50 nM. D Combined with CD28.

[0182] This disclosure also includes antibodies and antigen-binding fragments of the present disclosure that bind to CD28 with a dissociation half-life (t½) greater than about 3 minutes, as measured by surface plasmon resonance at 25°C or 37°C, for example, using the assay format defined in the examples herein or a substantially similar assay. In some embodiments, the antibodies or antigen-binding fragments of the present disclosure bind to CD28 with a t½ greater than about 5 minutes, greater than about 10 minutes, greater than about 20 minutes, greater than about 30 minutes, greater than about 40 minutes, or greater than about 50 minutes, as measured by surface plasmon resonance at 25°C or 37°C, for example, using the assay format defined in Example 2 herein or a substantially similar assay.

[0183] According to certain implementations, this disclosure includes K with a value less than about 1 nM. D Antibodies binding to human PD-L1 (e.g., at 25°C) and antigen-binding fragments of the antibodies are measured, for example, by surface plasmon resonance, using the assay format defined in Example 2 herein. In some embodiments, the antibodies or antigen-binding fragments of this disclosure are expressed in Kc values ​​less than about 1 nM, less than about 0.9 nM, less than about 0.8 nM, less than about 0.6 nM, less than about 0.4 nM, less than about 0.3 nM, less than 0.2 nM, less than 0.1 nM, or less than 0.05 nM. D Combined with PD-L1, such as that measured by surface plasmon resonance, for example, using an assay format defined as in Example 2 herein or a substantially similar assay. In some embodiments, the antibody or antigen-binding fragment of this disclosure is expressed at a Kc of about 0.05 nM to about 0.2 nM. D Combined with PD-L1.

[0184] This disclosure also includes antibodies and antigen-binding fragments of the present disclosure that bind to PD-L1 with a dissociation half-life (t½) greater than about 30 minutes, as measured by surface plasmon resonance at 25°C or 37°C, for example, using the assay format defined in the examples herein or a substantially similar assay. In some embodiments, the antibodies or antigen-binding fragments of the present disclosure bind to PD-L1 with a t½ greater than about 30 minutes, greater than about 60 minutes, greater than about 120 minutes, greater than about 150 minutes, greater than about 180 minutes, or greater than about 200 minutes, as measured by surface plasmon resonance at 25°C or 37°C, for example, using the assay format defined in Example 2 herein or a substantially similar assay.

[0185] This disclosure includes bispecific antigen-binding molecules (e.g., bispecific antibodies) capable of binding simultaneously to human CD28 and human PD-L1. According to certain embodiments, the bispecific antigen-binding molecules of this disclosure specifically interact with cells expressing CD28 and / or PD-L1. The extent to which the bispecific antigen-binding molecules bind to cells expressing CD28 and / or PD-L1 can be assessed by fluorescence activated cell sorting (FACS), as shown in Example 3 herein. For example, this disclosure includes bispecific antigen-binding molecules that specifically bind to human cell lines expressing CD28 but not PD-L1 (e.g., Jurkat cells) and HEK cell lines that do not express PD-L1 and CD28. In some embodiments, the bispecific antigen-binding molecules are expressed at a concentration of less than 1 × 10⁻⁶. -5 M's EC 50 The value binds to human or cynomolgus monkey T cells expressing CD28. In some embodiments, the bispecific antigen-binding molecule is expressed at a concentration of 1 × 10⁻⁶. -12 M to 1×10 -5 M's EC 50 The value binds to human or cynomolgus monkey T cells expressing CD28. In some embodiments, the bispecific antigen-binding molecule is expressed at a concentration of 1 × 10⁻⁶. -12 M to 1×10 -9 M's EC 50 The value binds to human or cynomolgus monkey T cells expressing CD28. In some embodiments, the bispecific antigen-binding molecule is at a value less than about 2.5 × 10⁻⁶. -8 M's EC 50 It binds to the surface of cell lines expressing PD-L1. The binding of bispecific antigen-binding molecules to the surface of cells or cell lines can be measured by in vitro FACS binding assays, as described in the examples.

[0186] This disclosure includes a PD-L1xCD28 bispecific antigen-binding molecule capable of depleting tumor cells in a subject. For example, according to certain embodiments, a PD-L1xCD28 bispecific antigen-binding molecule is provided, wherein a single administration of the antigen-binding molecule to a subject at a therapeutically effective dose results in a reduction in the number of tumor cells in the subject.

[0187] This disclosure includes an anti-PD-L1 x anti-CD28 bispecific antigen-binding molecule capable of binding to PD-L1 expressed on cell surfaces. A variety of tumor cells express PD-L1, including breast cancer cells (e.g., HeLa, MCF-7, and MDA-MB-231), melanoma cells (e.g., A375), lung cancer cells (e.g., HCC44), ovarian cancer cells (e.g., ES-2, SNU-8, MCAS), pancreatic cancer cells (e.g., SNU-324), and prostate cancer cells (e.g., DU145). Therefore, the bispecific antibody of this disclosure may prove useful in treating a variety of cancer indications.

[0188] This disclosure includes an anti-PD-L1 x anti-CD28 bispecific antigen-binding molecule capable of activating T cells by binding to PD-L1 on target cells and CD28 on T cells (see Example 3). For example, binding of the anti-PD-L1 x anti-CD28 bispecific antigen-binding molecule to T cells can lead to increased IL-2 release (see Example 4). Therefore, the bispecific antigen-binding molecule of this disclosure may be useful in promoting T cell-mediated immune responses.

[0189] This disclosure includes an anti-PD-L1 x anti-CD28 bispecific antigen-binding molecule capable of blocking the interaction between PD-L1 and PD-1 (see Example 5). Therefore, the bispecific antigen-binding molecule of this disclosure can be used to inhibit the immune checkpoint pathway and reduce T cell exhaustion, thereby promoting T cell-mediated immune responses.

[0190] This disclosure includes an anti-PD-L1 x anti-CD28 bispecific antigen-binding molecule that can enhance the cytotoxic efficacy of anti-tumor-associated antigen (TAA) x anti-CD3 bispecific antibodies against a variety of tumor cells (see Example 4). In some embodiments, TAA is selected from AFP, ALK, BAGE protein, BCMA, BIRC5 (survival protein), BIRC7, β-linkin, brc-abl, BRCA1, BORIS, CA9, carbonic anhydrase IX, caspase-8, CALR, CCR5, CD19, CD20 (MS4A1), CD22, CD30, CD40, CDK4, CEA, CTLA4, cyclin-B1, CYP1B1, EGFR, EGFRvIII, ErbB2 / Her2, ErbB3, ErbB4, ETV6-AML, EpCAM, EphA2, Fra-1, FOLR1, GAGE ​​protein (e.g., GAGE-1, -2), GD2, GD3, GloboH, phosphatidylinositol proteoglycan-3, GM3, gp100, Her2, HLA / B-raf, HLA / k-ras, HLA / MAG E-A3, hTERT, LMP2, MAGE proteins (e.g., MAGE-1, -2, -3, -4, -6, and -12), MART-1, mesothelin, ML-IAP, Muc1, Muc2, Muc3, Muc4, Muc5, Muc16 (CA-125), MUM1, NA17, NY-BR1, NY-BR62, NY-BR85, NY-ESO1, OX40, p15, p53, PAP, PAX3, PAX5, PCTA-1, PLAC1, PRLR, PRAME, PSMA (FOLH1), RAGE protein, Ras, RGS5, Rho, SART-1, SART-3, STEAP1, STEAP2, TAG-72, TGF-β, TMPRSS2, Thompson-nouvelle antigen (Tn), TRP-1, TRP-2, tyrosinase, and urinary plaque protein-3. The anti-PD-L1 X anti-CD28 antibody can also be useful when combined with checkpoint inhibitors (e.g., antibodies against PD-1) or any other checkpoint inhibitor.

[0191] Epitope Mapping and Related Technologies

[0192] The epitopes on CD28 or PD-L1 that the antigen-binding molecules of this disclosure bind to may consist of a single continuous sequence of three or more amino acids (e.g., 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 or more) of the CD28 or PD-L1 protein. Alternatively, the epitope may consist of multiple non-continuous amino acids (or amino acid sequences) of CD28 or PD-L1. The antibodies of this disclosure may interact with amino acids contained in a CD28 monomer or with amino acids on two different CD28 chains of a CD28 dimer. As used herein, the term "epitope" refers to an antigenic determinant that interacts with a specific antigen-binding site in the variable region of an antibody molecule, called a complementary site. A single antigen may have more than one epitope. Thus, different antibodies may bind to different regions on an antigen and may have different biological effects. Epitopes may be conformational or linear. Conformational epitopes are generated by spatially juxtaposed amino acids from different segments of a linear polypeptide chain. Linear epitopes are epitopes formed from adjacent amino acid residues in a polypeptide chain. In some cases, epitopes may contain sugar, phosphoryl, or sulfonyl groups on the antigen.

[0193] Various techniques known to those skilled in the art can be used to determine whether an antibody's antigen-binding domain interacts with "one or more amino acids" in a peptide or protein. Exemplary techniques that can be used to determine epitopes or binding domains of a particular antibody or antigen-binding domain include, for example, conventional cross-blocking assays, such as those performed in... AntibodiesAs described in Harlow and Lane (Cold Spring Harbor Press, Cold Spring Harb., NY), methods include assays, site mutagenesis (e.g., alanine scan mutagenesis, arginine scan mutagenesis, etc.), Western blot analysis (Reineke, 2004, Methods Mol Biol 248:443-463), protease protection, and peptide cleavage analysis. Alternatively, methods such as epitope cleavage, epitope extraction, and antigen chemical modification can be used (Tomer, 2000, Protein Science 9:487-496). Another method for identifying amino acids in peptides to which antibodies interact is hydrogen / deuterium exchange detected by mass spectrometry. Generally, hydrogen / deuterium exchange methods involve deuterating the target protein, followed by binding the antibody to the deuterated protein. The protein / antibody complex is then transferred to water to allow hydrogen-deuterium exchange to occur at all residues except those protected by the antibody (which retains the deuterium label). After antibody dissociation, the target protein is subjected to protease cleavage and mass spectrometry analysis to reveal deuterium-labeled residues corresponding to the specific amino acids with which the antibody interacts. See, for example, Ehring (1999) Analytical Biochemistry 267(2):252-259; Engen and Smith (2001) Anal. Chem. 73:256A-265A. Alternatively, in some embodiments, the target protein binds to the antibody, followed by hydrogen-deuterium exchange. After antibody dissociation, the target protein is subjected to protease cleavage and mass spectrometry analysis to reveal non-deuterium-labeled residues corresponding to the specific amino acids with which the antibody interacts. X-ray crystal structure analysis can also be used to identify amino acids in the peptide with which the antibody interacts.

[0194] This disclosure also includes anti-CD28 and anti-PD-L1 antibodies that bind to the same epitopes as any particular exemplary antibody described herein (e.g., antibodies containing any amino acid sequences shown in Tables 1, 3, 6, and 9). Similarly, this disclosure also includes anti-CD28 and / or anti-PD-L1 antibodies that compete with any particular exemplary antibody described herein (e.g., antibodies containing any amino acid sequences shown in Tables 1, 3, 6, and 9) for binding to CD28 and / or PD-L1.

[0195] This disclosure also includes bispecific antigen-binding molecules comprising a first antigen-binding domain specifically binding to human CD28 and a second antigen-binding fragment specifically binding to human PD-L1, wherein the first antigen-binding domain binds to the same epitope on CD28 as any particular exemplary CD28-specific antigen-binding domain described herein, and / or wherein the second antigen-binding domain binds to the same epitope on PD-L1 as any particular exemplary PD-L1-specific antigen-binding domain described herein. Similarly, this disclosure also includes bispecific antigen-binding molecules comprising a first antigen-binding domain specifically binding to human CD28 and a second antigen-binding fragment specifically binding to human PD-L1, wherein the first antigen-binding domain competes with any particular exemplary CD28-specific antigen-binding domain described herein for binding to CD28, and / or wherein the second antigen-binding domain competes with any particular exemplary PD-L1-specific antigen-binding domain described herein for binding to PD-L1.

[0196] By using conventional methods known in the art, it is readily possible to determine whether a particular antigen-binding molecule (e.g., an antibody) or its antigen-binding domain binds to the same epitope as a reference antigen-binding molecule of this disclosure, or whether it competes with the reference antigen-binding molecule of this disclosure for binding. For example, to determine whether a test antibody binds to the same epitope on CD28 (or PD-L1) as a reference bispecific antigen-binding molecule of this disclosure, the reference bispecific molecule is first allowed to bind to the CD28 protein (or PD-L1 protein). Next, the ability of the test antibody to bind to the CD28 (or PD-L1) molecule is evaluated. After saturation binding with the reference bispecific antigen-binding molecule, if the test antibody is able to bind to CD28 (or PD-L1), it can be concluded that the test antibody does not compete with the reference bispecific antigen-binding molecule for binding to CD28 (or PD-L1), and / or there is steric interference between antibodies binding to different sites on the antigen. On the other hand, after saturation binding with the reference bispecific antigen-binding molecule, if the test antibody cannot bind to the CD28 (or PD-L1) molecule, the test antibody competes with the reference bispecific antigen-binding molecule of this disclosure for binding to CD28 (or PD-L1). Further routine experiments (e.g., peptide mutation and binding assays) can then be performed to determine whether the observed lack of test antibody binding is actually due to binding to the same epitope as the reference bispecific antigen-binding molecule, or whether steric hindrance (or other phenomena) is the cause of the observed lack of binding. Such experiments can be performed using ELISA, RIA, Biacore, flow cytometry, or any other quantitative or qualitative antibody binding assay available in the art. According to certain embodiments of this disclosure, if, for example, an excess of 1, 5, 10, 20, or 100 times, one antigen-binding protein inhibits the binding of another antigen-binding protein by at least 50%, but preferably 75%, 90%, or even 99% (as measured in a competitive binding assay), then the two antigen-binding proteins compete for binding to the antigen (see, for example, Junghans et al., Cancer Res. 1990:50:1495-1502). Alternatively, if substantially all amino acid mutations in the antigen that reduce or eliminate the binding of one antigen-binding protein also reduce or eliminate the binding of the other antigen-binding protein, then the two antigen-binding proteins may bind to the same epitope. If only a subset of amino acid mutations that reduce or eliminate the binding of one antigen-binding protein reduces or eliminates the binding of the other antigen-binding protein, then the two antigen-binding proteins may have “overlapping epitopes.”

[0197] To determine whether an antibody or its antigen-binding domain competes with a reference antigen-binding molecule for binding, the binding method described above is performed in two orientations: In the first orientation, the reference antigen-binding molecule is allowed to bind to the CD28 protein (or PD-L1 protein) under saturation conditions, and then the binding of the test antibody to the CD28 (or PD-L1) molecule is evaluated. In the second orientation, the test antibody is allowed to bind to the CD28 (or PD-L1) molecule under saturation conditions, and then the binding of the reference antigen-binding molecule to the CD28 (or PD-L1) molecule is evaluated. If, in both orientations, only the first (saturated) antigen-binding molecule is able to bind to the CD28 (or PD-L1) molecule, the following conclusion is drawn: the test antibody and the reference antigen-binding molecule compete for binding to CD28 (or PD-L1). As will be understood by those skilled in the art, the antibody competing for binding with the reference antigen-binding molecule may not necessarily bind to the same epitope as the reference antibody, but may spatially block the binding of the reference antibody by binding to overlapping or adjacent epitopes.

[0198] Preparation of antigen-binding domains and construction of bispecific molecules

[0199] Antigen-binding domains specific to a particular antigen can be prepared using any antibody production technique known in the art. Once obtained, two different antigen-binding domains specific to two different antigens (e.g., CD28 and PD-L1) can be appropriately arranged relative to each other using conventional methods to produce the bispecific antigen-binding molecule of this disclosure. (Exemplary forms of bispecific antibodies that can be used to construct the bispecific antigen-binding molecule of this disclosure are discussed elsewhere herein.) In some embodiments, individual components (e.g., heavy and light chains) of one or more multispecific antigen-binding molecules of this disclosure are derived from chimeric antibodies, humanized antibodies, or fully human antibodies. Methods for preparing such antibodies are known in the art. For example, VELOCIMMUNE can be used. TM The technique described herein involves preparing one or more of the heavy chain and / or light chain of the bispecific antigen-binding molecule. Using VELOCIMMUNE... TM The technique (or any other human antibody production technique) initially isolates a high-affinity chimeric antibody against a specific antigen (e.g., CD28 or PD-L1) having a human variable region and a mouse constant region. The antibody is characterized and selected for desired characteristics, including affinity, selectivity, epitopes, etc. The mouse constant region is replaced with the desired human constant region to produce a fully human heavy chain and / or light chain in a bispecific antigen-binding molecule that can be incorporated into this disclosure.

[0200] Genetically modified animals can be used to prepare human bispecific antigen-binding molecules. For example, genetically modified mice that cannot rearrange and express endogenous mouse immunoglobulin light chain variable sequences can be used, wherein the mice express only one or two human light chain variable domains encoded by human immunoglobulin sequences operatively linked to a mouse κ constant gene at an endogenous mouse κ locus. Such genetically modified mice can be used to generate fully human bispecific antigen-binding molecules comprising two distinct heavy chains associated with the same light chain containing a variable domain derived from one of the two distinct human light chain variable region gene segments. (See, for example, US 2011 / 0195454, which discusses in detail such modified mice and their use in generating bispecific antigen-binding molecules).

[0201] bioequivalent

[0202] This disclosure covers antigen-binding molecules having an amino acid sequence different from that of the antibodies described herein but retaining the ability to bind CD28 and PD-L1. When compared to the parental sequence, such variant molecules contain the addition, deletion, or substitution of one or more amino acids but exhibit substantially equivalent biological activity to the antigen-binding molecule described herein. Similarly, the DNA sequences encoding antigen-binding molecules in this disclosure cover sequences that contain one or more nucleotide additions, deletions, or substitutions when compared to the disclosed sequences, but encode antigen-binding molecules that are substantially bioequivalent to the antigen-binding molecules of this disclosure. Some examples of such variant amino acid and DNA sequences are discussed above.

[0203] This disclosure includes antigen-binding molecules that are bioequivalent to any of the exemplary antigen-binding molecules described herein. If, for example, two antigen-binding proteins or antibodies are administered at the same molar dose under similar experimental conditions, and their absorption rates and extents do not show significant differences, they are considered bioequivalent. If some antibodies are equivalent in extent of absorption but not in rate of absorption, said antibodies will be considered equivalents or drug substitutes and may still be considered bioequivalent because such differences in absorption rates are intentional and reflected in labeling, are not necessary to achieve effective in vivo drug concentrations (e.g., for long-term use), and are considered medically insignificant for the specific pharmaceutical product under investigation.

[0204] In one implementation, two antigen-binding proteins are bioequivalent if there are no clinically significant differences in their safety, purity, and potency.

[0205] In one implementation, the two antigen-binding proteins are bioequivalent if a patient can switch between the reference product and the biological product once or more, and there is no increased risk of anticipated adverse effects (including clinically significant alterations in immunogenicity or reduced efficacy) compared to continuous treatment without such switching.

[0206] In one implementation, if two antigen-binding proteins function for one or more conditions through a common mechanism of action, to the extent that such mechanism is known.

[0207] Bioequivalence can be demonstrated through in vivo and in vitro methods. Bioequivalence measurements include, for example, (a) in vivo testing in humans or other mammals, wherein the concentration of the antibody or its metabolites is measured over time in blood, plasma, serum, or other biological fluids; (b) in vitro testing that correlates with and reasonably predicts bioavailability data in humans; (c) in vivo testing in humans or other mammals, wherein the appropriate acute pharmacological effect of the antibody (or its target) is measured over time; and (d) in clinical trials in which a well-controlled trial has established the safety, potency, or bioavailability or bioequivalence of the antibody.

[0208] Bioequivalent variants of the exemplary bispecific antigen-binding molecules shown herein can be constructed, for example, by making multiple substitutions to residues or sequences or by deleting terminal or internal residues or sequences that are not required for biological activity. For example, cysteine ​​residues that are not essential for biological activity can be deleted or substituted with other amino acids to prevent the formation of unnecessary or incorrect intramolecular disulfide bridges during renaturation. In other cases, bioequivalent antibodies may comprise exemplary bispecific antigen-binding molecules shown herein that include amino acid changes that alter the glycosylation characteristics of the antibody, such as by eliminating or removing glycosylation mutations.

[0209] Species selectivity and species cross-reactivity

[0210] According to certain embodiments, this disclosure provides antigen-binding molecules that bind to human CD28 but not to CD28 from other species. This disclosure also provides antigen-binding molecules that bind to human PD-L1 but not to PD-L1 from other species. This disclosure further includes antigen-binding molecules that bind to both human CD28 and CD28 from one or more non-human species; and / or antigen-binding molecules that bind to both human PD-L1 and PD-L1 from one or more non-human species.

[0211] According to certain exemplary embodiments, an antigen-binding molecule is provided that binds to human CD28 and human PD-L1, and may or may not bind to (as appropriate) one or more of CD28 and / or PD-L1 in mice, rats, guinea pigs, hamsters, gerbils, pigs, cats, dogs, rabbits, goats, sheep, cattle, horses, camels, cynomolgus monkeys, marmosets, rhesus monkeys, or chimpanzees. For example, in one exemplary embodiment, a bispecific antigen-binding molecule is provided that comprises a first antigen-binding domain that binds to human CD28 and cynomolgus monkey CD28 and a second antigen-binding domain that specifically binds to human PD-L1.

[0212] Immunoconjugates

[0213] This disclosure covers PD-L1xCD28 antigen-binding proteins, such as antibodies or antigen-binding fragments, such as REGN6194 (“immunoconjugates”), that are conjugated to another portion (e.g., a therapeutic portion). In one embodiment, a PD-L1xCD28 antigen-binding protein (e.g., an antibody or antigen-binding fragment) is conjugated to any other therapeutic agent described herein. The term “immunoconjugate” as used herein refers to an antigen-binding protein, such as an antibody or antigen-binding fragment, that is chemically or biologically linked to another antigen-binding protein, drug, radioactive agent, reporter portion, enzyme, peptide, protein, or therapeutic agent.

[0214] In some embodiments, the therapeutic component may be a cytotoxic agent, a chemotherapeutic agent, an immunosuppressant, or a radioactive isotope. Cytotoxic agents include any agent that is harmful to cells. Examples of suitable cytotoxic agents and chemotherapeutic agents for forming immune conjugates are known in the art (see, for example, WO 05 / 103081).

[0215] Therapeutic uses of antigen-binding molecules

[0216] The bispecific antibodies and antigen-binding molecules (and therapeutic compositions comprising them) of this disclosure are particularly useful for treating any disease or condition in which stimulation, activation, and / or targeting of an immune response would be beneficial. Specifically, the PD-L1xCD28 bispecific antigen-binding molecule of this disclosure can be used to treat, prevent, and / or improve hyperproliferative diseases, such as cancer. In some embodiments, this disclosure provides a method for treating cancer in a subject, comprising administering to the subject a therapeutically effective dose of a PD-L1xCD28 antigen-binding protein, such as REGN6194.

[0217] For the purposes of this document, hyperproliferative disorders are diseases characterized by abnormal, excessive, and / or uncontrolled cell growth, such as cancers in which cells express PD-L1. Hyperproliferative disorders include, for example, cancers. Exemplary cancers include, but are not limited to, esophageal cancer, squamous cell carcinoma of the lung, lung adenocarcinoma, squamous cell carcinoma of the cervix, glioma, thyroid cancer, lung cancer (e.g., non-small cell lung cancer), colorectal cancer, colon cancer, bladder cancer, rectal cancer, head and neck cancer, gastric cancer, liver cancer, pancreatic cancer, kidney cancer, urothelial carcinoma, prostate cancer, testicular cancer, breast cancer, cervical cancer, endometrial cancer, ovarian cancer, gastroesophageal cancer (e.g., gastroesophageal adenocarcinoma), basal cell carcinoma, diffuse large B-cell lymphoma, or multiple myeloma and melanoma. Therefore, the antibodies and bispecific antigen-binding molecules of this disclosure can be used to treat a wide range of cancers.

[0218] Cancers characterized by solid tumor cells or cancerous blood cells can be PD-L1-expressing cancers, for example, in which PD-L1 expression has been identified in the cells of a specific subject to be treated, including esophageal cancer, lung squamous cell carcinoma, lung adenocarcinoma, cervical squamous cell carcinoma, endometrial adenocarcinoma, bladder urothelial carcinoma, lung cancer (e.g., non-small cell lung cancer), colorectal cancer, rectal cancer, endometrial cancer, skin cancer (e.g., head and neck squamous cell carcinoma), brain cancer (e.g., glioblastoma multiforme), breast cancer, gastroesophageal cancer (e.g., gastroesophageal adenocarcinoma), prostate cancer, ovarian cancer, melanoma, basal cell carcinoma, cervical cancer, diffuse large B-cell lymphoma, and / or multiple myeloma.

[0219] The antigen-binding molecules disclosed herein may also be used to treat primary and / or metastatic tumors (or cancers discussed herein) arising in, for example, the colon, lung, breast, ovary, kidney, and bladder.

[0220] The antigen-binding protein disclosed herein can also be used to treat residual cancer in subjects. As used herein, the term "residual cancer" means one or more types of cancer cells present or persistent in a subject after treatment with anticancer therapies.

[0221] As used herein, the term "subject" refers to a mammal (e.g., rat, mouse, cat, dog, cow, sheep, horse, goat, rabbit), preferably a human, for example, who needs to prevent and / or treat cancer. The subject may have cancer, may be susceptible to such a condition, and / or will benefit from the administration of the bispecific antibody or its antigen-binding fragment of the present disclosure. In one embodiment, the subject may have a hyperproliferative disease or be at risk of developing a hyperproliferative disease.

[0222] The method for treating or preventing cancer (e.g., PD-L1-expressing cancer) in a subject requiring treatment or prevention by administering a therapeutically effective dose of the PD-L1xCD28 antigen-binding protein associated with an additional therapeutic agent is part of this disclosure. Other therapeutic agents are disclosed elsewhere herein.

[0223] An “effective” or “therapeuticly effective” amount of PD-L1xCD28 antigen-binding protein (e.g., antibody or antigen-binding fragment) used to treat or prevent hyperproliferative diseases (e.g., cancer) is an amount of antigen-binding protein sufficient to alleviate one or more signs and / or symptoms of the disease in the treated subject, whether by inducing the remission or elimination of such signs and / or symptoms or by inhibiting the progression of such signs and / or symptoms. In one embodiment of this disclosure, the therapeutically effective dose of PD-L1xCD28 antigen-binding protein is from 0.1 to 2000 mg. The amount of dose may vary depending on the age and size of the subject to be administered, the target disease, the condition, the route of administration, etc. In some embodiments, a second or more subsequent doses of antigen-binding protein may be administered after the initial dose in an amount substantially the same as, less than, or more than the initial dose, wherein the subsequent doses may be spaced 1 to 8 weeks apart.

[0224] The dosage of antigen-binding molecules administered to a patient can vary depending on the patient's age and body size, target disease, symptoms, route of administration, etc. The preferred dosage is usually calculated based on body weight or body surface area. The frequency and duration of treatment can be adjusted according to the severity of the condition. Effective dosages and timing of bispecific antigen-binding molecules can be determined empirically; for example, patient progression can be monitored through regular assessments, and the dosage adjusted accordingly.

[0225] Combination therapy

[0226] The bispecific antigen-binding molecule of this disclosure can be used in combination with one or more pharmaceutical agents, for example, for treating cancer in a subject. In some embodiments, the bispecific antigen-binding molecule can be administered in combination with one or more pharmaceutical agents (e.g., corticosteroids) to reduce or improve one or more adverse side effects, such as cytokine storms. In some embodiments, the bispecific antigen-binding molecule can be administered in combination with one or more therapeutic agents or treatments to enhance the efficacy of cancer treatment. Exemplary additional therapeutic agents or treatments that can be administered in combination with or in combination with the antigen-binding molecules of this disclosure include, for example, chemotherapy (e.g., anticancer chemotherapy, such as paclitaxel, docetaxel, vincristine, cisplatin, carboplatin, or oxaliplatin), radiation therapy, surgery, checkpoint inhibitors, PD-1 inhibitors (e.g., anti-PD-1 antibodies, such as pembrolizumab, nivolumab, or cimipril), CTLA-4 inhibitors, LAG3 inhibitors, TIM3 inhibitors, GITR agonists, OX40 agonists, and 4-1B. B agonists, oncolytic viruses, cancer vaccines, CAR-T cells, nucleic acid therapeutics, stem cell transplantation, modified IL2, modified IL12, IL15, IL6 inhibitors (e.g., sarilumab or tocilizumab), IL4R inhibitors (e.g., dupilumab), EGFR inhibitors, Ang2 inhibitors, VEGF inhibitors, corticosteroids, bispecific antibodies binding to CD3 and tumor-associated antigens (TAAs) (e.g., MUC16, PSMA, STEAP2, or any TAA disclosed herein), or antigen-binding fragments thereof. Exemplary bispecific antibodies containing an antigen-binding domain that binds to CD3 include, but are not limited to, those described in, for example, WO2017 / 053856A1, WO2014 / 047231A1, WO2018 / 067331A1, and WO2018 / 058001A1. PD-L1 is expressed in a wide range of cancers. Therefore, the bispecific anti-PD-L1xCD28 antibody of this disclosure can be combined with a wide range of bispecific antibodies containing an antigen-binding domain that binds to CD3 for the treatment of various cancers.

[0227] Other agents may be administered shortly before, simultaneously with, or shortly after the administration of the antigen-binding molecule of this disclosure (for the purposes of this disclosure, such administration regimens are considered to be administration of the antigen-binding molecule in combination with other agents or therapeutics or treatments).

[0228] Pharmaceutical preparation and administration

[0229] This disclosure provides compositions comprising PD-L1xCD28 antigen-binding protein and one or more components; and methods of using such compositions and methods of preparing such compositions. Pharmaceutical formulations (e.g., aqueous pharmaceutical formulations comprising the PD-L1xCD28 antigen-binding protein of this disclosure and pharmaceutically acceptable carriers or excipients) are part of this disclosure.

[0230] The pharmaceutical compositions disclosed herein can be formulated with suitable carriers, excipients, and other agents that provide improved transfer, delivery, tolerability, etc. A variety of suitable formulations are available in all formulation sets known to medicinal chemists: Remington's Pharmaceutical Sciences, Mack Publishing Company, Easton, PA. These formulations include, for example, powders, pastes, ointments, gels, waxes, oils, lipids, and lipid-containing vesicles (e.g., LlPOFECTIN). TM Life Technologies, Carlsbad, CA), DNA conjugates, anhydrous absorbent pastes, oil-in-water and water-in-oil emulsions, carbowax emulsions (polyethylene glycol in various molecular weights), semi-solid gels, and semi-solid mixtures containing carbowax. See also Powell et al., "Compendium of excipients for parenteral formulations" PDA (1998) J Pharm Sci Technol 52:238-311.

[0231] To prepare pharmaceutical formulations of PD-L1xCD28 antigen-binding proteins, such as antibodies and their antigen-binding fragments (e.g., REGN6194), the antigen-binding protein is mixed with a pharmaceutically acceptable carrier or excipient. See, for example, Remington's Pharmaceutical Sciences and US Pharmacopeia: NationalFormulary, Mack Publishing Company, Easton, Pa. (1984); Hardman et al., (2001) Goodman and Gilman's The Pharmacological Basis of Therapeutics, McGraw-Hill, New York, NY; Gennaro (2000) Remington: The Science and Practice ofPharmacy, Lippincott, Williams, and Wilkins, New York, NY; Avis et al., (eds.) (1993) Pharmaceutical Dosage Forms: Parenteral Medications, Marcel Dekker, NY; Lieberman et al., (eds.) (1990) Pharmaceutical Dosage Forms: Tablets, MarcelDekker, NY; Lieberman et al., (eds.) (1990) Pharmaceutical Dosage Forms: DisperseSystems, Marcel Dekker, NY; Weiner and Kotkoskie (2000) Excipient Toxicity and Safety, Marcel Dekker, Inc., New York, NY. In one embodiment of this disclosure, the pharmaceutical preparation is sterile. Such compositions are part of this disclosure.

[0232] The pharmaceutical formulations disclosed herein contain PD-L1xCD28 antigen-binding protein and pharmaceutically acceptable carriers, including, for example, water, buffers, preservatives, and / or detergents.

[0233] The scope of this disclosure includes dried, such as lyophilized, compositions containing PD-L1xCD28 antigen-binding protein (e.g., an antibody or an antigen-binding fragment thereof), or pharmaceutical formulations containing a pharmaceutically acceptable carrier but substantially lacking water.

[0234] Various delivery systems are known and can be used to administer the pharmaceutical compositions disclosed herein, such as encapsulation in liposomes, microparticles, microcapsules, recombinant cells capable of expressing mutant viruses, and receptor-mediated endocytosis (see, for example, Wu et al., 1987, J. BioI. Chem. 262:4429-4432). Delivery methods include, but are not limited to, intradermal, intramuscular, intraperitoneal, intravenous, subcutaneous, intranasal, rectal, intestinal, epidural, and oral routes. The composition can be administered via any convenient route, such as by infusion or bolus injection, by absorption through an epithelial or mucocutaneous lining (e.g., oral mucosa, rectal and intestinal mucosa), and can be administered with other bioactive agents. Administration can be systemic or local.

[0235] As discussed herein, this disclosure provides containers (e.g., plastic or glass vials) or injection devices (e.g., syringes, pre-filled syringes, or autoinjectors) that contain any of the PD-L1xCD28 antigen-binding proteins described herein, such as antibodies or antigen-binding fragments thereof, or pharmaceutical formulations that contain pharmaceutically acceptable carriers or excipients.

[0236] The pharmaceutical compositions of this disclosure can be delivered subcutaneously or intravenously using standard needles and syringes. Regarding subcutaneous delivery, pen-type delivery devices known in the art can be used to deliver the pharmaceutical compositions of this disclosure. Such pen-type delivery devices can be reusable or disposable.

[0237] Many reusable and disposable pen-type and auto-injector-type delivery devices are used for subcutaneous delivery of the pharmaceutical compositions disclosed herein. See, for example, AUTOPEN TM (Owen Mumford, Inc., Woodstock, UK) or HUMIRA TM Pen (Abbott Labs, Abbott Park, IL).

[0238] In some cases, the pharmaceutical composition can be delivered using a controlled-release system. In one embodiment, a pump can be used (see Langer, ibid.; Sefton, 1987, CRC Crit. Ref. Biomed. Eng. 14:201). In another embodiment, polymeric materials can be used; see Medical Applications of Controlled Release, Langer and Wise (ed.), 1974, CRC Pres., Boca Raton, Florida. In yet another embodiment, the controlled-release system can be placed near the target of the composition, thus requiring only a portion of the systemic dose (see, for example, Goodson, 1984, in Medical Applications of Controlled Release, ibid., Vol. 2, pp. 115–138). Other controlled-release systems are discussed in a review in Langer, 1990, Science 249:1527–1533.

[0239] Injectable formulations may include dosage forms for intravenous, subcutaneous, intradermal, and intramuscular injection, infusion, etc. These injectable formulations can be prepared by known methods. For example, they can be prepared, for instance, by dissolving, suspending, or emulsifying the aforementioned antibodies or their salts in a sterile aqueous or oily medium conventionally used for injection. Aqueous media for injection include, for example, physiological saline and other isotonic solutions that can be used in combination with suitable solubilizers. Injectable oily media are also part of this disclosure. Such oily media can be combined with solubilizers.

[0240] Advantageously, the above-described pharmaceutical compositions for oral or parenteral use are prepared into dosage forms suitable for matching the dosage of the active ingredient. Such dosage forms include, for example, tablets, pills, capsules, injections (ampoules), suppositories, etc. The amount of the antibody contained is typically from about 0.1 to about 2000 mg per unit dose dosage form; particularly in the form of injections.

[0241] Diagnostic uses

[0242] The bispecific antibodies of this disclosure can also be used to detect and / or measure CD28 or PD-L1 in samples, or cells expressing CD28 or PD-L1, for example, for diagnostic purposes. For example, a PD-L1xCD28 antibody or its antigen-binding fragment can be used to diagnose conditions or diseases characterized by abnormal expression of CD28 or PD-L1 (e.g., overexpression, low expression, lack of expression, etc.). Exemplary diagnostic assays for CD28 or PD-L1 may include, for example, contacting a sample obtained from a patient with an antibody of this disclosure, wherein the antibody is labeled with a detectable marker or reporter molecule. Alternatively, an unlabeled antibody may be combined with a secondary antibody that is itself a detectable marker for diagnostic applications. The detectable marker or reporter molecule may be a radioisotope, for example... 3 H, 14 C 32 p、 35 S or 125 I; a fluorescent or chemiluminescent component, such as fluorescein isothiocyanate or rhodamine; or an enzyme, such as alkaline phosphatase, β-galactosidase, horseradish peroxidase, or luciferase. Specific exemplary assays that can be used to detect or measure CD28 or PD-L1 in a sample include enzyme-linked immunosorbent assay (ELISA), radioimmunoassay (RIA), and fluorescence-activated cell sorting (FACS). Samples that can be used for CD28 or PD-L1 diagnostic assays according to this disclosure include any tissue or fluid sample obtainable from a patient that contains detectable amounts of CD28 or PD-L1 protein or fragments thereof under normal or pathological conditions. Typically, the level of CD28 or PD-L1 in a specific sample obtained from a healthy patient (e.g., a patient without a disease or condition associated with abnormal CD28 or PD-L1 levels or activity) will be measured to initially establish a baseline or standard level of CD28 or PD-L1. The baseline level of CD28 or PD-L1 can then be compared with the level of CD28 or PD-L1 measured in samples obtained from individuals suspected of having CD28 or PD-L1-related diseases or conditions.

[0243] Example

[0244] The following embodiments are provided to provide those skilled in the art with a complete disclosure and description of how to prepare and use the compositions and methods of this disclosure, and are not intended to limit the scope of what the inventors consider to be their invention.

[0245] Example 1: Construction of anti-PD-L1xCD28 antibody

[0246] Production of anti-PD-L1 antibodies

[0247] Anti-PD-L1 antibodies were obtained by immunizing genetically modified mice containing DNA encoding the variable regions of the heavy and κ light chains of human immunoglobulins with human PD-L1 antigen.

[0248] Following immunization, spleen cells were harvested from each mouse and B cells were sorted using a human PD-L1 fragment as a sorting agent to bind to and identify reactive antibodies (antigen-positive B cells) (as described in US 2007 / 0280945).

[0249] Antibodies are characterized and selected for desired characteristics, including affinity and selectivity. Antibodies may have desired constant regions, such as wild-type or modified hIgG1 or hIgG4 constant regions. As those skilled in the art will understand, antibodies with a specific constant region (e.g., modified hIgG1) can be converted into antibodies with different constant regions (e.g., modified hIgG4). While constant regions may vary depending on the specific application, high-affinity antigen binding and target-specific characteristics are present in variable regions.

[0250] Table 1 shows the amino acid sequence identifiers of the variable regions and CDRs of the heavy and light chains of the selected anti-PD-L1 antibodies of this disclosure. The corresponding nucleic acid sequence identifiers are shown in Table 2.

[0251] Table 1: Amino acid sequence identifiers of the selected parental PD-L1 monoclonal antibodies

[0252]

[0253] Table 2: Nucleic acid sequence identifiers of the selected parental PD-L1 monoclonal antibodies

[0254]

[0255] Production of anti-CD28 antibodies

[0256] Anti-CD28 antibodies were obtained by immunizing VELOCIMMUNE® mice (i.e., modified mice containing DNA encoding the variable regions of the heavy and universal light chains of human immunoglobulins) with human CD28 protein fused with the Fc portion of mouse IgG2a, or with CD28-expressing cells, or with DNA encoding CD28.

[0257] Antibody immune responses are monitored using a CD28-specific immunoassay. When the desired immune response is achieved, anti-CD28 antibodies are isolated directly from antigen-positive B cells, as described in US7,582,298.

[0258] Antibodies are characterized and selected for desired characteristics, including affinity and selectivity. Antibodies may have desired constant regions, such as wild-type or modified hIgG1 or hIgG4 constant regions. As those skilled in the art will understand, antibodies with a specific constant region (e.g., modified hIgG1) can be converted into antibodies with different constant regions (e.g., modified hIgG4). While constant regions may vary depending on the specific application, high-affinity antigen binding and target-specific characteristics are present in variable regions.

[0259] Table 3 shows the amino acid sequence identifiers of the variable regions and CDRs of the heavy and light chains of the selected anti-CD28 antibodies of this disclosure. The corresponding nucleic acid sequence identifiers are shown in Table 4.

[0260] Table 3: Amino acid sequence identifiers of the selected parental CD28 monoclonal antibodies

[0261]

[0262] Table 4: Nucleic acid sequence identifiers of the selected parental CD28 antibodies

[0263]

[0264] Production of bispecific antibodies (bsAb) binding to CD28 and PD-L1

[0265] Bispecific antibodies containing an anti-PD-L1 specific binding domain and an anti-CD28 specific binding domain are constructed using standard methods, wherein each of the anti-PD-L1 antigen-binding domain and the anti-CD28 antigen-binding domain contains a different, unique HCVR that pairs with a common LCVR. In some cases, bispecific antibodies are constructed using the heavy chain from an anti-CD28 antibody, the heavy chain from an anti-PD-L1 antibody, and a common light chain (from either the anti-PD-L1 or anti-CD28 antibody). Table 5 summarizes the components (parental antibody names) of the selected bispecific PD-L1xCD28 antibodies. Tables 6 and 7 show the amino acid and nucleic acid identifiers of the selected bispecific antibodies, respectively. Table 8 shows the full-length heavy and light chain sequences of the selected bispecific antibodies. Additional bispecific antibodies binding to PD-L1 and CD28 can be prepared using parental monoclonal antibodies with the names shown in Table 9.

[0266] Table 5: Summary of the components of the selected anti-PD-L1 x anti-CD28 bispecific antibodies

[0267]

[0268] Table 6: Amino acid sequence identifiers of the selected anti-PD-L1 x anti-CD28 bispecific antibodies

[0269]

[0270] Table 7: Nucleic acid sequence identifiers of selected anti-PD-L1 x anti-CD28 bispecific antibodies

[0271]

[0272] Table 8. Amino acid and nucleotide sequences of the full-length immunoglobulin chains of bispecific antibodies REGN6192, REGN6193, and REGN6194

[0273]

[0274] D = nucleotide sequence of DNA encoding the specified sequence

[0275] P = amino acids of the polypeptide with the specified sequence.

[0276] The number refers to the SEQ ID NO of the specified sequence.

[0277] HC is the full-length heavy chain of the specified antibody.

[0278] LC refers to the full-length light chain of a specified antibody.

[0279] Additional bispecific antibodies comprising one HCVR arm from a parental PD-L1 antibody and another HCVR arm from a parental CD28 antibody can be prepared using the techniques described herein. The parental PD-L1 antibody used to generate these additional anti-PD-L1 X anti-CD28 bispecific antibodies has the HCVR sequence described in Table 1 above. The parental CD28 antibody used to generate these additional anti-PD-L1 X anti-CD28 bispecific antibodies has the amino acid sequence described in Table 3 above and WO 2020 / 198009. These anti-PD-L1 and anti-CD28 binding domains (paired) are shown in Table 9 below.

[0280] Table 9: Summary of parental antibody names for the HCVR arm of other anti-PD-L1 x anti-CD28 bispecific antibodies

[0281] Anti-PD-L1 antigen-binding domain (parental antibody names and sequences are shown in Table 1) Anti-CD28 antigen-binding domain (parental antibody names and sequences are shown in Table 3 of this article). mAb9364 mAb14193 mAb9364 mAb14216 mAb9364 mAb14226 mAb9373 mAb14193 mAb9373 mAb14216 mAb9373 mab14226

[0282] The bispecific antibodies described in the following examples consist of antigen-binding arms that bind to human hCD28 protein and human PD-L1 (see, for example, the following Biacore binding data). Exemplary bispecific antibodies comprise a modified (chimeric) IgG4 Fc domain as shown in U.S. Patent No. 9,359,437.

[0283] The bispecific antibody generated according to this embodiment comprises two separate antigen-binding domains (i.e., binding arms). The first antigen-binding domain comprises a heavy chain variable region (“CD28-VH”) derived from the anti-CD28 antibody, and the second antigen-binding domain comprises a heavy chain variable region (“PD-L1-VH”) derived from the anti-PD-L1 antibody. Both anti-PD-L1 and anti-CD28 share a common light chain. The CD28-VH / PD-L1-VH pairing produces an antigen-binding domain that can be used to target CD28 on T cells and PD-L1 on, for example, tumor cells and antigen-presenting cells.

[0284] Example 2: Characterization of bispecific antibodies binding to CD28 and PD-L1 by surface plasmon resonance

[0285] PD-L1 kinetics: Using real-time surface plasmon resonance (SPR) biosensor technology and a Biacore S200 instrument, the equilibrium dissociation constant (K0) for the binding of human PD-L1 (hPD-L1.mmH, SEQ ID NO: 71) expressed with the C-terminal myc-myc-hexahistine tag to purified anti-PD-L1xCD28 antibody was determined. D The CM5 Biacore sensor surface was derivatized by amine conjugation with a monoclonal mouse anti-human Fc antibody. All Biacore binding studies were performed in a buffer (HBS-EP run buffer) consisting of 0.01 M HEPES pH 7.4, 0.15 M NaCl, 3 mM EDTA, and 0.05% v / v surfactant P20. Different concentrations of hPD-L1.mmH (0.37 nM to 30 nM in 3-fold serial dilutions), prepared in the HBS-EP run buffer, were injected onto the captured antibody at a flow rate of 50 µL / min. Antibody-reagent association was monitored for 5 min, while dissociation in the HBS-EP run buffer was monitored for 10 min. At the end of each cycle, the antibody capture surface was regenerated by a 12-second injection of 20 mM phosphate. All binding kinetics experiments were performed at 25 °C.

[0286] CD28 kinetics: Using real-time surface plasmon resonance biosensor technology and a Biacore T200 instrument, the equilibrium dissociation constant (K0) for the binding of human CD28 (hCD28.mFc, SEQ ID NO: 72) expressed with a C-terminal mouse Fc tag to purified anti-PD-L1xCD28 antibody was determined. DThe CM5 Biacore sensor surface was derivatized by amine conjugation with a polyclonal goat anti-mouse antibody (anti-mFc, Cytiva). All Biacore binding studies were performed in a buffer (HBS-EP run buffer) consisting of 0.01 M HEPES pH 7.4, 0.15 M NaCl, 3 mM EDTA, and 0.05% v / v surfactant P20. Different concentrations of antibody prepared in the HBS-EP run buffer (10 nM to 90 nM in a 3-fold serial dilution) were injected into the captured hCD28.mFc at a flow rate of 50 µL / min. Antibody-reagent association was monitored for 4 min, and dissociation in the HBS-EP run buffer was monitored for 5 min. At the end of each cycle, the antibody capture surface was regenerated by injecting 10 mM glycine (pH 1.5) for 40 s. All binding kinetics experiments were performed at 25 °C.

[0287] Data Analysis: Specific SPR-Biacore sensor maps were obtained using a double-reference procedure. Double referencing was performed by subtracting the signal from the reference surface (anti-hFc or anti-mFc) for each injection from the signal on the experimental surface (anti-hFc captured by anti-PD-L1xCD28 antibody or anti-mFc captured by hCD28.mFc), thus removing the influence of refractive index variations. Furthermore, run buffer injections were performed to allow subtraction of signal changes caused by the dissociation of the captured antibody or antigen from the conjugated anti-hFc or anti-mFc surface. The kinetic association rate constant (ka) and dissociation rate constant (k) were also analyzed. d The real-time sensor image was fitted to a 1:1 model using Scrubber v2.0c curve fitting software to determine the model. This was combined with the dissociation equilibrium constant (K). D The dissociation half-life (t½) and the dissociation half-life are calculated from the kinetic rate constant as follows:

[0288]

[0289] Biacore analysis showed that PD-L1xCD28 has a K0 ratio of approximately 4.5E-11 to 1.2E-10. D Combined with hPD-L1, and with K at approximately 4.7E-08 to 1.3E-08 D Combined with hCD28 (Tables 10 and 11).

[0290] Table 10: Kinetic and equilibrium binding parameters of hPD-L1 to surface-captured anti-PD-L1xCD28 antibody at 25 °C

[0291]

[0292] NB: No binding

[0293] Table 11: Kinetic and equilibrium binding parameters of hCD28 to surface-captured anti-PD-L1xCD28 antibody at 25 °C

[0294]

[0295] NB: No binding

[0296] Example 3: Characterization of the binding of bispecific PD-L1xCD28 antibody to cells expressing PD-L1 or CD28

[0297] The binding of the hPD-L1xCD28 bispecific antibody to cells was characterized using flow cytometry. Binding to the PD-L1 arm was tested using HEK293 cells modified to express PD-L1 (HEK293 / hCD20 / hPSMA / hPD-L1). Binding to the CD28 arm was assessed using Jurkat / NFkB-Luc cells expressing endogenous CD28. Nonspecific binding was evaluated using HEK293 cells lacking both PD-L1 and CD28 expression (HEK293 / hCD20 / hPSMA). Binding was detected by using labeled secondary antibodies and measuring fluorescence on a flow cytometer. The results are shown in Table 12.

[0298] Cell lines: Jurkat / NFkB-Luc (ACL12421) are Jurkat cells stably transduced using the nuclear factor kappa-light-chain enhancer of activated B cells (NF-κB)-luciferase reporter construct; the cell line was maintained in RPMI-1640 + 10% FBS + L-Glu / PS + 1ug / mL puromycin. HEK293 / hCD20 / hPSMA / hPD-L1 (ACL11384) are HEK293 / hCD20 / hPSMA cells stably transduced using the human programmed death-ligand-1 construct; the cell line was maintained in DMEM high glucose + 10% FBS + L-Glu / PS + 1ug / mL puromycin + 500ug / mL G418. HEK293 / hCD20 / hPSMA (ACL11383) was a cell line obtained by transfecting HEK293 cells with an hCD20 construct and selecting them with neomycin (G418), followed by flow cytometry sorting to obtain cells with high CD20 expression. The transduced cells were then stabilized with a human prostate-specific membrane antigen construct and subsequently sorted for high hPSMA expression. The cell line was maintained in DMEM with high glucose + 10% FBS + L-Glu / PS + 500ug / mLG418.

[0299] Assay Setup: HEK293 + / - PD-L1 cells were lifted with trypsin, washed, and resuspended in staining buffer (2% FBS in PBS). Similarly, Jurkat / NFkB-Luc cells were centrifuged and resuspended in staining buffer. Cells were counted, and 80 μL of 200,000 cells / well was added to each well. Subsequently, 20 μL of 5x antibody was added to the cells at a 9-point 1:4 titration from 100 nM to 6 pM (final concentration in the well). Cells and antibody were incubated on ice for 30 min, followed by washing in staining buffer. Cells were resuspended in 2 μg / ml allophycocyanin (APC) conjugated goat anti-human secondary antibody. A separate secondary antibody control was included. Cells and secondary antibody were incubated on ice for 30 min, followed by washing in staining buffer. After washing with PBS (FBS-free), cells were resuspended in viability dye (reconstituted in DMSO according to the manufacturer's protocol and diluted 1:1000 in PBS) and incubated on ice for 30 min. They were then washed with staining buffer and resuspended in 2% PFA, and incubated overnight. After washing with staining buffer, cells were filtered and analyzed by flow cytometry. EC50 of the antibody was determined using a 4-parameter logistic equation on a 9-point dose-response curve (including secondary antibody control only) using GraphPad Prism software. 50 value.

[0300] Controls: In addition to the isotype control, the following controls were used: Control 1: Bispecific antibody with the following: one arm binding to CD28 (derived from parental antibody mAb14226) and the other arm binding to unrelated antigen; Control 2: Bispecific antibody with the following: one arm binding to CD28 (derived from parental antibody mAb14216) and the other arm binding to unrelated antigen; and Control 3: Bispecific antibody with the following: one arm binding to CD28 (derived from parental antibody mAb14193) and the other arm binding to unrelated antigen.

[0301] Binding on Jurkat / NFkB-Luc cells: Dose-dependent binding of PD-L1xCD28 antibodies (REGN6192, REGN6193, and REGN6194) was observed in the presence of Jurkat / NFkB-Luc cells, with REGN6194 exhibiting the strongest potency and highest maximum binding. Notably, only REGN6194 showed the ability to generate EC. 50The binding values ​​were not high because binding with REGN6192 and REGN6193 did not saturate at the highest tested concentration. Similar binding was observed for the bispecific control antibodies (Control 1, Control 2, and Control 3), with the antibody having the same CD28 arm as REGN6194 (Control 1) showing the strongest potency and highest maximum binding. In contrast, the CD28 bivalent control antibodies mAb14193, mAb14216, and mAb14226, corresponding to the PD-L1xCD28 antibodies REGN6192, REGN6193, and REGN6194, respectively, showed stronger potency and higher maximum binding. The PD-L1 bivalent control antibodies mAb9364 and mAb9373 showed low maximum binding on Jurkat / NFkB-Luc cells, however, it was dose-dependent, suggesting that endogenous expression of PD-L1 may be low on these cells. The allotype control antibody did not bind to Jurkat / NFkB-Luc cells.

[0302] Binding on HEK293 and HEK293 / PDL1 cells: Low-dose-dependent binding of PD-L1xCD28 and the PD-L1 bivalent control was observed on HEK293 cells lacking PD-L1 overexpression, potentially indicating low endogenous PD-L1 expression levels. Notably, the maximum binding was not significantly higher than that observed in the isotype control. Conversely, for HEK293 cells modified to overexpress hPD-L1, both PD-L1xCD28 and the PD-L1 control antibody showed high maximum dose-dependent binding in the low to sub-nanomolar range. The CD28 bivalent and bispecific control antibodies did not bind to HEK293 cells in a dose-dependent manner.

[0303] Table 12: Maximum binding and EC50 of antibody binding 50 value

[0304]

[0305] Geometric mean fluorescence intensity (gMFI) values ​​were plotted using GraphPad Prism, and the EC50 values ​​of the antibody were determined on a 9-point dose-response curve (1:4 titration of the antibody ranging from 100 nM to 6 pM) and under antibody-free conditions using a four-parameter, variable slope, nonlinear regression equation.

[0306] The maximum gMFI is the highest MFI along the dose-response curve, and the fold change is the maximum MFI divided by the MFI value of the secondary antibody alone.

[0307] Abbreviations: ND: Undetermined because dose-dependent binding was not observed; NC: Not calculated because the data did not conform to the 4-parameter logistic equation.

[0308] Example 4: Activation of T cells by PD-L1xCD28 antibody

[0309] Two signals, “Signal 1” and “Signal 2”, are essential for proper T cell activation. “Signal 1” is induced by the binding of the T cell receptor (TCR) on the T cell to a peptide on the antigen-presenting cell (APC) via the major histocompatibility complex (MHC) molecule. “Signal 2” is provided by the binding of a co-stimulatory receptor on the T cell (e.g., CD28) to its ligand (e.g., differentiation clusters 80 or 86 on the APC (CD80 / CD86)) (Martin et al., 1986; June et al., 1987; Harding et al., 1992). Therefore, activation of CD28 signaling provides a targeted approach to enhance existing TCR signaling.

[0310] The PD-L1xCD28 bispecific antibody was designed to mimic the natural ligand of CD28 by bridging PD-L1+ target cells with CD28+ T cells to provide “signal 2” to enhance T cell activation in the presence of “signal 1” provided by tumor-associated antigen (TAA) x CD3.

[0311] T cell activation was characterized using reported assays.

[0312] The ability of a PD-L1xCD28 bispecific antibody to activate human primary T cells by binding PD-L1 on target cells to CD28 on T cells to deliver “signal 2” was evaluated in a modified reporter assay. In this assay, Jurkat cells (endogenously expressing CD28) were modified to express the reporter gene luciferase (NFkB-Luc) under the control of the transcription factor NF-κB. Target cells were modified to express CD20 and PSMA alone or in combination with PD-L1. Reporter Jurkat cells were incubated with target cells (HEK293 / hCD20 / hPSMA and HEK293 / hCD20 / hPSMA / hPD-L1) and a bispecific CD20xCD3 antibody to provide “signal 1”, and the ability of the PD-L1xCD28 antibody to specifically bind to PD-L1 on target cells and promote CD28 aggregation / NFkB-mediated activation of luciferase production was evaluated.

[0313] Experimental procedure: One day prior to the experiment, Jurkat reporter cells were divided into 5 × 10⁶ cells in RPMI + 10% FBS + penicillin / streptomycin / L-glutamine (P / S / G) + 1 μg / ml puromycin growth medium. 5 Cells / mL. On the day of the experiment, Jurkat / NFkB-luc cells were resuspended in assay medium (RPMI supplemented with 10% FBS + P / S / G) and cultured at 5 × 10⁻⁶ cells / mL. 4 The final concentration of cells / well was added to 96-well white plates. HEK293 / hCD20 / hPSMA and HEK293 / hCD20 / hPSMA / hPD-L1 were detached with trypsin, washed, and resuspended in assay medium. Cells were added at 1×10⁻⁶ cells / well. 4 The final concentration of cells / well was added to the appropriate well of a 96-well white plate. A constant concentration of CD20xCD3 was prepared in assay medium and added to the appropriate well at a final concentration of 0.1 nM. Subsequently, bispecific PD-L1xCD28, monospecific or bispecific controls, or allotype controls were titrated at a 1:5 dilution from 51.2 fM to 500 nM, with the endpoint of the 12-point dilution being free of titrating antibody. All titrations were performed in duplicate and added to the appropriate well. Antibody dilutions were generated in assay medium. The plate was incubated at 37°C and 5% CO2 for 5 hours, and then ONE-Glo luciferase substrate was added to each well according to the manufacturer's instructions. Luciferase activity was recorded as a luminescent signal using an ENVISION plate reader and expressed as a relative light unit (RLU). GraphPad Prism was used. TM EC is determined using a 4-parameter logic equation on the 12-point response curve. 50 Value. The signal recorded at point 12 on the dilution curve (without titration antibody) is plotted at 10 fM. The maximum RLU is given as the average maximum response detected within the range of reagent volumes.

[0314] The controls used included: Control 4: a bispecific antibody with one arm binding to CD28 (derived from parental antibody mAb14226) and another arm binding to an unrelated antigen; Control 5: a bispecific antibody with one arm binding to PD-L1 (derived from parental antibody mAb9373) and another arm binding to an unrelated antigen; and Control 6: a bispecific antibody with one arm binding to PD-L1 (derived from parental antibody mAb9364) and another arm binding to an unrelated antigen.

[0315] Results: In the presence of Jurkat / NFkB-Luc cells and CD20-expressing target cells, the addition of CD20xCD3 resulted in increased baseline activity compared to the absence of CD20xCD3. Furthermore, in the absence of CD20xCD3, the bivalent CD28 antibody (mAb14226) caused a slight dose-dependent increase in NFkB activity, which was significantly increased in the presence of "signal 1" (CD20xCD3).

[0316] In the presence of HEK293 / hCD20 / hPSMA / hPD-L1 target cells and CD20xCD3 primary stimulation, PD-L1xCD28 molecules (REGN6192, REGN6193, and REGN6194) resulted in a dose-dependent increase in luciferase activity, while the allotype control antibody did not. In the absence of CD20xCD3 primary stimulation, only the bivalent CD28 molecule (mAb14226) and one PD-L1xCD28 molecule (REGN6192) resulted in a slight increase in activity at the highest tested antibody concentration, but the potency could not be determined due to the lack of a signal plateau phase.

[0317] In the presence of primary CD20xCD3 stimulation and in HEK293 / hCD20 / hPSMA target cells lacking PD-L1, REGN6193 and Control 5 did not increase NFkB activity, while REGN6192 and REGN6194 caused a slight increase at the highest tested concentration, but the response was weaker than the non-TAA x CD28 control (Control 4), and failed to produce EC50 values ​​due to the lack of a signal plateau phase. Furthermore, the allotype control antibody did not produce any signal. In the absence of primary CD20xCD3 stimulation, PDL1xCD28 molecules did not produce a dose-response curve in the presence of HEK293 / hCD20 / hPSMA cells.

[0318] The luciferase activity and potency values ​​of the antibodies are shown in Table 13.

[0319] Table 13: Maximum luciferase activity and potency of antibodies

[0320]

[0321] Abbreviations: ND: Undetermined because no dose-dependent response was observed; NC: Not calculated because the data do not conform to the 4-parameter logistic equation.

[0322] Characterization of T cell activation using human T cells

[0323] In the presence of a human embryonic renal cell carcinoma line (HEK293 / HCD20 / HPSMA / hPD-L1) modified to express hCD20, hPSMA, and hPD-L1, the ability of the PD-L1xCD28 bispecific antibody to activate human primary T cells by conjugating PD-L1 to CD28 to deliver "signal 2," as determined by IL2 release, was evaluated using a bispecific CD20xCD3 antibody as "signal 1." HEK293 cells expressing only hCD20 and hPSMA were included as a control to measure the activity that could occur in the absence of PD-L1 on APCs.

[0324] Experimental Procedure: Human peripheral blood mononuclear cells (PBMCs) were isolated from healthy donor leukocyte packs from Precision for Medicine (donor 555060) using the following protocol: 15 mL of density gradient medium (FicollPaque Plus) was added to a 50 mL test tube. The leukocyte pack was diluted 1:1 with PBS + 2% FBS. The diluted leukocyte pack was added to the top of the density gradient medium. The tube was centrifuged at 400G for 30 minutes at room temperature, then braked off. The upper plasma layer was removed and discarded. The mononuclear cell layer at the plasma-density gradient medium interface was removed and retained. PBS + 2% FBS was added to the collected mononuclear cell layer, and the tube was centrifuged at 300G for 8 minutes. The supernatant was discarded, and the resulting PBMCs were resuspended in an appropriate culture medium. CD4+ was used. + T-cell isolation kit (Miltenyi Biotech.) and CD4 isolation according to the manufacturer's recommended instructions. + T cells.

[0325] IL2 release assay: CD4+ enriched cells resuspended in stimulation medium were... + T cells at 1×10 5 A concentration of cells / well was added to 96-well round-bottom plates. Growth-inhibiting (arresting) HEK293 / hCD20 / hPSMA or HEK293 / hCD20 / hPSMA / hPD-L1 was added at a concentration of 1×10⁻⁶ cells / well. 4 The final concentration of cells / well added to CD4 +T cells. After cell addition, a constant 0.2 nM of CD20xCD3 or its matched allotype control was added to wells containing HEK293 / hCD20 / hPSMA or HEK293 / hCD20 / hPSMA / hPD-L1. Subsequently, bispecific PD-L1xCD28, bivalent CD28, bivalent PD-L1, bispecific control, or allotype control antibody were titrated at 1:3 dilutions from 15 pM to 100 nM and added to the wells. The endpoint of the 10-point dilutions was free of titrating antibody (CD20xCD3 or allotype control antibody only). The plates were incubated at 37°C, 5% CO2 for 72 hours, and 5 μL of total supernatant was removed for IL2 measurement. The amount of cytokines in the supernatant was determined according to the manufacturer's protocol using the AlphaLisa kit from PerkinElmer. Cytokine measurements were obtained on the Perkin Elmer Envision multi-label plate reader and reported as RFU (relative fluorescence units). All serial dilutions were tested in duplicate. GraphPad Prism was used. TM The software determines the antibody's EC50 using a four-parameter logistic equation on a 10-point dose-response curve. 50 Values, where point 10 (antibody without titration) is represented by 5.1 pM. Maximum IL2 is given as the average maximum response detected within the range of reagent titrations.

[0326] Comparison used: as disclosed in the above embodiments.

[0327] Results: In the presence of T cells and CD20xCD3, target cells expressing PD-L1 resulted in decreased baseline activity compared to target cells lacking PD-L1. Furthermore, in the absence of CD20xCD3, none of the tested antibodies led to a dose-dependent increase in IL-2 release.

[0328] In the presence of HEK293 / hCD20 / hPSMA / hPD-L1 target cells and CD20xCD3 primary stimulation, PD-L1xCD28 molecules (REGN6192, REGN6193, and REGN6194) resulted in a dose-dependent increase in IL-2 release, while allotype control antibodies did not. Of all tested molecules, REGN6194 resulted in the highest maximal cytokine release. Increased IL-2 was observed in xCD28 bispecific controls (Control 1, Control 2, and Control 3), however, this was only observed at the highest tested concentrations; and EC was not generated because the signal did not reach a plateau phase. 50Value. In the presence of CD20xCD3, the bivalent CD28 antibody mAb14226 also led to a dose-dependent increase in IL-2 release, while other CD28 bivalent antibodies mAb14193 and mAb14216 did not.

[0329] In the presence of primary CD20xCD3 stimulation and in target cells lacking PD-L1, REGN6192 and REGN6194 led to a dose-dependent increase in IL-2 release, comparable to the corresponding xCD28 controls (Control 2 and Control 3, respectively), and to a much weaker degree compared to the response observed in target cells with high PD-L1 expression. In the presence of the CD20xCD3 antibody, the bivalent CD28 molecule resulted in a slight dose-dependent increase in IL-2 release.

[0330] The IL2 release and potency values ​​of the antibodies are shown in Table 14.

[0331] Table 14: Maximum IL2 Release and Potential Value of Antibodies

[0332]

[0333] Abbreviations: ND: Undetermined because no dose-dependent response was observed; NC: Not calculated because the data do not conform to the 4-parameter logistic equation.

[0334] Example 5: Blocking of PD-L1-PD1 binding by PD-L1xCD28 antibody

[0335] Characterization of PD-L1 blockade using ELISA

[0336] An ELISA-based blocking assay was developed to determine the ability of the PD-L1xCD28 bispecific antibody to block the binding of human programmed death-ligand 1 (hPD-L1) to human programmed cell death protein 1 (hPD1).

[0337] The recombinant human PD-L1 protein used in the experiment contains the hPD-L1 extracellular domain (amino acids F19 to T239), which is expressed at the C-terminus along with the Fc portion of mouse IgG2a (amino acids E98 to K330) (hPD-L1-mFc, accession number NP_054862.1). The human PD1 protein contains the hPD1 extracellular domain (amino acids L25 to V170; C93S), which is expressed at the C-terminus along with the Fc portion of human IgG1 (amino acids D104 to K330) (hPD1-hFc, accession number NP_005009.2).

[0338] In the PD-L1 blocking assay, hPD1-hFc protein (2 μg / ml in PBS) was coated onto 96-well microtiter plates overnight at 4°C. Non-specific binding sites were then blocked using 0.5% (w / v) BSA in PBS. In separate 96-well microtiter plates, a fixed amount of 0.7 nM hPD-L1-mFc was bound for 1 hour to PD-L1xCD28 bispecific antibody, its parental bivalent anti-PD-L1, anti-CD28, or related human allotype antibody diluted in PBS + 0.5% BSA in the range of 1.7 pM to 100 nM. The fixed concentration of hPD-L1-mFc was selected to be close to the 50% maximum binding (EC50) of the hPD1 plate. 50 The concentration of hPD-L1-mFc antibody complex was then transferred to an hPD1-coated plate. After incubation at room temperature for 1 hour, the plate was washed and the plate-bound hPD-L1-mFc protein was detected using a horseradish peroxidase (HRP)-conjugated goat anti-mouse Fcγ fragment-specific antibody. The plate was then developed using TMB substrate solution (BD Biosciences) according to the manufacturer's recommendations, and the absorbance was measured at 450 nm on a SpectraMax I3x plate reader.

[0339] Combining data with an S-shaped (four-parameter logic) dose-response model using GraphPad Prism TM The software performs the analysis. The calculated IC... 50 The value (defined as the antibody concentration required to block 50% binding of hPD-L1-mFc to plate-coated hPD1-hFc) is used as an indicator of blocking efficacy. The percentage of blockade by the test antibody at the highest test concentration of 100 nM is calculated based on the following formula:

[0340]

[0341] Antibodies that block more than 50% binding at the highest tested concentration are classified as blocking agents, and their IC50 is reported. 50 value.

[0342] Results: Three PD-L1xCD28 bispecific antibodies (REGN6192, REGN6193, and REGN6194) showed concentration-dependent blocking of hPD-L1-hPD1 binding, with 97% to 99% blockade at the highest tested antibody concentration (100 nM). IC50 values ​​for these bispecific antibodies were also measured. 50The values ​​ranged from 0.2 nM to 1.2 nM (Table 15). Their parental bivalent PD-L1 antibodies (mAb9364 and mAb9373) showed similar 99% blockade percentages and IC50 values ​​of 0.49 nM and 0.52 nM, respectively. 50 The antibodies inhibited the binding of hPD-L1 to hPD1. All parental bivalent CD28 (mAb14226, mAb14193, and mAb14216) antibodies and isotype control antibodies showed no or low blocking on the binding of hPD-L1 to hPD1, with blocking percentages ranging from -1% to 19%. These antibodies were classified as non-blocking agents.

[0343] Table 15: Summary of PD-L1xCD28 bispecific Abs and their parental Abs blocking the binding of human PD-L1 to human PD1

[0344]

[0345] NBL: Non-blocking, %blocking less than or equal to 50%.

[0346] NA: Unavailable

[0347] Characterization of PD-L1 blockade in cells

[0348] Characterization of the PD-L1xCD28 bispecific antibody was performed using WSU-DLCL2, WSU-DLCL2 / hPD-L1, and Jurkat / AP1-luc / hPD1 cells in a PD-L1 blocking assay. This study investigated whether bivalent PD-L1, PD-L1xCD28, bispecific control antibodies, or related isotype controls could block the interaction between PD-L1 and PD-1. The CD20xCD3 bispecific antibody provided "signal 1" by binding CD20 (endogenously expressed) on WSU-DLCL2 target cells to CD3 on Jurkat / AP1-luc / hPD1 cells, leading to increased AP-1-Luc activity, as observed by the enhanced luminescent signal. WSU-DLCL2 cells modified to express PD-L1 resulted in decreased AP-1 activity. The ability of the PD-L1 antibody to block the PD-L1 interaction and subsequently restore luciferase signaling was evaluated.

[0349] Cell lines: Jurkat / AP1-Luc / hPD1 (ACL8709) were generated from Jurkat cells stably transduced using the human programmed cell death protein 1 construct; the cell line was maintained in RPMI-1640 + 10% FBS + L-Glu / PS + 1ug / mL puromycin. WSU-DLCL2 / hPD-L1 (ACL17386) were generated from WSU-DLCL2 cells stably transduced using the human programmed cell death ligand-1 construct; the cell line was maintained in RPMI-1640 + 10% FBS + L-Glu / PS + 1ug / mL puromycin. WSU-DLCL2 cells (HCT883) were maintained in RPMI-1640 + 10% FBS + L-Glu / PS.

[0350] Experimental setup: One day prior to the experiment, Jurkat reporter cells were divided into 5 × 10⁶ cells in RPMI + 10% FBS + penicillin / streptomycin / L-glutamine (P / S / G) + 1 μg / ml puromycin growth medium. 5 Cells / ml. Jurkat / AP1-Luc / PD1 was resuspended in assay medium (RPMI supplemented with 10% FBS + P / S / G) and incubated at 5 × 10⁻⁶ cells / ml. 4 A concentration of 10 cells / well was added to 96-well white plates. Antigen-presenting cells with or without PD-L1 expression (WSU-DLCL2 / PD-L1 or WSU-DLCL2, respectively) were also resuspended in assay medium at 2.5 × 10⁻⁶ cells / well. 4 Cell / well concentrations were added to the plate. Bispecific CD20xCD3 antibody was added to all wells at a constant concentration of 1 nM. Subsequently, PD-L1xCD28, associated monospecific or bispecific control, or isotype control antibodies were titrated from 7.6 pM to 500 nM at a 1:4 dilution, with the endpoint of 10-point dilutions being free of titrated antibody (constant CD20xCD3 antibody only). All titrations were performed in duplicate and added to the appropriate wells. Antibody dilutions were generated in the assay medium. The plate was incubated at 37°C and 5% CO2 for 5 hours, and then ONE-Glo luciferase substrate was added to each well according to the manufacturer's instructions. Luciferase activity was recorded as a luminescent signal using an ENVISION plate reader and expressed as relative light units (RLU). GraphPad Prism was used. TM EC is determined using a 4-parameter logic equation on the 10-point response curve. 50Values. The signal recorded at point 10 was plotted on a dilution curve at 1.9 pM. Maximum RLU is given as the average maximum response detected within the range of reagent doses. Anti-PD1 antibody (REGN2810; cimipril, LIBTAYO®) was used in the experiment along with other controls.

[0351] Results: In the presence of CD20xCD3 antibody, PD-L1-expressing WSU-DLCL2 cells showed reduced AP-1 activity compared to PD-L1-negative WSU-DLCL2 cells.

[0352] In the absence of PD-L1 expression in WSU-DLCL2 cells, PD-L1 bispecific antibodies and bivalent antibodies and controls did not affect AP-1 activity.

[0353] In the presence of PD-L1 expression on WSU-DLCL2 cells, PD-L1 bispecific, PD-L1 bivalent, REGN2810, and xPD-L1 bispecific control (Control 5 and Control 6) antibodies resulted in dose-dependent recovery of the luminescent signal, which was inhibited by the interaction between PD1 and PD-L1. The isotype control did not result in increased AP-1 activity (Table 16).

[0354] Table 16: Antibody potency and maximum RLU (AP-1 activity)

[0355]

[0356] Abbreviations: ND: Undetermined because no dose-dependent response was observed; NC: Not calculated because the data do not conform to the 4-parameter logistic equation.

[0357] Example 6: Enhancement of MUC16+ cell killing by combination of bispecific PD-L1xCD28 antibody and Muc16xCD3 antibody

[0358] The ability of co-stimulatory PD-L1xCD28 bispecific antibody REGN6194 to enhance the cytotoxic potency of REGN4019 (a bispecific antibody targeting CD3 and MUC16) (WO 2018 / 067331) was tested. MUC16 is a tumor antigen expressed on the surface of OVCAR-3 tumor cells modified to overexpress hPD-L1. The experiment included a control PD-L1x4-1BBbsAb (Comparator 1), which contains the variable regions “CD137-009-HC7LC2” and “PD-L1-547” (WO2019 / 025545).

[0359] To monitor the killing of MUC16+ cells by flow cytometry in the presence of a combination of hPBMCs and a bispecific antibody against MUC16xCD3 and PD-L1xCD28 (bsAb), OVCAR-3 / hPD-L1 cells were labeled with the fluorescent tracking dye Violet cell tracker. After labeling, cells were plated overnight at 37°C. Human PBMCs were separately labeled with 1 × 10⁻⁶ cells / mL. 6 Cells / mL plates were seeded in supplemented RPMI medium and incubated overnight at 37°C to enrich lymphocytes by depleting adherent macrophages, dendritic cells, and some monocytes. The next day, target cells were co-incubated at 37°C for 72 hours with initially depleted adherent PBMCs (effective cells / target cells ratio 4:1), serially diluted MUC16xCD3 bispecific antibody REGN4019 or xCD3 bispecific isotype control (control xCD3) (concentration range: 66.7 nM to 150 pM), and a fixed concentration of PD-L1xCD28 co-stimulatory bispecific antibody (2.5 ug / ml (16.7 nM)). Cells were removed from the culture plates using trypsin-EDTA dissociation buffer and analyzed by flow cytometry on a BD Celesta cell counter. For flow cytometry analysis, cells were stained with dead / live Near IR Reactive (Invitrogen) dye. Immediately before FACS analysis, 2E04 counting beads were added to each well. 1E04 beads were collected for each sample. To assess the specificity of killing, cells were gated to a viable, Violet-tagged population. The percentage of the viable population was recorded and used to calculate survival.

[0360] T cell activation and upregulation of PD1 markers were assessed by incubating cells with antibodies directly conjugated to CD2, CD4, CD8, CD25, and PD1, and by reporting the percentage of activated (CD25+ / CD8+, CD25+ / CD4+) T cells and PD1+ / CD4+, PD1+ / CD8+ T cells as a percentage of total T cells (CD2+).

[0361] The release of Th1 / Th2 cytokines from the test well supernatant of human PBMC assays was evaluated using the BD Cell Counting Bead Array Human Kit and according to the manufacturer’s protocol.

[0362] Results: In the absence of bispecific agents targeting xCD3, such as MUC16xCD3 REGN4019, both REGN6194 and Comparative Compound 1 were inert. However, in the presence of human PBMCs, a fixed concentration of PD-L1xCD28 bsAb REGN6194 successfully enhanced the cytotoxic potency of REGN4019 against OVCAR-3 / hPD-L1 cells, and cells were killed in a dose-dependent manner. REGN6194 was more potent than Comparative Compound 1 in enhancing the cytotoxicity of REGN4019. In addition to the observed enhanced target cell lysis, T cell activation was also increased, and CD25 and PD1 expression on CD4+ and CD8+ T cells was upregulated (see Table 17).

[0363] Table 17: T cell activation and killing of ECGs via a combination of bispecific PD-L1xCD28 antibody and Muc16xCD3 antibody (REGN4019) 50

[0364]

[0365] N / C: Not calculable

[0366] Cytokine release from the supernatant of the cytotoxicity assay was assessed. The cytotoxic potency of REGN4019 was correlated with the accumulation of IFNg in the culture medium, and IFNg accumulation was enhanced when REGN4019 was combined with PD-L1xCD28. Combining REGN4019 with REGN6194 (PD-L1xCD28) resulted in higher levels of cytokine accumulation in the culture medium than Comparative 1 (see Table 18).

[0367] Table 18: Cytokine release induced by the combination of bispecific PD-L1xCD28 antibody and Muc16xCD3 antibody (REGN4019)

[0368]

[0369] N / C: Not calculable

[0370] Example 7: High antitumor efficacy of bispecific PD-L1xCD28 antibody in MC38 / hPDL1 tumor model

[0371] This embodiment relates to an in vivo study demonstrating the efficacy of a bispecific PD-L1xCD28 antibody in treating tumors in a mouse model of colon cancer. Mice were humanized against PD1, PD-L1, CD28, and CD3γδε (CD3-γ-δ-ε) by knocking out the mouse genes and replacing them with their human homologs (hPD1 / hPD-L1 / hCD28 / hCD3 mice). Tumors were generated from MC38 colon cancer cells (M38-hPDL1-mPDL1KO cells) modified to knock out mouse PD-L1 and overexpress human PD-L1. The following antibodies were used in this study: PD-L1xCD28 (REGN6194), PD-L1xBetV1 (a bispecific control with one arm binding PD-L1 and the other arm binding the unrelated antigen BetV1), BetV1xCD28 (a bispecific control with one arm binding CD28 and the other arm binding BetV1), and an isotype control.

[0372] On day 0, M38-hPDL1-mPDL1KO cells were implanted into 6 to 7 hPD1 / hPD-L1 / hCD28 / hCD3 mice in each group. When the tumor reached approximately 100 mm on day 8... 3 When the average tumor volume was determined, the antibody was administered at a dose of 10 mg / kg on days 9, 13, 16, 20, and 22. Tumors were measured approximately twice weekly until the end of the experiment on day 60.

[0373] Results: Compared with allotype control or bispecific control antibodies, REGN6194 significantly controlled tumor growth and improved survival. Figures 1A to 1B For example, on day 19, the average volume of tumors treated with the isotype control or BetV1xCD28 was approximately 800 mm. 3 In contrast, the average tumor volume treated with REGN6194 was significantly smaller, approximately 200 mm. 3 ( Figure 1A Furthermore, compared with mice administered isotype control or bispecific control antibodies, mice administered REGN6194 showed significantly improved survival – for example, by day 60, 30% of mice administered REGN6194 survived, while uncontrolled mice survived past day 30. Figure 1B ).exist Figure 1B In this study, REGN6194 is referred to as "PD-L1xCD28". In summary, the bispecific PD-L1xCD28 antibody is effective in inhibiting tumor growth and supporting survival.

[0374] Example 8: The potent antitumor efficacy of bispecific PD-L1xCD28 antibodies, alone and in combination with anti-PD1 antibodies, in an MC38 / hPDL1 tumor model.

[0375] This embodiment relates to an in vivo study demonstrating the efficacy of a bispecific PD-L1xCD28 antibody as a monotherapy and in combination with the anti-PD1 antibody cimiprimab in treating tumors in a mouse model of colon cancer. Mice were humanized targeting PD1, PD-L1, CD28, and CD3γδε (CD3-γ-δ-ε) by knocking out the mouse gene and replacing it with its human homolog (hPD1 / hPD-L1 / hCD28 / hCD3 mice), and tumors were generated from MC38 colon cancer cells (M38-hPDL1-mPDL1KO cells) modified to knock out mouse PD-L1 and overexpress human PD-L1. The following antibodies were used in this study: three PD-L1xCD28 antibodies (REGN6192, REGN6193, and REGN6194), cimiprimab, and an allotype control. Figure 2 and 4 In China, cimipril is referred to as "Cemi."

[0376] On day 0, MC38-hPDL1-mPDL1KO cells were implanted into 6 to 7 hPD1 / hPD-L1 / hCD28 / hCD3 mice in each group. When the tumor reached approximately 140 mm on day 11... 3 When the average tumor volume was determined, the antibody was administered at a dose of 10 mg / kg on days 11, 14, 17, 19, and 25. Tumors were measured approximately twice weekly until the end of the experiment on day 90.

[0377] Results: Compared with the isotype control, all bispecific PD-L1xCD28 antibodies showed potent antitumor efficacy. Figure 2 ), and this effect is enhanced when combined with cimipril. Figures 3A to 3E For example, on day 19, when the tumors in the isotype control group reached more than 1000 mm... 3 When considering the volume, the average volume of tumors treated with REGN6192, REGN6193, or REGN6194 was significantly lower, i.e., less than 200 mm. 3 ( Figure 2 Furthermore, although no tumor-free mice were observed in the isotype control mice by day 55 (…),… Figure 3A However, on day 55, 4 out of 7 mice that received REGN6192 ( Figure 3B ), 1 out of 7 mice treated with REGN6193 ( Figure 3C ) and 2 out of 6 mice that were administered REGN6194 ( Figure 3DNo tumors were observed. The combination therapy of REGN6194 and cimipril produced greater antitumor efficacy than REGN6194 monotherapy; on day 55, 6 out of 6 mice were tumor-free. Figure 3E Furthermore, when treated with REGN6194 alone, the mean tumor volume eventually increased, approaching 1000 mm by day 38. 3 When REGN6194 was used in combination with cimipril, the average tumor size significantly decreased to 0 mm by day 25. 3 This condition was maintained until the end of the experiment on day 90, and was associated with long-term survival. Figure 2 and Figure 4 This complete and durable tumor regression, achieved using a combination of PD-L1xCD28 bispecific antibody and anti-PD1 therapy, is a remarkable outcome.

[0378] Example 9: In a B16F10 / hPDL1 tumor model, tumor growth was effectively controlled by combination therapy with a bispecific PD-L1xCD28 antibody and an anti-PD1 antibody.

[0379] This embodiment relates to an in vivo study demonstrating the efficacy of a combination of a bispecific PD-L1xCD28 antibody and an anti-PD1 antibody, cimiprimab, in treating tumors in a mouse model of melanoma known to be resistant to α-PD1 therapy. Mice were humanized against PD1, PD-L1, CD28, and CD3γδε (CD3-γ-δ-ε) by knocking out mouse genes and replacing them with their human homologs (hPD1 / hPD-L1 / hCD28 / hCD3 mice), and tumors were generated from B16F10 cells modified to knock out mouse PD-L1 and overexpress human PD-L1 (B16F10-hPDL1-mPDL1KO cells). The following antibodies were used in this study: PD-L1xCD28 (REGN6194), α-PD-1 (cimiprimab), and an isotype control. Figure 5A and 5B In this context, REGN6194 is referred to as "PD-L1xCD28", and cimiprimab is referred to as "a-PD-1".

[0380] On day 0, B16F10-hPDL1-mPDL1KO cells were implanted into six hPD1 / hPD-L1 / hCD28 / hCD3 mice in each group. Tumors reached approximately 100 mm on day 10. 3 The average tumor volume was measured, and antibodies were administered at a dose of 10 mg / kg on days 10, 14, 16, 21, and 24. Tumors were measured approximately twice weekly until day 39. At the end of the experiment (day 55), tumor-free mice were identified.

[0381] Results: In the B16F10 / hPD-L1 tumor model, REGN6194 monotherapy showed significant antitumor efficacy compared with allotype control and monotherapy with a-PD1 cimiprimab. Figures 5A to 5B Interestingly, compared with REGN6194 or cimiprimab alone, the combination of REGN6194 and cimiprimab effectively controlled tumor growth and improved survival. Figures 5A to 5B For example, on day 24, although the average volume of tumors treated with the isotype control was approximately 2000 mm. 3 However, the average tumor volume treated with the combination of REGN6194 and cimipril was significantly lower (i.e., less than 250 mm). 3 In contrast, also on day 24, tumors treated with cimiprimab monotherapy showed a size of approximately 1200 mm. 3 The average tumor volume was [missing information], and tumors treated with REGN6194 monotherapy showed a size of approximately 600 mm. 3 Mean tumor volume ( Figure 5A Furthermore, while one of the six mice receiving combination therapy was tumor-free by day 55, none of the other treatment groups were tumor-free by day 55. Additionally, mice treated with the combination of REGN6194 and cimiprimab had significantly better survival than those treated with either REGN6194 or cimiprimab as monotherapy. At day 30, all mice receiving combination therapy survived, while only 50% of mice treated with REGN6194 and less than 20% of mice treated with cimiprimab as monotherapy survived. Figure 5B The data showed that combination therapy, which includes PD-L1xCD28 and anti-PD1, provides potent tumor control for B16F10 tumors.

[0382] Example 10: Induction of cytokine release in response to administration of bispecific PD-L1xCD28 antibody alone or in combination with anti-PD1 antibody

[0383] This embodiment relates to an in vivo study demonstrating an improved cytokine expression profile in response to PD-L1xCD28 monotherapy or a combination of PD-L1xCD28 and an anti-PD1 antibody, compared to the anti-CD28 superagonist TGN1412. Mice were humanized against PD1, PD-L1, CD28, and CD3γδε (CD3-γ-δ-ε) by knocking out the mouse genes and replacing them with their human homologs (hPD1 / hPD-L1 / hCD28 / hCD3γδε (CD3-γ-δ-ε) mice), and tumors were generated from parental M38 cells. The following antibodies were used in this study: PD-L1xCD28 (REGN6194), anti-PD1 (cimipril), the anti-CD28 superagonist TGN1412, and an isotype control. Figures 6A to 6C In this context, REGN6194 is referred to as "PD-L1xCD28", and cimipril is referred to as "Cemi."

[0384] M38 cells were implanted into hPD1 / hPD-L1 / hCD28 / hCD3 mice on day 0. Tumors reached approximately 100 mm on day 7. 3 The mean tumor volume was determined, and antibodies were administered at a dose of 10 mg / kg on days 7, 10, 14, 18, and 21. The concentrations of three different cytokines were measured from blood samples collected 4, 24, and 96 hours after the first antibody administration on day 7.

[0385] Results: Although TGN1412 induces high levels of IL-2, IL-5, and IL-4 release, REGN6194 does not induce cytokine release when administered as a monotherapy or in combination with cimipril. Figures 6A to 6C ).

[0386] Example 11: Enhanced combination therapy efficacy of MAGE-A4xCD3 and anti-PD1 when combined with PD-L1xCD28 in the SK-MEL-37 tumor model.

[0387] This embodiment relates to an in vivo study demonstrating the ability of PD-L1xCD28 to improve the efficacy of combination therapy comprising a TAAxCD28 bispecific antibody (i.e., a bispecific antibody with one arm binding to a tumor-specific antigen (TAA) and one arm binding to CD28) and an anti-PD1 antibody. The mice used in this study were the highly immunodeficient NSG strain. TM(NOD, SCID, and IL-2Rγ were ineffective (null)), human peripheral blood mononuclear cells (hPBMCs) were introduced as a source of human immune cells. The tumor originated from the human skin cancer cell line SK-MEL-37, which had low copy numbers of TAA MAGE-A4 and endogenous levels of human PD-L1. The following antibodies were used in this study: three bispecific antibodies (PD-L1xCD28 (REGN6194), MAGE-A4xCD3, and MUC16xCD28 as unrelated TAA bispecific controls); anti-PD1 (cimipril); and HLA-A2xCD3. Figure 7 In this study, REGN6194 was referred to as “PD-L1xCD28”, and cimipril was referred to as “Cemi.” The antibodies were administered to mice at the doses shown in Table 19 and the combinations shown in Table 20.

[0388] Table 19: Antibody Dosage Administered

[0389]

[0390] Table 20: Antibody Combinations Administered

[0391]

[0392] On day 0, 5×10⁻⁶ cells were implanted into NSG mice. 6 One SK-MEL-37 cell. The tumor reached approximately 200 mm on day 8. 3 The average tumor volume was 4 × 10⁻⁶ m², at which point the mouse transplanted 4 × 10⁻⁶ m². 6 One hPBMC. Antibodies were administered on days 18, 21, 26, 29, and 33. Tumors were measured approximately twice a week from day 13 to day 43.

[0393] Results: Mice treated with a combination of MAGE-A4xCD3, cimiprimab, and REGN6194 showed potent control of tumor growth, with a mean tumor volume of less than 300 mm on day 43. 3 ( Figure 7 In contrast, mice treated with a combination of MAGE-A4xCD3, cimipril, and the bispecific control antibody MUC16xCD28 showed significantly poorer control of tumor growth, with a mean tumor volume of approximately 650 mm² at day 43. 3 ( Figure 7Therefore, in controlling tumor growth, triple combination therapy containing PD-L1xCD28, MAGE-A4xCD3, and anti-PD1 is more than 50% more effective than dual combination therapy of MAGE-A4xCD3 and anti-PD1.

[0394] Example 12: Antitumor efficacy of bispecific PD-L1xCD28 antibody against human PD-L1 in tumors + Cell ratio dependence

[0395] This embodiment involves in vivo studies, which demonstrate the presence of human PD-L1. + Cells and human PD-L1 - A higher proportion of tumor cells were treated more effectively by REGN6194. Mice were humanized to target PD1, PD-L1, CD28, and CD3γδε (CD3-γ-δ-ε) by knocking out mouse genes and replacing them with their human homologs (hPD1 / hPD-L1 / hCD28 / hCD3 mice), and tumors were generated from a mixture of MC38 colon cancer cells and parental M38 cells, the MC38 colon cancer cells being modified to knock out mouse PD-L1 and overexpress human PD-L1 (M38-hPDL1-mPDL1KO cells), the parental M38 cells expressing mouse PD-L1 instead of human PD-L1, i.e., human PD-L1. - The following antibody was used in this study: PD-L1xCD28 (REGN6194) and an isotype control. Figures 8A to 8E In China, REGN6194 is referred to as "PD-L1xCD28".

[0396] hPD1 / hPD-L1 / hCD28 / hCD3 mice were engrafted with MC38 parental cells at ratios of 0:100, 50:50, 90:10, or 99:1 to MC38-hPDL1-mPDL1KO cells, and the tumor volume reached approximately 100 mm. 3 Subsequently, mice were administered 10 mg / kg REGN6194 or an isotype control on day 0. Tumors were measured on days 4, 7, and 11, and survival was assessed until the end of the experiment 10 weeks after treatment.

[0397] Results: REGN6194 monotherapy showed efficacy against hPD-L1 antibodies derived from a 50:50 ratio. - Cells: hPD-L1 + The cells showed moderate anti-tumor efficacy against tumors. Figure 8B However, for hPD-L1 derived at a 90:10 ratio... - Cells: hPD-L1 + Cellular tumors ( Figure 8C) and hPD-L1 derived at a 99:1 ratio - Cells: hPD-L1 + Cellular tumors ( Figure 8D The product lost its effectiveness. hPD-L1 derived from a 50:50 ratio was treated with REGN6194. - Cells: hPD-L1 + Mice with tumors derived from cells had significantly higher survival rates than those with hPD-L1 derived from 90:10 or 99:1 ratios. - Cells: hPD-L1 + Mouse cells with tumors ( Figure 8E ).

[0398] Example 13: Strong tumor growth control in the absence of hPD-L1 expression by combining PD-L1xCD28 with anti-PD1.

[0399] This embodiment relates to an in vivo study demonstrating that tumors lacking human PD-L1 expression were effectively treated with a combination of PD-L1xCD28 and cimipril, indicating the role of PD-L1 expression in non-tumor cells. Mice were humanized against PD1, PD-L1, CD28, and CD3γδε (CD3-γ-δ-ε) by knocking out mouse genes and replacing them with their human homologs (hPD1 / hPD-L1 / hCD28 / hCD3 mice), and tumors were generated from parental M38 cells expressing mouse PD-L1 instead of human PD-L1. The following antibodies were used in this study: PD-L1xCD28 (REGN6194), cimipril, and an allotype control. Figures 9A to 9E In this context, REGN6194 is referred to as "PD-L1xCD28", and cimipril is referred to as "Cemi."

[0400] Parental M38 cells were implanted into hPD1 / hPD-L1 / hCD28 / hCD3 mice, and the tumor volume reached an average of approximately 100 mm. 3 Subsequently (identified as D0), antibodies were administered at a dose of 10 mg / kg on days 0, 4, 7, 11, and 14. Tumors were measured approximately twice weekly until the end of the experiment, 3 weeks after treatment.

[0401] Results: Compared with REGN6194 or cimiprimab monotherapy, combination therapy including REGN6194 and cimiprimab effectively controlled tumor growth. Figures 9A to 9E For example, on day 14 post-treatment, the mean tumor volume treated with the combination of REGN6194 and cimipril was significantly reduced to 15 mm. 3 under( Figure 9AIn contrast, on day 14 post-treatment, the average tumor volume in mice treated with the isotype control alone was approximately 1000 mm. 3 The average tumor volume in mice treated with REGN6194 monotherapy was approximately 690 mm. 3 The average tumor volume in mice treated with cimiprimab monotherapy was approximately 185 mm. 3 ( Figure 9A Furthermore, in mice treated with a combination of REGN6194 and cimiprimab, 6 out of 7 mice were tumor-free at the end of the experiment (4 weeks after implantation), while only 5 out of 8 mice treated with cimiprimab monotherapy were tumor-free at the end of the experiment. Therefore, the combination therapy of PD-L1xCD28 and anti-PD1 effectively controlled tumor growth in the absence of PD-L1 expression, indicating the role of PD-L1 in non-tumor cells (e.g., antigen-presenting cells).

[0402] Example 14: Dose-response potency and formal pK of REGN6194

[0403] This prospective example relates to in vivo studies demonstrating the dose-response potency and formal pK of the bispecific PD-L1xCD28 antibody (REGN6194), alone and in combination with the anti-PD1 antibody cimiprimab, in treating tumors in a mouse model of colon cancer.

[0404] PD-L1xCD28 Single therapy The study involved humanizing mice with PD1, PD-L1, CD28, and CD3γδε (CD3-γ-δ-ε) by knocking out the mouse genes and replacing them with their human homologs (hPD1 / hPD-L1 / hCD28 / hCD3 mice). Tumors were generated from MC38 colon cancer cells (M38-hPDL1-mPDL1KO cells) that were modified to knock out mouse PD-L1 and overexpress human PD-L1. The following antibody, PD-L1xCD28 (REGN6194), and an isotype control were used in this study.

[0405] On day 0, M38-hPDL1-mPDL1KO cells were implanted into 7 to 8 hPD1 / hPD-L1 / hCD28 / hCD3 mice per group. By day 7, tumors had reached 75 to 100 mm. 3The average tumor volume was determined, and mice were randomly assigned to 6 groups based on tumor size. Antibodies were administered intraperitoneally (IP) on days 11, 14, 21, and 28 at doses of 0.1, 1, 5, and 10 mg / kg (REGN6194) and a dose of 10 mg / kg (isotype control). Tumor size and body weight were measured, and survival was assessed twice weekly. Blood was collected on days 7, 11, and 14, immediately before antibody administration, and 4 hours after antibody administration, as well as on days 21 and 28, to measure serum cytokines and perform FACS analysis.

[0406] Results: REGN6194 is expected to control tumor growth without reducing body weight and will also improve survival. Effective antitumor activity is expected even at low doses.

[0407] Combination therapy of PD-L1xCD28 and cimiprimab The study involved humanizing mice with PD1, PD-L1, CD28, and CD3γδε (CD3-γ-δ-ε) by knocking out the mouse genes and replacing them with their human homologs (hPD1 / hPD-L1 / hCD28 / hCD3 mice). Tumors were generated from MC38 colon cancer cells (M38-hPDL1-mPDL1KO cells) that were modified to knock out mouse PD-L1 and overexpress human PD-L1. The following antibodies were used in this study: PD-L1xCD28 (REGN6194), anti-PD1 (cimiprilmab), and isotype controls of both REGN6194 and cimiprilmab.

[0408] On day 0, M38-hPDL1-mPDL1KO cells were implanted into 6 to 8 hPD1 / hPD-L1 / hCD28 / hCD3 mice per group. By day 9, the tumors had reached 100 to 150 mm. 3Mice were randomly assigned to 7 groups based on tumor size at the mean tumor volume, and antibodies were administered intraperitoneally (IP) on days 13, 16, 23, and 30. Combination therapy consisted of 0.1, 1, 5, or 10 mg / kg doses (REGN6194) and 10 mg / kg doses (cimiprimab). REGN6194 monotherapy consisted of either 5 mg / kg of REGN6194 and 10 mg / kg of cimiprimab (an isotype control), or 10 mg / kg of REGN6194 and cimiprimab (an isotype control). Cimiprimab monotherapy consisted of 10 mg / kg of REGN6194 (an isotype control) and 10 mg / kg of cimiprimab. Two isotype control antibodies were administered to the control group. Tumor size and body weight were measured, and survival was assessed twice weekly. Blood was collected on days 9, 13, and 16, immediately before antibody administration and 4 hours after antibody administration. Blood was collected immediately prior to antibody administration on days 23 and 30. Blood samples were processed by measuring serum cytokine levels and performing FACS analysis.

[0409] Results: Combination therapy with REGN6194 and cimiprimab is expected to result in better tumor growth control and improved survival than either antibody monotherapy. It is also expected that the combination therapy will not lead to weight loss. Furthermore, it is anticipated that combination therapy with REGN6194 and cimiprimab will achieve effective antitumor activity at a lower dose than that required for REGN6194 monotherapy.

[0410] Compared with the combination of bispecific control antibodies, the combination of PD-L1xCD28 and cimiprilmab for the treatment of tumors... tumor The study involved humanizing mice for PD1, PD-L1, CD28, and CD3γδε (CD3-γ-δ-ε) by knocking out mouse genes and replacing them with their human homologs (hPD1 / hPD-L1 / hCD28 / hCD3 mice). Tumors were generated from MC38 colon cancer cells (M38-hPDL1-mPDL1KO cells) that were modified to knock out mouse PD-L1 and overexpress human PD-L1. The following antibodies were used in this study, all at a dose of 10 mg / kg: PD-L1xCD28 (REGN6194); an unrelated bispecific control antibody, PD-L1xBetV1; anti-PD1 (cimiprizumab); an isotype control antibody for cimiprizumab; and an isotype control antibody applicable to both bispecific antibodies.

[0411] On day 0, M38-hPDL1-mPDL1KO cells were implanted into 6 to 8 hPD1 / hPD-L1 / hCD28 / hCD3 mice per group. By day 7, the tumors had reached 130 to 150 mm. 3The average tumor volume was measured, and mice were randomly assigned to 6 groups based on tumor size. Antibodies were administered via IP injection on days 7, 11, and 14. Combination therapy consisted of REGN6194 and cimiprimab. Combination therapy control consisted of PD-L1xBetV1 and cimiprimab. Monotherapy consisted of (i) an isotype control antibody of REGN6194 and cimiprimab, (ii) an isotype control antibody of PD-L1xBetV1 and cimiprimab, and (iii) a bispecific isotype control antibody and cimiprimab. Negative control consisted of two isotype control antibodies. Tumor size and body weight were measured, and survival was assessed twice weekly. Blood was collected on day 9 at 4 hours and 24 hours after injection. Blood was collected on days 13, 15, and 21 immediately before antibody administration. Blood samples were processed by measuring serum cytokine levels and performing FACS analysis.

[0412] Results: The combination therapy of REGN6194 and cimiprimab is expected to control tumor growth without weight loss and is more effective in improving survival than the combination of PD-L1xBetV1 and cimiprimab, and also more effective than any one of REGN6194, PD-L1xBetV1 or cimiprimab monotherapy.

[0413] This disclosure is not limited by the specific embodiments described herein. In fact, based on the preceding description and accompanying drawings, various modifications to this disclosure, other than those described herein, will become apparent to those skilled in the art. Such modifications are intended to fall within the scope of the appended claims. All patents, applications, and non-patent publications referenced in this specification are incorporated herein by reference in their entirety.

Claims

1. An isolated bispecific antigen-binding molecule, comprising: (a) The first antigen-binding domain, which has a size of less than about 3 × 10 -8 M of K D Specific binding to human CD28, as measured by surface plasmon resonance at 25°C; and (b) The second antigen-binding domain, which has a capacitance of less than about 2 × 10⁻⁶. -10 M of K D It specifically binds to human programmed death ligand 1 (PD-L1) and is measured at 25°C by surface plasmon resonance.

2. The isolated bispecific antigen-binding molecule of claim 1, wherein the bispecific antigen-binding molecule has a concentration of less than about 2 × 10⁻⁶. -8 M's EC 50 It binds to the surface of human T cells and is detected by in vitro FACS binding assay.

3. The isolated bispecific antigen-binding molecule according to claim 1 or 2, wherein the bispecific antigen-binding molecule has a concentration of less than about 3 × 10⁻⁶. -9 M's EC 50 It binds to the surface of cells expressing PD-L1 and is detected by in vitro FACS binding assay.

4. The isolated bispecific antigen-binding molecule according to any one of claims 1 to 3, wherein the bispecific antigen-binding molecule is at an IC50 concentration of less than about 1.3 nM. 50 Blocking the binding of PD-L1 to PD-1 was detected by an ELISA-based blocking assay.

5. The isolated bispecific antigen-binding molecule according to any one of claims 1 to 4, wherein the bispecific antigen-binding molecule is combined with a bispecific MUC16xCD3 antibody at a concentration of less than about 10 -10 M's EC 50 Mediates in vitro T cell killing of OVCAR-3 cells expressing PD-L1.

6. The isolated bispecific antigen-binding molecule according to any one of claims 1 to 5, wherein the first antigen-binding domain comprises: (a) Three heavy chain complementarity-determining regions (HCDR1, HCDR2, and HCDR3) contained within a heavy chain variable region (HCVR), the heavy chain variable region (HCVR) comprising an amino acid sequence selected from SEQ ID NO: 50, 32, and 10 or variants thereof; and (b) Three light chain complementarity-determining regions (LCDR1, LCDR2 and LCDR3) contained within a light chain variable region (LCVR), the light chain variable region (LCVR) containing an amino acid sequence selected from SEQ ID NO: 58 and 18 or a variant thereof.

7. The isolated bispecific antigen-binding molecule of claim 6, comprising: HCDR1, comprising an amino acid sequence selected from SEQ ID NO: 52, 34 and 12; HCDR2, comprising an amino acid sequence selected from SEQ ID NO: 54, 36 and 14; and HCDR3, comprising an amino acid sequence selected from SEQ ID NO: 56, 38 and 16.

8. The isolated bispecific antigen-binding molecule of claim 6 or 7, comprising: LCDR1, comprising an amino acid sequence selected from SEQ ID NO: 60 and 20; LCDR2, comprising an amino acid sequence selected from SEQ ID NO: 62 and 22; and LCDR3, comprising an amino acid sequence selected from SEQ ID NO: 64 and 24.

9. The isolated bispecific antigen-binding molecule of claim 7 or 8, wherein the first antigen-binding domain comprises HCVR and LCVR, the HCVR comprising the amino acid sequence of SEQ ID NO: 50 or a variant thereof, and the LCVR comprising the amino acid sequence of SEQ ID NO: 58 or a variant thereof.

10. The isolated bispecific antigen-binding molecule of claim 7 or 8, wherein the first antigen-binding domain comprises HCVR and LCVR, wherein the HCVR comprises the amino acid sequence of SEQ ID NO: 32 or a variant thereof, and the LCVR comprises the amino acid sequence of SEQ ID NO: 18 or a variant thereof.

11. The isolated bispecific antigen-binding molecule of claim 7 or 8, wherein the first antigen-binding domain comprises HCVR and LCVR, wherein the HCVR comprises the amino acid sequence of SEQ ID NO: 10 or a variant thereof, and the LCVR comprises the amino acid sequence of SEQ ID NO: 18 or a variant thereof.

12. The isolated bispecific antigen-binding molecule of any one of claims 1 to 11, wherein the second antigen-binding domain comprises: (a) Three heavy chain complementarity-determining regions (HCDR1, HCDR2, and HCDR3) contained within a heavy chain variable region (HCVR), the heavy chain variable region (HCVR) comprising an amino acid sequence selected from SEQ ID NO: 42 and 2 or variants thereof; and (b) Three light chain complementarity-determining regions (LCDR1, LCDR2 and LCDR3) contained within a light chain variable region (LCVR), the light chain variable region (LCVR) containing an amino acid sequence selected from SEQ ID NO: 58 and 18 or a variant thereof.

13. The isolated bispecific antigen-binding molecule of claim 12, wherein the second antigen-binding domain comprises: (a) HCDR1, which contains the amino acid sequence of SEQ ID NO: 44 or SEQ ID NO: 4; (b) HCDR2, comprising the amino acid sequence of SEQ ID NO: 46 or SEQ ID NO: 6; and (c) HCDR3, which contains the amino acid sequence of SEQ ID NO: 48 or SEQ ID NO:

8.

14. The isolated bispecific antigen-binding molecule of claim 12 or 13, wherein the second antigen-binding domain comprises: LCDR1, comprising an amino acid sequence selected from SEQ ID NO: 60 and 20; LCDR2, comprising an amino acid sequence selected from SEQ ID NO: 62 and 22; and LCDR3, comprising an amino acid sequence selected from SEQ ID NO: 64 and 24.

15. The isolated bispecific antigen-binding molecule of claim 14, wherein the second antigen-binding domain comprises: (a) HCDR1, HCDR2, and HCDR3 domains containing the amino acid sequences of SEQ ID NO: 44, 46, and 48, respectively; and LCDR1, LCDR2, and LCDR3 domains containing the amino acid sequences of SEQ ID NO: 60, 62, and 64, respectively; or (b) HCDR1, HCDR2, and HCDR3 domains containing the amino acid sequences of SEQ ID NO: 4, 6, and 8, respectively; and LCDR1, LCDR2, and LCDR3 domains containing the amino acid sequences of SEQ ID NO: 20, 22, and 24, respectively.

16. The isolated bispecific antigen-binding molecule of claim 15, wherein the second antigen-binding domain comprises: (a) HCVR containing the amino acid sequence of SEQ ID NO: 42 or a variant thereof, and LCVR containing the amino acid sequence of SEQ ID NO: 58 or a variant thereof; or (b) HCVR containing the amino acid sequence of SEQ ID NO: 2 or a variant thereof, and LCVR containing the amino acid sequence of SEQ ID NO: 18 or a variant thereof.

17. An isolated bispecific antigen-binding molecule, comprising: (a) A first antigen-binding domain that specifically binds to human CD28, wherein the first antigen-binding domain comprises: HCDR1, HCDR2, and HCDR3 domains containing the amino acid sequences of SEQ ID NO: 52, 54, and 56, respectively, and LCDR1, LCDR2, and LCDR3 domains containing the amino acid sequences of SEQ ID NO: 60, 62, and 64, respectively; and (b) A second antigen-binding domain that specifically binds to human PD-L1, wherein the second antigen-binding domain comprises: HCDR1, HCDR2, and HCDR3 domains containing the amino acid sequences of SEQ ID NO: 44, 46, and 48, respectively, and LCDR1, LCDR2, and LCDR3 domains containing the amino acid sequences of SEQ ID NO: 60, 62, and 64, respectively.

18. An isolated bispecific antigen-binding molecule comprising: (a) A first antigen-binding domain that specifically binds to human CD28, wherein the first antigen-binding domain comprises: HCDR1, HCDR2, and HCDR3 domains containing the amino acid sequences of SEQ ID NO: 34, 36, and 38, respectively, and LCDR1, LCDR2, and LCDR3 domains containing the amino acid sequences of SEQ ID NO: 20, 22, and 24, respectively; and (b) A second antigen-binding domain that specifically binds to human PD-L1, wherein the second antigen-binding domain comprises: HCDR1, HCDR2, and HCDR3 domains containing the amino acid sequences of SEQ ID NO: 4, 6, and 8, respectively, and LCDR1, LCDR2, and LCDR3 domains containing the amino acid sequences of SEQ ID NO: 20, 22, and 24, respectively.

19. An isolated bispecific antigen-binding molecule comprising: (a) A first antigen-binding domain that specifically binds to human CD28, wherein the first antigen-binding domain comprises: HCDR1, HCDR2, and HCDR3 domains containing the amino acid sequences of SEQ ID NO: 12, 14, and 16, respectively, and LCDR1, LCDR2, and LCDR3 domains containing the amino acid sequences of SEQ ID NO: 20, 22, and 24, respectively; and (b) A second antigen-binding domain that specifically binds to human PD-L1, wherein the second antigen-binding domain comprises: HCDR1, HCDR2, and HCDR3 domains containing the amino acid sequences of SEQ ID NO: 4, 6, and 8, respectively, and LCDR1, LCDR2, and LCDR3 domains containing the amino acid sequences of SEQ ID NO: 20, 22, and 24, respectively.

20. The isolated bispecific antigen-binding molecule of claim 17, comprising: (a) A first antigen-binding domain comprising an HCVR containing the amino acid sequence of SEQ ID NO: 50 and an LCVR containing the amino acid sequence of SEQ ID NO: 58; and (b) A second antigen-binding domain comprising an HCVR containing the amino acid sequence of SEQ ID NO: 42 and an LCVR containing the amino acid sequence of SEQ ID NO:

58.

21. The isolated bispecific antigen-binding molecule of claim 18, comprising: (a) A first antigen-binding domain comprising an HCVR containing the amino acid sequence of SEQ ID NO: 32 and an LCVR containing the amino acid sequence of SEQ ID NO: 18; and (b) A second antigen-binding domain comprising an HCVR containing the amino acid sequence of SEQ ID NO: 2 and an LCVR containing the amino acid sequence of SEQ ID NO:

18.

22. The isolated bispecific antigen-binding molecule of claim 19, comprising: (a) A first antigen-binding domain comprising an HCVR containing the amino acid sequence of SEQ ID NO: 10 and an LCVR containing the amino acid sequence of SEQ ID NO: 18; and (b) A second antigen-binding domain comprising an HCVR containing the amino acid sequence of SEQ ID NO: 2 and an LCVR containing the amino acid sequence of SEQ ID NO:

18.

23. An isolated bispecific antigen-binding molecule that competes with a reference antibody for binding to PD-L1 or to the same epitope on PD-L1, wherein the reference antibody comprises a first antigen-binding domain and a second antigen-binding domain, the first antigen-binding domain comprising an HCVR / LCVR pair containing an amino acid sequence selected from SEQ ID NO: 50 / 58, 32 / 18, and 10 / 18, and the second antigen-binding domain comprising an HCVR / LCVR pair containing an amino acid sequence selected from SEQ ID NO: 42 / 58 and 2 / 18.

24. An isolated bispecific antigen-binding molecule that competes with a reference antibody for binding to human CD28 or to the same epitope on human CD28, wherein the reference antibody comprises a first antigen-binding domain and a second antigen-binding domain, the first antigen-binding domain comprising an HCVR / LCVR pair containing an amino acid sequence selected from SEQ ID NO: 50 / 58, 32 / 18, and 10 / 18, and the second antigen-binding domain comprising an HCVR / LCVR pair containing an amino acid sequence selected from SEQ ID NO: 42 / 58 and 2 / 18.

25. The isolated bispecific antigen-binding molecule according to any one of claims 1 to 24, wherein it is a human bispecific antigen-binding molecule.

26. The isolated bispecific antigen-binding molecule according to any one of claims 1 to 25, wherein it is a bispecific antibody. 27.26 The isolated bispecific antigen-binding molecule, wherein the antibody comprises a human IgG heavy chain constant region linked to HCVR of each of the first antigen-binding domain and the second antigen-binding domain, respectively.

28. The isolated bispecific antigen-binding molecule of claim 27, wherein the heavy chain constant region is isotype IgG1.

29. The isolated bispecific antigen-binding molecule of claim 27, wherein the heavy chain constant region is isotype IgG4.

30. The isolated bispecific antigen-binding molecule of any one of claims 27 to 29, wherein the heavy chain constant region connected to the HCVR of the first antigen-binding domain or the heavy chain constant region connected to the HCVR of the second antigen-binding domain contains an amino acid modification that reduces protein A binding, but does not simultaneously contain, the reduction in protein A binding being relative to the same isotype of heavy chain without said modification.

31. The isolated bispecific antigen-binding molecule of claim 30, wherein the modification comprises an H435R substitution (EU number) in the heavy chain of isotype IgG1 or IgG4.

32. The isolated bispecific antigen-binding molecule of claim 30, wherein the modification comprises an H435R substitution and a Y436F substitution (EU number) in the heavy chain of isotype IgG1 or IgG4.

33. The isolated bispecific antigen-binding molecule of any one of claims 28 to 32, wherein the bispecific antibody comprises a chimeric hinge that reduces the binding of the Fcγ receptor relative to a wild-type hinge of the same isotype.

34. The isolated bispecific antigen-binding molecule of any one of claims 27 to 33, wherein the antibody comprises a first heavy chain of HCVR containing the first antigen-binding domain and a second heavy chain of HCVR containing the second antigen-binding domain, wherein the first heavy chain comprises an amino acid sequence selected from SEQ ID NO: 68, 40 and 28; and the second heavy chain comprises an amino acid sequence selected from SEQ ID NO: 66 and 26.

35. The isolated bispecific antigen-binding molecule of claim 34, wherein the antibody comprises a common light chain of LCVRs containing the first antigen-binding domain and the second antigen-binding domain, wherein the common light chain comprises an amino acid sequence selected from SEQ ID NO: 70 and 30.

36. The isolated bispecific antigen-binding molecule of any one of claims 27 to 33, wherein the antibody comprises a first heavy chain of HCVR containing the first antigen-binding domain and a second heavy chain of HCVR containing the second antigen-binding domain, wherein the first heavy chain comprises the amino acid sequence of SEQ ID NO: 68 and the second heavy chain comprises the amino acid sequence of SEQ ID NO:

66.

37. The isolated bispecific antigen-binding molecule of claim 36, wherein the antibody comprises a common light chain of LCVR containing the first antigen-binding domain and the second antigen-binding domain, wherein the common light chain comprises the amino acid sequence of SEQ ID NO:

70.

38. A bispecific antibody comprising a first antigen-binding domain specifically binding to human CD28 and a second antigen-binding domain specifically binding to human PD-L1, wherein the bispecific antibody comprises a first heavy chain containing the amino acid sequence of SEQ ID NO: 68, which is paired with a common light chain containing the amino acid sequence of SEQ ID NO: 70; and a second heavy chain containing the amino acid sequence of SEQ ID NO: 66, which is paired with a common light chain containing the amino acid sequence of SEQ ID NO:

70.

39. A bispecific antibody comprising a first antigen-binding domain that specifically binds to human CD28 and a second antigen-binding domain that specifically binds to human PD-L1, wherein the bispecific antibody comprises a first heavy chain containing the amino acid sequence of SEQ ID NO: 40, which is paired with a common light chain containing the amino acid sequence of SEQ ID NO: 30; and a second heavy chain containing the amino acid sequence of SEQ ID NO: 26, which is paired with a common light chain containing the amino acid sequence of SEQ ID NO:

30.

40. A bispecific antibody comprising a first antigen-binding domain specifically binding to human CD28 and a second antigen-binding domain specifically binding to human PD-L1, wherein the bispecific antibody comprises a first heavy chain containing the amino acid sequence of SEQ ID NO: 28, which is paired with a common light chain containing the amino acid sequence of SEQ ID NO: 30; and a second heavy chain containing the amino acid sequence of SEQ ID NO: 26, which is paired with a common light chain containing the amino acid sequence of SEQ ID NO:

30.

41. The bispecific antibody according to any one of claims 38 to 40, wherein it is a human antibody.

42. A pharmaceutical composition comprising the bispecific antigen-binding molecule of any one of claims 1 to 37 and a pharmaceutically acceptable carrier or diluent.

43. A pharmaceutical composition comprising the bispecific antibody of any one of claims 38 to 41 and a pharmaceutically acceptable carrier or diluent.

44. A method for preparing the bispecific antigen-binding molecule according to any one of claims 1 to 37, comprising: (a) Introducing one or more nucleic acid molecules comprising a nucleic acid sequence encoding an immunoglobulin chain of the bispecific antigen-binding molecule into a host cell; (b) culturing the host cell under conditions favorable for expression of the nucleic acid molecule; and (c) optionally, isolating the bispecific antigen-binding molecule or immunoglobulin chain from the host cell and / or the culture medium in which the host cell is cultured.

45. The method of claim 44, wherein the host cell is a Chinese hamster ovary (CHO) cell.

46. ​​The method of claim 44 or 45, further comprising formulating the antigen-binding molecule into a pharmaceutical composition comprising an acceptable carrier.

47. An antigen-binding molecule or immunoglobulin chain, which is the product of the method of claim 44 or 45.

48. A nucleic acid molecule comprising a nucleotide sequence encoding a bispecific antigen-binding molecule as described in any one of claims 1 to 37; or a group of nucleic acid molecules comprising a nucleotide sequence encoding a first antigen-binding domain specifically binding to human CD28, a second antigen-binding domain specifically binding to human PD-L1, and an LCVR as described in any one of claims 1 to 37.

49. An expression vector comprising the nucleic acid molecule of claim 48; or a set of expression vectors comprising the set of nucleic acid molecules of claim 48.

50. A host cell comprising the expression vector or group of expression vectors as described in claim 49.

51. The host cell of claim 50, wherein the host cell is a Chinese hamster ovary (CHO) cell.

52. A method for generating a bispecific antigen-binding molecule that binds to PD-L1 and CD28, comprising: (a) The host cells of claim 50 are cultured under conditions conducive to the production of the bispecific antigen-binding molecules; (b) Optionally, the antigen-binding molecule or immunoglobulin chain is isolated from the host cell and / or the culture medium in which the host cell is cultured.

53. The method of claim 52, wherein the host is a CHO cell.

54. The method of claim 52 or 53, further comprising formulating the antigen-binding molecule into a pharmaceutical composition comprising an acceptable carrier.

55. An antigen-binding molecule or immunoglobulin chain, which is the product of the method of claim 52 or 53.

56. A nucleic acid molecule comprising a nucleotide sequence encoding a bispecific antibody as described in any one of claims 38 to 41; or a group of nucleic acid molecules comprising a nucleotide sequence encoding a heavy chain of a first antigen-binding domain specifically binding to human CD28, a heavy chain of a second antigen-binding domain specifically binding to human PD-L1, and a light chain as described in any one of claims 38 to 41.

57. An expression vector comprising the nucleic acid molecule of claim 56; or a set of expression vectors comprising the set of nucleic acid molecules of claim 56.

58. A host cell comprising the expression vector or group of expression vectors as described in claim 57.

59. The host cell of claim 58, wherein the host cell is a Chinese hamster ovary (CHO) cell.

60. A method for generating a bispecific antibody that binds to PD-L1 and CD28, comprising: (a) Culture the host cells of claim 58 under conditions favorable to the production of the bispecific antibodies; (b) Optionally, the bispecific antibody is isolated from the host cell and / or the culture medium in which the host cell is cultured.

61. The method of claim 60, wherein the host is a CHO cell.

62. The method of claim 60 or 61, further comprising formulating the bispecific antibody into a pharmaceutical composition comprising an acceptable carrier.

63. An antibody, which is the product of the method of claim 60 or 61.

64. A method for inhibiting tumor growth in a subject, comprising administering to the subject an isolated bispecific antigen-binding molecule of any one of claims 1 to 37, or a bispecific antibody of any one of claims 38 to 41, or a pharmaceutical composition of claims 42 or 43.

65. The method of claim 64, wherein the tumor is esophageal cancer, squamous cell carcinoma of the lung, adenocarcinoma of the lung, squamous cell carcinoma of the cervix, adenocarcinoma of the endometrium, urothelial carcinoma of the bladder, lung cancer, non-small cell lung cancer, colorectal cancer, rectal cancer, endometrial cancer, skin cancer, squamous cell carcinoma of the head and neck, brain cancer, glioblastoma multiforme, breast cancer, gastroesophageal cancer, gastroesophageal adenocarcinoma, hepatocellular carcinoma, prostate cancer, ovarian cancer, B-cell cancer, T-cell cancer, leukemia, pancreatic cancer, colon cancer, melanoma, basal cell carcinoma, cervical cancer, diffuse large B-cell lymphoma, or multiple myeloma.

66. The method of claim 64 or 65, wherein the tumor expresses PD-L1.

67. The method of any one of claims 64 to 66, further comprising administering a second therapeutic agent or treatment regimen.

68. The method of claim 67, wherein the second therapeutic agent or treatment regimen comprises a chemotherapeutic agent, a DNA alkylating agent, an immunomodulator, a proteasome inhibitor, a histone deacetylase inhibitor, radiotherapy, surgery, stem cell transplantation, a bispecific antibody interacting with tumor-associated antigens (TAAs) and T-cell or immune cell antigens, an antibody-drug conjugate, an oncolytic virus, a bispecific antibody conjugated to an antitumor agent, a VEGF inhibitor, a checkpoint inhibitor, a GITR agonist, a CD27 agonist, a 4-1BB activator, a PD-1 inhibitor, a CTLA-4 inhibitor, an EGFR inhibitor, an Ang2 inhibitor, a MUC16 inhibitor, a cancer vaccine, a cytokine, modified IL2, modified IL12, an IL4 inhibitor, an IL6 inhibitor, a corticosteroid, or a combination thereof.

69. The method of claim 68, wherein the T cell or immune cell antigen is CD3.

70. The method of claim 68 or 69, wherein the TAA is selected from AFP, ALK, BAGE protein, BCMA, BIRC5 (survival protein), BIRC7, β-linkin, brc-abl, BRCA1, BORIS, CA9, carbonic anhydrase IX, caspase-8, CALR, CCR5, CD19, CD20 (MS4A1), CD22, CD30, CD40, CDK4, CEA, CTLA4, cyclin-B1, CYP1B1, EGFR, EGFRvIII, ErbB2 / Her2, ErbB3, ErbB4, ETV6-AML, EpCAM, EphA2, Fra-1, FOLR1, GAGE ​​protein (e.g., GAGE-1, -2), GD2, GD3, GloboH, phosphatidylinositol proteoglycan-3, GM3, gp100, Her2, HLA / B-raf, HLA / k-ras, H LA / MAGE-A3, hTERT, LMP2, MAGE proteins (e.g., MAGE-1, -2, -3, -4, -6, and -12), MART-1, mesothelin, ML-IAP, Muc1, Muc2, Muc3, Muc4, Muc5, Muc16 (CA-125), MUM1, NA17, NY-BR1, NY-BR62, NY-BR85, NY-ESO1, OX40, p15, p53, P AP, PAX3, PAX5, PCTA-1, PLAC1, PRLR, PRAME, PSMA (FOLH1), RAGE protein, Ras, RGS5, Rho, SART-1, SART-3, STEAP1, STEAP2, TAG-72, TGF-β, TMPRSS2, Thompson-nouvelle antigen (Tn), TRP-1, TRP-2, tyrosinase, and urinary plaque protein-3.

71. Use of the bispecific antigen-binding molecule of any one of claims 1 to 37, or the bispecific antibody of any one of claims 38 to 41, or the pharmaceutical composition of claims 42 or 43 in the treatment of tumors.

72. The use according to claim 71, wherein the tumor is esophageal cancer, squamous cell carcinoma of the lung, adenocarcinoma of the lung, squamous cell carcinoma of the cervix, adenocarcinoma of the endometrium, urothelial carcinoma of the bladder, lung cancer, non-small cell lung cancer, colorectal cancer, rectal cancer, endometrial cancer, skin cancer, squamous cell carcinoma of the head and neck, brain cancer, glioblastoma multiforme, breast cancer, gastroesophageal cancer, gastroesophageal adenocarcinoma, hepatocellular carcinoma, prostate cancer, ovarian cancer, B-cell cancer, T-cell cancer, leukemia, pancreatic cancer, colon cancer, melanoma, basal cell carcinoma, cervical cancer, diffuse large B-cell lymphoma, or multiple myeloma.

73. The use according to claim 71 or 72, wherein the tumor expresses PD-L1.

74. The use according to any one of claims 71 to 73, wherein the antigen-binding molecule or pharmaceutical composition is used in combination with a second therapeutic agent or treatment regimen, the second therapeutic agent or treatment regimen comprising a chemotherapeutic agent, a DNA alkylating agent, an immunomodulator, a proteasome inhibitor, a histone deacetylase inhibitor, radiotherapy, surgery, stem cell transplantation, a bispecific antibody interacting with a tumor-associated antigen (TAA) and a T-cell or immune cell antigen, an antibody-drug conjugate, an oncolytic virus, a bispecific antibody conjugated to an antitumor agent, a VEGF inhibitor, a checkpoint inhibitor, a GITR agonist, a CD27 agonist, a 4-1BB activator, a PD-1 inhibitor, a CTLA-4 inhibitor, an EGFR inhibitor, an Ang2 inhibitor, a MUC16 inhibitor, a cancer vaccine, a cytokine, a modified IL2, a modified IL12, an IL4 inhibitor, an IL6 inhibitor, a corticosteroid, or a combination thereof.

Images