Binding molecules that modulate biological activity expressed by cells

Abstract

The present invention relates to binding molecules that modulate the biological activity expressed by a cell, and particularly provides means and methods for inhibiting the biological activity of a cell. In one embodiment, the invention relates to a method of inhibiting biological activity mediated by binding of two membrane proteins that are binding partners to each other in a first or second cell. The biological activity is inhibited by providing a cell with an antibody or antibody-like molecule that can bind to each of the binding partners, and the binding blocks the binding of both binding partners, thereby inhibiting the biological activity.

Description

[0001] This application is a divisional application of the Chinese patent application with application number 201880055254.X and invention name “Binding molecules for regulating biological activity expressed by cells”. The original application is the PCT international application PCT / NL2018 / 050449 filed on July 6, 2018, which entered the Chinese national phase on February 25, 2020. Technical Field

[0002] The present invention relates to the field of binding molecules. In particular, the present invention relates to the field of therapeutic binding molecules for treating diseases involving abnormal cells. More particularly, the present invention relates to binding molecules that bind to the extracellular portions of two or more different membrane-associated proteins and thereby modulate a biological activity expressed by a cell. Background Art

[0003] Despite many advances in the treatment of the disease and an increasing understanding of the molecular events that lead to cancer, cancer remains the leading cause of death in the world. For example, colorectal cancer (CRC) is the third most common cancer worldwide. In 2008, 1.23 million people were diagnosed with the disease. It is the second most common cancer in Europe, with approximately 447,000 new cases diagnosed in 2012 (13% of the total). Colorectal cancer is the fourth most common cause of cancer death, estimated to cause 608,000 deaths per year (EU 148,000). Although some new treatments have been proposed in CRC, many have failed clinical testing; metastatic CRC remains largely incurable with conventional treatments. Melanoma is another example of a very frequent cancer. When detection is not early enough, cancer is likely to metastasize to a stage that is difficult to treat. Immunotherapy has been shown to be effective for at least some patients with metastatic melanoma. Non-small cell lung cancer is a type of cancer that is rarely found early enough for surgery. These types of cancer have also been successfully treated with immunotherapy.

[0004] Traditionally, the discovery of most cancer drugs has focused on agents that block key cell functions and kill dividing cells. However, in the case of advanced cancer, no matter how aggressively applied, even to the extent that the patient suffers from the life-threatening side effects of the treatment, chemotherapy rarely achieves complete cure. In most cases, the tumor in the patient stops growing or temporarily shrinks (called remission), just starts to proliferate again, sometimes more rapidly (called relapse), and becomes increasingly difficult to treat. Recently, the focus of cancer drug development has shifted from extensive cytotoxic chemotherapy to targeted cell inhibition therapy with less toxicity. The treatment of advanced cancer has been clinically verified in leukemia and other cancers. However, in most cancers, targeted approaches are still proven to be insufficiently effective for completely eliminating cancer in most patients.

[0005] Targeting of cancer has been achieved using a variety of different approaches, including, for example, small molecules directed against signaling proteins that cancers depend on for survival and / or growth; vaccines with tumor-specific proteins; cell therapy with immune cells that actively kill tumor cells and antibodies that target cytotoxic molecules to tumors; interfering with signaling and / or (re)directing the host's immune system to tumor cells. Monoclonal antibodies that block CTLA-4 or PD-1 have been shown to induce durable clinical responses in patients with melanoma, NSCLC, renal cell carcinoma, and urothelial carcinoma.

[0006] The present invention provides novel means and methods for (re)directing components of the immune system.The present invention also relates to means and methods for modulating biological activities expressed by cells. Summary of the invention

[0007] In one aspect, the invention provides a method of inhibiting a biological activity in a first or second cell mediated by the binding of a first membrane protein on the first cell to a second membrane protein on the second cell, wherein the first and second membrane proteins are binding partners (i.e., a ligand and receptor pair), the method comprising:

[0008] - providing an antibody or a variant of the antibody maintaining the binding specificity of the antibody to a system comprising the first and second cells, wherein the antibody or the variant of the antibody comprises a variable domain that can bind to the extracellular part of the first membrane protein and a variable domain that can bind to the extracellular part of the second membrane protein; and

[0009] - incubating the system in the absence of the antibody or variant thereof under conditions that allow the first or second cell to express the biological activity mediated by the binding of the first membrane protein to the second membrane protein;

[0010] - wherein binding of the variable domain that can bind to the extracellular part of the first membrane protein blocks binding of the first membrane protein to the second membrane protein, and / or binding of the variable domain that can bind to the extracellular part of the second membrane protein blocks binding of the first membrane protein to the second membrane protein.

[0011] In some embodiments, the method may be an in vitro or ex vivo method.

[0012] The binding of the antibody or its variant preferably reduces the activity of the binding of the receptor-ligand pair in the first cell. The binding preferably reduces the inhibitory activity of the binding of the receptor-ligand pair in the first cell. Receptor-ligand pairs of the CD28 family and the B7 family that exhibit inhibitory activity are also so-called co-inhibitory receptor-ligand pairs.

[0013] The present invention also provides a method for inhibiting a biological activity in a first or second cell mediated by the binding of a CD28 family member (first membrane protein) on the first cell to a B7 family member (second membrane protein) on the second cell, wherein the first and second membrane proteins are binding partners (i.e., a ligand and receptor pair), the method comprising:

[0014] - providing an antibody or a variant of the antibody maintaining the binding specificity of the antibody to a system comprising the first and second cells, wherein the antibody or the variant of the antibody comprises a variable domain that can bind to the extracellular part of the first membrane protein and a variable domain that can bind to the extracellular part of the second membrane protein; and

[0015] - incubating the system in the absence of the antibody or variant thereof under conditions that allow the first or second cell to express the biological activity mediated by the binding of the first membrane protein to the second membrane protein;

[0016] - wherein binding of the variable domain that can bind to the extracellular part of the first membrane protein blocks binding of the first membrane protein to the second membrane protein, and / or binding of the variable domain that can bind to the extracellular part of the second membrane protein blocks binding of the first membrane protein to the second membrane protein.

[0017] In some embodiments, the method is an in vitro or ex vivo method.

[0018] The binding of the antibody or its variant preferably reduces the activity of the binding of the receptor-ligand pair in the first cell. The binding preferably reduces the inhibitory activity of the binding of the receptor-ligand pair in the first cell. Receptor-ligand pairs that exhibit inhibitory activity are so-called co-inhibitory receptor-ligand pairs.

[0019] Also provided is a method of enhancing a biological activity in a first or second cell mediated by binding of a first membrane protein on a first cell to a second membrane protein on a second cell, the method comprising:

[0020] - providing an antibody or a variant of the antibody maintaining the binding specificity of the antibody to a system comprising the first and second cells, wherein the antibody or the variant of the antibody comprises a variable domain that can bind to the extracellular part of the first membrane protein and a variable domain that can bind to the extracellular part of the second membrane protein; and

[0021] - incubating the system in the absence of the antibody or variant thereof under conditions that allow the first or second cell to express the biological activity mediated by the binding of the first membrane protein to the second membrane protein;

[0022] - wherein binding of the variable domain that can bind to the extracellular part of the first membrane protein does not block binding of the first membrane protein to the second membrane protein, and binding of the variable domain that can bind to the extracellular part of the second membrane protein does not block binding of the first membrane protein to the second membrane protein.

[0023] In some embodiments, the method is an in vitro or ex vivo method.

[0024] The first membrane protein is preferably programmed cell death 1 protein (PD-1); cytotoxic T-lymphocyte-associated protein 4 (CTLA-4); B and T lymphocyte attenuator (BTLA); or transmembrane and immunoglobulin domain containing protein 2 (TMIGD2).

[0025] The second membrane protein is preferably programmed cell death 1 ligand 1 protein (Programmed Cell Death 1 Ligand 1 protein, PD-L1); programmed cell death 1 ligand 2 protein (Programmed Cell Death 1 Ligand 2 protein, PD-L2); ICOSL; CD80; CD86; B7-H3; B7-H4; TNFRSF14; B7-H6 or B7-H7. In a preferred embodiment, the second membrane protein is PD-L1; PD-L2; CD80; CD86; B7-H4; TNFRSF14; or B7-H7. In a particularly preferred embodiment, the second membrane protein is PD-L1 or PD-L2, preferably PD-L1. In a particularly preferred embodiment, the first membrane protein is PD-1 and the second membrane protein is PD-L1. The molecule of the present invention is preferably a bispecific antibody capable of blocking the interaction of PD-1 with PD-L1 and / or PD-L2 and PD-L1 with PD-1 and / or CD80. Preferably, the molecules of the invention are capable of blocking the complete PD-1 axis, including PD-1 with PD-L1 or PD-L2, and PD-L1 with PD-1 and CD80. Fig.33 .

[0026] The present invention also provides an antibody or a variant of the antibody that maintains the binding specificity of the antibody, which comprises a variable domain that can bind to the extracellular part of a first membrane protein and a variable domain that can bind to the extracellular part of a second membrane protein, wherein the first and second membrane proteins are binding partners (i.e., a ligand and a receptor pair), and wherein the binding of the variable domain that can bind to the extracellular part of the first membrane protein blocks the binding of the first membrane protein to the second membrane protein, and / or the binding of the variable domain that can bind to the extracellular part of the second membrane protein blocks the binding of the first membrane protein to the second membrane protein.

[0027] The present invention also provides an antibody or a variant of the antibody that maintains the binding specificity of the antibody, comprising:

[0028] - a variable domain that can bind to the extracellular part of a protein of the CD28 family (the first membrane protein), and

[0029] - a variable domain that can bind to the extracellular part of a protein of the B7 family (second membrane protein);

[0030] Wherein the first and second membrane proteins are binding partners (i.e., ligand and receptor pairs), and wherein the binding of the variable domain that can bind to the extracellular portion of the first membrane protein blocks the binding of the first membrane protein to the second membrane protein, and / or the binding of the variable domain that can bind to the extracellular portion of the second membrane protein blocks the binding of the first membrane protein to the second membrane protein. The binding of the antibody or its variant preferably reduces the activity of the binding of the receptor-ligand pair in the first cell. The binding preferably reduces the inhibitory activity of the binding of the receptor-ligand pair in the first cell. Receptor-ligand pairs that exhibit inhibitory activity are so-called co-inhibitory receptor-ligand pairs.

[0031] The antibody or variant thereof is preferably a bispecific antibody.

[0032] In a preferred embodiment, the antibody or a variant of said antibody that maintains the binding specificity of the antibody comprises:

[0033] - a variable domain that can bind to the extracellular portion of PD-1, CTLA-4, BTLA or TMIGD2; and

[0034] - a variable domain that can bind to the extracellular portion of PD-L1, PD-L2, CD80, CD86, B7-H4, TNFRSF14 or B7-H7.

[0035] In a preferred embodiment, the antibody or a variant of said antibody that maintains the binding specificity of the antibody comprises:

[0036] - a variable domain that can bind to the extracellular portion of PD-1; and

[0037] - A variable domain that can bind to the extracellular portion of PD-L1 or PD-L2.

[0038] Preferably, binding of the variable domain that can bind to PD-1 blocks the binding of PD-1 to PD-L1.

[0039] Preferably, binding of the variable domain that can bind to PD-1 also blocks the binding of PD-1 to PD-L2.

[0040] Preferably, binding of the variable domain that can bind to PD-1 also blocks the binding of PD-1 to PD-L1 and PD-L2.

[0041] Binding of the variable domain that can bind to PD-L1 can block the binding of PD-L1 to PD-1.

[0042] Binding of the variable domain that can bind to PD-L1 can block the binding of PD-L1 to CD80.

[0043] Binding of the variable domain that can bind to PD-L1 can block the binding of PD-L1 to PD-1, and can also block the binding of PD-L1 to CD80.

[0044] Also provided are compositions or kits of parts or pharmaceutical compositions comprising one or more antibodies or variants thereof of the invention.

[0045] Also provided is a nucleic acid molecule encoding at least one CDR region of an antibody of the present invention or a variant of the antibody that maintains the binding specificity of the antibody, preferably encoding a heavy chain variable region. Also provided is a nucleic acid molecule or a collection of nucleic acid molecules encoding an antibody of the present invention or a variant of the antibody that maintains the binding specificity of the antibody. Also provided is a vector comprising the nucleic acid molecule of the present invention.

[0046] Also provided is a cell or non-human animal comprising one or more nucleic acid molecules encoding an antibody of the present invention or a variant thereof that maintains the binding specificity of the antibody, either alone or together. Also provided is a method for using the cell to produce an antibody of the present invention or a variant thereof that maintains the binding specificity of the antibody, preferably together with a method for harvesting the antibody or its variant from a cell culture.

[0047] Also provided are cellular systems comprising an antibody of the invention or a variant of the antibody that maintains the binding specificity of the antibody.

[0048] Also provided is a method for treating an individual with a disease involving abnormal cells (e.g., cancer) or with a chronic infection (e.g., a viral or parasitic infection), the method comprising administering an antibody of the present invention or a variant of the antibody that maintains the binding specificity of the antibody to an individual in need thereof. The antibody or its variant may preferably bind to PD-1 and PD-L1. The variable domain that binds to PD-1 preferably blocks the binding of PD-1 to PD-L1. The variable domain that binds to PD-L1 preferably blocks the binding of PD-1 to PD-L1. The antibody or its variant is preferably a bispecific antibody that can block the interaction of PD-1 with PD-L1 and / or PD-L2 and PD-L1 with PD-1 and / or CD80. Preferably, the molecule of the present invention is capable of blocking the complete PD-1 axis, including PD-1 with PD-L1 or PD-L2, and PD-L1 with PD-1 and CD80.

[0049] Also provided is an antibody or variant thereof or a nucleic acid molecule of the invention for use as a medicament.

[0050] The invention also provides antibodies of the invention or variants thereof that maintain the binding specificity of the antibodies of the invention for use in treating individuals suffering from a disease involving abnormal cells (eg, cancer) or suffering from an infection (eg, a viral or parasitic infection).

[0051] Also provided is the use of an antibody or variant according to the invention for the preparation of a medicament for the treatment or prevention of cancer and / or infection (eg viral or parasitic infection).

[0052] The parasite may be an intracellular parasite.

[0053] Also provided is a method for inducing and / or stimulating an immune response in an individual against abnormal cells in the individual, the method comprising providing the individual with an antibody of the present invention or a variant of the antibody that maintains the binding specificity of the antibody. The abnormal cell is preferably a cancer cell, a virus-infected cell, a parasite, or a parasite-infected cell. In a preferred embodiment, the cell is a cancer cell or a neoplastic cell. DETAILED DESCRIPTION OF THE INVENTION

[0055] The CD28 receptor family plays a role in controlling the adaptive immune response. Members of the CD28 family include PD-1; CTLA-4; BTLA; TMIGD2; ICOS; CD28; and NKp30. While some members of this family (e.g., CD28 and ICOS) induce co-stimulatory signals upon binding to the corresponding B7 family members (ligands), others (notably CTLA4 and PD-1) induce inhibitory signals upon binding to B-7 family members (CD80; CD86; PD-L1 or PD-L2).

[0056] Programmed cell death 1 protein (PD-1) is a cell surface receptor belonging to the CD28 receptor family and is expressed on T cells and pro-B cells. It is currently known that PD-1 binds two ligands, PD-L1 and PD-L2. PD-1 acts as an immune checkpoint and plays an important role in downregulating the immune system by inhibiting the activation of T cells, which in turn reduces autoimmunity and enhances self-tolerance. The inhibitory effect of PD-1 is believed to be achieved by a dual mechanism of promoting apoptosis (programmed cell death) of antigen-specific T cells in the lymph nodes while reducing apoptosis of regulatory T cells (suppressor T cells). PD-1 is also known by many different aliases, such as PDCD1; programmed cell death 1; systemic lupus erythematosus susceptibility 2; protein PD-1; HPD-1; PD1; programmed cell death 1 protein; CD279 antigen; CD279; HPD-L; HSLE1; SLEB2; and PD-1. The external IDs of PD-1 are HGNC: 8760; Entrez Gene: 5133; Ensembl: ENSG00000188389; OMIM: 600244; and UniProtKB: Q15116. A new class of drugs that block PD-1 activity, PD-1 inhibitors, activate the immune system to attack tumors and are therefore successfully used to treat some types of cancer.

[0057] Cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) is a protein receptor that functions as an immune checkpoint. Members of the CD28 receptor family are involved in downregulating immune responses in animals. The protein acts as an "off" switch when bound to CD80 or CD86 on the surface of antigen-presenting cells. CTLA4 is known by many other names, such as cytotoxic T-lymphocyte-associated protein 4; insulin-dependent diabetes mellitus 12; Celiac Disease 3; CD152; ligand and transmembrane splicing cytotoxic T-lymphocyte-associated antigen 4; cytotoxic T-lymphocyte-associated antigen 4; CD152 antigen; CELIAC3; IDDM12; ALPS5; GRD4; GSE; and CD. The external IDs for CTLA4 are HGNC: 2505; Entrez Gene: 1493; Ensembl: ENSG00000163599; OMIM: 123890; and UniProtKB: P16410.

[0058] B and T lymphocyte attenuation factor (BTLA) is another member of the CD28 protein family. It is induced during T cell activation. BTLA is expressed on Th1 cells but not Th2 cells. BTLA is a binding partner of B7-H4. Unlike other members of this family, BTLA can interact with non-B7 family members. BTLA shows T cell inhibitory effects by interacting with tumor necrosis family receptors (TNF-R). BTLA is a ligand for tumor necrosis factor (receptor) superfamily, member 14 (TNFRSF14) (also known as herpesvirus entry mediator (HVEM)). The BTLA-HVEM complex negatively regulates T cell immune responses. Other names for BTLA are B and T lymphocyte-associated protein; CD272 antigen; BTLA1; and CD272. The external IDs of BTLA are HGNC: 21087; Entrez Gene: 151888; Ensembl: ENSG00000186265; OMIM: 607925; and UniProtKB: Q7Z6A9.

[0059] Transmembrane and immunoglobulin domain-containing protein 2 (TMIGD2) is a member of the CD28 protein family. The protein can be detected in cells of epithelial and endothelial origin, and when overexpressed by endothelial cell lines, it can enhance angiogenesis in vitro. It is reported that TGMID2 is a stimulatory receptor expressed on naive T cells. The ligand of this receptor is HHLA2 (B7-H7). The latter is expressed on a wide variety of cancer cells. TMIGD2 is known by many other names, such as transmembrane and immunoglobulin domain-containing protein 2; immunoglobulin and proline-rich receptor-1; CD28 homolog 2; CD28 homolog; IGPR-1; CD28H; and IGPR1. The external Id of TMIGD2 is HGNC: 28324; Entrez Gene: 126259; Ensembl: ENSG00000167664; OMIM: 614715; and UniProtKB: Q96BF3.

[0060] ICOS (Inducible T-cell Costimulator) or CD278 is a CD28 family co-stimulatory molecule expressed on activated T cells. It is believed to be particularly important for Th2 cells. The protein forms homodimers and plays an important role in cell-cell signaling, immune response, and cell proliferation regulation. Compared with wild-type initial T cells, the proliferation and IL-2 secretion of ICOS knockout T cells activated by plate-bound anti-CD3 are reduced. The number of ICOS+T cells in tumor tissue and blood of patients treated with Ipilimumab (a monoclonal antibody that binds to CTLA-4) increases. This increase is used as a pharmacodynamic biomarker for anti-CTLA-4 treatment. ICOS is known by many different names, such as activation-induced lymphocyte immunoregulatory molecule; AILIM; inducible co-stimulatory molecule; CD278 antigen; CD278; and CVID1. The ICOS external ID is HGNC: 5351; Entrez Gene: 29851; Ensembl: ENSG00000163600; OMIM: 604558, UniProtKB: Q9Y6W8.

[0061] CD28 (Cluster of Differentiation 28) is a protein expressed on T cells that provides co-stimulatory signals required for T cell activation and survival. In addition to the T-cell receptor (TCR), T cell stimulation by CD28 can provide an effective signal for the production of a variety of interleukins (especially IL-6). CD28 is a receptor for CD80 (B7.1) and CD86 (B7.2) proteins. When activated by Toll-like receptor ligands, CD80 expression is upregulated in antigen presenting cells (APCs). CD86 expression on antigen presenting cells is constitutive (expression is independent of environmental factors). CD28 is a B7 receptor that is constitutively expressed on naive T cells. It is believed that the association of the TCR of naive T cells with the MHC: antigen complex in the absence of CD28: B7 interaction will result in anergic T cells. CD28 is known by many different names, such as CD28 molecule; CD28 antigen; Tp44; T cell-specific surface glycoprotein; and CD28 antigen (Tp44). The external IDs of CD28 are HGNC: 1653; Entrez Gene: 940; Ensembl: ENSG00000178562; OMIM: 186760 and UniProtKB: P10747.

[0062] NKp30 is a member of the CD28 family. It is a member of the natural cytotoxicity receptor (NCR) including NKp30, NKp44 and NKp46. Natural Killer (NK) cells recognize and destroy tumors and virus-infected cells in an antibody-independent manner. The regulation of NK cells is mediated by activating and inhibiting receptors on the surface of NK cells. One family of activating receptors is the NCR family. NCR initiates tumor targeting by recognizing heparan sulfate on cancer cells. NKp30 interacts with highly charged HS / heparin structures. One ligand of NKp30 is B7-H6. B7-H6 is a co-stimulatory molecule. B7-H6 is not usually expressed on normal cells, but can be highly expressed on tumor cells. It can also be expressed on antigen presenting cells (APC) by induction. Due to the interaction of NKp30 with B7-H6, NK cells can be activated to release TNFα and IFNγ. NKp30 is known by many different names, such as Natural Cytotoxicity Triggering Receptor 3 (NCR3); Natural Killer Cell P30-related protein; Activating Natural Killer Receptor P30; Lymphocyte Antigen 117; NK-P30; LY117; 1C7; Activating NK-A1 Receptor; CD337 Antigen; CD337; or MALS. The external IDs of NKp30 are HGNC: 19077; Entrez Gene: 259197; Ensembl: ENSG00000204475; OMIM: 611550; and UniProtKB: O14931.

[0063] The B7 family contains many structurally related cell surface proteins that bind to receptors on lymphocytes that regulate immune responses. The activation of lymphocytes is caused by the engagement of cell surface antigen-specific T cell receptors or B cell receptors. Additional signals delivered simultaneously by the B7 ligands further determine the immune response of these cells. These so-called "co-stimulatory" or "co-inhibitory" signals are delivered by B7 family members through the CD28 receptor family on lymphocytes. The binding of B7 family members to co-stimulatory receptors enhances the immune response, and the binding to co-inhibitory receptors weakens the immune response. Currently, seven members of the family are known: B7.1 (CD80), B7.2 (CD86), inducible co-stimulatory ligand (ICOS-L), programmed death 1 ligand (PD-L1), programmed death 2 ligand (PD-L2), B7-H3 and B7-H4. B7 family members are expressed in lymphoid and non-lymphoid tissues. In mice with B7 family gene mutations, the occurrence of immunodeficiency and autoimmune diseases indicates the role of members in regulating immune responses. Manipulation of the signals delivered by the B7 ligand has shown potential in the treatment of autoimmunity, inflammatory diseases and cancer.

[0064] CD80 is a protein found on activated B cells and monocytes that provides the co-stimulatory signal necessary for T cell activation and survival. It is a ligand for two different proteins on the surface of T cells, CD28 and CTLA-4. When bound to CD28, it is associated with co-stimulation, while binding to CTLA4 is associated with a reduction in the immune response. CD80 works in conjunction with CD86 to activate T cells. CD80 is also reported to bind PD-L1. CD80 is known by many other names, such as CD80 molecule; CD80 antigen; CD28 antigen ligand 1; B7-1 antigen; B lymphocyte activation antigen B7; CTLA-4 counter-receptor B7.1; activated B7-1 antigen; CD28LG1; CD28LG; LAB7; BB1; B7; co-stimulatory factor CD80; CD80 antigen; and B7-1. The external IDs of CD80 are HGNC: 1700; Entrez Gene: 941; Ensembl: ENSG00000121594; OMIM: 112203; and UniProtKB: P33681.

[0065] CD86 is a protein expressed on antigen presenting cells. It can provide co-stimulatory signals for T cell activation and survival. It is a ligand for two different proteins on the surface of T cells: It is a ligand for two different proteins on the surface of T cells, CD28 and CTLA-4. When bound to CD28, it is associated with co-stimulation, while binding to CTLA4 is associated with a reduction in the immune response. CD86 works in conjunction with CD80 to activate T cells. It is known by many different names, such as CD86 molecule; CD86 antigen; CD28 antigen ligand 2; B7-2 antigen; CTLA-4 counter-receptor B7.2; CD28LG2; FUN-1; BU63; B70; B lymphocyte activation antigen B7-2; B lymphocyte antigen B7-2; Activating B7-2 antigen; CD86 antigen; LAB72; and B7-2. The external IDs of CD86 are HGNC:1705; Entrez Gene:942; Ensembl:ENSG00000114013; OMIM:601020; and UniProtKB:P42081.

[0066] PD-L1 is a type 1 transmembrane protein that plays a role in suppressing immune responses during specific events (e.g., pregnancy, tissue allografts, autoimmune diseases, and other disease states such as hepatitis). PD-L1 is expressed in various types of cancers including NSCLC, melanoma, renal cell carcinoma, gastric cancer, hepatocytes, and various leukemias and multiple myeloma. PD-L1 is present in the cytoplasm and plasma membrane of cancer cells, but not all cancers or all cells in tumors express PD-L1. Multiple tumor microenvironment cells promote immunosuppression by upregulating PD-L1 expression. This effect is called "adaptive immune resistance" because tumors protect themselves by inducing PD-L1 in response to IFN-γ produced by activated T cells. PD-L1 can also be regulated by oncogenes, a mechanism known as inherent immune resistance. Within the tumor microenvironment, PD-L1 is also expressed on myeloid cells and activated T cells. The expression of PD-L1 is induced by a variety of proinflammatory molecules including type I and type II IFN-γ, TNF-α, LPS, GM-CSF and VEGF, and cytokines IL-10 and IL-4. The binding of PD-L1 to PD-1 or B7.1 (CD80) transmits an inhibitory signal that reduces the proliferation of T cells expressing PD-1. It is believed that PD-1 can control the accumulation of foreign antigen-specific T cells by apoptosis. PD-L1 is expressed by a variety of cancer cells and its expression is believed to be at least partly responsible for the suppression of immune responses against cancer cells. PD-L1 is a member of the B7 protein family and is known by a variety of other names, such as CD274 molecule; CD274 antigen; B7 homolog 1; PDCD1 ligand 1; PDCD1LG1; PDCD1L1; B7H1; PDL1; programmed cell death 1 ligand 1; programmed death ligand 1; B7-H1; and B7-H. The external ID of CD274 is HGNC: 17635; Entrez Gene: 29126; Ensembl: ENSG00000120217; OMIM: 605402; UniProtKB: Q9NZQ7.

[0067] PD-L2 is the second ligand of PD-1. Engagement of PD-L2 with PD-1 inhibits T cell receptor (TCR)-mediated proliferation and cytokine production by T cells. At low antigen concentrations, PD-L2 / PD-1 binding inhibits B7-CD28 signaling. At high antigen concentrations, PD-L2 / PD-1 binding reduces cytokine production. PD-L1 expression is upregulated on antigen presenting cells by interferon gamma treatment. It is expressed in some normal tissues and a variety of tumors. PD-L1 and PD-L2 are thought to have overlapping functions and regulate T cell responses. The protein PD-L2 is known by many other names, such as programmed cell death 1 ligand 2; B7 dendritic cell molecule; programmed death ligand 2; Butyrophilin B7-DC; PDCD1 ligand 2; PD-1 ligand 2; PDCD1L2; B7-DC; CD273; B7DC; PDL2; PD-1-ligand 2; CD273 antigen; BA574F11.2; and Btdc. The external IDs of PD-L2 are HGNC: 18731; Entrez Gene: 80380; Ensembl: ENSG00000197646; OMIM: 605723; and UniProtKB: Q9BQ51.

[0068] Inducible T cell co-stimulatory ligand (ICOSL or CD275) is constitutively expressed by APCs as well as many non-hematological tissues. Expression can be downregulated with ongoing inflammation. ICOSL is currently known to interact with ICOS, CD28, and CTLA-4 in humans. The ICOSL / CD28 interaction has been shown to co-stimulate the primary response of human T cells to alloantigens as well as memory recall responses. ICOSL / CTLA-4 is thought to generate a co-inhibitory signal. ICOSL is also known as ICOSLG; B7-related protein 1; B7 homolog 2; B7-like protein Gl50; B7 homolog 2; B7RP-1; B7-H2; B7RP1; B7H2; transmembrane protein B7-H2 ICOS ligand; CD275 antigen; KIAA0653; ICOS-L; LICOS; and GL50. The external IDs of ICOSL are HGNC: 17087; Entrez Gene: 23308; Ensembl: ENSG00000160223; OMIM: 605717; and UniProtKB: O75144.

[0069] CD276 (or B7-H3) expression is elevated in a variety of malignancies and distinguishes normal from tumor-derived circulating endothelial cells (Kraan et al British Journal of Cancer (2014) 111, 149–156). Stimulation of the receptor directs differentiation of human bone marrow stromal cells into osteoblasts (Xu et al 2011; Immunobiology 216 (2011) 1311–1317). The protein contains four Ig-like domains in humans, while the mouse protein was shown to have two such domains. The protein is believed to be the first identified ligand for triggering receptor expressed on myeloid cells (TREM)-like transcript 2 (TLT-2 or TREMML2). The latter protein binds B7-H3 (4Ig-B7-H3) and co-stimulates activation of CD8 T cells (Hofmeyer et al 2009 PNAS 105; 10277-10278). CD276 is widely expressed. It acts as a T cell co-stimulator. CD276 is also known by many other names, such as CD276 molecule; co-stimulatory molecule; CD276 antigen; B7 homolog 3; 4Ig-B7-H3; B7-H3; B7H3; and B7RP-2. The external IDs of CD276 are HGNC: 19137; Entrez Gene: 80381; Ensembl: ENSG00000103855; OMIM: 605715; and UniProtKB: Q5ZPR3.

[0070] B7-H4 (VTCN1) mRNA is shown to be widely expressed, but only a few cells actively express the protein on the membrane. The expression of this protein on the membrane and the binding to activated T cells lead to inhibition of T cell effector function through cell cycle arrest, reduced proliferation and reduced IL-2 production. In a variety of human cancers, B7-H4 is upregulated on the surface of cancer cells and immunosuppressive tumor-associated macrophages (TAMs). B7-H4 is a binding partner of BTLA. Signaling through the B7-H4 pathway leads to inhibition of TCR-mediated CD4+ and CD8+ T cell proliferation, cell cycle progression and IL-2 production. B7-H4 is also known by many other names, such as V-Set domain-containing inhibitor of T cell activation 1; immune co-stimulatory protein B7-H4; T cell co-stimulatory molecule B7x; B7 superfamily member 1; B7 homolog 4; B7h.5; B7H4; T cell co-stimulatory molecule B7x; B7 family member H4; protein B7S1; PRO1291; VCTN1; B7S1; B7X; and H4 2. The external IDs of B7-H4 are HGNC: 28873; Entrez Gene: 79679; Ensembl: ENSG00000134258; OMIM: 608162 and UniProtKB: Q7Z7D3.

[0071] B7-H6 belongs to the B7 family (see MIM 605402) and is selectively expressed on tumor cells. The binding of B7-H6 to NKp30 (NCR3; MIM 611550) leads to natural killer (NK) cell activation and cytotoxicity (Brandt et al., 2009 J Exp Med. 2009 Jul 6; 206 (7): 1495-503). Natural killer (NK) cells are lymphocytes of the innate immune system that are involved in eliminating tumors. B7-H6 is a tumor cell surface molecule that binds to NKp30, which is a human receptor that causes anti-tumor NK cell cytotoxicity and cytokine secretion. Other names for B7-H6 are NCR3LG1; natural killer cell cytotoxicity receptor 3 ligand 1; B7 homolog 6; B7H6; putative Ig-like domain-containing protein DKFZp686O24166 / DKFZp686I21167; and DKFZp686O24166. The external IDs of B7-H6 are HGNC: 42400; Entrez Gene: 374383; Ensembl: ENSG00000188211; OMIM: 613714; and UniProtKB: Q68D85.

[0072] B7-H7 (HHLA2) protein is detected in trophoblastic cells of the placenta and epithelial cells of the intestine, kidney, gallbladder and breast, but not in most other organs. HHLA2 protein is widely expressed in human breast cancer, lung cancer, thyroid cancer, melanoma, pancreatic cancer, ovarian cancer, liver cancer, bladder cancer, colon cancer, prostate cancer, kidney cancer and esophageal cancer. High HHLA2 expression is associated with regional lymph node metastasis and stage (Janakiram et al. Clin Cancer Res; 21 (10): 2359-66; May 15, 2015). TMIGD2 has been identified as a receptor for HHLA2. B7-H7 is known by many different names, such as HERV-H LTR-associated 2; human endogenous retrovirus-H long terminal repeat-associated protein 2; B7H7 and B7y. The external IDs of B7-H7 are HGNC:4905; Entrez Gene:11148; Ensembl:ENSG00000114455; OMIM:604371 and UniProtKB:Q9UM44.

[0073] Tumor necrosis factor receptor superfamily member 14 (TNFRSF14) is a human cell surface receptor of the TNF receptor superfamily. The protein was originally called herpesvirus entry mediator A (HveA). It is also called CD270 in the differentiation group classification. TNFRSF14 was identified as a cellular mediator of invasion by herpes simplex virus (HSV). Binding of HSV viral envelope glycoprotein D (gD) to this receptor protein has been shown to be part of the viral invasion mechanism. The cytoplasmic region of the receptor was found to bind to several TRAF family members, which may mediate signal transduction pathways that activate immune responses. TNFRSF14 is known by many different names, such as tumor necrosis factor receptor superfamily member 14 (herpesvirus entry mediator). Herpesvirus invasion mediator A; HVEA; HVEM; TR2; tumor necrosis factor receptor-like gene 2; tumor necrosis factor receptor-like 2; herpesvirus invasion mediator A; herpesvirus invasion mediator; CD40-like protein; CD270 antigen; LIGHTR; CD270; and ATAR. The external IDs of TNFRSF14 are: HGNC: 11912; EntrezGene: 8764; Ensembl: ENSG00000157873; OMIM: 602746; and UniProtKB: Q92956.

[0074] Reference sequence identifiers are used to identify which protein is targeted. Antibodies of the present invention or variants thereof, such as variants thereof, generally also recognize at least some of their variants, such as allelic variants thereof, splice variants and mutant variants thereof, as long as the epitope identified by the corresponding variable domains is unaffected. Some alternative names may or may not have been used to refer to other proteins. These names are given for reference only. Antibodies of the present invention or variants thereof, such as variants thereof, are combined with proteins such as those expressed on cells. It may also be combined with variants of proteins, as long as the epitope to which the antibody is combined is available. Therefore, as long as the epitope is available, splice variants or mutant proteins (if any) will also be combined with antibodies or variants thereof. The fact that an antibody or its variant is combined with a specified protein means that it can be combined with proteins as a characteristic and does not necessarily imply that the antibody or its variant is actually combined with the target. This does not necessarily mean that the variable domains cannot be combined with other proteins.

[0075] The present invention provides a method for inhibiting a biological activity in a first or second cell mediated by the binding of a first membrane protein on the first cell to a second membrane protein on the second cell, the method comprising:

[0076] - providing an antibody or a variant thereof to a system comprising the first and second cells, the antibody or a variant thereof comprising a variable domain that can bind to the extracellular portion of the first membrane protein and a variable domain that can bind to the extracellular portion of the second membrane protein, and

[0077] - incubating said system in the absence of said antibody or variant thereof, under conditions allowing expression of said biological activity;

[0078] -Wherein the binding of the variable domain that can bind to the extracellular part of the first membrane protein blocks the binding of the first membrane protein to the second membrane protein, and / or the binding of the variable domain that can bind to the extracellular part of the second membrane protein blocks the binding of the first membrane protein to the second membrane protein. In some embodiments, the method is an in vitro method. The antibody used in the method binds to two binding partners and blocks the binding between the two. In the method, two cells expressing the binding partners are brought into close proximity and / or kept in close proximity but at the same time the binding of the binding partners is inhibited. This is different from the combination of two monoclonal monospecific antibodies comprising variable domains. These also block the binding of the two binding partners, but do not bring the cells into close proximity and / or keep them in close proximity. The special activity of the antibodies of the present invention (particularly the bispecific antibodies of the present invention) is particularly evident in complex environments comprising three or more different cell types.

[0079] The present invention also provides a method for enhancing a biological activity in a first or second cell mediated by the binding of a first membrane protein on the first cell to a second membrane protein on the second cell, the method comprising:

[0080] - providing an antibody or a variant thereof to a system comprising the first and second cells, the antibody or a variant thereof comprising a variable domain that can bind to the extracellular portion of the first membrane protein and a variable domain that can bind to the extracellular portion of the second membrane protein; and

[0081] - incubating said system in the absence of said antibody or variant thereof under conditions allowing the cells to express said biological activity;

[0082] - wherein binding of the variable domain that can bind to the extracellular part of the first membrane protein does not block binding of the first membrane protein to the second membrane protein, and binding of the variable domain that can bind to the extracellular part of the second membrane protein does not block binding of the first membrane protein to the second membrane protein.

[0083] In some embodiments, the method is an in vitro or ex vivo method.

[0084] The present invention also provides an antibody or variant thereof, comprising a variable domain that can bind to the extracellular portion of a first membrane protein and a variable domain that can bind to the extracellular portion of a second membrane protein, wherein the first and second membrane proteins are binding partners (i.e., members of a binding pair or ligand and receptor pair), and wherein the binding of the variable domain that can bind to the extracellular portion of the first membrane protein blocks the binding of the first membrane protein to the second membrane protein, and / or the binding of the variable domain that can bind to the extracellular portion of the second membrane protein blocks the binding of the first membrane protein to the second membrane protein. As described herein, such antibodies can be used in methods of inhibiting biological activity mediated by the binding of the first membrane protein to the second membrane protein.

[0085] The present invention also provides an antibody or a variant thereof, wherein the antibody or a variant thereof comprises a variable domain that can bind to the extracellular portion of a first membrane protein and a variable domain that can bind to the extracellular portion of a second membrane protein, wherein the first and second membrane proteins are binding partners (i.e., members of a binding pair or a ligand and receptor pair), and wherein the binding of the variable domain that can bind to the extracellular portion of the first membrane protein does not block the binding of the first membrane protein to the second membrane protein and the binding of the variable domain that can bind to the extracellular portion of the second membrane protein does not block the binding of the first membrane protein to the second membrane protein. Such an antibody can be used in a method for enhancing a biological activity mediated by the binding of the first membrane protein to the second membrane protein.

[0086] The first membrane protein and the second membrane protein are cell membrane proteins having at least an extracellular portion. The combination of the two can result in the expression of biological activity in the first cell, the second cell, or both. In the method of the present invention, the first and second cells are preferably different cells. The different cells may have the same cell type, but they are usually different. The biological activity is an activity expressed by the cell that can be measured in the first and / or second cell or the culture medium surrounding the cell in response to the combination of the first and second membrane proteins. If both cells express biological activity, the activity may be the same but is usually different. Such a combination of the first and second proteins is generally referred to as receptor-ligand binding. In the present invention, both the receptor and the ligand are associated with the extracellular side of the cell membrane of the corresponding cell and are at least partially accessible. The term "receptor" is generally used for proteins that trigger the biological activity of the cell when bound to the ligand. The term "ligand" is generally used for proteins that bind to the receptor. The ligand in the present invention is also a membrane protein, and the binding of the ligand to the receptor can trigger the biological activity of the cell containing the ligand. A non-limiting example of such a so-called bidirectional effect is the interaction of the HVEM / BTLA binding pair. The binding of HVEM to BTLA induces biological activity in cells expressing HVEM, and induces biological activity in cells expressing BTLA. Therefore, in the present invention, the use of the term "ligand" does not necessarily mean that the binding of a receptor to a ligand cannot induce biological activity in a cell containing a ligand on a cell membrane. If a protein has a transmembrane region that is present in the cell membrane of a cell in which it is located, the protein is called a membrane protein on a cell. The protein may have additional transmembrane regions. In such a case, all transmembrane regions present in the cell membrane are present in the cell membrane of the same cell.

[0087] Receptors and ligands are specific binding partners that usually interact (bind) with each other through non-covalent bonds. Binding is specific because under physiological conditions, receptors and ligands usually only bind to each other and not to other proteins. Some receptors and ligands can specifically bind to a limited number of other binding partners. A non-limiting example is the receptor PD-1 that can interact with binding partners PD-L1 and PD-L2.

[0088] The type of biological activity induced depends on the binding partner. Binding can induce growth; change the activation state of the cell; induce or inhibit the efflux of one or more cytokines; affect the cytolytic function of the cell; etc. Biological activity is usually measured by measuring the response of the receiving (receptor-expressing) cell to binding. For example, PD-1 / PD-L1 binding is usually measured by detecting the biological activity of cells expressing PD-1. The interaction of PD-1 with its ligands PD-L1 or PD-L2 induces biological activities, such as inhibition of T cell proliferation, inhibition of cytokine production, and inhibition of cytolytic function.

[0089] When the biological activity measured in the presence of an antibody or its variant is lower than the biological activity measured under otherwise identical conditions in the absence of an antibody or variant, the biological activity is inhibited. The biological activity can be inhibited by at least 10%, 20%, 30%, 40%, 50%, or preferably at least 60%. The biological activity is preferably inhibited by at least 70%, preferably at least 80%, preferably at least 90%. When the biological activity measured in the presence of an antibody or its variant is correspondingly 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80% or 90% lower than the biological activity measured under otherwise identical conditions in the absence of an antibody or variant, the biological activity is inhibited by at least 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80% or 90%. When one of the variable domains of an antibody of the present invention can block the binding partner of its target to the target, the biological activity is usually inhibited. When the antibody further comprises a variable domain that binds to said binding partner of the target and blocks binding of the binding partner to the target, the biological activity is generally further inhibited.

[0090] When the biological activity measured in the presence of an antibody or its variant is higher than the biological activity measured under otherwise identical conditions in the absence of an antibody or variant, the biological activity is triggered and / or enhanced. The biological activity can be enhanced by at least 10%, 20%, 30%, 40%, 50%, or preferably at least 60%. The biological activity is preferably enhanced by at least 70%, preferably at least 80%, preferably at least 90%. When the biological activity measured in the presence of an antibody or its variant is 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80% or 90% higher than the biological activity measured under otherwise identical conditions in the absence of an antibody or variant, the biological activity is enhanced by at least 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80% or 90%. When both variable domains of the antibody of the present invention do not block the binding of their targets to each other, the biological activity is generally enhanced.

[0091] A suitable system is a cell culture in which the first cell and the second cell are provided. Another suitable system is an animal comprising the first cell and the second cell. Other suitable systems are in vitro systems in which the cells are maintained in an active form, but in which the growth of the cells is not necessarily promoted. The first cell and the second cell can be incubated together, for example, under assay conditions that do not necessarily promote growth but allow for measurement of biological activity.

[0092] Incubating the system under conditions that allow cells to express the biological activity mediated by the binding of the first membrane protein to the second membrane protein means maintaining the system under conditions where the first and second cells can exhibit biological activity due to the binding partners. In vivo incubation does not have to be performed for more than a time sufficient for biological activity to become apparent.

[0093] A variable domain that "blocks" the binding of the first membrane protein to the second membrane protein interferes with the binding of the first membrane protein to the second membrane protein. Such a variable domain may bind to the first membrane or the second membrane protein. A blocking variable domain that binds, for example, the first membrane protein may bind to an epitope on the first membrane protein and compete with the second membrane protein for binding to the epitope. Such a blocking variable domain and the second membrane protein may also bind to different epitopes on the first membrane protein. In such a case, the blocking activity may be, for example, due to reduced binding of the second membrane protein, displacement of the second membrane protein when the second membrane protein is already bound to the first membrane protein, or may be prevented from binding to the first membrane protein by steric hindrance. All of these and other mechanisms may at least partially prevent the second membrane protein from binding to the first membrane protein. A variable domain that binds to the first membrane protein or the second membrane protein may block the binding of a binding partner.

[0094] When compared with the combination in the absence of variable domains, the variable domains of the combination of the specific binding pair of the blocking membrane protein as described herein generally reduce the combination of the pair. This is usually measured with antibodies comprising variable domains. This is preferably measured in an in vitro assay. Usually, this is done by incubating the variable domains with the membrane protein to which it can bind, and then incubating the mixture with another member of the pair. The pair of binding is then compared with the pair of binding in the absence of variable domains. The variable domains can completely prevent the combination of the first membrane protein and the second membrane protein. It can also partially prevent the pair of binding. Compared with the combination in the absence of variable domains, the variable domains of the specific binding pair of the blocking membrane protein preferably reduce the pair of binding by at least 50%, preferably at least 60%, preferably at least 70%, preferably at least 80%, and more preferably at least 90%. The blocking of the combination by the variable domain is defined herein as the blocking obtained using a bivalent monoclonal antibody comprising the two identical variable domains. The variable domains of course also block binding when present in an antibody comprising said variable domains and a variable domain that binds to a second target, wherein the second target may be the same or different than the target bound by the first variable domain. A specific variable domain that can bind to the extracellular domain of PD-1 and at least partially block the binding of PD1 to PD-L1 is a variable domain comprising the amino acid sequence of the following VH: MF6076; MF6236; MF6256; MF6932; MF6935; MF6936; MF6972; MF6974; MF6982; MF6929; MF7699; MF7698; MF7687; MF7686; MF7685; or MF7684 ( Figure 3 and / or Fig.13 ).

[0095] The specific variable domain that can bind to the extracellular domain of PD-L1 and block the binding of PD1 to PD-L1 is a variable domain comprising the amino acid sequence of the following VH: MF5359; MF5377; MF5382; MF5424; MF5426; MF5439; MF5442; MF5553; MF5557; MF5561; MF5576; MF5594; MF5708; MF5442; MF7691; MF7690; MF7689; MF7688; MF7700; MF7701; MF7703; MF7694; MF7693; MF7692; MF7697; MF7696; or MF7695 ( Figure 3 and / or Fig.13 ).

[0096] When compared with the combination in the absence of variable domains, the variable domains of the combination of the specific binding pair of membrane proteins as described herein that do not block do not generally reduce the combination of the pair. This is generally measured with antibodies comprising variable domains. This is preferably measured in an in vitro assay. Generally, this is accomplished by incubating the variable domains with the membrane protein to which it can bind, and then incubating the mixture with another member of the pair. The combination of the pair is then compared with the combination of the pair in the absence of variable domains.

[0097] The specific variable domain that can bind to the extracellular domain of PD-L1 and does not block the binding of PD1 to PD-L1 is a protein containing MF5361 ( Figure 3 )'s VH amino acid sequence.

[0098] Functional aspects of the variable domains (in kind but not necessarily in quantity) can be determined in a variety of ways, such as binding to an antigen, blocking ability of receptor ligand interactions, biological activity of the variable domains, etc. Suitable formats are FAB fragments or antibodies. Suitable antibody formats are monospecific bivalent antibodies comprising two variable domains. Another suitable format is, for example, a bispecific antibody comprising a variable domain to be tested and another variable domain. Another variable domain is preferably a variable domain with neutral specificity relative to the assay to be performed. Suitable neutral variable domains are variable domains that can bind to tetanus toxoid.

[0099] The antibody or variant thereof of the present invention preferably comprises a variable domain that blocks the binding of its target membrane protein to its binding partner. In this preferred embodiment, another variable domain of the antibody or variant thereof binds to the binding partner of the target membrane protein. The variable domain that binds to the binding partner can block the binding of the binding partner to the target membrane protein, or it does not block the binding of the binding partner to the target membrane protein. In a preferred embodiment, the variable domain that binds to the binding partner can block the binding of the binding partner to the target membrane protein.

[0100] In another embodiment, the antibody of the present invention or its variant comprises a variable domain that does not block the binding of its target membrane protein to its binding partner. In this preferred embodiment, another variable domain of the antibody or its variant binds to the binding partner of the target membrane protein. The variable domain that binds to the binding partner can block the binding of the binding partner to the target membrane protein, or it does not block the binding of the binding partner to the target membrane protein. In a preferred embodiment, the variable domain that binds to the binding partner does not block the binding of the binding partner to the target membrane protein.

[0101] The present invention also provides a method for inducing or stimulating an immune response of an immune cell, comprising providing:

[0102] An immune cell (first cell) having a first membrane protein (preferably a CD28 family member) on its cell membrane;

[0103] - a second cell having a second membrane protein (preferably a B7 family member or TNFRSF14) on its cell membrane;

[0104] - providing an antibody or a variant thereof, wherein the antibody or the variant thereof comprises a variable domain that can bind to the extracellular part of the first membrane protein and a variable domain that can bind to the extracellular part of the second membrane protein, and the method further comprises incubating the first cell and the second cell with the antibody, thereby inducing or stimulating an immune response of the first cell.

[0105] The present invention also provides a method for inducing or stimulating an immune response of an immune cell, comprising providing:

[0106] An immune cell (first cell) having a first membrane protein (preferably a CD28 family member) on its cell membrane;

[0107] - a second cell having a second membrane protein (preferably a B7 family member or TNFRSF14) on its cell membrane;

[0108] - providing an antibody or a variant thereof (a first antibody), wherein the first antibody comprises a variable domain that can bind to the extracellular portion of the first membrane protein and a variable domain that can bind to the extracellular portion of the second membrane protein;

[0109] - providing another antibody or a variant thereof (a second antibody), wherein the second antibody comprises a variable domain that can bind to the extracellular portion of the first membrane protein and a variable domain that can bind to the extracellular portion of the second membrane protein;

[0110] Wherein the primary and secondary antibodies bind:

[0111] - a different epitope on said first membrane protein;

[0112] - a different epitope on the second membrane protein; or

[0113] - different epitopes on the first membrane protein and

[0114] different epitopes on the first membrane protein;

[0115] The method further includes incubating the first cell and the second cell with the first and second antibodies, thereby inducing or stimulating an immune response in the first cell.

[0116] The present invention also provides a method for inducing or stimulating an immune response of an immune cell, comprising providing:

[0117] - immune cells (first cells) that have PD-1 on their cell membranes;

[0118] - a second cell having PD-L1 on its cell membrane;

[0119] - providing an antibody or a variant thereof, the antibody or variant thereof comprising a variable domain that can bind to the extracellular portion of PD-1 and a variable domain that can bind to the extracellular portion of PD-L1, the method further comprising incubating the first cell and the second cell with the antibody, thereby inducing or stimulating an immune response of the first cell. In some embodiments, the method is an in vitro or ex vivo method.

[0120] The present invention also provides a method for inducing or stimulating an immune response of an immune cell, comprising providing:

[0121] - immune cells (first cells) that have PD-1 on their cell membranes;

[0122] - a second cell having PD-L1 on its cell membrane;

[0123] - providing an antibody or a variant thereof (first antibody), the first antibody comprising a variable domain that can bind to the extracellular part of PD-1 and a variable domain that can bind to the extracellular part of PD-L1;

[0124] - providing another antibody or a variant thereof (second antibody), the second antibody comprising a variable domain that can bind to the extracellular part of PD-1 and a variable domain that can bind to the extracellular part of PD-L1;

[0125] Wherein the primary and secondary antibodies bind:

[0126] -Different epitopes on PD-1;

[0127] - a different epitope on PD-L1; or

[0128] - Different epitopes on PD-1 and different epitopes on PD-L1;

[0129] The method further comprises incubating the first cell and the second cell with the first and second antibodies, thereby inducing or stimulating an immune response in the first cell. In some embodiments, the method is an in vitro method.

[0130] The immune response may or may not be mediated by a T cell receptor (TCR). In a preferred embodiment, the immune response is mediated by a TCR receptor. The immune response is preferably measured by measuring proinflammatory cytokines released by immune cells. In a preferred embodiment, the cytokine is IL-2. When compared with the absence of an antibody, changes in IL-2 levels indicate whether the immune response is affected by the antibody. Increases indicate that the immune response is stimulated. Changes from undetectable to detectable IL-2 indicate an induced immune response.

[0131] The first cell is preferably an immune cell. The immune cell is preferably a T cell or a NK cell. In one embodiment, the immune cell is a T cell. The second cell is preferably an antigen presenting cell, a neoplastic cell, a virus-infected cell, or a cell infected by an intracellular parasite. The first cell is a cell expressing the first membrane protein on its cell membrane. The second cell is a cell expressing the second membrane protein on its cell membrane. The cell membrane is also called a plasma membrane or a cytoplasmic membrane and is a biological membrane that separates the interior of the cell from the external environment.

[0132] The present invention also provides a composition or a kit, which comprises two or more antibodies or functional parts, derivatives and / or analogs thereof, wherein the antibodies or functional parts, derivatives and / or analogs thereof comprise: an antibody or a variant thereof (a first antibody), wherein the first antibody comprises a variable domain that can bind to the extracellular portion of a first membrane protein (preferably a CD28 family member) and a variable domain that can bind to the extracellular portion of a second membrane protein (preferably a B7 family member or TNFRSF14); and another antibody or a variant thereof (a second antibody), wherein the second antibody comprises a variable domain that can bind to the extracellular portion of the first membrane protein and a variable domain that can bind to the extracellular portion of the second membrane protein.

[0133] Wherein the primary and secondary antibodies bind:

[0134] - a different epitope on said first membrane protein;

[0135] - a different epitope on the second membrane protein; or

[0136] - different epitopes on the first membrane protein and

[0137] A different epitope on the second membrane protein.

[0138] Embodiments of methods, uses, compositions or kits comprising two or more antibodies having variable domains that bind to the same first and second membrane proteins, or functional parts, derivatives and / or analogs thereof are also referred to as "Oligoclonics" embodiments. Some examples of such oligoclonal embodiments are embodiments having said first and second antibodies. "Oligoclonal" is a registered trademark. For making such General methods for the production of the product are disclosed in WO 2013 / 157953 and WO 2004 / 009618 and are incorporated herein by reference.

[0139] In some oligoclonal embodiments, the first and second antibodies comprise variable domains that bind to the same CD28 family member (e.g., PD-1) and the same B7 family member or TNFRSF14 (e.g., PD-L1). A membrane comprising a family member as described herein typically comprises many and usually a large number of individual proteins of the member on the membrane. An antibody having a variable domain that binds to the same family member can bind to the same individual protein, but this is not necessarily the case. The antibody of the present invention that binds to the TNa membrane protein binds to an epitope on the membrane protein. An epitope is a part of an antigen, in this case a membrane protein recognized by an antibody. The first and second antibodies that bind to different epitopes on a membrane protein can bind to the same individual protein on the membrane. To this end, different epitopes are preferably non-overlapping epitopes. In other words, different epitopes are sufficiently separated on a membrane protein that two antibodies can bind to the same individual protein at the same time. It was unexpectedly found that oligoclonal (a combination of a first antibody and a second antibody or more antibodies) can be more effective than each antibody alone of the same amount.

[0140] The variable domains of the first antibody that can bind to a CD28 family member preferably block the binding of the member to its binding partner in the B7 family or TNFRSF14. The variable domains of the first antibody that can bind to a B7 family member or TNFRSF14 preferably block the binding of the member to its binding partner in the CD28 family. The variable domains of the second antibody that can bind to a CD28 family member preferably block the binding of the member to its binding partner in the B7 family or TNFRSF14. The variable domains of the second antibody that can bind to a B7 family member or TNFRSF14 preferably block the binding of the member to its binding partner in the CD28 family. Some preferred combinations of blocking and non-blocking variable domains in the first and second antibodies are indicated below.

[0141]

[0142] The above combinations are preferred combinations. These combinations specify that the variable domains of the first antibody are variable domains that block the binding of CD28 family members to their binding partners in the B7 family or TNFRSF14. The second antibody may have one or more variable domains that do not block this interaction. When analyzing the oligoclonal embodiment, it is specified that the antibody comprising two blocking variable domains has the "first antibody" qualification. When the oligoclone comprises two or more antibodies with two blocking variable domains, one of them is specified to have the "first antibody" qualification. The known binding partners of CD28 family members or B7 family members or TNFRSF14 are indicated above.

[0143] The first membrane protein is preferably a member of the CD28 family. The CD28 family member has a single extracellular immunoglobulin variable-like (IgV) domain followed by a short cytoplasmic tail. The member is expressed on cells of the constitutive or inductive immune system. Members of the family include CD28, CTLA-4, PD-1, ICOS, BTLA, NKp30 and TMIGD2. The first membrane protein is preferably PD-1; CTLA-4; BTLA; or TMIGD2, preferably PD-1.

[0144] The second membrane protein is preferably a B7 family member or TNFRSF14. The phrase "the B7 family member or TNFRSF14" means a member selected from the protein group consisting of proteins of the B7 family and TNFRSF14. In a preferred embodiment, the second membrane protein is a B7 family member. The B7 family is a collection of structurally related cell surface proteins that bind to proteins (CD28 family members) on lymphocytes that regulate immune responses. B7 family members are generally referred to as ligands, while CD28 family members are referred to as receptors. The activation of T and B lymphocytes is caused by the engagement of cell surface antigen-specific T cell receptors or B cell receptors, but the final immune response is determined by additional signals delivered simultaneously by one or more B7 ligands. B7 ligands deliver these "co-stimulation" or "co-inhibition" signals through the CD28 receptor family on lymphocytes. The interaction of B7 family members with co-stimulatory receptors enhances the immune response, while the interaction with co-inhibitory receptors weakens the immune response. Some preferred B7 family members are CD80; CD86; ICOS-L; PD-L1; PD-L2; B7-H3, B7-H4; B7-H6 and B7-H7. B7 ligands are expressed in lymphoid and non-lymphoid tissues. B7 ligands that transmit co-inhibitory signals are associated with neoplasms, virally infected cells, and cells infected with intracellular parasites and provide them with the ability to escape or at least suppress immune responses against them. Manipulation of signals delivered by B7 ligands has shown activity in the treatment of autoimmune diseases, inflammatory diseases, and cancer. In a preferred embodiment, the second membrane protein is PD-L1; PD-L2; ICOSL, CD80; CD86; B7-H3; B7-H4; TNFRSF14; B7-H6 or B7-H7. In a preferred embodiment, the second membrane protein is PD-L1; PD-L2; CD80; CD86; B7-H4; TNFRSF14; or B7-H7. In a particularly preferred embodiment, the second membrane protein is PD-L1; or PD-L2, preferably PD-L1.

[0145] In a particularly preferred embodiment, the first membrane protein is PD-1 and the second membrane protein is PD-L1.

[0146] When the first membrane protein is CD28, it is preferred that the second membrane protein is CD80, CD86 or ICOSL, preferably CD80. Therefore, the antibody or variant thereof is preferably an antibody or variant thereof comprising a variable domain that can bind to the extracellular part of CD28 and a variable domain that can bind to the extracellular part of CD80, the extracellular part of CD86, or the extracellular part that binds to ICOSL (preferably the extracellular part of CD80). CD28 and CD80 are binding partners. CD28 and CD86 are binding partners and CD28 and ICOSL are binding partners.

[0147] When the first membrane protein is CTLA-4, it is preferred that the second membrane protein is CD80, CD86 or ICOSL, preferably CD80. Therefore, the antibody or variant thereof is preferably an antibody or variant thereof comprising a variable domain that can bind to the extracellular portion of CTLA-4 and a variable domain that can bind to the extracellular portion of CD80 or the extracellular portion of CD86 (preferably the extracellular portion of CD80). CTLA-4 and CD80 are binding partners. CTLA-4 and CD86 are binding partners and CTLA-4 and ICOSL are binding partners.

[0148] When the first membrane protein is ICOS, it is preferred that the second membrane protein is ICOSL. Therefore, the antibody or variant thereof is preferably an antibody or variant thereof comprising a variable domain that can bind to the extracellular part of ICOS and a variable domain that can bind to the extracellular part of ICOSL. ICOS and ICOSL are binding partners.

[0149] When the first membrane protein is BTLA, it is preferred that the second membrane protein is B7-H4 or TNFRSF14, preferably TNFRSF14. Therefore, the antibody or variant thereof is preferably an antibody or variant thereof comprising a variable domain that can bind to the extracellular portion of BTLA and a variable domain that can bind to the extracellular portion of B7-H4 or the extracellular portion of TNFRSF14 (preferably the extracellular portion of TNFRSF14). BTLA and B7-H4 are binding partners and BTLA and TNFRSF14 are binding partners.

[0150] When the first membrane protein is NKp30, it is preferred that the second membrane protein is B7-H6. Therefore, the antibody or variant thereof is preferably an antibody or variant thereof comprising a variable domain that can bind to the extracellular portion of NKp30 and a variable domain that can bind to the extracellular portion of B7-H6. NKp30 and B7-H6 are binding partners.

[0151] When the first membrane protein is TMIGD2, it is preferred that the second membrane protein is B7-H7 (HHLA2). Therefore, the antibody is preferably an antibody comprising a variable domain that can bind to the extracellular portion of TMIGD2 and a variable domain that can bind to the extracellular portion of B7-H7 (HHLA2). TMIGD2 and B7-H7 are binding partners.

[0152] When the first membrane protein is PD-1, it is preferred that the second membrane protein is PD-L1 or PD-L2, preferably PD-L1. PD-1 and PD-L1 are binding partners. PD-1 and PD-L2 are binding partners. PD-L1 and CD80 are binding partners. The antibody or variant preferably comprises a variable domain that can bind to the extracellular portion of PD-1 and a variable domain that can bind to the extracellular portion of PD-L1 or the extracellular portion of PD-L2 (preferably the extracellular portion of PD-L1). The PD-1 / PD-L1 antibody or variant thereof and the PD-1 / PD-L2 antibody or variant thereof are preferably antibodies or variants thereof having a PD-1 binding variable domain, which, when provided as a bivalent monoclonal antibody comprising two said variable domains that bind to PD-1, inhibits PD-1 / PD-L1-mediated inhibition of T cell receptor-mediated Jurkat cell activation by 20% to 150% compared to the inhibition obtained with the antibody nivolumab on Jurkat cells. Some embodiments provide antibodies or variants according to the present invention, wherein the antibody or variant comprises a PD1 binding variable domain, which, when present in a bivalent monospecific antibody format, counteracts PD-1 / PD-L1-mediated inhibition of T cell receptor-mediated Jurkat cell activation to a greater extent than the antibody nivolumab. The variable domain that binds to the extracellular portion of PD-1 is defined as a variable domain that, when in a bivalent monospecific antibody format comprising two of said variable domains that bind PD-1, inhibits PD-1 / PD-L1-mediated inhibition of T cell receptor-mediated activation of Jurkat cells in the range of 20% to 150% compared to the inhibition obtained with the antibody nivolumab on Jurkat cells.

[0153] The inhibition of PD-1 inhibition of TCR-mediated Jurkat cell activation is preferably 50% to 150%, preferably 80% to 150%, more preferably 100% to 150%, compared to the inhibition obtained with the antibody nivolumab on the Jurkat cells. In a preferred embodiment, the inhibition is at least 100% compared to the inhibition obtained with the antibody nivolumab on the Jurkat cells. PD-1 inhibition of TCR-mediated Jurkat cell activation is preferably measured by measuring the immunosuppressive effect of PD-1 / PD-L1 binding in Jurkat cells, which are incubated under conditions of activation by T cell receptors if the antibody or its variant is not present. Suitable assays for determining inhibition are described in the Examples. Suitable Jurkat cell lines are described in the Examples.

[0154] The variable domain that can bind to the extracellular portion of PD-L1 is preferably a variable domain that, when present in a monovalent monoclonal antibody, binds to the antibody with a K of 0.1 to 14 nM as measured by surface plasmon resonance (SPR). D Binds to PD-L1, preferably with a K of 0.5 to 14 nM as measured by SPR D , preferably a K of 1 to 14 nM D , preferably a K of 1 to 12 nM D , preferably a K of 2 to 12 nM D In some preferred embodiments, the antibody or variant according to the invention comprises a variable domain that can bind to the extracellular portion of PD-L1, wherein when present in a bispecific antibody with a second variable domain that binds an unrelated antigen (e.g., tetanus toxoid), the variable domain binds to the extracellular portion of PD-L1 with a K of less than or equal to 4.27 nM, preferably less than or equal to 1.31 nM, preferably less than or equal to 1.27 nM, as measured by surface plasmon resonance (SPR). D Preferably, when present in a bispecific antibody with a second variable domain that binds an unrelated antigen (e.g., tetanus toxoid), the PD-L1-specific variable domain binds to PD-L1 with a K of 0.9 to 4.27 nM. D , preferably with a K of 0.94 to 4.27 nM D , or with a K of 0.9 to 1.31 nM D , preferably with a K of 0.94 to 1.31 nM D Binding to PD-L1. The antibody or variant according to the present invention comprising a variable domain that can bind to the extracellular portion of PD-1 and a variable domain that can bind to the extracellular portion of PD-L1 or the extracellular portion of PD-L2 is preferably an antibody or variant thereof comprising a variable domain that can bind to the extracellular portion of PD-1 and block the binding of PD-1 to PD-L1. Preferably, the variable domain that can bind to the extracellular portion of PD-L1 or the extracellular portion of PD-L2 is a variable domain that blocks the binding of PD-1 to PD-L1 or PD-L2, respectively. The variable domain that can bind to PD-1 and block the binding of PD-1 to PD-L1 is preferably a domain that also blocks the binding of PD-1 to PD-L2. The variable domain of the present invention can be a variable domain that can bind to PD-L1 and block the binding of PD-L1 to PD-1 and / or PD-L1 to CD80. This provides the advantage that these antibodies are also able to counteract the resistance of tumors to treatment through the PD-1 / PD-L2 pathway.

[0155] The antibodies or variants thereof according to the present invention preferably reduce the activity of the binding pair. The reduced activity is usually achieved by blocking the ability of the binding pair to bind to each other. The ability can be blocked by blocking any one or preferably both members of the binding pair. When the binding pair is co-inhibitory, the inhibition or co-inhibitory activity is reduced by blocking. When the binding pair is co-stimulatory, the co-stimulatory activity is reduced by blocking.

[0156] In oligoclonal embodiments, the preference for CD28 family members and B7 family members or TNFRSF14 is the same.

[0157] The present invention also provides a method for engaging and / or activating T cells, comprising providing a system comprising T cells and cells to be engaged or activated with the T cells; and providing one or more antibodies to the system, each antibody comprising a variable domain that can bind to a CD28 family member and a variable domain that can bind to an extracellular portion of a B7 family member or TNFRSF14, and incubating the system under conditions that allow T cells to be engaged and / or activated. In some embodiments, the method is an in vitro method. Some embodiments provide a method for engaging and / or activating T cells, comprising providing a system comprising T cells and cells to be engaged or activated with the T cells; and providing one or more antibodies to the system, each antibody comprising a variable domain that can bind to an extracellular portion of PD-1 and a variable domain that can bind to an extracellular portion of PD-L1, and incubating the system under conditions that allow T cells to be engaged and / or activated. The cells to be engaged or activated with the T cells are preferably immune cells, such as antigen presenting cells, macrophages, neoplastic cells, virus-infected cells, or cells infected with intracellular parasites. Engagement and / or activation of T cells directs T cells to specific targets. Activating T cells is to activate the T cell receptor of the T cells. Engaging T cells is generally to activate T cells. Engagement can also direct activated T cells to targets specified by antibodies. The conditions that allow the T cells to be engaged and / or activated are generally culture conditions, but can also be incubated in animals, and therefore particularly methods of treatment can be encompassed. The conditions are generally such that in the absence of antibodies, T cells are not engaged. If a collection of T cells is measured, some of these have been engaged or activated, provided that the collection contains enough T cells that are not engaged or activated.

[0158] The antibodies of the invention can bring two cells into close proximity, thereby allowing interactions between the cells mediated by proteins other than the receptor-ligand pair bound by the antibodies of the invention. One such interaction is the interaction of a T cell receptor on one cell with an MHC on another cell.

[0159] The antibody or variant thereof comprising a variable domain that can bind to the extracellular portion of PD-1 preferably comprises a heavy chain variable region having a CDR3 region comprising the amino acid sequence of the CDR3 region of the following heavy chain variable region: MF6076; MF6236; MF6256; MF6226; MF6930; MF6932; MF6935; MF6936; MF6972; MF6974; MF6982; MF6929; MF7699; MF7698; MF7687; MF7686; MF7685; or MF7684 ( Figure 3 and / or Fig.13 ).

[0160] The antibody or variant thereof comprising a variable domain that can bind to the extracellular portion of PD-1 preferably comprises a heavy chain variable region having CDR1, CDR2 and CDR3 regions, wherein the CDR1, CDR2 and CDR3 regions comprise the amino acid sequences of CDR1, CDR2 and CDR3 of the heavy chain variable region of one of the VHs shown below: MF6076; MF6236; MF6256; MF6226; MF6930; MF6932; MF6935; MF6936; MF6972; MF6974; MF6982; MF6929; MF7699; MF7698; MF7687; MF7686; MF7685; or MF7684 ( Figure 3 and / or Fig.13 ). The CDR1, CDR2 and CDR3 sequences are preferably selected from the same VH region.

[0161] The antibody or variant thereof comprising a variable domain that can bind to the extracellular part of PD-1 preferably comprises the amino acid sequence of the heavy chain variable region of MF6076; MF6256; MF6226; MF6930; MF6932; MF6935; MF6936; MF6972; MF6974; MF6982; MF6929; MF7699; MF7698; MF7687; MF7686; MF7685; or MF7684 ( Figure 3 and / or Fig.13 ). Amino acid insertions, deletions, substitutions or combinations thereof (if any) are preferably not in the amino acid sequences of the CDR regions.

[0162] The antibody or variant thereof comprising a variable domain that can bind to the extracellular portion of PD-L1 preferably comprises a heavy chain variable region having a CDR3 region comprising the amino acid sequence of the CDR3 region of the following heavy chain variable region: MF5359; MF5361; MF5377; MF5382; MF5424; MF5426; MF5439; MF5442; MF5553; MF5557; MF5561; MF5576; MF5594; MF5708; MF5442; MF7691; MF7690; MF7689; MF7688; MF7700; MF7701; MF7703; MF7694; MF7693; MF7692; MF7697; MF7696; or MF7695( Figure 3 and / or Fig.13 ).

[0163] The antibody or variant thereof comprising a variable domain that can bind to the extracellular portion of PD-L1 preferably comprises a heavy chain variable region having CDR1, CDR2 and CDR3 regions, wherein the CDR1, CDR2 and CDR3 regions comprise the amino acid sequences of CDR1, CDR2 and CDR3 of the heavy chain variable region of one of the VHs shown below: MF5359; MF5361; MF5377; MF5382; MF5424; MF5426 ;MF5439;MF5442;MF5553;MF5557;MF5561;MF5576;MF5594;MF5708;MF5442;MF7691;MF7690;MF7689;MF7688;MF7700;MF7701;MF7703;MF7694;MF7693;MF7692;MF7697;MF7696; or MF7695( Figure 3 and / or Fig.13 ). The CDR1, CDR2 and CDR3 sequences are preferably selected from the same VH region.

[0164] The antibody or variant thereof comprising a variable domain that can bind to the extracellular part of PD-L1 preferably comprises the following amino acid sequences of the heavy chain variable region having up to 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the amino acid sequence of the specified MF: MF5359; MF5361; MF5377; MF5382; MF542 4; MF5426; MF5439; MF5442; MF5553; MF5557; MF5561; MF5576; MF5594; MF5708; MF5442; MF7691; MF7690; MF7689; MF7688; MF7700; MF7701; MF7703; MF7694; MF7693; MF7692; MF7697; MF7696; or MF7695( Figure 3 and / or Fig.13 ). Amino acid insertions, deletions, substitutions or combinations thereof (if any) are preferably not in the amino acid sequences of the CDR regions.

[0165] The antibody or variant thereof preferably comprises a variable domain that blocks the binding of PD-1 to PD-L1 and that binds to the extracellular portion of PD-1 and a variable domain that blocks the binding of PD-1 to PD-L1 and that binds to the extracellular portion of PD-L1. The variable domain that binds to PD-1 and blocks the binding of PD-1 to PD-L1 preferably also blocks the binding of PD-1 to PD-L2. The variable domain that binds to PD-L1 and blocks the binding of PD-1 to PD-L1 preferably also blocks the binding of PD-L1 to CD80. This provides the advantage that the immunosuppression of tumor cells by the interaction between PD-1 and CD80 can also be counteracted.

[0166] The antibodies or variants thereof of the present invention, such as bispecific antibodies or variants thereof, preferably comprise a variable domain that can bind to the extracellular portion of PD-1 and block the interaction of PD-1 with PD-L1 and / or PD-L2, and comprise a variable domain that can bind to the extracellular portion of PD-L1 and block the interaction of PD-L1 with PD-1 and / or CD80. Preferably, such molecules of the present invention are capable of blocking the complete PD-1 axis, including PD-1 with PD-L1 or PD-L2, and PD-L1 with PD-1 and CD80.

[0167] The variable domain that binds to the extracellular portion of PD-L1 in such an antibody or variant thereof preferably comprises a VH region having the amino acid sequence of CDR3 or the amino acid sequence of CDR1, CDR2 and CDR3 of one of the following VHs: MF5359; MF5377; MF5382; MF5424; MF5426; MF5439; MF5442; MF5553; MF5557; MF5561; MF5576; MF5594; MF5708; MF5442; MF7691; MF7690; MF7689; MF7688; MF7700; MF7701; MF7703; MF7694; MF7693; MF7692; MF7697; MF7696; or MF7695( Figure 3 and / or Fig.13 ). In a preferred embodiment, the variable domain that binds to the extracellular portion of PD-L1 comprises a VH region having an amino acid sequence of the following VH with up to 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the amino acid sequence of a specified MF: MF5359; MF5377; MF5382; MF5424 ;MF5426;MF5439;MF5442;MF5553;MF5557;MF5561;MF5576;MF5594;MF5708;MF5442;MF7691;MF7690;MF7689;MF7688;MF7700;MF7701;MF7703;MF7694;MF7693;MF7692;MF7697;MF7696; or MF7695( Figure 3 and / or Fig.13 ).

[0168] The variable domain that binds to the extracellular portion of PD-1 in the antibody or variant thereof preferably comprises a VH region having the amino acid sequence of CDR3 or the amino acid sequence of CDR1, CDR2 and CDR3 of one of the following VHs: MF6076; MF6236; MF6256; MF6226; MF6930; MF6932; MF6935; MF6936; MF6972; MF6974; MF6982; MF6929; MF7699; MF7698; MF7687; MF7686; MF7685; or MF7684 ( Figure 3 and / or Fig.13In a preferred embodiment, the variable domain that binds to the extracellular portion of PD-1 comprises a VH region having an amino acid sequence of the following VH with up to 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the amino acid sequence of the VH of a specified MF: MF6076; MF6236; MF6256; MF6226; MF6930; MF6932; MF6935; MF6936; MF6972; MF6974; MF6982; MF6929; MF7699; MF7698; MF7687; MF7686; MF7685; or MF7684 ( Figure 3 and / or Fig.13 ). Amino acid insertions, deletions, substitutions or combinations thereof (if any) are preferably not in the amino acid sequence of the CDR region. A particularly preferred combination in the antibody or variant thereof is a combination of variable domains comprising the specified sequences of MF5382 and MF6256 or variants thereof.

[0169] Some embodiments provide an antibody or variant according to the invention, comprising a variable domain that can bind to the extracellular portion of PD-1 and a variable domain that can bind to the extracellular portion of PD-L1, wherein the antibody or variant has a stronger CD4+ T cell activation potential in a Staphylococcus enterotoxin B (SEB) assay compared to the following equimolar mix:

[0170] - a bivalent monospecific antibody comprising two of said variable domains that bind to PD-1, and

[0171] - a bivalent monospecific antibody comprising two of said variable domains binding to PD-L1. The antibodies or variants according to these embodiments are preferred over an equimolar mixture of the parent antibodies in view of their stronger CD4+ T cell activation potential.

[0172] Some embodiments provide an antibody or variant according to the present invention, which comprises a variable domain that can bind to the extracellular portion of PD-1 and a variable domain that can bind to the extracellular portion of PD-L1, and the antibody or variant can activate T cells more strongly in an antigen-specific CD4+T cell assay compared to the benchmark antibody 5C4 or the benchmark antibody YW243.55.S70 or the combination of the benchmark antibody 5C4 and YW243.55.S70. In view of its stronger antigen-specific CD4+T cell activation potential, the antibody or variant according to this embodiment is preferred over the anti-PD-1 benchmark antibody Nivolumab and the anti-PD-L1 benchmark antibody Atezolizumab. It is noteworthy that the antibodies or variants according to these embodiments are even preferred over the combination of Nivolumab and Atezolizumab.

[0173] Some embodiments provide an antibody or variant according to the present invention, comprising a variable domain that can bind to the extracellular portion of PD-1 and a variable domain that can bind to the extracellular portion of PD-L1, and the antibody or variant has a stronger CD4+T cell activation potential in a mixed lymphocyte reaction (MLR) assay compared to the benchmark antibody 5C4 based on nivolumab or the benchmark antibody YW243.55.S70 based on atezolizumab. It is noteworthy that the antibodies or variants according to these embodiments are preferred over nivolumab and atezolizumab in view of their stronger CD4+T cell activation potential.

[0174] Some embodiments provide an antibody or variant according to the invention, comprising a variable domain that can bind to the extracellular portion of PD-1 and a variable domain that can bind to the extracellular portion of PD-L1, and the antibody or variant is capable of enhancing the proliferation of CD4+ and / or CD8+ tumor-infiltrating T cells. In view of its tumor-infiltrating T cell activation potential, the antibodies or variants according to these embodiments are particularly suitable for inducing or enhancing T cell-mediated anti-tumor responses.

[0175] Some embodiments provide antibodies or variants according to the present invention, which comprise a variable domain that can bind to the extracellular portion of PD-1 and a variable domain that can bind to the extracellular portion of PD-L1, and the antibodies or variants are capable of inducing a stronger T cell-mediated anti-tumor response in vivo compared to a combination of a benchmark antibody MK-3475 based on Pembrolizumab and YW243.55.S70 based on Atezolizumab. Therefore, the antibodies or variants according to these embodiments are particularly suitable for inducing or enhancing T cell-mediated anti-tumor responses in vivo. It is noteworthy that the antibodies or variants according to these embodiments are preferred over Pembrolizumab and Atezolizumab in view of their stronger T cell activation potential.

[0176] As shown in Table 4, when the PD-1-specific Fab arm 6076 was combined with the PD-L1-specific Fab arm MF5553, MF5359, MF5424, MF5561, MF5442 or MF5382, a bispecific antibody was formed that showed at least 60% blocking in a PD-1 / PD-L1 reporter assay.

[0177] The antibody or variant thereof as described herein preferably comprises:

[0178] - a PD-1 binding variable domain comprising a VH region having the amino acid sequence of CDR3 or the amino acid sequences of CDR1, CDR2 and CDR3 of VH of MF6076; and

[0179] - a PD-L1 binding variable domain comprising a VH region having the amino acid sequence of CDR3 or the amino acid sequences of CDR1, CDR2 and CDR3 of the following VH: MF5553, MF5359, MF5424, MF5561, MF5442 or MF5382, preferably MF5359, MF5424, MF5561, MF5442 or MF5382, preferably MF5359, MF5424, MF5561, MF5442 or MF5382, preferably MF5359, MF5424, MF5561 or MF5442, preferably MF5359, MF5424 or MF5442, preferably MF5359 or MF5442, preferably MF5442 ( Figure 3 ).

[0180] The antibody or variant thereof preferably comprises:

[0181] - a PD-1 binding variable domain comprising a VH region having an amino acid sequence of VH of MF6076 having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the amino acid sequence of VH of MF6076; and

[0182] - a PD-L1 binding variable domain comprising a VH region having an amino acid sequence of the following VH with up to 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the amino acid sequence of the VH of a specified MF: MF5553, MF5359, MF5424, MF5561, MF5442 or MF5382, preferably MF5359, MF5424, MF5561, MF5442 or MF5382, preferably MF5359, MF5424, MF5561, MF5442 or MF5382 ( Figure 3 ).

[0183] As shown in Table 4, when the PD-1 specific Fab arm MF6236 was combined with the PD-L1 specific Fab arm MF5561, MF5442 or MF5382, a bispecific antibody was formed that showed at least 60% blocking in the PD-1 / PD-L1 reporter assay. The antibody or variant thereof as described herein preferably comprises:

[0184] -PD-1 binding variable domain, comprising a VH region having MF6236 ( Figure 3 )'s VH CDR3 amino acid sequence or the amino acid sequences of CDR1, CDR2 and CDR3; and

[0185] - a PD-L1 binding variable domain comprising a VH region having the amino acid sequence of CDR3 or the amino acid sequences of CDR1, CDR2 and CDR3 of the following VH: MF5561; MF5442 or MF5382, preferably MF5561 or MF5442, preferably MF5442 ( Figure 3 ).

[0186] The antibody or variant thereof preferably comprises:

[0187] -PD-1 binding variable domain comprising a VH region having a Figure 3) having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof; and

[0188] - a PD-L1 binding variable domain comprising a VH region having an amino acid sequence of the following VH with up to 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the amino acid sequence of the VH of a specified MF: MF5561; MF5442 or MF5382, preferably MF5561 or MF5442, preferably MF5442 ( Figure 3 ).

[0189] As shown in Table 4, when the PD-1 specific Fab arm MF6974 is combined with the PD-L1 specific Fab arm MF5424, MF5561, MF5442 or MF5382, a bispecific antibody is formed that shows at least 60% blocking in the PD-1 / PD-L1 reporter assay. The antibody or variant thereof as described herein preferably comprises:

[0190] -PD-1 binding variable domain comprising a VH region having MF6974 ( Figure 3 )'s VH CDR3 amino acid sequence or the amino acid sequences of CDR1, CDR2 and CDR3; and

[0191] - a PD-L1 binding variable domain comprising a VH region having the amino acid sequence of CDR3 or the amino acid sequences of CDR1, CDR2 and CDR3 of the following VH: MF5424, MF5561, MF5442 or MF5382, preferably MF5424, MF5561 or MF5442, preferably MF5424 or MF5442, preferably MF5442 ( Figure 3 ).

[0192] The antibody or variant thereof preferably comprises:

[0193] -PD-1 binding variable domain comprising a VH region having a Figure 3 ) having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof; and

[0194] - a PD-L1 binding variable domain comprising a VH region having an amino acid sequence of VH of MF5424, MF5561, MF5442 or MF5382, preferably MF5424, MF5561 or MF5442, preferably MF5424 or MF5442, preferably MF5442( Figure 3 ).

[0195] As shown in Table 4, when the PD-1 specific Fab arm MF6935 was combined with the PD-L1 specific Fab arm MF5424, MF5561, MF5442 or MF5382, a bispecific antibody was formed that showed at least 60% blocking in the PD-1 / PD-L1 reporter assay. The antibody or variant thereof as described herein preferably comprises:

[0196] -PD-1 binding variable domain, comprising a VH region having MF6935 ( Figure 3 )'s VH CDR3 amino acid sequence or the amino acid sequences of CDR1, CDR2 and CDR3; and

[0197] - a PD-L1 binding variable domain comprising a VH region having the amino acid sequence of CDR3 or the amino acid sequences of CDR1, CDR2 and CDR3 of the following VH: MF5424, MF5561, MF5442 or MF5382, preferably MF5561, MF5442 or MF5382, preferably MF5442 or MF5382, preferably MF5382 ( Figure 3 ).

[0198] The antibody or variant thereof preferably comprises:

[0199] -PD-1 binding variable domain comprising a VH region having a Figure 3 ) having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof; and

[0200] - a PD-L1 binding variable domain comprising a VH region having an amino acid sequence of the following VH with up to 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the amino acid sequence of the VH of a specified MF: MF5424, MF5561, MF5442 or MF5382, preferably MF5561, MF5442 or MF5382, preferably MF5442 or MF5382, preferably MF5382( Figure 3 ).

[0201] As shown in Table 4, when the PD-1 specific Fab arm MF6936 was combined with the PD-L1 specific Fab arm MF5576, MF5424, MF5561, MF5557, MF5442 or MF5382, a bispecific antibody was formed that showed at least 60% blocking in the PD-1 / PD-L1 reporter assay. The antibody or variant thereof as described herein preferably comprises:

[0202] -PD-1 binding variable domain, comprising a VH region having MF6936 ( Figure 3 )'s VH CDR3 amino acid sequence or the amino acid sequences of CDR1, CDR2 and CDR3; and

[0203] - a PD-L1 binding variable domain comprising a VH region having the amino acid sequence of CDR3 or the amino acid sequences of CDR1, CDR2 and CDR3 of the following VH: MF5576, MF5424, MF5561, MF5557, MF5442 or MF5382, preferably MF5424, MF5561, MF5557, MF5442 or MF5382, preferably MF5424, MF5561, MF5442 or MF5382, preferably MF5561, MF5442 or MF5382, preferably MF5442 or MF5382, preferably MF5382 ( Figure 3 ).

[0204] The antibody or variant thereof preferably comprises:

[0205] -PD-1 binding variable domain comprising a VH region having a Figure 3 ) having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof; and

[0206] - a PD-L1 binding variable domain comprising a VH region having an amino acid sequence of the following VH with up to 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the amino acid sequence of the VH of a specified MF: MF5576, MF5424, MF5561, MF5557, MF5442 or MF5382, preferably MF5424, MF5561, MF5557, MF5442 or MF5382, preferably MF5424, MF5561, MF5442 or MF5382, preferably MF5561, MF5442 or MF5382, preferably MF5382 ( Figure 3 ).

[0207] As shown in Table 4, when the PD-1 specific Fab arm MF6256 is combined with the PD-L1 specific Fab arm MF5576, MF5424, MF5561, MF5557, MF5439, MF5442 or MF5382, a bispecific antibody is formed that shows at least 60% blocking in the PD-1 / PD-L1 reporter assay. The antibody or variant thereof as described herein preferably comprises:

[0208] -PD-1 binding variable domain comprising a VH region having MF6256 ( Figure 3 )'s VH CDR3 amino acid sequence or the amino acid sequences of CDR1, CDR2 and CDR3; and

[0209] - a PD-L1 binding variable domain comprising a VH region having the amino acid sequence of CDR3 or the amino acid sequences of CDR1, CDR2 and CDR3 of the following VH: MF5576, MF5424, MF5561, MF5557, MF5439, MF5442 or MF5382, preferably MF5576, MF5424, MF5561, MF5557, MF5442 or MF5382, preferably MF5576, MF5561, MF5557, MF5442 or MF5382, preferably MF5576, MF5442 or MF5382, preferably MF5442 or MF5382, preferably MF5382 ( Figure 3 ).

[0210] The antibody or variant thereof preferably comprises:

[0211] -PD-1 binding variable domain comprising a VH region having a Figure 3 ) having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof; and

[0212] - a PD-L1 binding variable domain comprising a VH region having an amino acid sequence of VH of the following MF5576, MF5424, MF5561, MF5557, MF5439, MF5442 or MF5443 with up to 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the amino acid sequence of VH of a specified MF: MF5576, MF5424, MF5561, MF5557, MF5439, MF5442 or MF5443 MF5382, preferably MF5576, MF5424, MF5561, MF5557, MF5442 or MF5382, preferably MF5576, MF5561, MF5557, MF5442 or MF5382, preferably MF5576, MF5561, MF5442 or MF5382, preferably MF5576, MF5442 or MF5382, preferably MF5442 or MF5382, preferably MF5382 ( Figure 3 ).

[0213] In some preferred embodiments, the antibody or variant thereof as described herein comprises:

[0214] - a PD-1 binding variable domain comprising a VH region having the amino acid sequence of CDR3 or the amino acid sequences of CDR1, CDR2 and CDR3 of VH of MF6076; and

[0215] - a PD-L1 binding variable domain comprising a VH region having the amino acid sequence of CDR3 or the amino acid sequences of CDR1, CDR2 and CDR3 of the following VHs: MF5442, MF7691, MF7690, MF7689, MF7688, MF7700, MF7701, MF7703, MF7694, MF7693, MF7692, MF7697, MF7696, or MF7695, preferably MF7703 or MF7689 ( Figure 3 and / or Fig.13 ).

[0216] The antibody or variant thereof preferably comprises:

[0217] - a PD-1 binding variable domain comprising a VH region having an amino acid sequence of VH of MF6076 having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the amino acid sequence of VH of MF6076; and

[0218] - a PD-L1 binding variable domain comprising a VH region having an amino acid sequence of the following VH with up to 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the amino acid sequence of the VH of a specified MF: MF5442, MF7691, MF7690, MF7689, MF7688, MF7700, MF7701, MF7703, MF7694, MF7693, MF7692, MF7697, MF7696, or MF7695, preferably MF7703 or MF7689 ( Figure 3 and / or Fig.13 ).

[0219] In some preferred embodiments, the antibody or variant thereof as described herein comprises:

[0220] -PD-1 binding variable domain comprising a VH region having MF6974 ( Figure 3 )'s VH CDR3 amino acid sequence or the amino acid sequences of CDR1, CDR2 and CDR3; and

[0221] - A PD-L1 binding variable domain comprising a VH region having the amino acid sequence of CDR3 or the amino acid sequences of CDR1, CDR2 and CDR3 of the following VH: MF7703 or MF7689, preferably MF7703 or MF7689, more preferably MF7689 ( Figure 3 and / or Fig.13 ).

[0222] The antibody or variant thereof preferably comprises:

[0223] -PD-1 binding variable domain comprising a VH region having a Figure 3 ) having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof; and

[0224] - a PD-L1 binding variable domain comprising a VH region having an amino acid sequence of the following VH with up to 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the amino acid sequence of the VH of a specified MF: MF7703 or MF7689, preferably MF7703 or MF7689, more preferably MF7689( Figure 3 and / or Fig.13 ).

[0225] In some preferred embodiments, the antibody or variant thereof as described herein comprises:

[0226] -PD-1 binding variable domain comprising a VH region having MF7686 ( Fig.13 )'s VH CDR3 amino acid sequence or the amino acid sequences of CDR1, CDR2 and CDR3; and

[0227] - A PD-L1 binding variable domain comprising a VH region having the amino acid sequence of CDR3 or the amino acid sequences of CDR1, CDR2 and CDR3 of the following VH: MF5359, MF5361, MF5377, MF5382, MF5424, MF5426, MF5439, MF5442, MF5553, MF5557, MF5561, MF5576, MF5594, MF5708, MF7703 or MF7689, preferably MF7703 or MF7689, more preferably MF7703 ( Figure 3 and / or Fig.13 ).

[0228] The antibody or variant thereof preferably comprises:

[0229] -PD-1 binding variable domain comprising a VH region having a Fig.13 ) having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof; and

[0230] - a PD-L1 binding variable domain comprising a VH region having an amino acid sequence of the following VH with at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the amino acid sequence of the VH of a specified MF: MF5359, MF5361, MF5377, MF5382, MF5424, MF5426, MF5439, MF5442, MF5553, MF5557, MF5561, MF5576, MF5594, MF5708, MF7703 or MF7689, preferably MF7703 or MF7689, more preferably MF7703( Figure 3 and / or Fig.13 ).

[0231] Some preferred embodiments provide an antibody or a variant thereof according to the present invention, wherein the antibody or variant thereof comprises a variable domain that can bind to the extracellular portion of PD-1, and wherein the variable domain that can bind to the extracellular portion of PD-1 comprises:

[0232] - a heavy chain CDR3 sequence that is identical to the heavy chain CDR3 sequence of the variable region selected from MF6974, MF6076 and MF7686; or

[0233] - a heavy chain CDR3 sequence that deviates from the heavy chain CDR3 sequence of MF6974 or MF6076 or MF7686 by no more than three, preferably by no more than two, more preferably by no more than one amino acid; or

[0234] - a heavy chain CDR1, CDR2 and CDR3 sequence that is identical to the heavy chain CDR1, CDR2 and CDR3 sequence of the variable region selected from MF6974, MF6076 and MF7686; or

[0235] - a heavy chain CDR1 , CDR2 and CDR3 sequence that deviates by no more than three, preferably by no more than two, more preferably by no more than one amino acid from the heavy chain CDR1 , CDR2 and CDR3 sequence of MF6974 or MF6076 or MF7686.

[0236] Some preferred embodiments provide an antibody or a variant thereof according to the present invention, wherein the antibody or variant thereof comprises a variable domain that can bind to the extracellular portion of PD-1, and wherein the variable domain that can bind to the extracellular portion of PD-1 comprises:

[0237] - a heavy chain variable region comprising the amino acid sequence of MF6974 or MF6076 or MF7686, or

[0238] - a heavy chain variable region having a sequence that is at least 80%, or at least 85%, or at least 90%, or at least 91%, or at least 92%, or at least 93%, or at least 94%, or at least 95%, or at least 96%, or at least 97%, or at least 98%, or at least 99% identical to the amino acid sequence of MF6974 or MF6076 or MF7686.

[0239] Some preferred embodiments provide an antibody or a variant thereof according to the present invention, comprising a variable domain that can bind to the extracellular portion of PD-L1, and wherein the variable domain that can bind to the extracellular portion of PD-L1 comprises:

[0240] - a heavy chain CDR3 sequence that is identical to the heavy chain CDR3 sequence of the variable region selected from MF7689 and MF7703; or

[0241] - a heavy chain CDR3 sequence that deviates from the heavy chain CDR3 sequence of MF7689 or MF7703 by no more than three, preferably by no more than two, more preferably by no more than one amino acid; or

[0242] - a heavy chain CDR1, CDR2 and CDR3 sequence that is identical to the heavy chain CDR1, CDR2 and CDR3 sequence of the variable region selected from MF7689 and MF7703; or

[0243] - A heavy chain CDR1 and CDR2 and CDR3 sequence that deviates from the heavy chain CDR1 and CDR2 and CDR3 sequence of MF7689 or MF7703 by no more than three, preferably by no more than two, more preferably by no more than one amino acid.

[0244] Some preferred embodiments provide an antibody or a variant thereof according to the present invention, comprising a variable domain that can bind to the extracellular portion of PD-L1, and wherein the variable domain that can bind to the extracellular portion of PD-L1 comprises:

[0245] - a heavy chain variable region comprising the amino acid sequence of MF7689 or MF7703, or

[0246] - A heavy chain variable region having a sequence that is at least 80%, or at least 85%, or at least 90%, or at least 91%, or at least 92%, or at least 93%, or at least 94%, or at least 95%, or at least 96%, or at least 97%, or at least 98%, or at least 99% identical to the amino acid sequence of MF7689 or MF7703.

[0247] Some preferred embodiments provide antibodies or variants thereof according to the present invention, which antibodies or variants thereof comprise a variable domain that can bind to the extracellular portion of PD-1 and a variable domain that can bind to the extracellular portion of PD-L1, wherein the variable domain that can bind to the extracellular portion of PD-1 comprises heavy chain CDR1, CDR2 and CDR3 sequences that are identical to the heavy chain CDR1, CDR2 and CDR3 sequences of the variable regions selected from MF6974, MF6076 and MF7686, and wherein the variable domain that can bind to the extracellular portion of PD-L1 comprises heavy chain CDR1, CDR2 and CDR3 sequences that are identical to the heavy chain CDR1, CDR2 and CDR3 sequences of the variable regions selected from MF7689 and MF7703.

[0248] Some preferred embodiments provide an antibody or variant thereof according to the present invention, the antibody or variant thereof comprising a variable domain that can bind to the extracellular portion of PD-1 and a variable domain that can bind to the extracellular portion of PD-L1, wherein the variable domain that can bind to the extracellular portion of PD-1 comprises a heavy chain CDR1, CDR2 and CDR3 sequence identical to the heavy chain CDR1, CDR2 and CDR3 sequence of MF6974, and wherein the variable domain that can bind to the extracellular portion of PD-L1 comprises a heavy chain CDR1, CDR2 and CDR3 sequence identical to the heavy chain CDR1, CDR2 and CDR3 sequence of MF7689. As shown in the Examples, the antibody or variant according to this embodiment has good PD-1 / PD-L1 blocking activity and strong T cell activation potential.

[0249] Some embodiments provide an antibody or variant thereof according to the present invention, the antibody or variant thereof comprising a variable domain that can bind to the extracellular portion of PD-1 and a variable domain that can bind to the extracellular portion of PD-L1, wherein the variable domain that can bind to the extracellular portion of PD-1 comprises a heavy chain CDR1, CDR2, and CDR3 sequence identical to the heavy chain CDR1, CDR2, and CDR3 sequence of MF7686, and wherein the variable domain that can bind to the extracellular portion of PD-L1 comprises a heavy chain CDR1, CDR2, and CDR3 sequence identical to the heavy chain CDR1, CDR2, and CDR3 sequence of MF7703. As shown in the Examples, the antibody or variant according to this embodiment has good PD-1 / PD-L1 blocking activity and strong T cell activation potential.

[0250] In another embodiment, the antibody or variant thereof comprises a variable domain that can bind to the extracellular portion of PD-L1 and does not block the binding of PD-1 to PD-L1. The variable domain that binds to the extracellular portion of PD-L1 in the antibody or variant thereof preferably comprises MF5361 ( Figure 3) or the amino acid sequences of CDR3 of VH of MF5361. In a preferred embodiment, the variable domain that binds to the extracellular portion of PD-L1 comprises the amino acid sequence of VH of MF5361 having up to 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the amino acid sequence of a specified MF.

[0251] The up to 15, preferably 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10 and preferably 1, 2, 3, 4 or 5 amino acid substitutions in the H, VH, L and VL regions are preferably conservative amino acid substitutions, insertions, deletions, substitutions or a combination thereof are preferably not in the CDR3 region of the H, VH, L or VL chain, preferably not in the CDR1, CDR2 or CDR3 region of the VH or VL chain and preferably not in the FR4 region.

[0252] The methods as described herein preferably use an antibody or variant thereof as described herein.

[0253] The antibody or variant thereof according to the present invention, such as a part, derivative or analog thereof, preferably comprises two of the variable domains. Such an antibody is preferably a bispecific antibody or a variant thereof. Two or more antibodies or variants thereof can be linked together. Various methods are known in the art. A suitable method is conjugation. In addition, the technology for making multispecific antibodies has been developed to also include such bispecific antibodies, which have the same overall structure as a normal monospecific antibody, but in which the two arms of the antibody each bind to a different target. The bispecific antibody or variant thereof preferably has two heavy chains with compatible heterodimerization domains. The light chain is preferably a common light chain. Some embodiments provide antibodies or variants according to the present invention, wherein the antigen binding site of the antibody or variant is composed of an immunoglobulin variable domain that can bind to the extracellular portion of PD-1 and an immunoglobulin variable domain that can bind to the extracellular portion of PD-L1 or PD-L2. The antibody according to the present invention is preferably a full-length bispecific antibody, which consists of two heavy chains with compatible heterodimerization domains. In some embodiments, the antibody according to the invention is an IgG, preferably IgG1 or IgG3 or IgG4, more preferably IgG1 or IgG4, most preferably IgG1. The IgG format generally offers the advantages of longer half-life and / or reduced immunogenicity compared to other formats.

[0254] In some embodiments, the antibody or variant according to the invention is monovalent for PD-1 and monovalent for PD-L1.

[0255] The light chain of the antibody or variant according to the invention is preferably a common light chain.

[0256] The term "conjugate" as used herein refers to two or more molecules that have been covalently linked optionally through a connecting region. For example, in some embodiments, a conjugate is a first protein or non-protein part that is connected to a second protein or non-protein part through a connecting region. For example, in some embodiments of the binding molecules of the present invention, it comprises or is composed of two or more antibodies that have been covalently linked. The conjugate is not limited to the first and second parts, but in some embodiments, it may also have a third, fourth or more parts connected by another connecting region. As described elsewhere in this application, some examples of protein parts include but are not limited to: polypeptides, peptide mimetics or antibodies (or antibody parts, derivatives or analogs, as described elsewhere in this application). Some examples of non-protein parts include but are not limited to aptamers. Various types of joints can be used, and suitable joints will be selected according to the desired properties of the molecule type and joint in the conjugate (length, flexibility, resistance to protease activity and other similar features). Such joints may include nucleotides, polypeptides or suitable synthetic materials. For example, the joint may be a flexible peptide joint. In certain embodiments, the joint may be a cleavable joint that allows the parts of the conjugate to be separated from each other. In other embodiments, the peptide joint may be a spiral joint. Various examples and kits for connecting proteins to other molecules are well known in the art. As used herein, the term "fusion protein" refers to a protein that contains two or more polypeptides or proteins that have been connected at the DNA level by recombination and expressed together as a single polypeptide. The fusion protein may also include a peptide linker region that is also encoded by the DNA and expressed together with the fusion protein. The peptide linker that is part of the fusion protein can be designed to have specific characteristics, such as flexibility, hydrophilicity, protease resistance, cleavage, etc. All of these characteristics can be designed within the DNA sequence, and methods for designing linkers are well known in the art. For example, antibodies can be linked together by methods well known in the art, and as described herein, to form bispecific or multi-targeted antibodies. In addition, various methods known in the art, such as by using, for example Bispecific antibodies are constructed using the technology of (see, for example, WO2013 / 157954). Bispecific monoclonal antibodies (BsMAbs, BsAbs) generally contain the binding domains of two different monoclonal antibodies and thus bind to two different epitopes. Molecules and other full-length IgG bispecific antibodies have two different antigen-binding specificities encoded by two different variable regions of the full-length IgG molecule of scFv and Fab. They can be produced by co-transfecting a single cell with genetic constructs encoding two different common light chain (cLC) antibodies. As described in detail elsewhere herein, CH3 engineering ensures efficient heterodimerization and formation of substantially pure bispecific antibodies.

[0257] The antibodies of the present invention are preferably bispecific antibodies. Antibodies typically bind to their targets via a so-called antigen binding site. Unmodified antigen binding sites are typically formed by and present in the variable domains of the antibody. The variable domains comprise the antigen binding site. A variable domain that can bind to an antigen is a variable domain that comprises an antigen binding site that can bind to an antigen.

[0258] The antibody variable domain generally comprises a heavy chain variable region (VH) and a light chain variable region (VL). The antigen binding site may be present in a combined VH / VL variable domain, or in only the VH region or only the VL region. When the antigen binding site is present in one of the two regions of the variable domain, the paired variable regions may promote the folding and / or stability of the bound variable regions, but do not significantly promote the binding of the antigen itself.

[0259] Antigen binding as used herein refers to the typical binding ability of an antibody to its antigen. The binding of an antibody to an antigen can be assessed in a variety of ways. One method is to incubate the antibody with an antigen (preferably a cell expressing the antigen), remove unbound antibody (preferably by a washing step) and detect the bound antibody by a labeled antibody that is bound to the bound antibody.

[0260] In contrast to the random, nonspecific adhesion (sticking) of proteins, the combination of antigens and antibodies is usually mediated by the complementarity determining region (complementarity determining region, CDR) of the antibody and the specific three-dimensional structure of both antigens and variable domains, thereby allowing the two structures to be accurately (similar to the interaction of locks and keys) combined together. Since antibodies usually recognize the part of the antigen known as the epitope of the antigen, and such epitopes may also be present in other compositions, if such other compositions comprise the same epitope, antibodies according to the present invention may also recognize other proteins. Therefore, the term "combination" does not exclude the combination of antibodies with other one or more proteins comprising the same epitope. Such other one or more proteins are preferably not human proteins.

[0261] In humans after birth, preferably in adults, antibodies generally do not bind to proteins other than the specific target protein on the cell membrane.

[0262] The variable domains in the antibodies of the present invention or their variants that can bind to the extracellular portion of a CD28 family member bind to a specified member, and under otherwise identical conditions, are at least 100 times lower than the extracellular portion of another CD28 family member of the same species. For example, the variable domains of an antibody or its variant that binds to PD-1 bind to PD-1, and under otherwise identical conditions, are at least 100 times lower than CD28, CTLA4, ICOS, BTLA, NKp30, and TMIGD2 of the same species. Of course, when an antibody or its variant is designed to bind to two or more members of a family, the binding to the two or more members may be substantially the same. In the present invention, it is preferred that the corresponding antibodies each bind to only one member of the CD28 family member. Considering that the CD28 family is a family of cell surface molecules, binding is usually assessed on cells expressing members on the cell surface.

[0263] The variable domains in the antibodies of the present invention or their variants that can bind to the extracellular portion of a B7 family member or TNFRSF14 bind to a specified molecule, and under otherwise identical conditions, are at least 100 times lower than the extracellular portion of another B7 family member or TNFRSF14 of the same species. For example, the variable domains of an antibody or its variant that binds to PD-L1 bind to PD-L1, and under otherwise identical conditions, are at least 100 times lower than CD80, CD86, ICOSL, PD-L2, B7-H3, B7-H4, B7-H6, and B7-H7 of the same species. Of course, when an antibody or its variant is designed to bind to two or more members of a family, the binding to the two or more members may be substantially the same. In the present invention, it is preferred that the corresponding antibodies each bind to only one member of the B7 family member or TNFRSF14. Considering that the B7 family is a family of cell surface molecules, binding is usually assessed on cells expressing members on the cell surface.

[0264] The CD28 family member binding variable domain of the antibody or variant thereof as described herein blocks the binding of the member to one or more of its binding partners; when bound to the binding partner, inhibits the biological activity that would otherwise be exhibited by a cell containing the member. Antibodies containing the variable domain have the same activity in kind but not necessarily in amount.

[0265] B7 family member or TNFRSF14 binding variable domain of an antibody or variant thereof as described herein, which blocks the binding of the member or TNFRSF14 to one or more of its binding partners; when bound to the binding partner, the biological activity that would otherwise be exhibited by a cell containing the member or TNFRSF14 is generally inhibited. Antibodies containing the variable domain have the same activity in kind but not necessarily in amount.

[0266] An antibody comprising a variable domain that can bind to the extracellular portion of a CD28 family member and a variable domain that can bind to its binding partner selected from the B7 family or TNFRSF14, wherein both variable domains do not block the binding of the CD28 family member to the binding partner; when bound to the binding partner, it generally does not inhibit the biological activity that would otherwise be exhibited by cells containing CD28 family members, B7 family members or TNFRSF14.

[0267] As used herein, the term "antibody" means a protein molecule preferably belonging to the immunoglobulin class of proteins, which comprises one or more variable domains that bind to an epitope on an antigen, wherein such domains are derived from the variable domains of an antibody or have sequence homology thereto. Antibodies for therapeutic use are preferably as close as possible to the natural antibodies of the subject to be treated (e.g., human antibodies for human subjects). Antibody binding can be represented by specificity and affinity. Specificity determines which antigen or its epitope is specifically bound by the binding domain. Affinity is a measure of the intensity of binding to a specific antigen or epitope. Antibodies, such as the bispecific antibodies of the present invention, generally comprise the constant domains (Fc parts) of natural antibodies, which can be engineered as described elsewhere herein, for example, to reduce ADCC and / or CDC activity. The antibodies of the present invention are generally bispecific full-length antibodies, preferably human IgG subclasses.

[0268] The terms "variable domain", "VH / VL pair", "VH / VL" are used interchangeably herein. The variable domain consists of the variable region of the heavy chain and the variable region of the light chain. The variable region of the heavy chain is usually formed by a rearranged VDJ region. The variable region of the light chain is usually formed by a rearranged VJ region. VDJ / VJ regions can also now be artificially generated using the large amount of sequence information available for example for functional antibodies.

[0269] The antibody of the present invention is preferably a "full-length" antibody. The term "full-length" according to the present invention is defined as comprising a substantially complete antibody without one or more artificially added parts, the size of which is greater than 20 amino acid residues, such as additional antigen binding sites or additional activation sites or additional ligands or additional ligand binding moieties. However, a full-length antibody does not necessarily have all the functions of a complete antibody. In order to avoid confusion, a full-length antibody comprises two heavy chains and two light chains. Each chain comprises a constant (C) region and a variable (V) region, which can be divided into domains referred to as CH1, CH2, CH3, VH for the heavy chain and CL, VL for the light chain. The domains of the heavy chain preferably exist in the order of natural antibodies (VH-CH1-CH2-CH3, meaning that the VH domain is adjacent to the CH1 domain, followed by the CH2 domain, and followed by the CH3 domain). The domains of the light chain also preferably exist in the order of natural antibodies (VL-CL: meaning that the VL domain is adjacent to the CL domain). The antibody binds to the antigen through the variable domains contained in the Fab fragment portion. Antibodies can interact with molecules and cells of the immune system through their constant domains, primarily through the Fc portion.

[0270] In some embodiments, the antibody of the present invention is an IgG, preferably a full-length IgG. Preferably, a full-length IgG antibody is preferred, because of its particularly advantageous half-life, and because of immunogenicity reasons, it is desirable to keep close to a complete autologous (human) molecule. In some embodiments, the antibody of the present invention is a full-length IgG1, full-length IgG2, full-length IgG3, or full-length IgG4 antibody.

[0271] Full-length antibodies according to the present invention encompass such antibodies, wherein mutations may exist to provide desired features or are only replacements of those in the original chain. Such mutations should not be the disappearance of the substantial part of any region. However, antibodies in which one or more amino acid residues are acid inserted, deleted, replaced or a combination thereof and do not substantially change the binding characteristics of the resulting antibody are included in the term "full-length antibody". For example, an IgG antibody may have 1 to 20 amino acid residues inserted, replaced, deleted or a combination thereof in the constant region.

[0272] The antibody of the present invention or its variant is preferably a bispecific antibody or its variant. In a preferred embodiment, it is a bispecific IgG antibody with reduced effector function. In a preferred embodiment, the antibody of the present invention is a bispecific full-length antibody. The antibody of the present invention is preferably a bispecific full-length IgG antibody, preferably mutated in the CH2 / lower hinge region to reduce effector function. Based on its long circulation half-life in humans, it is advantageous to mutate in the CH2 / lower hinge region to reduce effector function IgG1. In order to prevent any immunogenicity in humans, it is preferred that the bispecific antibody according to the present invention is a human antibody.

[0273] The term "bispecific" (bispecific, bs) means that one part of the antibody (as defined above) binds to an epitope on an antigen, while the second part binds to a different epitope on the same antigen or a different antigen. Different epitopes are usually present on different antigens. However, different epitopes can also be present on the same antigen. According to the present invention, the first and second antigens are actually two different proteins. Preferred bispecific antibodies are antibodies that contain parts of two different monoclonal antibodies and can therefore bind to two different epitopes, preferably two different epitopes on two different antigens. Depending on the expression level, (sub) cellular localization and stoichiometry of the two antigens recognized by the bispecific antibody, the two Fab arms of the antibody may or may not bind to their epitopes at the same time. One arm of a bispecific antibody usually contains the variable domain of one antibody, while the other arm contains the variable domain of another antibody (i.e., one arm of the bispecific antibody is formed by pairing a heavy chain with a light chain, while the other arm is formed by pairing a different heavy chain with a light chain). The heavy chain variable regions of the bispecific antibodies of the present invention are usually different from each other, while the light chain variable regions are preferably the same in the bispecific antibodies of the present invention. Bispecific antibodies in which different heavy chain variable regions are associated with the same or common light chain variable regions are also referred to as bispecific antibodies with a common light chain variable region (cLcv). Preferably, the light chain constant regions are also the same. Such bispecific antibodies are referred to as having a common light chain (cLc). Therefore, a bispecific antibody according to the present invention is also provided, wherein the two arms comprise a common light chain.

[0274] As described herein, bispecific antibodies preferably include a common light chain variable domain, preferably a common light chain. According to the term "common light chain" of the present invention, it refers to a light chain that can be identical or have some amino acid sequence differences but the binding specificity of the full-length antibody is not affected. For example, within the definition of the common light chain used in this article, it is possible to prepare or find different but functionally equivalent light chains, such as by introducing and testing conservative amino acid changes, not causing or only partially causing amino acid changes in the region of binding specificity when paired with the heavy chain. In the case of adding or not adding the term "rearrangement", the terms "common light chain", "common LC", "cLC" and "single light chain" are all used interchangeably herein. In the case of adding or not adding the term "rearrangement", the terms "common light chain variable region", "common VL", "common LCv", "cLCv" and "single VL" are all used interchangeably herein. A preferred aspect of the present invention is that the bispecific antibody has a common light chain (variable region), which can be combined with at least two, and preferably multiple, heavy chains (variable regions) with different binding specificities to form an antibody with a functional antigen binding domain (WO2004 / 009618, WO2009 / 157771, Merchant et al. 1998 and Nissim et al. 1994). The common light chain (variable region) is preferably a human light chain (variable region). The common light chain (variable region) preferably has a germline sequence. A preferred germline sequence is a light chain variable region that is frequently used in the human repertoire and has good thermodynamic stability, yield and solubility. The preferred germline light chain is O12. The common light chain is preferably a rearranged germline human κ light chain IgVκ1-39*01 / IGJκ1*01 ( Figure 1 A). The common light chain variable region is preferably the variable region of the rearranged germline human kappa light chain IgVκ1-39*01 / IGJκ1*01. The common light chain preferably comprises Figure 1 B or 1D, wherein the light chain variable region has 0 to 5 amino acid insertions, deletions, substitutions, additions or combinations thereof. The common light chain preferably further comprises a light chain constant region, preferably a kappa light chain constant region. The nucleic acid encoding the common light chain can be codon optimized for a cell system for expressing the common light chain protein. The encoding nucleic acid can deviate from the germline nucleic acid sequence.

[0275] In a preferred embodiment, the light chain comprises a light chain region comprising Figure 1 The amino acid sequence of the O12 / IgVκ1-39*01 gene fragment shown in A, which has 0 to 10, preferably 0 to 5 amino acid insertions, deletions, substitutions, additions, or a combination thereof. The phrase "O12 light chain" will be used throughout the specification as "a light chain comprising a light chain variable region, the light chain variable region comprising Figure 1A "O12 / IgVκ1-39*01 gene fragment amino acid sequence shown in Figure 1, and having 0 to 10, preferably 0 to 5 amino acid insertions, deletions, substitutions, additions, or combinations thereof". IgVκ1-39 is the abbreviation for immunoglobulin variable κ1-39 gene. The gene is also known as immunoglobulin κ variable 1-39; IGKV139; IGKV1-39; O12a or O12. The external ID of the gene is HGNC: 5740; Entrez Gene: 28930; Ensembl: ENSG00000242371. A preferred amino acid sequence of IgVκ1-39 is Figure 1 E. It lists the sequence of the V region. The V region can be combined with one of the five J regions. Figure 1 B and 1D depict two preferred sequences of IgVκ1-39 in combination with the J region. The linked sequences are indicated as IGKV1-39 / jk1 and IGKV1-39 / jk5; alternative names are IgVκ1-39*01 / IGJκ1*01 or IgVκ1-39*01 / IGJκ5*01 (according to the nomenclature of the IMGT database on the World Wide Web at imgt.org).

[0276] Preferably, O12 / IgVκ1-39*01 comprising the light chain variable region is a germline sequence. Further preferably, IGJκ1*01 or / IGJκ5*01 comprising the light chain variable region is a germline sequence. In a preferred embodiment, the IGKV1-39 / jk1 or IGKV1-39 / jk5 light chain variable region is a germline sequence.

[0277] In a preferred embodiment, the light chain variable region comprises germline O12 / IgVκ1-39*01. In a preferred embodiment, the light chain variable region comprises a κ light chain IgVκ1-39*01 / IGJκ1*01 or IgVκ1-39*01 / IGJκ5*01. In a preferred embodiment, IgVκ1-39*01 / IGJκ1*01. The light chain variable region preferably comprises a germline κ light chain IgVκ1-39*01 / IGJκ1*01 or a germline κ light chain IgVκ1-39*01 / IGJκ5*01, preferably a germline IgVκ1-39*01 / IGJκ1*01.

[0278] Mature B cells that produce antibodies with O12 light chains typically produce light chains that have undergone one or more mutations relative to the germline sequence (i.e., the normal sequence in the non-lymphocytes of the organism). The process responsible for these mutations is generally referred to as somatic (hyper)mutation. The light chains produced are referred to as affinity-matured light chains. When derived from the O12 germline sequence, such light chains are O12-derived light chains. In this specification, the phrase "O12 light chain" will include O12-derived light chains. Of course, mutations introduced by somatic hypermutation can also be introduced artificially in the laboratory. Other mutations can also be introduced in the laboratory without affecting the properties of the light chain (in kind but not necessarily in quantity). If the light chain contains such Figure 1 A. Figure 1 B; Figure 1 D or Figure 1 E and having 0 to 10, preferably 0 to 5 amino acid insertions, deletions, substitutions, additions, or combinations thereof, the light chain is at least an O12 light chain. In a preferred embodiment, the O12 light chain is a light chain comprising Figure 1 A; 1B; 1D or 1E and having 0 to 9, 0 to 8, 0 to 7, 0 to 6, 0 to 5, 0 to 4 amino acid insertions, deletions, substitutions, additions, or a combination thereof. In a preferred embodiment, the O12 light chain is a light chain comprising Figure 1 A. Figure 1 B; Figure 1 D or Figure 1 E and having 0 to 5, preferably 0 to 4, more preferably 0 to 3 amino acid insertions, deletions, substitutions, additions or a combination thereof. In a preferred embodiment, the O12 light chain is a light chain comprising Figure 1 A. Figure 1 B; Figure 1 D or Figure 1 E and having 0 to 2, more preferably 0 to 1, and most preferably 0 amino acid insertions, deletions, substitutions, additions, or a combination thereof. In a preferred embodiment, the O12 light chain is a light chain comprising Figure 1 A or Figure 1 B and having the amino acid insertion, deletion, substitution, addition, or a combination thereof. In a preferred embodiment, the light chain comprises Figure 1 In a preferred embodiment, the light chain variable region comprises Figure 1 The sequence of B.

[0279] The common light chain (variable region) may be a lambda light chain, and therefore also provided in the context of the present invention, but preferably a kappa light chain. The constant portion of the common light chain of the present invention may be a constant region of a kappa or lambda light chain. It is preferably a constant region of a kappa light chain, preferably wherein the common light chain is a germline light chain, preferably a rearranged germline human kappa light chain comprising an IgVκ1-39 gene segment, most preferably a rearranged germline human kappa light chain IgVκ1-39*01 / IGJκ1*01( Figure 1 ). The terms rearranged germline human κ light chain IgVκ1-39*01 / IGJκ1*01, IGKV1-39 / IGκJ1, huVκ1-39 light chain or abbreviated huVκ1-39, or simply 1-39 are used interchangeably throughout the application. Those skilled in the art will recognize that "common" also refers to functional equivalents of light chains that are not identical in amino acid sequence. There are many variants of the light chain in which there are mutations (deletions, substitutions, additions) that do not affect the formation of a functional binding region.

[0280] IgVκ1-39 is the abbreviation of immunoglobulin variable κ1-39 gene. This gene is also called immunoglobulin κ variable 1-39; IGKV139; IGKV1-39; O12a or O12. The external ID of this gene is HGNC: 5740; Entrez Gene: 28930; Ensembl: ENSG00000242371. A preferred amino acid sequence of IgVκ1-39 is Figure 1 . It lists the sequences of the V regions. The V regions can be combined with one of the five J regions. Figure 1 Two preferred sequences of IgVκ1-39 in combination with the J region are described. The linked sequences are indicated as IGKV1-39 / jk1 and IGKV1-39 / jk5; alternative names are IgVκ1-39*01 / IGJκ1*01 or IgVκ1-39*01 / IGJκ5*01 (according to the nomenclature of the IMGT database on the World Wide Web at imgt.org).

[0281] The common light chain variable region is preferably connected to a kappa light chain constant region. In a preferred embodiment, the light chain variable region comprises a kappa light chain IgVκ1-39*01 / IGJκ1*01 or IgVκ1-39*01 / IGJκ5*01. In a preferred embodiment, IgVκ1-39*01 / IGJκ1*01.

[0282] A common light chain producing cell may produce, for example, a rearranged germline human kappa light chain IgVκ1-39*01 / IGJκ1*01 and a light chain comprising said light chain variable region fused to a lambda constant region. Where germline sequences are referred to herein, preferably the variable region is a germline sequence.

[0283] The bispecific antibodies or variants thereof as described herein preferably have a heavy chain variable region / light chain variable region (VH / VL) combination that binds to the extracellular portion of a CD28 family member, and a second VH / VL combination that binds to the extracellular portion of a B7 family member or TNFRSF14. In a preferred embodiment, the VL in the first VH / VL combination is similar to the VL in the second VH / VL combination. In a more preferred embodiment, the VL in the first and second VH / VL combinations are the same. In a preferred embodiment, the bispecific antibody is a full-length antibody having a heavy / light (H / L) chain combination that binds to the extracellular portion of a CD28 family member, and a H / L chain combination that binds to the extracellular portion of a B7 family member or TNFRSF14. In a preferred embodiment, the light chain in the first H / L chain combination is similar to the light chain in the second H / L chain combination. In a more preferred embodiment, the light chains in the first and second H / L chain combinations are the same.

[0284] Several methods have been disclosed to facilitate the production of bispecific antibodies, or vice versa, the production of monospecific antibodies. In the present invention, preferably, relative to the production of corresponding monospecific antibodies, cells are more conducive to the production of bispecific antibodies. This is usually achieved by modifying the constant region of the heavy chain so that it is more conducive to heterodimerization (i.e., the dimerization of the heavy chain combined with another heavy chain / light chain) relative to homodimerization. In a preferred embodiment, the bispecific antibody of the present invention comprises two different immunoglobulin heavy chains with a compatible heterodimerization domain. A variety of compatible heterodimerization domains have been described in the art. The compatible heterodimerization domain is preferably a compatible immunoglobulin heavy chain CH3 heterodimerization domain. When using a wild-type CH3 domain, the co-expression of two different heavy chains (A and B) and a common light chain will result in three different antibody species, AA, AB and BB. AA and BB are the names of two monospecific bivalent antibodies, and AB is the name of a bispecific antibody. In order to increase the percentage of the desired bispecific product (AB), CH3 engineering can be used, or in other words, a heavy chain with a compatible heterodimerization domain as defined below can be used. The art describes a variety of ways to achieve such heavy chain heterodimerization. One way is to produce a "knob into hole" bispecific antibody. See WO1998 / 050431 (Arathoon et al.).

[0285] The term "compatible heterodimerization domain" as used herein refers to a protein domain that is engineered such that engineered domain A' will preferentially form heterodimers with engineered domain B' and vice versa, while homodimerization between A'-A' and B'-B' is reduced.

[0286] In US13 / 866,747 (now published as US 9,248,181), US14 / 081,848 (now published as US 9,358,286); WO2013 / 157953 and PCT / NL2013 / 050294 (published as WO2013 / 157954); incorporated herein by reference), methods and means for producing bispecific antibodies using compatible heterodimerization domains are disclosed. These means and methods can also be advantageously used in the present invention. In particular, the bispecific antibodies of the present invention preferably comprise mutations that substantially only produce bispecific full-length IgG molecules. Preferred mutations are amino acid substitutions L351K and T366K (EU numbering) in the first CH3 domain ("KK variant" heavy chain) and amino acid substitutions L351D and L368E in the second domain ("DE variant" heavy chain), or vice versa. Previously, it was demonstrated in our US 9,248,181 and US 9,358,286 patents and WO2013 / 157954 PCT application that DE variants and KK variants prefer pairing to form heterodimers (so-called "DEKK" bispecific molecules). Due to the repulsion between charged residues in the CH3-CH3 interface between the same heavy chains, homodimerization of the DE variant heavy chains (DEDE homodimers) hardly occurs.

[0287] Bispecific antibodies can be produced by (transient) transfection of a plasmid encoding a CH3 engineered light chain and two different heavy chains to ensure efficient heterodimerization and the formation of bispecific antibodies. The production of these chains in a single cell results in the formation of bispecific antibodies that are more favorable than the formation of monospecific antibodies.

[0288] Therefore, also provided is a method for producing an antibody or variant according to the invention from a single cell, wherein the antibody or variant thereof comprises two CH3 domains capable of forming an interface, the method comprising providing:

[0289] -cell, the cell having a) a first nucleic acid molecule encoding an IgG heavy chain, the IgG heavy chain specifically recognizing the extracellular part of PD-1 and comprising a first CH3 domain, and b) a second nucleic acid sequence encoding an IgG heavy chain, the IgG heavy chain specifically recognizing the extracellular part of PD-L1 and comprising a second CH3 domain, wherein the nucleic acid sequence has a mutation for preferentially pairing the first and second CH3 domains, the method further comprising the steps of culturing the cell and allowing the nucleic acid sequence to be expressed and harvesting the antibody or variant thereof from the culture. In some preferred embodiments, the cell has a third nucleic acid sequence encoding a common light chain, preferably a rearranged germline human κ light chain IgVκ1-39*01 / IGJκ1*01.

[0290] Some preferred mutations that essentially only generate bispecific full-length IgG1 molecules are amino acid substitutions at positions 351 and 366, such as L351K and T366K (numbered according to EU numbering) in the first CH3 domain ("KK variant" heavy chain) and amino acid substitutions at positions 351 and 368, such as L351D and L368E, in the second CH3 domain ("DE variant" heavy chain), or vice versa. Therefore, a method according to the invention for producing an antibody or variant according to the invention from a single cell is also provided, wherein the first CH3 domain comprises the amino acid substitutions L351K and T366K (numbered according to EU numbering), and wherein the second CH3 domain comprises the amino acid substitutions L351D and L368E (numbered according to EU numbering), the method further comprising the steps of culturing the cell and allowing expression of the nucleic acid sequence and harvesting the antibody or variant thereof from the culture.

[0291] In one embodiment, the heavy chain / light chain combination comprising a variable domain that binds to PD-1 comprises a DE variant of the heavy chain. In this embodiment, the heavy chain / light chain combination comprising a variable domain that can bind to an antigen other than PD-1 comprises a KK variant of the heavy chain.

[0292] The effector function of Fc region-mediated antibody, for example complement dependent cytotoxicity (complement dependentcytotoxicity, CDC), antibody-dependent cellular cytotoxicity (antibody-dependent cellularcytotoxicity, ADCC) and antibody-dependent cellular phagocytosis (antibody-dependent cellularphagocytosis, ADCP). According to the application of therapeutic antibodies or Fc fusion proteins, it is expected to reduce or improve the effector function. The effector function reduced in the present invention is preferred. As in some embodiments of the present invention, when activating, enhancing or stimulating immune response, it is expected to reduce the effector function. The antibody with reduced effector function can be used for cell surface molecules of targeted immune cells, etc.

[0293] It was found that IgG binding to FcγR or C1q requires residues located in the hinge region and CH2 domain. Two regions of the CH2 domain ( Figure 2 D) is associated with FcγR and C1q binding. It has been shown that replacing IgG2 residues at positions 233 to 236 and IgG4 residues at positions 327, 330 and 331 with human IgG1 can greatly reduce ADCC and CDC (Armour et al., 1999. Eur J Immunol. 29 (8): 2613-24; Shields et al., 2001. J Biol Chem. 276 (9): 6591-604). In addition, Idusogie et al. demonstrated that alanine substitutions at different positions (including K322) significantly reduced complement activation (Idusogie et al., 2000. J Immunol. 164 (8): 4178-84).

[0294] Due to their reduced effector function, IgG4 antibodies represent an IgG subclass for receptor blocking without cell depletion. IgG4 molecules can exchange half molecules in a dynamic process known as Fab arm exchange. This phenomenon can occur between therapeutic antibodies and endogenous IgG4. The S228P mutation is an example of a mutation that ensures reduced Fab arm exchange capacity. (Labrijn. et al., 2009. Nat Biotechnol. 27(8): 767-71).

[0295] The antibody with reduced effector function is preferably an IgG antibody comprising a modified CH2 / lower hinge region, for example for reducing Fc-receptor interactions or for reducing C1q binding. In some embodiments, the antibody of the present invention is an IgG antibody with a mutant CH2 and / or lower hinge domain, such that the interaction of the bispecific IgG antibody with the Fc-γ receptor is reduced. The antibody comprising a mutant CH2 region is preferably an IgG1 antibody. Such a mutant IgG1 CH2 and / or lower hinge domain preferably comprises an amino substitution at position 235 and / or position 236 (EU numbering), preferably L235G and / or G236R substitution ( Figure 2 E).

[0296] Variants of antibodies or bispecific antibodies as described herein may comprise functional parts, derivatives and / or analogs of antibodies or bispecific antibodies. Variants generally maintain the binding specificity of antibodies (e.g., bispecific antibodies). Variants generally maintain the binding specificity of antibodies (e.g., bispecific antibodies). Variants are preferably functional parts or derivatives of antibodies or bispecific antibodies as described herein. It is preferably a functional part.

[0297] Binding specificity is defined by the ability to bind to the extracellular portion of a CD28 family member and the extracellular portion of a B7 family member or TNFRSF14, where the members are binding partners (ie, a receptor-ligand pair).

[0298] The functional part of the antibody as described herein, or preferably the functional part of the bispecific antibody, is a part comprising the following variable domains: a variable domain that binds to the extracellular part of a CD28 family member (preferably PD-1), and a variable domain that binds to the extracellular part of a B7 family member or TNFRSF14 (preferably PD-L1). Suitable parts are, for example, F(ab')2 fragments produced by digesting a bispecific antibody with pepsin. Other parts comprising the variable domains are included in the present invention.

[0299] Functional derivatives of antibodies as described herein, or preferably functional derivatives of bispecific antibodies, are proteins comprising the following variable domains: variable domains that bind to the extracellular portion of CD28 family members (preferably PD-1), and variable domains that bind to the extracellular portion of B7 family members or TNFRSF14 (preferably PD-L1), which are connected by a connecting region. The variable domain can be such a variable domain, or a Fab fragment or a variable domain-like molecule, such as a single-chain Fv fragment comprising VH and VL connected together via a linker. Other examples of variable domain-like molecules are so-called single domain antibody fragments. Single-domain antibody fragments (single-domain antibody fragment, sdAb) are antibody fragments with a single monomer variable antibody region. Like a complete antibody, it can selectively bind to a specific antigen. The molecular weight of a single domain antibody fragment is only 12 to 15 kDa, much smaller than a normal antibody (150 to 160 kDa) consisting of two heavy protein chains and two light chains, and even smaller than a Fab fragment (about 50 kDa, one light chain and half a heavy chain) and a single chain variable fragment (about 25 kDa, two variable regions, one from the light chain and the other from the heavy chain). Single domain antibodies themselves are not much smaller than normal antibodies (usually 90 to 100 kDa). Single domain antibody fragments are mainly engineered from heavy chain antibodies found in camelids; these fragments are called VHH fragments. Some fish also have heavy chain-only antibodies (IgNAR, "immunoglobulin new antigen receptor"), from which single domain antibody fragments called VNAR fragments can be obtained. An alternative approach is to split the dimeric variable domains from common immunoglobulin G (IgG) derived from humans or mice into monomers. Although most of the current research on single domain antibodies is based on heavy chain variable domains, nanobodies derived from light chains have also been shown to specifically bind to target epitopes. Other non-limiting examples of variable domain-like molecules are VHH, human domain antibodies (dAb) and Unibodies. Preferred functional parts are parts that contain variable domains containing heavy chain variable regions and light chain variable regions. Some non-limiting examples of such variable domains are F(ab) fragments and single-chain Fv fragments. The bispecific format of variable domain (like) linkage is, for example, human serum albumin (HSA) bound to two different scFvs; bispecific minibodies comprising two different scFvs, the scFvs being bound together by a dimerization motif or a self-associated secondary structure (e.g., a helical bundle or coiled coil), thereby achieving dimerization of the scFv fragments (Morrison (2007) Nat. Biotechnol 25: 1233-34). Some examples of suitable HSA linkers and methods for coupling scFvs to linkers are described in WO2009 / 126920.

[0300] Functional analogs of antibodies as described herein, or preferably functional analogs of bispecific antibodies, are molecules comprising the following binding sites: binding sites for the extracellular portion of CD28 family members (preferably PD-1), and binding sites for the extracellular portion of B7 family members or TNFRSF14 (preferably PD-L1). Functional derivatives can be antibody mimetics, polypeptides, aptamers, or a combination thereof. These proteins or aptamers are usually bound to one target. The protein of the present invention is bound to two or more targets. It should be understood that any combination of these antibodies, antibody mimetics, polypeptides, and aptamers can be linked together by methods known in the art. For example, in some embodiments, the binding molecules of the present invention are conjugates or fusion proteins. For antibodies, the technology for making multispecific antibodies has been developed to also include such bispecific antibodies, which have the same overall structure as normal monospecific antibodies, but wherein the two arms of the antibody each bind to a different target.

[0301] Antibody mimics are polypeptides that can specifically bind to antigens like antibodies, but are structurally unrelated to antibodies. Antibody mimics are usually artificial peptides or proteins with a molar mass of about 3 kDa to 20 kDa. Compared to antibodies, the general advantages are better solubility, tissue penetration, stability to heat and enzymes, and relatively low production costs. Some non-limiting examples of antibody mimetics are affibody molecules (usually based on the Z domain of protein A); affilins (usually based on gamma-B crystals or ubiquitin); affimers (usually based on cysteine ​​protease inhibitors (Cystatins)); affitins (usually based on Sac7d from Sulfolobus acidocaldarius); alphabodies (usually based on triple-helical coiled-coils); anticalins (usually based on Lipocalins); avimers (usually based on the A domains of various membrane receptors); DARPins (usually based on ankyrin repeat motifs); fynomers (usually based on the SH3 domain of Fyn 7); kunitz domain peptides (usually based on the Kunitz domains of various protease inhibitors); and monobodies (usually based on the type III domain of fibronectin).

[0302] Monoclonal antibodies are synthetic binding proteins constructed using the fibronectin type III domain (FN3) as a molecular scaffold. Monoclonal antibodies are a simple and robust alternative to antibodies for generating target binding proteins. The term "monoclonal antibody" was coined by the Koide group in 1998, who published the first paper demonstrating the monoclonal antibody concept using the tenth FN3 domain of human fibronectin.

[0303] Single antibodies and other antibody mimics are usually generated from combinatorial libraries in which parts of the scaffold are diversified using molecular display and directed evolution techniques such as phage display, mRNA display, and yeast surface display. A large number of antibody mimics have high affinity and high specificity for their respective targets.

[0304] Aptamers are oligonucleotide or peptide molecules that bind to a specific target molecule. Aptamers are usually produced by selecting them from a large random sequence pool, but natural aptamers also exist in riboswitches. Aptamers can be used as macromolecules for both basic research and clinical purposes.

[0305] The term "conjugate" as used herein refers to two or more molecules that have been covalently linked, optionally through a connecting region also referred to as a joint. For example, in some embodiments, a conjugate is a first protein or non-protein part that is connected to a second protein or non-protein part through a connecting region. For example, in some embodiments of the binding molecules of the present invention, it comprises or is composed of two or more antibodies that have been covalently linked. The conjugate is not limited to the first and second parts, but in some embodiments, it may also have a third, fourth or more parts connected by an additional connecting region. As described elsewhere in this application, some examples of protein parts include, but are not limited to: polypeptides, peptide mimetics or antibodies (or antibody parts, derivatives or analogs, as described elsewhere in this application). Some examples of non-protein parts include, but are not limited to aptamers. Various types of joints can be used, and suitable joints will be selected according to the desired properties of the molecule type and joint in the conjugate (length, flexibility, resistance to protease activity and other similar features). Such joints may include nucleotides, polypeptides or suitable synthetic materials. For example, the joint may be a flexible peptide joint. In certain embodiments, the joint may be a cleavable joint that allows the parts of the conjugate to be separated from each other. In other embodiments, the peptide joint may be a spiral joint. Various examples and kits for connecting proteins to other molecules are known in the art. As used herein, the term "fusion protein" refers to a protein that contains two or more polypeptides or proteins that have been linked at the DNA level by recombination and expressed together as a single polypeptide. The fusion protein may also contain a peptide linker region that is also encoded by the DNA and expressed together with the fusion protein. The peptide linker that is part of the fusion protein can be designed to have specific characteristics, such as flexibility, hydrophilicity, protease resistance, cleavage, etc. All of these characteristics can be designed within the DNA sequence, and methods for designing linkers are well known in the art. For example, antibodies can be linked together by methods well known in the art, and as described herein, to form bispecific or multi-targeted antibodies. In addition, a variety of methods known in the art, such as by using, for example Bispecific antibodies are constructed using the same technology. Bispecific monoclonal antibodies (BsMAbs, BsAbs) usually contain the binding domains of two different monoclonal antibodies and thus bind to two different epitopes. The scFv and other full-length IgG bispecific antibodies have two different antigen-binding specificities encoded by two different variable regions of the full-length IgG molecule of the Fab. They can be produced by co-transfecting a single cell with genetic constructs encoding two different common light chain (cLC) antibodies. As described in detail elsewhere herein, CH3 engineering ensures efficient heterodimerization and formation of substantially pure bispecific antibodies.

[0306] The invention also provides an antibody or variant according to the invention for use as a medicament.An antibody or variant according to the invention is also provided for use in a method of treating cancer or a pathogen infection (eg a viral or parasitic infection).

[0307] The present invention also provides a method for treating an individual with cancer, the method comprising administering an antibody (e.g., a bispecific antibody) or a variant (e.g., a variant) of the present invention to an individual in need thereof. The individual is preferably an individual with cancer. In some embodiments, the cancer is a cancer comprising cancer cells expressing the second membrane protein. In a preferred embodiment, the cancer is a cancer comprising cancer cells expressing a B7 family member or TNFRSF14. The cancer is preferably adenocarcinoma. Some preferred cancers are colorectal cancer; pancreatic cancer; lung cancer; breast cancer; liver cancer; prostate cancer; ovarian cancer; cervical cancer; endometrial cancer; head and neck cancer; melanoma; testicular cancer; urothelial cancer; kidney cancer; gastric cancer; or carcinoid. In a preferred embodiment, the cancer is colorectal cancer; pancreatic cancer; lung cancer; breast cancer; liver cancer; prostate cancer; ovarian cancer; cervical cancer; endometrial cancer; head and neck cancer; or melanoma. In a particularly preferred embodiment, the cancer is colorectal cancer; pancreatic cancer; lung cancer; breast cancer; liver cancer; prostate cancer; ovarian cancer; cervical cancer; endometrial cancer; head and neck cancer; or melanoma. In a particularly preferred embodiment, the cancer is colorectal cancer; pancreatic cancer; lung cancer; breast cancer; or liver cancer. In a particularly preferred embodiment, the cancer is gastrointestinal cancer. In a preferred embodiment, the cancer is colorectal cancer. In this embodiment, the antibody or variant thereof is preferably an antibody having a variable domain that can bind to PD-1 and a variable domain that can bind to PD-L1. The variable domains preferably each block the binding of PD-1 to PD-L1.

[0308] Also provided is an in vitro system comprising an antibody or variant according to the present invention and a first cell and a second cell. The first and second cells preferably express the first membrane protein and the second membrane protein, respectively, on the cell membrane. The system is preferably a cell system suitable for the maintenance and / or growth of the first cell. The cell system is preferably suitable for the maintenance and / or growth of the second cell. Such a system is, for example, suitable for cultivating immune cells against abnormal cells and / or for their proliferation. Such immune cells can then be administered to an individual in need thereof, such as a cancer patient. The immune cells preferably comprise T cells or NK cells, preferably cytotoxic T cells. The immune cells are preferably autologous for an individual in need thereof.

[0309] Also provided is a method for inducing and / or stimulating an immune response of an individual against abnormal cells in the individual, the method comprising providing the individual with an antibody of the present invention or a variant thereof. The abnormal cell is preferably a cancer cell, a virus-infected cell, a parasite, or a parasite-infected cell. In a preferred embodiment, the cell is a cancer cell or a neoplastic cell. In this embodiment, the antibody or variant thereof is preferably an antibody having a variable domain that can bind to PD-1 and a variable domain that can bind to PD-L1. The variable domains preferably each block the binding of PD-1 to PD-L1.

[0310] A neoplasm is an abnormal growth of tissue and is often referred to as a tumor when it also forms a mass. The neoplasm in the present invention often forms a mass. Neoplastic cells are cells from a neoplasm that has formed a mass. The World Health Organization (WHO) divides neoplasms into four major categories: benign neoplasms, in situ neoplasms, malignant neoplasms, and neoplasms of uncertain or unknown behavior. Malignant neoplasms are also referred to as cancer.

[0311] Inducing and / or stimulating an immune response encompasses both inducing an immune response and enhancing an already existing immune response.The immune response in an individual may be measured by measuring the individual's tumor load; the individual's viral load; the individual's parasite load, where applicable.

[0312] The virus-infected cells are preferably cells infected with the following viruses: immunodeficiency virus, herpes virus (preferably herpes simplex virus), varicella-zoster virus, cytomegalovirus or Epstein-Barr virus, papillomavirus, hepatitis virus (preferably hepatitis A, B or C virus), measles virus or adenovirus. The virus is preferably a virus known to persist in an individual. Persistent infection is characterized by those infections in which the virus is not cleared and remains in the specific cells of the infected individual. Persistent infection can involve both silent infection and productive infection stages without rapidly killing the host cell or even causing excessive damage to it. Persistent virus-host interaction may be a latent, chronic and / or slow infection.

[0313] The cells infected by parasites are cells infected with intracellular parasites. Such parasites are parasitic microorganisms that can grow and reproduce in host cells. Some intracellular parasites can also live outside the cell. Such parasites are so-called facultative intracellular parasites. Some non-limiting examples are Listeria monocytogenes, Legionella, mycobacteria of certain species, and Cryptococcus neoformans. Preferred intracellular parasites are those that are unable to grow outside of a host cell, some preferred examples being Chlamydia and closely related species; certain species of mycobacteria (e.g. Mycobacterium leprae); certain protozoa including: Apicomplexans (Plasmodium spp., Toxoplasma gondii and Cryptosporidium parvum) and trypanosomatids.

[0314] The antibodies or variants thereof or the preferred bispecific antibodies or variants thereof of the present invention are preferably used in humans. For this reason, the antibodies or variants thereof of the present invention are preferably human antibodies or humanized antibodies. Human tolerance to polypeptides is controlled by many different aspects. Immunity, whether T cell mediated, B cell mediated or other, is one of the variables covered in human tolerance to polypeptides. The constant region of the bispecific antibody of the present invention preferably comprises: a human heavy chain constant region, preferably comprising Figure 2 and a human light chain constant region, preferably comprising Figure 1The sequence shown in C. Relative to the constant region of a naturally occurring human antibody, the constant region may contain one or more, preferably no more than 10, preferably no more than 5 amino acid differences. Preferably, the constant part is all derived from naturally occurring human antibodies. As described in WO2009 / 157771, the various antibodies produced herein are derived from common light chain mice immunized with corresponding targets. The various antibodies produced herein are derived from human antibody variable domain libraries. Therefore, these variable domains are human. The unique CDR region may be derived from humans, may be synthetic, or may be derived from other organisms. When the variable region has an amino acid sequence identical to the amino acid sequence of the variable region of a naturally occurring human antibody except for the CDR region, the variable region is at least a human variable region. In such an embodiment, relative to the variable region of a naturally occurring human antibody, the VH of the antibody variable domain of the present invention that binds to a CD28 family member or a membrane-associated member of the B7 family or TNFRSF14, or the light chain in the antibody may contain one or more, preferably no more than 10, preferably no more than 5 amino acid differences, without calculating the possible differences in the amino acid sequence of the CDR region. Such mutations also occur in nature in the context of somatic hypermutation.

[0315] Antibodies can be derived from a variety of animal species, at least in terms of the heavy chain variable region. It is common practice to humanize, for example, a murine heavy chain variable region. There are a variety of ways to achieve this, among which are: transplanting the CDRs into a human heavy chain variable region having a 3D structure that matches the 3D structure of the murine heavy chain variable region; de-immunizing the murine heavy chain variable region, preferably by removing known or suspected T cell epitopes or B cell epitopes from the murine heavy chain variable region. Removal is usually accomplished by replacing one or more amino acids in the epitope with another (usually conservative) amino acid, so that the sequence of the epitope is modified so that it is no longer a T cell epitope or a B cell epitope.

[0316] Compared with the original mouse heavy chain variable region, the immunogenicity of the de-immunized mouse heavy chain variable region in humans is lower. Preferably, the variable region or variable domain of the present invention is further humanized, for example, as venner. By using venning technology, external residues that the immune system easily encounters are selectively replaced with human residues to provide a hybrid molecule comprising a weakly immunogenic or substantially non-immunogenic venning surface. The animal used in the present invention is preferably a mammal, more preferably a primate, and most preferably a human.

[0317] The antibody or bispecific antibody according to the present invention or their variants preferably comprise the constant region of a human antibody. According to the differences in their heavy chain constant domains, antibodies can be divided into five categories or isotypes: IgG, IgA, IgM, IgD and IgE. These categories or isotypes comprise at least one of the heavy chains named with the corresponding Greek letters. In a preferred embodiment, the present invention provides antibodies according to the present invention, wherein the constant region is selected from the IgG constant region, i.e., selected from IgG1, IgG2, IgG3 and IgG4. Preferably, the constant region is an IgG4 or IgG1 constant region ( Figure 2 ), more preferably a mutated IgG1 constant region. Some variations in the IgG1 constant region are naturally occurring and / or are permitted without changing the immunological properties of the resulting antibody. Generally, about 1 to 10 amino acid insertions, deletions, substitutions, or combinations thereof are permitted in the constant region. To reduce effector function or for other reasons, including half-life, stability, etc., the constant region may be mutated as described herein to achieve effective heterodimerization.

[0318] Rational methods have been developed towards minimizing the content of non-human residues in the human background. A variety of methods can be used to successfully transplant the antigen binding properties of an antibody to another antibody. The binding properties of an antibody depend primarily on the exact sequence in the CDR3 region, which is usually supported by a combination of the sequence in the CDR1 and CDR2 regions in the variable domains and the appropriate structure of the entire variable domain. A variety of methods can be used to transplant the CDR region to the appropriate variable domain of another antibody. Some of these methods are reviewed in JC Almagro1 and J. Fransson (2008) Frontiers in Bioscience 13, 1619-1633, which are incorporated herein by reference.

[0319] Contains Figure 3 and / or Fig.13The light chain variable domain of the variable domain of the heavy chain variable sequence shown in is preferably a germline light chain of O12 or a germline light chain based on O12, preferably a rearranged germline human κ light chain IgVκ1-39*01 / IGJκ1*01 or a fragment or functional equivalent thereof (according to the nomenclature of the IMGT database World Wide Web imgt.org). The term rearranged germline human κ light chain IgVκ1-39*01 / IGJκ1*01, IGKV1-39 / IGKJ1, huVκ1-39 light chain or huVκ1-39 for short is used. The light chain may have 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof. The 1, 2, 3, 4 or 5 amino acid substitutions are preferably conservative amino acid substitutions, and the insertions, deletions, substitutions or a combination thereof are preferably not in the CDR3 region of the VL chain, preferably not in the CDR1, CDR2 or CDR3 region or FR4 region of the VL chain. The preferred sequence of the common light chain is Figure 1 as shown in .

[0320] A variety of methods can be used to produce bispecific antibodies. One method involves expressing two different heavy chains and two different light chains in a cell and collecting antibodies produced by the cell. Antibodies produced in this way will generally contain antibody collections with different combinations of heavy chains and light chains, some of which are desired bispecific antibodies. The bispecific antibodies can then be purified from the collection. The ratio of bispecific antibodies produced by cells to other antibodies can be increased in a variety of ways. In a preferred embodiment of the present invention, the ratio is increased by expressing two substantially identical light chains instead of two different light chains in the cell. The two substantially identical light chains can be light chains with substantially identical light chain variable regions and different light chain constant regions, or preferably, two substantially identical light chain constant regions. This concept is also referred to as the "common light chain" method in the art. Compared to expressing two substantially different light chains, when substantially identical light chains work together with two different heavy chains to allow the formation of variable domains with different antigen binding sites and accompanying different binding properties, the ratio of bispecific antibodies produced by cells to other antibodies is significantly improved. Compared to the pairing of two identical heavy chains, the ratio of bispecific antibodies produced by cells can be further increased by stimulating the pairing of two different heavy chains with each other. The art describes a variety of ways in which such heavy chain heterodimerization can be achieved. Preferred methods are described in U.S. Provisional Application 61 / 635,935, followed by U.S. Regular Application No. 13 / 866,747 and PCT Application No. PCT / NL2013 / 050294 (WO 2013 / 157954 A1), which are incorporated herein by reference. Methods and means for producing bispecific antibodies (from a single cell) are disclosed, thereby providing means for the formation of bispecific antibodies that are more inclined to the formation of monospecific antibodies. These methods can also be advantageously used in the present invention. Therefore, the present invention provides a method for producing a bispecific antibody according to the present invention (from a single cell), wherein the bispecific antibody comprises two CH3 domains capable of forming an interface, the method comprising providing in the cell: a) a first nucleic acid molecule encoding a heavy chain comprising a first CH3 domain, b) a second nucleic acid molecule encoding a heavy chain comprising a second CH3 domain, wherein the nucleic acid molecule has a means for preferentially pairing heavy chains comprising the first and second CH3 domains, the method further comprising culturing the host cell and allowing the two nucleic acid molecules to express and harvesting the bispecific antibody from the culture. The first and second nucleic acid molecules may be part of the same nucleic acid molecule, vector or gene delivery vehicle, and may be integrated at the same site of the host cell genome. Alternatively, the first and second nucleic acid molecules are provided to the cell separately. The host cell comprises at least one light chain, and preferably a common light chain.

[0321] A preferred embodiment provides a method for producing a bispecific antibody according to the present invention from a single cell, wherein the bispecific antibody comprises two CH3 domains capable of forming an interface, the method comprising providing:

[0322] - a cell having: a) a first nucleic acid molecule encoding a heavy chain comprising an antigen binding site that can bind to the extracellular portion of a membrane-associated member of the CD28 family and comprising a first CH3 domain, and b) a second nucleic acid molecule encoding a heavy chain comprising an antigen binding site that can bind to the extracellular portion of a membrane-associated member of the B7 family or TNFRSF14 and comprising a second CH3 domain, wherein the nucleic acid molecule has means for preferential pairing of the first with the second CH3 domain,

[0323] The method also includes culturing the cell and allowing the protein encoded by the two nucleic acid molecules to be expressed and the step of harvesting the bispecific IgG antibody from the culture. In a particularly preferred embodiment, the cell also has a third nucleic acid molecule encoding a common light chain. The first, second and third nucleic acid molecules can be part of the same nucleic acid molecule, vector or gene delivery vehicle, and can be integrated at the same site of the genome of the host cell. Alternatively, the first, second and third nucleic acid molecules are provided to the cell respectively. The preferred common light chain is based on O12, preferably the rearranged germline human κ light chain IgVκ1 39*01 / IGJκ1*01 as described above. The means for making the first and second CH3 domains preferentially paired are preferably corresponding mutations in the CH3 domains of the heavy chain coding region. The preferred mutations that produce essentially only bispecific antibodies are amino acid replacements L351K and T366K in the first CH3 domain (numbered according to EU numbering) and amino acid replacements L351D and L368E in the second CH3 domain, or vice versa ( Figure 2). Therefore, a method for producing a bispecific antibody according to the present invention is also provided, wherein the first CH3 domain comprises the amino acid substitutions L351K and T366K (numbered according to EU numbering), and wherein the second CH3 domain comprises the amino acid substitutions L351D and L368E, the method further comprising culturing the cell and allowing the protein encoded by the nucleic acid molecule to be expressed and the step of harvesting the bispecific antibody from the culture. A method for producing a bispecific antibody according to the present invention is also provided, wherein the first CH3 domain comprises the amino acid substitutions L351D and L368E (numbered according to EU numbering), and wherein the second CH3 domain comprises the amino acid substitutions L351K and T366K, the method further comprising culturing the cell and allowing the nucleic acid molecule to be expressed and the step of harvesting the bispecific antibody from the culture. Antibodies that can be produced by these methods are also part of the present invention. The CH3 heterodimerization domain is preferably an IgG1 heterodimerization domain. The heavy chain constant region comprising the CH3 heterodimerization domain is preferably an IgG1 constant region.

[0324] The present invention also provides nucleic acid molecules encoding at least a portion of the variable region of the heavy chain of an antibody according to the present invention. Nucleic acid molecules having a length of at least 15 nucleotides are provided herein, encoding at least one CDR region of an antibody or variant according to the present invention. Nucleic acid molecules encoding at least the variable region of the heavy chain of an antibody or variant according to the present invention are also provided.

[0325] The nucleic acid molecule (generally an in vitro, isolated or recombinant nucleic acid molecule) preferably encodes Figure 3 and / or Fig.13 Any one of the heavy chain variable regions shown in , or any ... Figure 3 and / or Fig.13 Some embodiments provide nucleic acid molecules encoding antibodies or variants according to the present invention. The nucleic acid molecules preferably use codons optimized for expression in the antibody-producing cells to be used. Preferably, the coding for Figure 3 and / or Fig.13 or a heavy chain variable region as shown in , ... Figure 3 and / or Fig.13 The nucleic acid of the heavy chain variable region shown in is codon optimized for expression in human cells, preferably Per.C6 TM or Chinese hamster cells, preferably CHO cells. The present invention also provides a heavy chain variable region encoding the Figure 2 A nucleic acid molecule containing a heavy chain constant region.

[0326] Nucleic acid molecules as used in the present invention are typically, but not exclusively, ribonucleic acid (RNA) or deoxyribonucleic acid (DNA).Alternative nucleic acids, such as, for example, locked nucleic acid (LNA) and peptide nucleic acid (PNA), are available to those skilled in the art.

[0327] Also provided are vectors comprising the nucleic acid molecules according to the invention.

[0328] The nucleic acid molecule according to the present invention is for example contained in a cell. When the nucleic acid molecule is expressed in the cell, the cell can produce an antibody or variant according to the present invention. Therefore, in one embodiment, the present invention provides a cell comprising an antibody or variant according to the present invention and / or a nucleic acid molecule according to the present invention and / or a vector according to the present invention. When the cell produces heavy chains and light chains, antibodies are produced. A cell that can produce an antibody of the present invention is provided. The cell preferably comprises a nucleic acid molecule encoding an antibody heavy chain, the antibody heavy chain comprising an antibody heavy chain variable region, which can bind to the first membrane protein when combined with a common light chain. The cell preferably also comprises a nucleic acid molecule encoding an antibody heavy chain, the antibody heavy chain comprising an antibody heavy chain variable region, which can bind to the second membrane protein when combined with a common light chain. The cell preferably also comprises a nucleic acid molecule encoding a common light chain. The cell is preferably an animal cell, more preferably a mammalian cell, more preferably a primate cell, and most preferably a human cell. For the purposes of the present invention, a suitable cell is any cell that can contain and preferably produce an antibody according to the present invention and / or a nucleic acid according to the present invention.

[0329] The present invention also provides cells comprising antibodies according to the present invention. Cells comprising one or more nucleic acid molecules encoding antibodies of the present invention alone or together are also provided. One or more nucleic acid molecules are expressible nucleic acid molecules, which means that they contain cis signals required for RNA transcription and protein coding domain translation. Preferably, the cell (usually an in vitro, isolated or recombinant cell) produces the antibody. In a preferred embodiment, the cell is a hybridoma cell, a Chinese hamster ovary (CHO) cell, a NS0 cell or a PER-C6 TM Cells. In a particularly preferred embodiment, the cells are CHO cells. Cell cultures comprising cells according to the invention are also provided. A number of institutions and companies have developed cell lines for large-scale production of antibodies, e.g. for clinical use. Some non-limiting examples of such cell lines are CHO cells, NS0 cells or PER.C6 TMCells. These cells are also used for other purposes, such as producing proteins. Cell lines developed for industrial-scale production of proteins and antibodies are further referred to herein as industrial cell lines. Therefore, in a preferred embodiment, the present invention provides the use of cell lines developed for large-scale production of antibodies for producing antibodies of the present invention. The present invention also provides cells for producing antibodies, which contain a gene encoding a Figure 3 , 13 , 1 and / or 2. Preferably, the nucleic acid molecule comprises a nucleic acid molecule of VH, VL and / or heavy chain. Figure 1 or Figure 2 The sequence shown in .

[0330] The present invention also provides a method for producing an antibody, comprising culturing the cell of the present invention and harvesting the antibody from the culture. Preferably, the cell is cultured in a serum-free medium. Preferably, the cell is adapted for growth in suspension. Also provided is an antibody obtainable by a method for producing an antibody according to the present invention. The antibody is preferably purified from the culture medium of the culture. Preferably, the antibody is affinity purified.

[0331] The cells of the invention are for example hybridoma cell lines, CHO cells, 293F cells, NS0 cells or any other cell type known in the art that is suitable for the production of antibodies for clinical purposes, in particular for the production of antibodies for administration in humans. In a particularly preferred embodiment, the cells are human cells, preferably cells transformed with the E1 region of an adenovirus or a functional equivalent thereof. A preferred example of such a cell line is PER.C6 TM Cell line or its equivalent. In a particularly preferred embodiment, the cell is a CHO cell or a variant thereof, preferably using a glutamine synthetase (GS) vector system to express the antibody variant.

[0332] Antibodies or variants according to the present invention may also be produced in non-human animals. Therefore, non-human animals are also provided, comprising antibodies according to the present invention and / or nucleic acid molecules according to the present invention and / or vectors according to the present invention. In some embodiments, the non-human animal comprises a rodent or a rabbit, preferably a mouse or a rat.

[0333] The present invention also provides a composition or kit comprising at least one antibody or variant according to the present invention. The present invention also provides a pharmaceutical composition comprising one or more antibodies or variants thereof according to the present invention. The pharmaceutical composition preferably comprises a pharmaceutically acceptable excipient, diluent or carrier.

[0334] The antibodies of the present invention or their variants may also comprise a label, preferably a label for in vivo imaging. Such a label is generally not necessary for therapeutic applications. Labels may be helpful in, for example, diagnostic situations. For example, target cells in vivo are visualized. A variety of labels are suitable, and many are well known in the art. In a preferred embodiment, the label is a radioactive label for detection. In another preferred embodiment, the label is an infrared label. Preferably, the infrared label is suitable for in vivo imaging. A variety of infrared labels are available to those skilled in the art. Preferred infrared labels are, for example, IRDye 800; IRDye 680RD; IRDye680LT; IRDye 750; IRDye 700DX; IRDye 800RS IRDye 650; IRDye 700 phosphoramidite; IRDye 800 phosphoramidite (LI-COR USA; 4647 Superior Street; Lincoln, Nebraska).

[0335] The amount of the antibody according to the invention to be administered to the patient is generally within the therapeutic window, which means that a sufficient amount is used to obtain a therapeutic effect while the amount does not exceed the threshold that causes an unacceptable degree of side effects. The lower the amount of antibody required to obtain the desired therapeutic effect, the larger the therapeutic window will generally be. Therefore, antibodies according to the invention that exert a sufficient therapeutic effect at low doses are preferred. The dosage may be within the dosing regimen of nivolumab. The dosage may also be lower.

[0336] Compared with the combination of bivalent monospecific antibodies with variable domains, antibodies according to the present invention or their variants, and particularly bispecific antibodies or their variants can have less side effects. The combination of antibodies blocking inhibitory and / or co-stimulatory molecules is beneficial to patients who are unresponsive to existing immunotherapy. However, the dual blocking of immunomodulatory receptors (iMOD) has been shown to improve immune-related toxicity. Antibodies according to the present invention or their variants, and particularly bispecific antibodies or their variants are suitable for solving the dual blocking of iMOD, because they can play a functional activity that cannot be replicated by a monoclonal antibody combination, and can more selectively target specific cell populations, which reduces the safety responsibility in patients.

[0337] By cloning the antibody into a complementary expression vector comprising a mutation in the CH3 region driving the heavy chain heterodimerization, an antibody as a bispecific antibody is produced. Many bispecific antibodies have been produced on a small scale and tested in binding and functional assays on cancer cell lines. The antibody of the present invention, particularly the bispecific antibody of the present invention, can combine low toxicity characteristics with high efficacy. The antibody of the present invention can be used for various types and types of immune targeted therapies. When compared with antibodies that bind to the same antigen with two arms, the antibody of the present invention can have an expanded therapeutic window.

[0338] Also provided is the use of the bispecific antibody or variant thereof according to the invention for the preparation of a medicament for treating or preventing abnormal cells, cancer cells, tumors and / or metastasis formation. The tumor from which the metastasis originates is preferably a tumor that is positive for a B7 family member or TNFRSF14.

[0339] After transient transfection in suspended 293F cells, the antibodies of the invention can be produced at levels >50 mg / L. The bispecific antibodies can be purified to greater than 98% purity at yields >70%. Analytical characterization studies showed a bispecific IgG1 antibody spectrum comparable to that of a bivalent monospecific IgG1.

[0340] The present invention also provides a bispecific antibody or a variant thereof, which can bind to the extracellular portion of a membrane-associated member of the CD28 family and the extracellular portion of a membrane-associated member of B7 or TNFRSF14.

[0341] Also provided is a method for treating an individual suffering from cancer or pathogen infection, the method comprising administering to an individual in need thereof a therapeutically effective amount of: an antibody or variant according to the invention, or a composition according to the invention, or a nucleic acid molecule according to the invention, or a vector according to the invention.

[0342] The present invention also provides a protein of the present invention or a bispecific antibody of the present invention for use in treating an individual suffering from cancer.

[0343] Also provided is a cell system comprising an antibody or bispecific antibody of the present invention or a variant thereof, and a first cell expressing a membrane-associated member of the CD28 family and a second cell expressing a membrane-associated member of the B7 family or TNFRSF14.

[0344] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3The amino acid sequence of the VH chain MF5708 is shown in A. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF5708 is shown in A. In a preferred embodiment of the antibody, the antibody is monospecific.

[0345] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF5594 is shown in B. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF5594 is shown in B. In a preferred embodiment of the antibody, the antibody is monospecific.

[0346] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF5576 is shown in C. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF5576 is shown in C. In a preferred embodiment of the antibody, the antibody is monospecific.

[0347] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF5561 is shown in D. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF5561 is shown in D. In a preferred embodiment of the antibody, the antibody is monospecific.

[0348] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF5557 is shown in E. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF5557 is shown in E. In a preferred embodiment of the antibody, the antibody is monospecific.

[0349] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF5553 is shown in F. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF5553 is shown in F. In a preferred embodiment of the antibody, the antibody is monospecific.

[0350] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF5442 is shown in G. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF5442 is shown in G. In a preferred embodiment of the antibody, the antibody is monospecific.

[0351] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF5439 is shown in FIG. 14A . In a preferred embodiment, the VH chain comprises Figure 3The amino acid sequence of MF5439 is shown in H. In a preferred embodiment of the antibody, the antibody is monospecific.

[0352] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF5426 is shown in FIG. 1. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF5426 is shown in Figure 1. In a preferred embodiment of the antibody, the antibody is monospecific.

[0353] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF5424 is shown in J. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF5424 is shown in J. In a preferred embodiment of the antibody, the antibody is monospecific.

[0354] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF5382 shown in K. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF5382 is shown in K. In a preferred embodiment of the antibody, the antibody is monospecific.

[0355] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3The amino acid sequence of the VH chain MF5377 is shown in FIG. L. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF5377 is shown in L. In a preferred embodiment of the antibody, the antibody is monospecific.

[0356] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF5359 shown in FIG. 4 is shown in FIG. 4 . In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF5359 is shown in M. In a preferred embodiment of the antibody, the antibody is monospecific.

[0357] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF5361 is shown in N. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF5361 is shown in N. In a preferred embodiment of the antibody, the antibody is monospecific.

[0358] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 In a preferred embodiment, the VH chain comprises the amino acid sequence of MF5442. Fig.13 The amino acid sequence of MF5442 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0359] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 The amino acid sequence of the VH chain MF7691 shown in . In a preferred embodiment, the VH chain comprises Fig.13 The amino acid sequence of MF7691 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0360] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 The amino acid sequence of the VH chain MF7690 shown in . In a preferred embodiment, the VH chain comprises Fig.13 The amino acid sequence of MF7690 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0361] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 The amino acid sequence of the VH chain MF7689 shown in . In a preferred embodiment, the VH chain comprises Fig.13 The amino acid sequence of MF7689 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0362] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 The amino acid sequence of the VH chain MF7688 shown in . In a preferred embodiment, the VH chain comprises Fig.13 The amino acid sequence of MF7688 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0363] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 The amino acid sequence of the VH chain MF7700 shown in . In a preferred embodiment, the VH chain comprises Fig.13 The amino acid sequence of MF7700 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0364] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 The amino acid sequence of the VH chain MF7701 shown in . In a preferred embodiment, the VH chain comprises Fig.13 The amino acid sequence of MF7701 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0365] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 The amino acid sequence of the VH chain MF7703 shown in . In a preferred embodiment, the VH chain comprises Fig.13 The amino acid sequence of MF7703 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0366] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 The amino acid sequence of the VH chain MF7694 shown in . In a preferred embodiment, the VH chain comprises Fig.13The amino acid sequence of MF7694 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0367] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 The amino acid sequence of the VH chain MF7693 shown in . In a preferred embodiment, the VH chain comprises Fig.13 The amino acid sequence of MF7693 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0368] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 The amino acid sequence of the VH chain MF7692 shown in . In a preferred embodiment, the VH chain comprises Fig.13 The amino acid sequence of MF7692 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0369] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 The amino acid sequence of the VH chain MF7697 shown in . In a preferred embodiment, the VH chain comprises Fig.13 The amino acid sequence of MF7697 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0370] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13The amino acid sequence of the VH chain MF7696 shown in . In a preferred embodiment, the VH chain comprises Fig.13 The amino acid sequence of MF7696 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0371] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-L1, wherein the VH chain of the variable domain comprises a VH chain having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 In a preferred embodiment, the VH chain comprises the amino acid sequence of MF7695. Fig.13 The amino acid sequence of MF7695 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0372] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF6982 is shown in FIG. 1. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF6982 is shown in O. In a preferred embodiment of the antibody, the antibody is monospecific.

[0373] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF6974 is shown in P. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF6974 is shown in P. In a preferred embodiment of the antibody, the antibody is monospecific.

[0374] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF6972 is shown in Q. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF6972 is shown in Q. In a preferred embodiment of the antibody, the antibody is monospecific.

[0375] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF6936 is shown in R. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF6936 is shown in R. In a preferred embodiment of the antibody, the antibody is monospecific.

[0376] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF6935 is shown in S. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF6935 is shown in S. In a preferred embodiment of the antibody, the antibody is monospecific.

[0377] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of VH chain MF6932 is shown in T. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF6932 is shown in T. In a preferred embodiment of the antibody, the antibody is monospecific.

[0378] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF6076 is shown in FIG. 1. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF6076 is shown in U. In a preferred embodiment of the antibody, the antibody is monospecific.

[0379] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF6236 is shown in V. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF6236 is shown in V. In a preferred embodiment of the antibody, the antibody is monospecific.

[0380] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF6256 is shown in W. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF6256 is shown in W. In a preferred embodiment of the antibody, the antibody is monospecific.

[0381] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF6226 shown in Y. In a preferred embodiment, the VH chain comprises Figure 3The amino acid sequence of MF6226 is shown in W. In a preferred embodiment of the antibody, the antibody is monospecific.

[0382] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Figure 3 The amino acid sequence of the VH chain MF6930 is shown in X. In a preferred embodiment, the VH chain comprises Figure 3 The amino acid sequence of MF6930 is shown in W. In a preferred embodiment of the antibody, the antibody is monospecific.

[0383] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 The amino acid sequence of the VH chain MF6929 shown in . In a preferred embodiment, the VH chain comprises Fig.13 The amino acid sequence of MF6929 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0384] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 The amino acid sequence of the VH chain MF7699 shown in . In a preferred embodiment, the VH chain comprises Fig.13 The amino acid sequence of MF7699 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0385] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13The amino acid sequence of the VH chain MF7698 shown in . In a preferred embodiment, the VH chain comprises Fig.13 The amino acid sequence of MF7698 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0386] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 The amino acid sequence of the VH chain MF7687 shown in . In a preferred embodiment, the VH chain comprises Fig.13 The amino acid sequence of MF7687 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0387] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 In a preferred embodiment, the VH chain comprises the amino acid sequence of MF7686. Fig.13 The amino acid sequence of MF7686 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0388] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 In a preferred embodiment, the VH chain comprises the amino acid sequence of MF7685. Fig.13 The amino acid sequence of MF7685 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0389] The present invention also provides a bivalent antibody or a variant thereof comprising a variable domain that binds to PD-1, wherein the VH chain of the variable domain comprises at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the VH. Fig.13 The amino acid sequence of the VH chain MF7684 shown in . In a preferred embodiment, the VH chain comprises Fig.13 The amino acid sequence of MF7684 is shown in . In a preferred embodiment of the antibody, the antibody is monospecific.

[0390] As used herein, a monospecific antibody is an antibody that can bind to only one epitope. Therefore, only one antigen is bound. As used herein, a bivalent antibody comprises two variable domains that can bind to the same epitope. A monospecific bivalent antibody is generally referred to as a monoclonal antibody. Such a monospecific bivalent antibody has two variable domains, each having a heavy chain variable region that comprises the amino acid sequence of a specified MF and a common light chain as defined herein.

[0391] Bispecific antibodies or their variants contain two variable domains that bind to different epitopes. The two different epitopes are usually present on two different antigens or proteins. Bispecific antibodies are monovalent for the ability to bind to a specific epitope. They are bivalent in the sense of having the potential for two binding interactions.

[0392] The present invention is further described in the following examples. These examples do not limit the scope of the present invention, but are only used to illustrate the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0393] Figure 1 : Common light chain used in monospecific IgG and bispecific IgG.

[0394] Figure 1 A: Common light chain amino acid sequence. Figure 1 B: Common light chain variable domain DNA sequence and translation (IGKV1-39 / jk1). Figure 1 C: Common light chain constant region DNA sequence and translation. Figure 1 D: Translation of the common light chain variable domain of IGKV1-39 / jk5. Figure 1 E: V region IGKV1-39A.

[0395] Figure 2 : IgG heavy chain used to generate bispecific molecules. Figure 2 A: VH is the code Figure 3 The nucleic acid of MF amino acid sequence is shown in . Figure 2B: CH1 area. Figure 2 C: Hinge region. Figure 2 D: CH2 region. Figure 2 E: CH2 containing L235G and G238R silent substitutions. Figure 2 F: CH3 domain containing substitutions L351K and T366K (KK). Figure 2 G: CH3 domain containing substitutions L351D and L368E (DE).

[0396] Figure 3 : amino acid sequence of the heavy chain variable region: The symbol MF refers to a fab comprising the heavy chain variable region and a common light chain. The amino acid sequence of the light chain is as follows Figure 1 A. The underlined sequences indicate the CDR1, CDR2 and CDR3 regions of each amino acid sequence, respectively.

[0397] Figure 4 :Vector map and characteristics of pIRES-Neo3(MV1363).

[0398] Figure 5 : Vector map and characteristics of pVAX1.

[0399] Figure 6 : Vector map and features of the phagemid vector MV1473 used to generate an “immune” phage display library.

[0400] Figure 7 : Vector map and features of the IgG expression vector MV1452, which is used to express PD-1 and PD-L1 specific Fab arms in the KK variant heavy chain for bispecific IgG production.

[0401] Figure 8 : The amino acid sequence of the VH gene, when combined with a common light chain (e.g., MF1337), is tetanus toxin specific and is present in the DE variant heavy chain used to generate PD-L1xTT and PD-1xTT bispecific IgG molecules. The underlined sequences indicate the CDR1, CDR2, and CDR3 regions of each amino acid sequence, respectively.

[0402] Fig. 9 : Vector map and features of the IgG expression vector MV1377, which is used to express the TT-specific Fab arm MF1337 in the DE variant heavy chain for bispecific IgG production.

[0403] Fig.10: PD-1 / PD-L1 blocking assay. The ability of anti-PD-L1 and anti-PD-1 antibody panels to block the interaction of PD-L1 with coated PD-1 was assessed at a concentration of 10 μg / ml bispecific IgG. Data were normalized to those obtained with the bivalent benchmark PD-L1 antibody MPDL3280A at a concentration of 10 μg / ml (100% blocking). Representative examples of PD-L1 (top) and PD-1 (bottom) panels are shown. Maximum binding (normalized to 0% blocking) was established by incubation with non-PD-1 / PD-L1 specific human isotype antibodies. Contains Figure 3 All PD-1 and PD-L1 variable domains of the MF sequences shown in and not represented here blocked the PD-1 / PD-L1 interaction by >70%.

[0404] Fig.11 : PD-1 / PD-L1 functional activity of a panel of antibodies in a PD-1 / PD-L1-luc reporter system in dose titration.

[0405] Fig.12 : SEB-induced IL-2 production in PBMCs was enhanced by PD-1xPD-L1 antibody in a dose-dependent manner. PB composition is listed in Table 5. IL-2 production is shown as a stimulation index relative to the negative control antibody (Control Ab).

[0406] Fig.13 : amino acid sequence of the heavy chain variable region: The symbol MF refers to a fab comprising the heavy chain variable region and a common light chain. The amino acid sequence of the light chain is as follows Figure 1 A. The underlined sequences indicate the CDR1, CDR2 and CDR3 regions of each amino acid sequence, respectively.

[0407] Fig.14 : Binding to PD-1. All bispecific IgGs bound to PD-1 in ELISA.

[0408] Fig.15 : Binding to PD-L1. All bispecific IgGs bind to PD-L1 in ELISA

[0409] Fig.16 : FACS assay confirms binding to antigens expressed by CHO cells. Upper panel: binding to CHO-PD-L1 cells; lower panel, right: binding to CHO-PD1 cells.

[0410] Fig.17 : FACS assay confirms binding of lead candidate PD-1xPD-L1 bispecific IgG to antigen on activated T cells.

[0411] Fig.18: PD-1 / PD-L2 blocking assay. The ability of anti-PD-1 / PD-L1 antibodies and their parental bivalent antibodies to block the interaction between PD-L2 and PD-1 was evaluated in an in vitro blocking reporter assay.

[0412] Fig.19 : PD-L1 blocking assay. The ability of anti-PD-1 / PD-L1 bispecific antibodies and their parental bivalent antibodies to block the interaction of PD-L1 with coated PD-1 or CD80 was evaluated in a PD-L1 blocking assay.

[0413] Fig. 20 : PD-1xPD-L1 antibody increases IL-2 production induced by SEB of PBMC relative to an equimolar mixture of the parent bivalent monospecific antibodies of PD-1xPD-L1 antibody. The data represent the average IL-2 production by PBMCs from 4 independent donors in the presence of increasing antibody concentrations. The PG code indicates the corresponding parent bivalent antibody of each of the two bispecific antibodies (indicated by PB). IL-2 production is shown as relative luminescence units (RLU) measured in AlphaLISA.

[0414] Fig.21 : Antigen-specific T cell IFNγ / IL-2 release is enhanced in vitro by a bispecific PD-1xPD-L1 antibody. The graphs show the levels of IL-2 (upper graph) or IFNγ (lower graph) present in the supernatant. All graphs show the average value from duplicate wells. The dotted line indicates the cytokine level in the presence of antigen (Ag). The exact concentration (Ag=pg / ml) is shown only if the dotted line is not shown

[0415] Fig. 22 : Bispecific PD-1xPD-L1 antibodies enhance the responsiveness of T cells in vitro. Bars show the average level of IFNγ present in supernatants collected from allogeneic iMLRs, and error bars indicate SEM. (n=6). One-way ANOVA with Sidaks post hoc test analysis compared the average levels obtained with the tested antibodies with those obtained with isotype control antibodies (*) or vehicle controls (#). *P<0.05, **P<0.01, ##P<0.01, ###P<0.001, ####P<0.0001. The dotted line indicates the average IFNγ level obtained with vehicle alone. iMLR: immature mixed lymphocyte reaction.

[0416] Fig.23 : Effect of PD-1xPD-L1 on the proliferation of tumor-infiltrating CD4+ (left) and CD8+ T cells (right) from five patients with hepatocellular carcinoma (HCC).

[0417] Fig.24 : MF7686xMF7703 induces anti-tumor responses.

[0418] Fig.25 : MF7686xMF7703 enhances T cell numbers in tumors.

[0419] Fig.26 : MF7686xMF7703 enhances IFNγ release from tumor-specific T cells co-cultured with tumor cells.

[0420] Fig. 27 : MF7686xMF7703 enhances the cytotoxicity of tumor-specific T cells co-cultured with tumor cells.

[0421] Fig.28 : MF7686xMF7703 induces anti-tumor responses.

[0422] Fig.29 MF7686xMF7703 enhances tumor-specific (NY-ESO-1-specific) CD8 + The number of T cells.

[0423] Fig.30 : In the co-stimulation-dependent naive human T cell activation assay, PD-1xPD-L1 antibodies blocked the inhibitory effects of human recombinant PD-L1-Fc. Data represent the average stimulation index (SI) of duplo cultures performed in experiments with a single PBMC donor (1058), as the concentrations of the antibodies shown gradually increased (from left to right). PG codes indicate the corresponding parental bivalent antibodies of the two bispecific antibodies.

[0424] Fig.31 : In the costimulation-dependent nascent human T cell activation assay, the blocking effect of PD-1xPD-L1 bispecific antibodies was higher than that of the parental antibody combination or the benchmark antibody combination. The data represent the average stimulation index (SI) of duplo cultures performed in experiments with a single PBMC donor (1058), as the concentration of the antibodies shown gradually increased (from left to right). The PG code indicates the corresponding parental bivalent antibodies of the two bispecific antibodies.

[0425] Fig.32: T cells in the co-stimulation-dependent naive human T cell activation assay express increased amounts of PD-1 and PD-L1 after activation with anti-CD3 and anti-CD28. The mean fluorescence index (MFI) represents the expression level of PD-1 or PD-L1 on T cells harvested 24, 48, and 72 hours after activation with anti-CD3 and anti-CD28 antibodies.

[0426] Fig.33 : A bispecific PD1xPDL1 antibody that can completely block the PD1 axis. DETAILED DESCRIPTION

[0427] Example

[0428] As used herein, "MFXXXX" (wherein X is independently a number from 0 to 9) refers to a Fab comprising a variable domain in which VH has an amino acid sequence identified by four numbers. Unless otherwise indicated, the light chain variable region of the variable domain generally has Figure 1 A (usually Figure 1 B). "MFXXXX VH" refers to the amino acid sequence of VH identified by 4 numbers. MF also comprises a light chain constant region and a heavy chain constant region that usually interacts with the light chain constant region. PG refers to a monospecific antibody comprising the same heavy chain and light chain. PB refers to a bispecific antibody with two different heavy chains. The variable regions of the heavy chains are different, and usually the CH3 regions are also different, with one heavy chain having a KK mutation in its CH3 domain and the other having a complementary DE mutation in its CH3 domain (see PCT / NL2013 / 050294 (published as WO2013 / 157954) as reference).

[0429] Example: 1

[0430] Generation of materials for selection and screening

[0431] Cell line culture

[0432] Human ES-2 cells (Catalog No. CRL-1978) were purchased from ATCC and routinely maintained in McCoy's 5A (Gibco) supplemented with 10% FBS (Lonza). Freestyle 293F cells (Catalog No. p / n51-0029) were obtained from Invitrogen and routinely maintained in 293 FreeStyle medium. HEK293T (Catalog No. ATCC-CRL-11268), CHO-K1 (Catalog No. DSMZACC110) cell lines were purchased from ATCC and routinely maintained in DMEM / F12 (Gibco) supplemented with L-glutamine (Gibco) and FBS (Lonza), and CHO-S (Catalog No. 11619-012) cell lines were purchased from Gibco and routinely maintained in Freestyle CHO expression medium (Invitrogen) supplemented with L-glutamine.

[0433] Generation of PD-1 and PD-L1 expression vectors for immunization and for generating stable cell lines

[0434] The full-length cDNA of each target was synthesized or obtained by PCR amplification on a commercially available expression construct, the full-length cDNA comprising the target cDNA and specific primers that introduced the unique restriction sites for cloning and the Kozak consensus sequence for efficient translation. The cDNA of each target was cloned into a eukaryotic expression construct (e.g., pIRES-Neo3 (Clontech; Figure 4 ) or pVAX1 (ThermoFisher Scientific; Figure 5 )) to obtain pIRES-Neo3_[target name] and pVAX1_[target name], respectively. The insert sequence was verified by comparison with the NCBI reference amino acid sequence. The pIRES-Neo3 construct was used to generate stable cell lines. The pVAX1 construct was used for immunization purposes. For an overview of the resulting construct names, see Table 1.

[0435] Amino acid sequence full-length huPD-1 insert for expression on the cell surface (both in pIRES-Neo3 and pVAX1) (same as GenBank: NP_005009.2):

[0436]

[0437] in:

[0438] Signal peptide.

[0439]

[0440] ECD of huPD-1. Predicted TM region.

[0441]

[0442] Intracellular tail region.

[0443] Amino acid sequence full length macaque ( Macaca fascicularis ) PD-1 insert for expression on the cell surface (both in pIRES-Neo3 and pVAX1) (same as GenBank: ABR15751.1):

[0444]

[0445] in:

[0446] Signal peptide.

[0447]

[0448] ECD of maPD-1.

[0449] Predicted TM region.

[0450]

[0451] Intracellular tail region.

[0452] Amino acid sequence full-length huPD-L1 insert for expression on the cell surface (both in pIRES-Neo3 and pVAX1) (same as GenBank: AAI13735.1):

[0453]

[0454] in:

[0455] Signal peptide.

[0456]

[0457] ECD of huPD-L1.

[0458] Predicted TM region.

[0459] Intracellular tail region.

[0460] Amino acid sequence full-length rhesus monkey (macaca mulatta) PD-L1 insert (both in pIRES-Neo3 and pVAX1) for expression on the cell surface (same as GenBank: ABO33161.1):

[0461]

[0462] in:

[0463] Signal peptide.

[0464]

[0465] ECD of maPD-L1.

[0466] Predicted TM region.

[0467] Intracellular tail region.

[0468] Generation of stable cell lines expressing PD-1 or PD-L1

[0469] The pIRES-Neo3_[target name] expression construct (Table 1) was used to generate CHO-S or CHO-K1 clones that stably expressed the corresponding protein. The construct was transiently transfected into CHO-K1 cells using lipofectamine transfection or, for CHO-S cells, PEI transfection, and screened by FACS using antibodies that reacted with the corresponding protein. After confirming expression, the transiently transfected cells were inoculated with limiting dilution and cultured under the selection pressure associated with the expression construct used to obtain stable cell clones. After 2 to 3 weeks of selection, clones were screened by FACS. The selected clones were amplified by continuous passage, retested in FACS and frozen to -150°C. The clones that stably expressed heterologous proteins were named CHO-K1_[target name] cells or CHO-S_[target name] cells. For an overview of the constructs used to generate stable cell lines and their resulting names, see Table 1.

[0470] Example 2

[0471] Immunization, selection and screening

[0472] Mice used for immunization

[0473] To generate human antibodies that bind to huPD-1 and huPD-L1, mice transgenic for human VK1-39 light chains (common light chain mice, see WO2009 / 157771) and human heavy chain (HC) minilocus (containing selected human V gene segments, all human D and all human J) were immunized with recombinant proteins or DNA encoding the proteins, as briefly described below. These mice are referred to as Mouse.

[0474] Protein immunization

[0475] The recombinant protein and Gerbu adjuvant MM (Gerbu Biotechnik c#3001) were injected subcutaneously into Mice were immunized. Recombinant huPD-L1-His (SinoBiological; catalog number 10084-H08H) and huPD-1-Fc (R&D; catalog number 1086-PD) proteins were used for immunization. Mice were immunized with a mixture of 40 μg of recombinant protein and 40 μl of adjuvant in PBS (total volume of 100 μl). The mice were then boosted with 20 μg of recombinant protein in PBS together with 20 μl of adjuvant (total volume of 50 μl) on days 14 and 28. Mouse serum was collected on day 35 to determine serum titers. Mice with low serum titers received additional booster immunizations and serum analysis cycles. Each cycle consisted of two immunizations per week, each using 20 μg of recombinant protein in 50 μl PBS, followed by serum collection one week later for titer analysis. Mice that showed high serum titers against humans but not against macaque homologs received booster immunizations with macaque antigen proteins. Mice showing high serum titers against human and macaque targets received a final boost consisting of daily injections of 20 μg of recombinant protein in 50 μl PBS for three consecutive days. One day after the last injection, mouse lymphoid tissues were collected.

[0476] DNA immunization

[0477] DNA tattooing using a micropigmentation device Mice were immunized. DNA tattoo immunization was performed with 20 μg of plasmid DNA encoding the target antigen (pVAX1_[target name], Table 1). Mice were immunized with DNA encoding only human targets (PD-1, PD-L1). For PD-L1 immunization, Treg cells were depleted to destroy tolerance by injecting 0.5 mg of anti-CD25 antibody PC61.5 (Bioceros) into mice four days before the start of immunization. Mice were immunized on days 0, 3, 6, 14, 17, 28, and 31. Mouse serum was collected on day 35 to determine serum titer. Mice with low serum reactivity received additional booster immunizations and serum analysis cycles. Each cycle consisted of two immunizations per week, followed by serum collection for titer analysis one week later. Mice that showed strong serum reactivity to cells expressing human and macaque targets received the final booster immunization, followed by lymphoid tissue collection 3 days later.

[0478] Determination of serum titer

[0479] Serum titers were determined by FACS analysis using cell lines expressing human and macaque target antigens (Table 1).

[0480] Generation of “immune” phage Fab libraries from tissues of immunized mice by RT-PCR

[0481] The spleen and draining lymph nodes were removed from the mice and significant humoral responses against the corresponding target proteins were observed.

[0482] Single cell suspensions were generated from both spleen and inguinal lymph nodes, and these tissues were subsequently lysed in Trizol LS reagent (Thermo Scientific c#10296028) and stored at -80°C until use.

[0483] From successfully immunized mice, inguinal lymph nodes were used to construct an "immune" phage antibody library. RNA was extracted from a single cell suspension of lymphoid tissue. 1 μg of total RNA was used in an RT reaction using IgG-CH1 specific primers. The resulting cDNA was then used to amplify a polyclonal library of VH encoding cDNAs using internally adapted VH specific primers, essentially as described in Marks et al. (J Mol Biol. 1991 Dec 5; 222(3): 581-97). The library was then amplified as described in de Haard et al. (J Biol Chem. 1999 Jun 25; 274(26): 18218-30) (with the exception of the light chain of each antibody ( Figure 1 A and 1B) are the same and encoded by the vector), and the resulting PCR product is cloned into the phagemid vector ( Figure 6) for displaying Fab fragments on phage. After ligation, the phagemid was used to transform E. coli TG1 bacteria, and the transformed bacteria were plated on LB agar plates containing ampicillin and glucose. All phage libraries contained >4×10 5 transformants were obtained and the insertion frequency was >90%. Bacteria were harvested after overnight growth and used to prepare phage according to established protocols (de Haard et al., J Biol Chem. 1999 Jun 25; 274(26): 18218-30).

[0484] Use recombinant proteins to select phage carrying Fab fragments that specifically bind to human target proteins from an “immune” phage Fab library

[0485] The generated phage Fab library was used to select target-specific Fab using phage display on directly coated recombinant protein. For PD-L1, huPD-L1-His (Sinobiological; Catalog No. 10084-H08H), huPD-L1-Fc (R&D; Catalog No. 156-B7) and maPD-L1-His (Sinobiological; Catalog No. 90251-C08H) were used. For PD-1, huPD-1-Fc (R&D; Catalog No. 1086-PD) and huPD-1 biotin (BPS bioscience; Catalog No. 71109) were used.

[0486] For selection of non-biotinylated recombinant proteins ("panning selections"), proteins were coated onto MAXISORP TM Place MAXISORP TM ELISA plates were blocked with 4% dry skim milk (Marvel) in PBS. Phage Fab libraries were also blocked with 4% Marvel and, when Fc-tagged recombinant proteins were used, were also blocked with excess human IgG to deplete Fc region binders before adding the phage library to the coated antigen.

[0487] The phage library was incubated with the coated protein for 1.5 hours at room temperature with shaking. The plate or tube was then washed fifteen times with 0.05% Tween-20 in PBS, followed by five washes with PBS. The bound phage was eluted using trypsin for 20 minutes, and then the trypsin was neutralized with AEBSF trypsin inhibitor (Sigma).

[0488] For selection of biotinylated proteins ("selection in solution"), neutravidin was coated onto MAXISORP TMPlace MAXISORP TM ELISA plates were blocked with 1% casein in PBS. In parallel, in separate Eppendorf tubes, biotinylated proteins and phage Fab libraries were blocked for 30 minutes in 0.5% casein in PBS containing excessive human IgG. Thereafter, the blocked phage and biotinylated proteins were mixed and incubated at room temperature for 2 hours. Thereafter, the mixture was added to the neutravidin-coated wells for 20 minutes to capture the phage Fab particles bound to the biotinylated proteins. The plate was then washed 15 times with 0.05% Tween-20 in PBS, followed by 5 times with PBS. The bound phage was eluted for 20 minutes using trypsin, and then trypsin was neutralized with AEBSF trypsin inhibitor (Sigma).

[0489] The eluates from both selection strategies ("panning and in-solution selection") were added to E. coli TG-1 and incubated at 37°C for phage infection. The infected bacteria were then plated on agar plates containing ampicillin and glucose and incubated overnight at 37°C. Depending on the target, individual clones from the selection output were screened for target binding in ELISA or FACS. Phage carrying Fab fragments that specifically bind to human targets were selected from the "immune" phage Fab library using cells stably expressing the target protein.

[0490] Phage display on cells expressing the corresponding target was used to select the phage Fab library produced from the target by immunization of mice. Stable cell lines expressing PD-1 or PD-L1 (Table 1) were used for the first round of selection. The cells were blocked with 10% FBS in PBS. After blocking, the rescued phage was incubated with the blocked cells. The cells plus phage were incubated at 4 ° C for 1 hour. The cells were washed with 10% FBS in 1 ml of PBS (5 times). The bound phage was eluted with trypsin for 20 minutes, and then the trypsin was neutralized with AEBSF trypsin inhibitor (Sigma). The eluate was added to Escherichia coli TG-1 and incubated at 37 ° C for phage infection. The phage-infected bacteria were then plated on agar plates containing ampicillin and glucose and incubated overnight at 37 ° C.

[0491] For PD-L1, a second round of selection of ES-2 cells endogenously expressing huPD-L1 was performed using the same protocol as used for the first round of selection. After selection, individual clones were screened for target binding in FACS.

[0492] Screening of target-specific Fab clones in ELISA

[0493] In individual clones, soluble Fab was prepared as described (J Mol Biol. 1991 Dec 5; 222 (3): 581-97; J Biol Chem. 1999 Jun 25; 274 (26): 18218-30). It was diluted 1:5 in 4% dry skim milk (Marvel) in PBS (blocking buffer) and tested in ELISA for binding to wells coated with the same antigen as that used for selection. The bound Fab was detected by staining with anti-myc antibody (Roche; catalog number 11667203001) diluted 1:1000 in blocking buffer, followed by staining with HRP-conjugated anti-mouse IgG antibody (Jackson Immunoresearch; catalog number 715-035-150) diluted 1:5000 in blocking buffer. After staining with each antibody, the wells were washed with PBS-T (PBS-0.05% v / v Tween 20). Bound secondary antibodies were visualized by TMB / H2O2 staining and OD 450nm The staining was quantified by measurement. A clone was considered to be bound to the target when the OD450nm was at least three times higher than the background signal obtained with the negative control Fab.

[0494] The VH encoding cDNA of all target-specific clones was sequenced. The clones obtained from the cell selection output were then analyzed in FACS for binding to PD-L1 expressed on cells, and the selected unique clones based on sequence features and cluster analysis were analyzed as follows. Screening of target-specific Fab clones in FACS

[0495] In individual clones, soluble Fabs selected on cells expressing the corresponding targets were prepared as described (J Mol Biol. 1991 Dec 5; 222(3): 581-97; J Biol Chem. 1999 Jun 25; 274(26): 18218-30). Binding to cells expressing human and macaque targets (Table 1) was tested in FACS by incubation with a mixture of Fab samples diluted 1:5 in FACS buffer (0.5% HI-FBS in PBS) and anti-myc antibodies (Gentaur; catalog number 04-CMYC-9E10) diluted 1:1000. Bound Fab / anti-myc complexes were detected by incubation with APC-conjugated goat anti-mouse IgG antibodies (BD Bioscience; catalog number 550826) diluted 1:500 in FACS buffer. After each antibody incubation, the wells were washed three times with FACS buffer. The stained cells were analyzed using a FACS Accuri C6 instrument (Becton and Dickinson). Clones were considered positive when the mean fluorescence intensity was at least three times higher than the background signal obtained with the negative control Fab.

[0496] Example 3

[0497] Characterization of huPD-L1 and huPD-1-specific Fab clones in IgG format

[0498] Human PD-L1 and PD-1 specific Fabs were recloned into IgG format

[0499] Unique clones selected based on CDR3 sequence and VH germline differences that bind to human and macaque target proteins expressed on cells were then recloned into a pool containing a common light chain ( Figure 1 ) of an IgG expression plasmid (e.g., MV1452) ( Figure 7 )middle.

[0500] Expression of bispecific IgG containing human PD-L1 or human PD-1 specific Fab and tetanus toxin specific Fab

[0501] Bispecific antibodies are produced by transient co-transfection of two plasmids encoding IgG with different VH domains using proprietary CH3 engineering technology to ensure efficient heterodimerization and formation of bispecific antibodies. The common light chain present on the two plasmids containing the heavy chain is also co-transfected into the same cell. In our co-pending applications (WO2013 / 157954 and WO2013 / 157953; incorporated herein by reference), we have disclosed methods and means for producing bispecific antibodies from a single cell, thereby providing a means for the formation of bispecific antibodies that is more inclined to the formation of monospecific antibodies. These methods can also be advantageously used in the present invention. In particular, some preferred mutations that essentially generate only bispecific full-length IgG molecules are amino acid substitutions at positions 351 and 366 in the first CH3 domain ("KK variant" heavy chain), such as L351K and T366K (numbering according to EU numbering) and amino acid substitutions at positions 351 and 368 in the second CH3 domain ("DE variant" heavy chain), such as L351D and L368E, or vice versa ( Figure 2 ). It has been previously demonstrated in our co-pending applications that negatively charged DE variant heavy chains preferentially pair with positively charged KK variant heavy chains to form heterodimers (so-called "DEKK" bispecific molecules). Homodimerization of DE variant heavy chains (DE-DE homodimers) or homodimerization of KK variant heavy chains (KK-KK homodimers) hardly occurs due to the strong repulsion between charged residues in the CH3-CH3 interface between the same heavy chains.

[0502] The VH gene encoding the antibody that binds to human PD-L1 and PD-1 was cloned into the MV1452 IgG expression vector encoding the positively charged CH3 domain. Figure 8 ) was cloned into the MV1377 IgG expression vector encoding the negatively charged CH3 domain ( Fig. 9 293F Freestyle cells adapted to suspension growth were cultured in T125 flasks on a rocking platform to a density of 3.0 × 10 6 cells / ml. 6 to 0.5×10 6 The density of 100 viable cells / ml was inoculated in each well of a 24-deep well plate. Cells were transiently transfected with a mixture of two plasmids encoding different antibodies, which were cloned into a proprietary vector system. Seven days after transfection, cell supernatants were harvested and filtered through a 0.22 μm filter (Sartorius). Sterile supernatants were stored at 4 °C until antibody purification.

[0503] Purification of bispecific IgG

[0504] IgG purification was performed on a small scale (<500 μg) using protein A affinity chromatography. Small-scale purification was performed in 24-well filter plates under aseptic conditions using filtration. First, the pH of the culture medium was adjusted to pH 8.0, and then the supernatant containing IgG was incubated for 2 hours at 25 ° C on a shaking platform at 600 rpm using Protein A Sepharose CL-4B beads (50% v / v) (Pierce). Next, the beads were harvested by filtration. The beads were washed twice with PBS (pH 7.4). Then, the bound IgG was eluted at pH 3.0 with 0.1M citrate buffer, and the eluent was neutralized using Tris (pH 8.0) immediately. Buffer exchange was performed by centrifugation using multi-screen Ultracel10 multi-well plates (Millipore). Samples were finally harvested in PBS (pH 7.4). Octet was used to measure IgG concentration. Protein samples were stored at 4 ° C.

[0505] IgG quantification using Octet

[0506] To determine the amount of purified IgG, total human IgG (Sigma Aldrich, catalog number 14506) was used as a standard and the concentration of the antibody was determined by Octet analysis using a protein A biosensor (Forte-Bio, according to the supplier's recommendations).

[0507] Specificity analysis of huPD-L1xTT and huPD-1xTT bispecific IgG

[0508] The bispecific antibodies were tested for binding in FACS to stable cell lines expressing the relevant human and macaque interspecies homologues (Table 1) and to wt cells. Therefore, cells were harvested and diluted to 10% in FACS buffer (PBS / 0.5% BSA / 0.5 mM EDTA). 6 1×10 cells / ml were added to each well of a U-bottom 96-well plate. 5 Up to 2×10 5Cells. The cells were centrifuged at 300g for 2 minutes at 4°C. The supernatant was discarded by inverting the plate. 50 μl of each IgG sample with a concentration of 10 μg / ml was added and incubated on ice for 1 hour. The cells were centrifuged once, the supernatant was removed, and the cells were washed twice with 150 μl of FACS buffer. 50 μl of diluted 1:400 goat anti-human IgG PE (Invitrogen) was added and incubated on ice for 30 minutes in the dark. After adding FACS buffer, the cells were centrifuged once, the supernatant was removed, and the cells were washed twice with FACS buffer. Cells were analyzed on a FACSCanto flow cytometer (Becton and Dickinson) under HTS settings. The binding of antibodies to cells was assessed by measuring the mean fluorescence intensity (MFI) of the stained cell population. When the MFI was at least five times the MFI of the same cell population stained with a (negative control) non-binding antibody (for tetanus toxoid), it was considered that the antibody was bound to its target. Identification of huPD-L1 and huPD-1 specific Fab arms in PD-L1xTT and PD-1xTT bispecific IgGs for ligand blocking activity

[0509] The ability of huPD-L1 and huPD-1 binding clones to block the interaction of PD-L1 with PD-1 was tested. For the PD-L1 Fab arm, the ability to block the interaction between PD-L1 and CD80 was also evaluated. Therefore, PD1-Fc (R&D systems; Catalog No. 1086-PD) or CD80-Fc (R&D systems; Catalog No. 140-B1) was coated to the maxisorp plate at 1 and 3 μg / ml, respectively. The coated wells were blocked with 4% BSA in PBS. Thereafter, 0.55 μg / ml of biotinylated PD-L1 (BPS bioscience; Catalog No. 71105) was added in the presence or absence of 0.15 to 20 μg / ml of IgG. The bound biotinylated PD-L1 was detected with HRP-conjugated streptavidin (BD bioscience: Catalog No. 554066) diluted 1:2000 in blocking buffer. After each incubation step, the ELISA plate was washed three times with PBS-T (PBS-0.05% v / v Tween 20). Bound streptavidin was visualized by TMB / H2O2 staining and the OD 450nmThe staining was quantified by measurement. A clone was considered to block the interaction of PD-1 with PD-L1 when the ELISA signal was reduced by more than 70% at a concentration of 10 μg / ml IgG (PD-L1xTT or PD-1xTT) compared to a control to which a TT-specific competing antibody was added. For results obtained with a representative selection of PD-1 and PD-L1 antibody panels tested as PD-1xTT or PD-L1xTT bispecific molecules, see Fig.10 .

[0510] Affinity ranking of huPD-L1 and huPD-1 specific Fab arms present in PD-L1xTT and PD-1xTT bispecific IgGs

[0511] The bispecific antibodies that showed binding to the corresponding human and macaque interspecies homologs in FACS were ranked for their apparent affinity against both interspecies homologs in FACS. Therefore, stable cell lines expressing the respective interspecies homologs (Table 1) were harvested and diluted to 10 μl in FACS buffer (PBS / 0.5% BSA / 0.5 mM EDTA). 6 Cells / ml. The cells were centrifuged at 300g for 2 minutes at 4°C. The supernatant was discarded by inverting the plate. Each IgG sample of 50μl was added in a 2-fold dilution series (from 10 to 0.01μg / ml) in 11 steps and incubated on ice for 1 hour. The cells were centrifuged once, the supernatant was removed, and the cells were washed twice with 150μl of FACS buffer. 50μl of diluted 1:400 goat anti-human IgG PE (Invitrogen) was added and incubated on ice for 30 minutes in the dark. After adding FACS buffer, the cells were centrifuged once, the supernatant was removed, and the cells were washed twice with FACS buffer. Cells were analyzed on a FACSCanto flow cytometer (Becton and Dickinson) under HTS settings. The binding of antibodies to cells was assessed by measuring the mean fluorescence intensity (MFI) of the stained cell population. An antibody is considered bound to its target when the MFI is at least five times the MFI of the same cell population stained with a (negative control) non-binding antibody (to tetanus toxoid).

[0512] PBMC isolation

[0513] Human whole blood was obtained from buffy coat (Sanquin) and diluted with PBS 1: 1. Leucosep tubes (Greiner Bio-One catalog number 227290) were filled with 17.5m Ficoll-Paque Plus (Amersham Biosciences catalog number 17-1440-02) warmed at room temperature (RT). Ficoll-Paque Plus was spun down at 1000 × g for 30 seconds at room temperature. 30ml of diluted whole blood was poured on top. The tube was spun at 1000 × g for 10 minutes at room temperature, and the mononuclear PBMC interface was harvested, washed twice in PBS, and resuspended in 250 μl PBS. PBMCs were counted and readjusted to 1 × 106 / ml in tissue culture medium (DMEM with 10% FCS), and frozen by adding an equal volume of ice-cold freezing medium (80% culture medium / 20% DMSO). The cells were stored in 1 ml aliquots at -150°C until further use.

[0514] SEB assay

[0515] The functional activity of the bispecific antibody was determined by using PBMC stimulated by Staphylococcus enterotoxin B (SEB). SEB specifically activates T cells expressing Vβ3, 12, 14, 15, 17 and 20 T cell receptor chains. PBMCs from 3 donors were thawed, washed, counted and resuspended in culture medium (RPMI1640 plus 10% heat-inactivated FBS) to a concentration of 2 × 106 cells / ml. In the presence of SEB (2000 or 125ng / ml), cells were seeded in flat-bottomed 96-well plates (2 × 105 cells / well). Serial dilutions of antibodies starting from 20 μg / ml were added. Each plate contained a negative control antibody (TT-specific PG1337) and a positive control antibody (ipilumumab, nivolumab (PD-1xPD-L1), LAG3.5 (LAG-3xPD-1)) that served as a reference control. Cells were stimulated for 3 days at 37°C, 5% CO2 in 95% relative humidity before testing for cytokine secretion and / or cell surface expression of antigens.

[0516] Cytokine assay

[0517] ELISA: After stimulation of T cells or PBMCs at various times, the plates were centrifuged and the medium removed. Cytokine levels were detected by AlphaLISA according to the manufacturer's instructions (Perkin Elmer). Concentrations were calculated based on the standard curve.

[0518] Luminex assay: Another method for determining in vitro cytokine production is to use a multiplex assay developed by eBioscience. Following the manufacturer's instructions, levels of IFN-γ, IL-2, and TNF-α were measured in culture supernatants. Results were analyzed by eBioscience analysis software.

[0519] Reference Antibodies

[0520] Antibodies that inhibit PD-1 and PD-L1 function are known in the art. Monoclonal bivalent antibodies were constructed based on public information and expressed in CHO-S cells. The anti-PD-1 antibody nivolumab was produced based on information disclosed in CA02607147. The anti-PD-L1 antibody MPDL3280A is based on information disclosed in WO2010077634A1 Genentech Inc).

[0521] PD-1 / PD-L1 Blockade Reporter Assay

[0522] The PD-1 / PD-L1 blocking reporter assay used was developed by Promega and is based on two cell systems; CHO cells and T cell activators expressing PD-L1, and Jurkat / NFAT-RE reporter cell lines overexpressing PD-1. Promega's thawing and use format was used for PD-1 / PD-L1 blocking reporter assays. Cells expressing PD-L1 (Catalog No. C187103) were thawed in 14.5 mL of cell recovery medium (DMEM / F12 containing 10% FBS). Next, 50 μL of cell suspension was added to the inner well of a 96-well half-area plate (Corning, Catalog No. 3688). The plate was incubated overnight at 37 ° C, 5% CO in 95% relative humidity. The next day, the culture medium was removed, and 20 μL of test antibodies in a serially diluted (starting concentration 10 μg / mL) assay medium (RPMI 1640 containing 4% FBS) were added to each well. Each plate contains a negative control antibody (TT-specific PG1337) and a positive control antibody (a nivolumab-based control, referred to herein as 5C4, and an atezolizumab-based control, referred to herein as MPDL3280A or YW243.55.S70) that are serially diluted and serve as reference controls. PD-1 effector cells (Cat. No. C187105) were thawed in 5.9 ml of assay medium, and 20 μL of cell suspension was added to each well. The plate was incubated at 37° C., 5% CO for 6 hours or overnight in 95% relative humidity. 40 μL of luciferase (Bio-Glo luciferase assay system, Cat. No. G794L) was added the next day, and the amount of luciferase activity was measured using a BioTek Synergy 2 multi-mode microplate reader. The potency was measured by luciferase activity compared to the negative control antibody.

[0523] Example 4

[0524] Screening of PD1xPD-L1 antibody panel

[0525] The VH from the PD-1 and PD-L1 antibody panels were recloned into a charged engineered Fc silencing vector, so that after the antibody heavy chain was expressed, the heavy chain was forced to undergo heterodimerization, thereby generating bispecific antibodies after transfection. The PD-1 Fab arm was cloned in the MV1625 vector, while the PD-L1 Fab arm was recloned in the MV1624 vector. The PD-1 and PD-L1 antibodies were combined with MF1337 (TT targeting Fab arm) to generate bispecific antibodies targeting PD-1 or PD-L1 in a monovalent manner. The bispecific antibodies were tested in a semi-logarithmic continuous titration (starting concentration 10 μg / ml) in a PD-1 / PD-L1 blocking reporter assay to rank the blocking potency of the antibodies. In the PD-1 / PD-L1 blocking reporter assay, the monovalent PD-L1 antibody panel could be ranked, while the monovalent PD-1 antibody panel showed insufficient blocking ability to be ranked. Therefore, the PD-1 antibody was generated in a bivalent format and retested in the PD-1 / PD-L1 blocking reporter assay.

[0526] Based on the activity data, antibodies were selected from the PD-1 or PD-L1 antibody group for subsequent PD1xPD-L1 bispecific screening. The activities of the selected candidates in the reporter assay are shown in Tables 2 and 3, respectively. The PD-1 Fab group consists of functionally active variants within three antibody clusters (i.e., A, B, and C) and a non-functional variant D, while the PD-L1 Fab group consists of antibodies derived from eleven antibody clusters. Both the PD-1 and PD-L1 antibody groups contain a functionally inactive antibody.

[0527] A total of 120 PD-1xPD-L1 bispecific antibodies containing 10 different PD-1 Fab arms and 12 different PD-L1 Fab arms were generated in a 24-well format and IgG purified. All antibodies were tested for their ability to induce dose-dependent luciferase expression in a PD-1 / PD-L1-luc reporter system with serial titration. Nivolumab or MPDL3280A were included as reference antibodies. Fig.11 Shown are luciferase-induced selected bispecific antibodies that displayed a broad range of responses.

[0528] Table 4 shows the percentage activity of the bispecific antibodies compared to MPDL3280A. The PD-1xPD-L1 combination containing one non-functional Fab arm had the lowest effect. In addition, the combination of the most effective PD-L1Fab with the most effective PD-1Fab did not produce the most effective bispecific PD-1xPD-L1 antibody. Some PD-L1 Fab arms (such as MF5561, MF5442, and MF5382) have high efficacy with multiple PD-1Fab arms, while MF5359, for example, is not. The PD-1Fab arm with the highest activity induced potent activity together with several PD-L1 Fab arms.

[0529] The most potent PD-1xPD-L1 bispecific antibodies were tested in a serial titration in the SEB assay. Fig.12 A representative experiment of the SEB assay performed on three donors stimulated with 2 μg / ml SEB is shown. All PD-1xPD-L1 combinations tested showed dose-related induction of IL-2 and γ-IFN (not shown). The maximum induced IL-2 levels obtained were 50 to 100 ng / ml. Most PD-1xPD-L1 bispecific antibodies showed higher potency than nivolumab. Below 10 μg / ml, most PD-1xPD-L1 bispecific antibodies were more effective than ipilimumab. Therefore, PD-1xPD-L1 antibodies are very effective in stimulating T cell responses.

[0530] Example 5

[0531] PD-1xPD-L1 Cloning and expression of Fab variants

[0532] A total of 15 PD-1 and 25 PD-L1 Fab arms were combined to generate 65 bispecific antibodies. The sequences of the Fab arms are as follows Figure 3 and 13 as shown in .

[0533] 65 VH regions from the PD-L1 and PD-1 Fab groups were re-cloned into an engineered expression vector so that the expression of the antibody forced the heavy chain to undergo heterodimerization, thereby producing a bispecific antibody after transfection. The VH region of the PD-1 Fab arm was cloned into the vector MV1625, and the VH region of the PD-L1 Fab arm was cloned into the vector MV1624. Both vectors carry additional mutations in the CH2 and CH3 coding regions of the IgG protein: both MV1625 and MV1624 contain L235G and G236R substitutions, which eliminate the Fcγ receptor and C1q interactions of the resulting antibody. MV1625 also contains amino acid substitutions L351D and L368E in the CH3 domain ("DE variant" heavy chain), while MV1624 contains amino acid substitutions L351K and T366K in the CH3 domain ("KK variant" heavy chain).

[0534] After cloning into the relevant vector, the DNA sequence was confirmed by PCR. Midiprep DNA was prepared for all constructs. Different vector pairs (one vector carrying anti-PD-1 clones and another vector carrying anti-PD-L1 clones) were co-transfected into FreeStyle 293-F cells in duplo to produce bispecific proteins (65 in total). The IgG produced in each well of the 24-well plate was then purified and buffer exchanged using a Zeba desalting column according to the manufacturer's instructions; the protein yield was then quantified by OD280 using NanoDrop.

[0535] Overview of PB numbers and their MF composition.

[0536]

[0537] Example 6

[0538] Confirmation of PD-1 and PD-L1 by ELISA Antigen binding of Fab

[0539] Restricted antigen ELISA for PD-1 or PD-L1

[0540] To confirm the binding of PD-1 and PD-L1 Fabs present in the bispecific IgG, an antigen titration ELISA was performed. In this ELISA, serially diluted PD-1 or PD-L1 antigens were coated to 96-well plates. The plates were then incubated with the test antibodies and detected using a secondary mouse anti-human antibody conjugated to horseradish peroxidase (HRP), which converts the colorless substrate into an easily visible dye. A negative TT-specific control antibody, PG1337, was included in all plates. The benchmark anti-huPD-1 antibody 5C4 (based on nivolumab) was included as a positive control on all plates coated with PD-1 protein, and the benchmark anti-huPD-L1 antibody YW243.55.S70 (based on atezolizumab) was included as a positive control on all plates coated with PD-L1 protein.

[0541] To this end, 96-well Nunc Maxisorp plates were coated overnight at 4°C with human PD-L1-Fc (R&D systems, catalog number 156-B7) or human PD-1-Fc (R&D systems, catalog number 1086-PD) in PBS in 3-fold 7-step serial dilutions from 10 μg / mL to 0.014 μg / mL in each column of the plate. The next day, the ELISA plate was washed 3 times with 300 μL PBST and blocked by filling the wells with 2% BSA in PBS and incubating for 1 hour at room temperature. The plate was emptied and PD-1xPD-L1 antibody and control antibodies PG1337, 5C4 and YW243.55.S70 (50 μL / well) were added at 5 μg / mL in PBS-2% BSA, one column for each antibody. After incubation at room temperature for 1 hour, the wells were washed three times with 300 μL PBST, and then 50 μL secondary antibodies in the form of HRP-conjugated mouse anti-human IgG (BD, catalog number 555788) diluted 1:2000 in PBS-2% BSA were added. After incubation at room temperature for 1 hour, the wells were washed 3 times with 300 μL PBST, and then TMB peroxidase substrates A (BD, catalog number 51-2506KC) and B (BD, catalog number 51-2607KC) were added in a ratio of 1:1, 50 μL per well. After up to 10 minutes, the reaction was stopped by adding 1M H2SO4 solution with 50 μL / well. Then, the ELx808 microplate reader measured the optical density (opticaldensity, OD) at 450nM wavelength.

[0542] The results of the restriction-specific human IgG ELISA against PD-1 are shown in Fig.14 The results of the restriction-specific human IgG ELISA against PD-L1 are shown in Fig.15The figure shows the optical density (OD) at 450 nm of increasing concentrations of coated antigen as measured by a microplate reader. All tested Fab arms bound well to PD-1 and PD-L1, and the binding levels were similar to those of the benchmark antibody, making all Fab arms suitable for use in bispecific PD-1xPD-L1 antibodies.

[0543] Example 7

[0544] Activity of Antibodies in PD-1 / PD-L1 Blocking Reporter Assays

[0545] In a physiologically relevant PD-1 / PD-L1 blocking reporter assay developed by Promega, the blocking activity of the generated bispecific PD-1xPD-L1 antibodies was tested in vitro. The assay is based on a dual-cell system in which CHO cells expressing PD-L1 and T cell receptor activators are co-cultured with Jurkat / NFAT-RE reporter cell lines overexpressing PD-1. Jurkat T cells contain luciferase reporters that can be activated by the NFAT (nuclear factor of activated T cells) pathway. The interaction of PD-1 with PD-L1 inhibits the activation of this pathway. However, blocking the interaction of PD-1 / PD-L1 with antibodies against PD-1 or PD-L1 can activate the NFAT pathway. Therefore, the greater the degree of PD-1 / PD-L1 blocking, the greater the activation of the luciferase reporter. To this end, each antibody was added to CHO cells expressing PD-L1 in serial dilutions before adding Jurkat / NFAT-RE reporter cells overexpressing PD-1.

[0546] method

[0547] PD-1 / PD-L1 Blockade Reporter Assay

[0548] The PD-1 / PD-L1 blocking reporter assay used was developed by Promega and is based on two cell systems; CHO cells and T cell activators expressing PD-L1, and Jurkat / NFAT-RE reporter cell lines overexpressing PD-1. Promega's thawing and use format was used for PD-1 / PD-L1 blocking reporter assays. Cells expressing PD-L1 (Catalog No. C187103) were thawed in 14.5 mL of cell recovery medium (DMEM / F12 containing 10% FBS). Next, 50 μL of cell suspension was added to the inner well of a 96-well half-area plate (Corning, Catalog No. 3688). The plate was incubated overnight at 37°C, 5% CO2 in 95% relative humidity. The next day, the culture medium was removed and 20 μL of test antibody in a serially diluted (starting concentration 10 μg / mL) assay medium (RPMI 1640 containing 4% FBS) was added to each well. Each plate contains a negative antibody (PG1337) against tetanus toxoid and a positive control anti-PD-L1 antibody (YW243.55.S70; based on atezolizumab) that serves as a reference control. PD-1 effector cells (Cat. No. C187105) were thawed in 5.9 ml of assay medium and 20 μL of cell suspension was added to each well. The plate was incubated at 37 ° C, 5% CO2 for 24 hours in 95% relative humidity. 40 μL of luciferase substrate (Bio-Glo Luciferase Assay System, Catalog No. G794L) was added the next day, and the amount of luciferase activity was measured using a BioTek Synergy 2 multi-mode microplate reader. The potency was measured by luciferase activity compared to the positive control antibody (YW243.55.S70).

[0549] result

[0550] The extent of blocking after 24 hours is shown in Table 6, where the data show the relative induction of luciferase activity compared to the activity measured in wells to which no antibody was added. The percentage activity was calculated based on the area under the curve (AUC) relative to the positive control anti-PD-L1 antibody YW243.55.S70. All tested bispecific PD-1xPD-L1 antibodies showed significant blocking activity.

[0551] Example 8

[0552] FACS binding of PD-1xPD-L1 bispecific IgG to antigen-expressing cells

[0553] The binding of two PD-1xPD-L1 bispecific antibodies to PD-1 and PD-L1 on CHO cell lines stably expressing each antigen was tested. The cell lines used for this FACS were CHO-huPD-L1, CHO-S-huPD-1, and untransfected CHO cells (negative cells). In brief, these cell lines were stained with increasing concentrations of bispecific IgG, parental IgG, or control IgG, followed by goat anti-human IgG-PE detection. The positive controls were the benchmark anti-huPD-1 antibody 5C1 (based on nivolumab) and the benchmark anti-huPD-L1 antibody YW243.55.S70 (based on atezolizumab); the negative control was the anti-tetanus toxin antibody PG1337.

[0554] To this end, stable CHO-huPD-L1 cells (MC0866) and CHO-S-huPD-1 cells (MC0617) were harvested and diluted to 10 μl in FACS buffer (PBS / 0.5% BSA / 2 mM EDTA). 6 0.5×10 cells / mL were added to each well of a U-bottom 96-well plate (BD, catalog number 353910). 5 Up to 2×10 5 Cells. The cells were centrifuged at 300g for 3 minutes at 4°C. The supernatant was discarded by inverting the plate. The cells were washed by adding 200 μL of ice-cold FACS buffer. The cells were centrifuged at 300g for 3 minutes at 4°C again, and the supernatant was discarded as before. Each IgG sample of 40 μL was added (3 times 9 steps of serial dilution starting from 10 μg / mL), and the cells were incubated on ice in the dark for 30 minutes. The cells were then washed twice, starting from directly adding 200 μL of ice-cold FACS buffer, followed by centrifugation at 300g for 3 minutes at 4°C, and the supernatant was removed. Secondary antibody staining was performed by adding 40 μL of goat anti-human IgG-PE (3 μg / mL; Invitrogen, catalog number H10104) and incubating the plate on ice for 30 minutes in the dark. The cells were washed twice with 200 μL of ice-cold FACS buffer again, and resuspended in 50 to 200 μL of FACS buffer. Cells were analyzed on a FACSCanto flow cytometer (Becton and Dickinson) in a high throughput sampler (HTS) setting. The binding of antibodies to cells was assessed by measuring the mean fluorescence intensity (MFI) of the stained cell population. Antibodies were considered to bind to their targets when the MFI was at least five times the MFI of the same cell population stained with a negative control antibody.

[0555] result

[0556] The binding of two bispecific IgGs to PD-L1 and PD-1 and the binding of the reference antibody are shown in Fig.16 The data show the MFI detected by FACS on CHO cells expressing PD-L1 and PD-1 stained with increasing concentrations of bispecific IgG or control IgG. These results show that both bispecific IgGs can recognize both targets when expressed on CHO cells, and their binding extent is similar or better than that of the reference antibody.

[0557] Example 9

[0558] FACS binding of PD-1xPD-L1 bispecific IgG to activated T cells

[0559] FACS binding assay using human activated T cells

[0560] The binding of two PD-1xPD-L1 bispecific antibodies to PD-1 and PD-L1 on activated T cells was also tested. Briefly, 96-well plates were coated with anti-CD3 antibodies overnight. Purified T cells from a single donor were then added and cultured for 3 days. After 3 days, activated T cells were harvested and pooled and used for FACS assays to compare the binding of the bispecific IgG with that of its parental monospecific bivalent IgG. The positive controls were the benchmark anti-huPD-1 antibody 5C4, the benchmark anti-huPD-L1 antibody YW243.55.S70; the negative control was the anti-tetanus toxin antibody PG1337.

[0561] method

[0562] T cell purification

[0563] Peripheral blood mononuclear cells (PBMC) from healthy donors were thawed and 9 volumes of culture medium (RPMI1640 with 10% heat-inactivated (hi) FBS) were added dropwise. The cells were centrifuged at 200g for 10 minutes at room temperature. The cell pellet was resuspended in 10mL of culture medium and the cells were allowed to rest by incubating overnight at 37°C, 5% CO2 in 95% relative humidity. The next day, T lymphocytes were isolated using the EasySep T cell enrichment (pan CD3) purification procedure (as described by the manufacturer (Stem cell Technologies catalog number 19051)). The EasySep procedure uses negative selection. In brief, PBMCs were centrifuged at 200g for 10 minutes at room temperature. The cell pellet was centrifuged at 5×10 7The cells were resuspended in EasySep buffer at a concentration of 10 cells / mL. 50 μL of EasySep human T cell enrichment mixture was added to each mL of cell volume, mixed and incubated at room temperature for 10 minutes. Next, 50 μL of EasySep D magnetic particles were added to each mL of cell volume and incubated at room temperature for 5 minutes. The total volume was brought to 2.5 mL with EasySep buffer, and the cell suspension was transferred to a 5 mL round-bottom Falcon tube (BD Biosciences, catalog number 352235) after mixing. Next, the tube was placed in a magnet so that the undesirable cell fraction was bound to the magnet at room temperature for 5 minutes. Next, the tube was inverted and the purified T cell fraction was poured into a new tube containing 7.5 mL of culture medium. The cells were harvested by centrifugation at 200 g for 10 minutes at room temperature, and then 1 × 10 6 Resuspend in culture medium at a concentration of 10 cells / mL.

[0564] FACS binding assay

[0565] One day before the start of the assay, 96-well flat-bottom plates (Cellstar, catalog number 655180) were coated overnight at 4°C with 5 μg / mL anti-CD3 (clone OKT3, eBioscience, catalog number: 16-0037-85). The next day, the culture plates were washed twice with PBS, and 100 μL of T cell suspension (100,000 cells / well) was added to each well. The plates were incubated at 37°C, 5% CO2 for 3 days. The activated T cells were then harvested by gently pipetting up and down several times using a multichannel pipette. The cells were merged, mixed, and plated at 0.2×10 5 Up to 5×10 5 The cells were transferred to a U-bottom 96-well FACS assay plate (BD, catalog number 353910).

[0566] In order to carry out FACS analysis, cells are centrifuged at 300g for 3 minutes at 4°C. Supernatant is discarded by inverting the plate. Cells are washed by adding 200 μL ice-cold FACS buffer. Cells are centrifuged at 300g for 3 minutes at 4°C again, and supernatant is discarded as previously described. Each bispecific IgG, parent IgG or control IgG sample (8 step half log titrations starting from 20 μg / mL) of 40 μL is added, and cells are incubated on ice for 30 minutes in the dark. The cells are then washed twice, starting from directly adding 200 μL ice-cold FACS buffer, then centrifuged at 300g for 3 minutes at 4°C, and then supernatant is removed. Secondary antibody staining is carried out by adding 40 μL goat anti-human IgG-PE (3 μg / mL; Invitrogen, catalog number H10104) and incubating the plate on ice for 30 minutes in the dark. Cells were washed twice with 200 μL ice-cold FACS buffers again and resuspended in 50 to 200 μL FACS buffers. Cells were analyzed on a FACSCanto flow cytometer (Becton and Dickinson) under a high throughput sampler (HTS) setting. The combination of antibodies and cells was assessed by measuring the mean fluorescence intensity (MFI) of the stained cell colony. When the MFI was at least five times the MFI of the same cell colony stained with the negative control antibody, it was considered that the antibody was combined with its target.

[0567] result

[0568] The results of the FACS binding assays are provided in Fig.17 In. The data show the MFI detected by FACS on human activated T cells stained with increasing concentrations of IgG. The figure compares the binding of two bispecific IgGs and their parental IgGs, and shows that all specific IgGs bind to activated T cells. The maximum level of binding of the bispecific IgG is higher than the maximum level of binding of both the parental monospecific bivalent IgG and the positive control IgG.

[0569] Example 10

[0570] Anti-PD-L1 by SPR Determination of affinity of Fab arms for PD-L1

[0571] The affinity of the anti-PD-L1 Fab arms to PD-L1 was determined using surface plasmon resonance (SPR). To avoid avidity effects, affinity was measured in the context of a bispecific IgG with only one arm specificity for PD-L1 (i.e., monovalent format).

[0572] To determine the kinetics of binding of the anti-PD-L1 Fab arm to the antigen, surface plasmon resonance (SPR) using BIAcore T100 was used. Recombinant, purified, Fc-tagged human PD-L1 (R&D systems, catalog number 156-B7-100) was coupled to the flow cell (FC) 2 of the CM5 sensor chip at a level of approximately 200 resonance units (RU) using NHS / EDC chemistry at pH 5.0 (NaAc buffer), 2 μg / ml antigen concentration, and 10 μl / min flow (FC1 served as a blank for subtraction and was activated and then directly inactivated using ethanolamine). Then, bispecific IgG consisting of an anti-PD-L1 Fab arm and an irrelevant Fab arm was run on the FC1 and FC2 surfaces at different concentrations (100 nM and serial 2-fold dilutions in HBS, 6 dilutions) in a 30 μl / min kinetic run. The irrelevant Fab arm is specific for PD-1. Regeneration was performed using pulses of 50 mM HCl in water (15 μl at a flow rate of 10 μl / min).The obtained sensorgrams were evaluated using the BIAevaluation software and the kinetic association and dissociation rate constants were determined.

[0573] Several measurements were performed over several days on different surfaces of different sizes. The different measurements gave very similar results, underlining its validity. All measurements were performed at 25°C.

[0574] result

[0575] The results of affinity determination are provided in Table 7, which show that all tested anti-PD-L1 Fab arms have good affinity.

[0576] Embodiment 11

[0577] PD-L2 blockade via PD-1xPD-L1 antibodies

[0578] While therapeutic antibodies against PD-1 and PD-L1 are known to be effective in cancer treatment, the role of PD-L2 in anticancer immunity is currently unclear. PD-L1 blockade may promote tumor resistance to therapy by upregulating PD-L2. Since PD-L1 and PD-L2 interact with a large overlapping region of PD-1, it is expected that anti-PD-1 antibodies that block the interaction between PD-1 and PD-L1 will also block the interaction between PD-1 and PD-L2. Therefore, it was decided to determine whether several PD-1 / PD-L1 bispecific antibodies could block the PD-1 / PD-L2 pathway.

[0579] This blocking activity was tested in vitro in a physiologically relevant PD-1 / PD-L2 blocking reporter assay developed by Promega, which is based on a two-cell system: CHO cells expressing PD-L2 and a T cell receptor activator, and a Jurkat / NFAT-RE reporter cell line overexpressing PD-1. Jurkat T cells contain a luciferase reporter that can be activated through the NFAT (nuclear factor of activated T cells) pathway. The interaction of PD-1 with PD-L2 inhibits the activation of this pathway. However, blocking the interaction of PD-1 / PD-L2 with antibodies against PD-1 or PD-L2, but not PD-L1, activates the NFAT pathway. This means that the greater the degree of PD-1 / PD-L2 blockade, the greater the activation of the luciferase reporter.

[0580] Promega's thawing and use format was used to perform PD-1 / PD-L2 blocking reporter assays. Cells expressing PD-L2 (catalog number CS187127) were thawed in 14.5 ml cell recovery medium (DMEM / F12 containing 10% FBS). Next, 100 μl of cell suspension was added to the inner hole of two 96-well assay plates (Costar, catalog number 3917). The plate was incubated overnight at 37 ° C, 5% CO2 at 95% relative humidity. The next day, the culture medium was removed, and 40 μl of test antibodies in the assay medium (RPMI 1640 containing 4% FBS) of serial dilutions (starting concentration 25 μg / ml) were added to each well. Each plate contained a negative control (RSV G specific antibody PG2708) and a positive control (anti-PD-1 therapeutic antibody based on nivolumab, referred to herein as 5C4) that served as a reference control. Thaw PD-1 effector cells (Cat. No. CS187105) in 5.9 ml of assay medium and add 40 μl of cell suspension to each well. Incubate the plate at 37°C, 5% CO2, 95% relative humidity for 6 hours. Then add 80 μl of Bio-Glo reagent (Bio-Glo TM Luciferase Assay System, catalog number G7941, G7940) and the amount of luciferase activity was measured using a BioTek Synergy 2 multi-mode microplate reader. The fold induction was calculated as the luciferase activity after induction relative to the luciferase activity measured in the wells without antibody.

[0581] result

[0582] The two PD-1xPD-L1 bispecific IgGs tested were able to reduce the interaction between PD-1 and PD-L2 ( Fig.18). This offers the advantage that these antibodies counteract tumor resistance to treatment via the PD-1 / PD-L2 pathway.

[0583] The negative control anti-RSV G antibody PG2708 was unable to prevent the interaction between PD-1 and PD-L2.

[0584] Example 12

[0585] CD80 blockade via PD-1xPD-L1 antibodies

[0586] In recent years, the immunosuppressive PD-1 / PD-L1 pathway has been extensively studied, and therapeutic antibodies that block PD-1 or PD-L1 are effective anticancer treatments. It is also believed that immunosuppression is induced by the interaction between PD-1 and its alternative ligand PD-L2, and by the binding of PD-L1 to CD80 (B7-1). Since the CD80 binding site on PD-L1 is shown to overlap with the PD-1 binding site, most commercial anti-PD-L1 antibodies have been shown to block the interaction of PD-L1 with both PD-1 and CD80. However, a prior art antibody (MIH3) was shown to block PD-1: PD-L1 interaction, but not CD80: PD-L1 interaction (Butte et al., 2008). Therefore, we decided to test whether our bispecific antibody could block the interaction of PD-L1 with PD-1 and CD80.

[0587] In the current example, the blocking activity of the bispecific antibody MF7686xMF7703 was tested in a PD-L1 blocking ELISA, which also contained parental anti-PD-1 (PG7686) and anti-PD-L1 (PG7703) IgGs, which could not be tested directly because the bispecific antibody (MF7686xMF7703) was specific for both PD-1 and PD-L1. The ability of these parental IgGs to block the interaction of PD-L1 with PD-1 or CD80 was compared with the ability of the anti-PD-1 benchmark antibody 5C4 (based on nivolumab) and the anti-PD-L1 benchmark antibody YW243.55.S70 (based on atezolizumab) to block the interaction of PD-L1 with PD-1 or CD80. Anti-RSV G antibody PG2708 was used as a negative competition control on each plate.

[0588] For this ELISA, PD1-Fc (R&D systems; catalog number 1086-PD) or CD80-Fc (R&D systems; catalog number 140-B1) was coated to maxisorp plates at 1 and 3 μg / ml, respectively. The coated wells were blocked with 4% BSA in PBS. Thereafter, 0.55 μg / ml biotinylated PD-L1 (BPS bioscience; catalog number 71105) diluted in 2% BSA in PBS was added in the presence or absence of 0.08 to 10 μg / ml (final concentration in the plate) of IgG. Bound biotinylated PD-L1 was detected with HRP-conjugated streptavidin (BD bioscience: catalog number 554066) diluted 1:2000 in 2% BSA in PBS. After each incubation step, the ELISA plate was washed three times with PBS-T (PBS-0.05% v / v Tween 20). Bound streptavidin was visualized by TMB / H2O2 staining and the staining was quantified by measuring the optical density (OD) at 450 nm using a microplate reader.

[0589] result

[0590] In the PD-1 / PD-L1 competition assay, it is not possible to determine the ability of the bispecific antibody to block the PD-L1:PD-1 interaction ( Fig.19 , left). However, when tested in a bivalent IgG format, both arms of the bispecific antibody (MF7686 and MF7703) blocked the PD-L1:PD-1 interaction. The PD-L1 parent IgG PG7703 blocked the interaction more strongly than the benchmark PD-L1 antibody (YW243.55.S70). Anti-PD-1 IgG inhibited the interaction to a greater extent, with the PD-1 parent IgG PG7686 performing better than the PD-1 benchmark IgG (5C4).

[0591] In the CD80 competition assay, the tested bispecific antibody (MF7686xMF7703) blocked the interaction between CD80 and PD-L1. This provides the advantage that the antibody counteracts the immunosuppression through the interaction between PD-1 and CD80.

[0592] PD-1 IgG was unable to inhibit this interaction.

[0593] Embodiment 13

[0594] SEB assay: dual specificity compared to mixed parent

[0595] Functional comparison of PD-1xPD-L1 bispecific IgG and an equimolar mixture of two parental IgGs

[0596] SEB assay was performed to compare a mixture of bispecific antibodies and their parental IgG. Peripheral blood mononuclear cells (PBMCs) were stimulated with Staphylococcal enterotoxin B (SEB). SEB specifically activates T cells expressing Vβ3, 12, 14, 15, 17 and 20 T cell receptor chains, and the level of IL-2 released by the cells is an indicator of T cell activation.

[0597] In the current experiments, PBMCs from two donors were thawed, washed, counted, and resuspended in culture medium (RPMI1640 plus 10% heat-inactivated FBS) to a concentration of 2 × 10 6 In the presence of SEB (2000 ml), cells were seeded in a flat-bottom 96-well plate (2×10 5 Cells / well) were added, followed by 6 steps of 10-fold serial dilutions of antibody, starting from 20 μg / ml. Before collecting the supernatant, the cells were stimulated for 3 days at 37°C, 5% CO2 at 95% relative humidity. The plate was centrifuged at 350 g for 5 minutes, and 140 μl of supernatant was collected for IL-2 AlphaLISA (PerkinElmer catalog number AL221C), which was performed according to the manufacturer's instructions. The following bispecific antibodies were tested: PB16666 (MF7686xMF7703) and PB16672 (MF6974x7689).

[0598] result

[0599] Fig. 20 The results in show that the tested bispecific antibodies are able to activate T cells. Notably, when cells were incubated with the PD-1 / PD-L1 bispecific antibody, SEB-induced IL-2 production by PBMCs was higher than when cells were incubated with an equimolar mixture of their parental bivalent monospecific antibodies. Thus, the bispecific antibodies have a stronger T cell activation potential than an equimolar mixture of each of the tested parental bivalent monospecific antibodies.

[0600] Embodiment 14

[0601] PD-1xPD-L1 T cell response

[0602] In the SEB assay of the previous example, it was found that all tested bispecific PD-1xPD-L1 antibodies induced a stronger immune response (IL-2 production) than that of the parental IgG, and even a stronger response than that of an equimolar mixture of the parental IgG. These results suggest that in this assay, simultaneous blocking of both PD-1 and PD-L1 is more effective than blocking only one target in the PD-1 / PD-L1 pathway. The subsequent step is to determine whether the bispecific antibodies are also more effective than existing PD-1 or PD-L1 therapeutic antibodies in an antigen-specific CD4+T cell assay.

[0603] In the SEB assay, T cells are strongly activated by cross-linking of T cell receptors with MHC-II molecules present on antigen presenting cells. However, the number of T cells that specifically recognize a certain antigen is usually much lower than the number of T cells that respond to SEB. In order to more closely simulate the activation of antigen-specific cells, an antigen-specific CD4+T cell assay is used. In this assay, PBMCs are stimulated with a mixture of antigens to which the immune system of the donor individual normally responds. The activation of T cells is assessed by measuring the production of IL-2 and IFNγ.

[0604] To this end, PBMCs from three healthy donors were isolated by density gradient and cultured at 2 × 10 cells per well in 96-well plates. 5 Cells were stimulated with mixed antigens (influenza and tetanus toxoids) in the presence or absence of the test antibodies. Supernatants were harvested on day 5 and stored at -80 °C until analysis of cytokine production by Luminex assay. Antibodies were tested on a four-point dose-response curve (10, 100, 1000 and 10000 ng / ml). For each donor, the effect of the bispecific PD-1xPD-L1 antibody (MF7686xMF7703) was compared with that of a bivalent anti-PD-L1 antibody (YW243.55S70, based on atezolizumab) and a bivalent anti-PD-1 antibody (5C4, based on nivolumab) or a 1:1 mixture of these two control antibodies. The negative control was the anti-RSV-G antibody PG2708p217.

[0605] result

[0606] like Fig.21As shown in , the bispecific PD-1xPD-L1 antibody induced IL-2 and IFNγ levels that were higher than those induced by the benchmark control antibody. Notably, the bispecific antibody also induced higher IL-2 and IFNγ levels, especially at lower concentrations, than the levels induced by the combination of the two benchmark antibodies. These results suggest that the bispecific PD-1xPD-L1 antibody is more effective than existing PD-1 or PD-L1 therapeutic antibodies or mixtures thereof in enhancing antigen-driven cytokine release by CD4+ T cells.

[0607] Embodiment 15

[0608] PD-1xPD-L1 mixed lymphocyte reaction

[0609] Bispecific PD-1xPD-L1 antibody enhances IFNγ production by T cells in mixed lymphocyte reactions

[0610] Mixed lymphocyte reaction (MLR) assays are often used to understand the effects of antibodies on T cell activation and proliferation. Such assays help understand whether such compounds affect the potential of T cells to mount such a response in the tumor microenvironment.

[0611] Here, we used an allogeneic MLR protocol with immature DCs to determine that the bispecific PD-1xPD-L1 antibody enhanced IFNγ production by T cells compared to that of a benchmark reference antibody. T cell responsiveness was quantified by measuring IFNγ levels in the culture supernatant.

[0612] To this end, human peripheral blood mononuclear cells (PBMCs) from healthy donors were prepared from buffy coats. Immature monocyte-derived dendritic cells (Mo-DCs) were prepared by separating CD14+ cells (EasySepStemcell, lot number 16C69672) using magnetic activated cell sorting (MACS) and culturing them in differentiation medium for 7 days. On the day of need, T cell isolation kit (EasySep Stemcell, lot number 16D70573) was used to prepare responding T cells from different donors for Mo-DCs from refrigerated PBMCs to obtain uncontacted T cells. Six separate MLRs were performed to provide biological replication.

[0613] For the assay, 1×10 4 Immature Mo-DCs and 1×10 5T cells were co-cultured for 4 days. The cultures were performed in triplicate. Supernatants were collected at the end of the culture period and assessed for IFNγ by ELISA (R&D BioTechne, lot number 342687) according to the manufacturer's instructions and the plate was read at 450 nm.

[0614] result

[0615] Fig. 22 The results of this assay are shown, illustrating the ability of the bispecific PD-1xPD-L1 antibody to enhance IFNγ production by T cells compared to IFNγ production by T cells of the bivalent anti-PD-L1 antibody (YW243.55.S70) and the bivalent anti-PD-1 antibody (5C4) antibodies. This increase in responsiveness can also be seen in the presence of the benchmark control antibody, but to a lesser extent. Therefore, compared with the benchmark antibodies YW243.55.S70 (based on atezolizumab) and 5C4 (based on nivolumab), the tested bispecific antibodies have a stronger potential for T cell activation. These results indicate that bispecific antibodies will increase the potential of T cells to establish an immune response in the tumor microenvironment, and the effect of bispecific antibodies will be more pronounced than the benchmark antibodies.

[0616] Example 16

[0617] Effect of bispecific PD-1xPD-L1 antibody on proliferation of tumor-infiltrating T cells

[0618] To test our bispecific antibodies in a tumor-relevant setting, we used a recently developed ex vivo assay based on T cells isolated from patient tumor material. Zhou et al. have developed methods to obtain fresh tumor material from patients with hepatocellular carcinoma (HCC) and isolate tumor-infiltrating cells (myeloid cells and lymphocytes), providing a method to test the effects of antibodies targeting immune checkpoint inhibitors on the function of t...

Claims

1. An antibody or a variant of the antibody that maintains the binding specificity of the antibody, comprising: - a variable domain that can bind to the extracellular part of a protein of the CD28 family (the first membrane protein), and - a variable domain that can bind to the extracellular part of a protein of the B7 family (second membrane protein); wherein the first and second membrane proteins are binding partners (i.e., a ligand and receptor pair), and wherein binding of the variable domain that can bind to the extracellular portion of the first membrane protein blocks binding of the first membrane protein to the second membrane protein, and / or binding of the variable domain that can bind to the extracellular portion of the second membrane protein blocks binding of the first membrane protein to the second membrane protein.

2. The antibody or variant thereof of claim 1, wherein the binding of the antibody or variant thereof reduces the inhibitory activity on the binding of the receptor-ligand pair.

3. The antibody or variant thereof according to claim 1 or claim 2, comprising: - a variable domain that can bind to the extracellular portion of PD-1, CTLA-4, BTLA or TMIGD2; and - a variable domain that can bind to the extracellular portion of PD-L1, PD-L2, CD80, CD86, B7-H4, TNFRSF14 or B7-H7.

4. The antibody or variant thereof according to claim 3, comprising: - a variable domain that can bind to the extracellular portion of PD-1; and - A variable domain that can bind to the extracellular portion of PD-L1 or PD-L2.

5. The antibody or variant thereof of claim 3 or claim 4, wherein binding of the variable domain that binds to PD-1 blocks binding of PD-1 to PD-L1, and / or binding of the variable domain that binds to PD-L1 blocks binding of PD-L1 to PD-1.

6. The antibody or variant thereof of claim 5, wherein binding of the variable domain that binds to PD-1 blocks binding of PD-1 to PD-L1, and binding of the variable domain that binds to PD-L1 blocks binding of PD-L1 to PD-1.

7. The antibody or variant thereof according to any one of claims 1 to 6, comprising a variable domain that can bind to the extracellular portion of PD-1, wherein the binding of the variable domain to PD-1 blocks the binding of PD-1 to PD-L2.

8. The antibody or variant thereof according to any one of claims 1 to 7, comprising a variable domain that can bind to the extracellular portion of PD-L1, wherein the binding of the variable domain to PD-L1 blocks the binding of PD-L1 to CD80.

9. The antibody or variant thereof according to any one of claims 1 to 8, comprising a variable domain that can bind to the extracellular portion of PD-1 and a variable domain that can bind to the extracellular portion of PD-L1, and having a stronger CD4+T cell activation potential in a Staphylococcal enterotoxin B (SEB) assay compared to an equimolar mixture of: - a bivalent monospecific antibody comprising two of said variable domains that bind to PD-1, and - a bivalent monospecific antibody comprising two of said variable domains binding to PD-L1.

10. The antibody or variant thereof according to any one of claims 1 to 9, comprising a variable domain that can bind to the extracellular portion of PD-1 and a variable domain that can bind to the extracellular portion of PD-L1, and the antibody or variant thereof can activate T cells more strongly in an antigen-specific CD4+T cell assay compared to the benchmark antibody 5C4 or the benchmark antibody YW243.55.S70 or the combination of the benchmark antibody 5C4 and YW243.55.S70.

11. The antibody or variant thereof according to any one of claims 1 to 10, comprising a variable domain that can bind to the extracellular portion of PD-1 and a variable domain that can bind to the extracellular portion of PD-L1, and the antibody or variant thereof has a stronger CD4+T cell activation potential in a mixed lymphocyte reaction (MLR) assay than the benchmark antibody 5C4 or the benchmark antibody YW243.55.S70.

12. The antibody or variant thereof of any one of claims 1 to 11, comprising a variable domain having binding affinity for the extracellular portion of PD-L1, wherein the equilibrium dissociation constant (K) is 0.001 % as measured by SPR. D ) is lower than or equal to 4.27 nM, preferably lower than or equal to 1.31 nM, preferably lower than or equal to 1.27 nM.

13. The antibody or variant thereof according to any one of claims 1 to 12, comprising a variable domain that can bind to the extracellular portion of PD-1 and a variable domain that can bind to the extracellular portion of PD-L1, and the antibody or variant thereof can enhance the proliferation of CD4+ and / or CD8+ tumor-infiltrating T cells.

14. The antibody or variant thereof according to any one of claims 1 to 12, comprising a variable domain that can bind to the extracellular portion of PD-1 and a variable domain that can bind to the extracellular portion of PD-L1, and the antibody or variant thereof can induce a stronger T cell-mediated anti-tumor response in vivo compared to the combination of the benchmark antibodies MK-3475 and YW243.55.S70.

15. The antibody or variant thereof of any one of claims 3 to 14, wherein the variable domain that binds the extracellular portion of PD-1 is defined as a variable domain that, when in a bivalent monospecific antibody format comprising the variable domain that binds PD-1, inhibits PD-1 / PD-L1-mediated inhibition of T cell receptor-mediated activation of Jurkat cells in the range of 20% to 150% compared to the inhibition obtained with the antibody nivolumab on Jurkat cells.

16. The antibody or variant thereof of any one of claims 3 to 15, wherein the variable domain that binds the extracellular portion of PD-L1 is defined as a variable domain that, when in a bispecific antibody with a second variable domain that binds an unrelated antigen, e.g., tetanus toxoid, provides the bispecific antibody with a Kd for PD-L1 binding of 0.1 to 14 nM (as measured by biacore).

17. The antibody or variant thereof of any one of claims 1 to 16, comprising a variable domain that can bind to the extracellular portion of PD-1 and comprising a heavy chain variable region having a CDR3 region, wherein the CDR3 region comprises the amino acid sequence of the CDR3 of the heavy chain variable region of one of the following VHs shown in Figures 3 and / or 13: MF6076; MF6236; MF6256; MF6226; MF6930; MF6932; MF6935; MF6936; MF6972; MF6974; MF6982; MF6929; MF7699; MF7698; MF7687; MF7686; MF7685; or MF7684.

18. The antibody or variant thereof of any one of claims 1 to 17, wherein the variable domain that binds to PD-1 comprises a heavy chain variable region, the heavy chain variable region comprising a heavy chain variable region having CDR1, CDR2 and CDR3 regions, the CDR1, CDR2 and CDR3 regions comprising the amino acid sequence of CDR1, CDR2 and CDR3 of the heavy chain variable region of one of the following VHs shown in Figures 3 and / or 13: MF6076; MF6236; MF6256; MF6226; MF6930; MF6932; MF6935; MF6936; MF6972; MF6974; MF6982; MF6929; MF7699; MF7698; MF7687; MF7686; MF7685; or MF7684.

19. The antibody or variant thereof according to any one of claims 1 to 18, comprising a variable domain that can bind to the extracellular part of PD-1 and comprising a heavy chain variable region, wherein the heavy chain variable region comprises an amino acid sequence of a heavy chain variable region shown below with up to 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the amino acid sequence of VH shown in the specified MF: MF6076; MF6236; MF6256; MF6226; MF6930; MF6932; MF6935; MF6936; MF6972; MF6974; MF6982; MF6929; MF7699; MF7698; MF7687; MF7686; MF7685; or MF7684 (as shown in Figure 3 and / or Figure 13).

20. The antibody or variant thereof of any one of claims 1 to 19, comprising a variable domain that can bind to the extracellular portion of PD-L1 and comprising a heavy chain variable region having a CDR3 region, wherein the CDR3 region comprises the amino acid sequence of the CDR3 region of the heavy chain variable region shown in Figure 3 and / or Figure 13 as follows: MF5359; MF5361; MF5377; MF5382; MF5424; MF5426; MF5439; MF5442; MF5553; MF5557; MF5561; MF5576; MF5594; MF5708; MF5442; MF7691; MF7690; MF7689; MF7688; MF7700; MF7701; MF7703; MF7694; MF7693; MF7692; MF7697; MF7696; or MF7695.

21. The antibody or variant thereof of any one of claims 1 to 20, comprising a variable domain that can bind to the extracellular portion of PD-L1 and comprising a heavy chain variable region having CDR1, CDR2 and CDR3 regions, wherein the CDR1, CDR2 and CDR3 regions comprise the amino acid sequences of CDR1, CDR2 and CDR3 of the heavy chain variable region of one of the following VHs shown in Figures 3 and / or 13: MF5359; MF5361; MF5377; MF5380; MF5390; MF5406; MF5410; MF5421; MF5430; MF5440 382;MF5424;MF5426;MF5439;MF5442;MF5553;MF5557;MF5561;MF5576;MF5594;MF5708;MF5442;MF7691;MF7690;MF7689;MF7688;MF7700;MF7701;MF7703;MF7694;MF7693;MF7692;MF7697;MF7696; or MF7695.

22. The antibody or variant thereof according to any one of claims 1 to 21, comprising a variable domain that can bind to the extracellular portion of PD-L1 and comprising a heavy chain variable region, the heavy chain variable region comprising the amino acid sequence of the heavy chain variable region shown below (of Figure 3 and / or Figure 13) having at most 15, preferably 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10, and preferably having 0, 1, 2, 3, 4 or 5 amino acid insertions, deletions, substitutions or a combination thereof relative to the amino acid sequence of a specified MF: MF5359; MF5361;MF5377;MF5382;MF5424;MF5426;MF5439;MF5442;MF5553;MF5557;MF5561;MF5576;MF5594;MF5708;MF5442;MF7691;MF7690;MF7689;MF7688;MF7700;MF7701;MF7703;MF7694;MF7693;MF7692;MF7697;MF7696; or MF7695.

23. The antibody or variant thereof according to any one of claims 1 to 22, comprising a variable domain that can bind to the extracellular portion of PD-1, and wherein the variable domain that can bind to the extracellular portion of PD-1 comprises: - a heavy chain CDR3 sequence that is identical to the heavy chain CDR3 sequence of the variable region selected from MF6974, MF6076 and MF7686; or - a heavy chain CDR3 sequence that deviates by no more than three, preferably by no more than two, more preferably by no more than one amino acid from the heavy chain CDR3 sequence of MF6974 or MF6076 or MF7686.

24. The antibody or variant thereof according to any one of claims 1 to 23, comprising a variable domain that can bind to the extracellular portion of PD-1, and wherein the variable domain that can bind to the extracellular portion of PD-1 comprises: - a heavy chain CDR1, CDR2 and CDR3 sequence that is identical to the heavy chain CDR1, CDR2 and CDR3 sequence of the variable region selected from MF6974, MF6076 and MF7686; or - a heavy chain CDR1 , CDR2 and CDR3 sequence that deviates by no more than three, preferably by no more than two, more preferably by no more than one amino acid from the heavy chain CDR1 , CDR2 and CDR3 sequence of MF6974 or MF6076 or MF7686.

25. The antibody or variant thereof according to any one of claims 1 to 24, comprising a variable domain that can bind to the extracellular portion of PD-1, and wherein the variable domain that can bind to the extracellular portion of PD-1 comprises: - a heavy chain variable region comprising the amino acid sequence of MF6974 or MF6076 or MF7686, or - A heavy chain variable region having a sequence that is at least 80% identical to the amino acid sequence of MF6974 or MF6076 or MF7686.

26. The antibody or variant thereof according to any one of claims 1 to 25, comprising a variable domain that can bind to the extracellular portion of PD-L1, and wherein the variable domain that can bind to the extracellular portion of PD-L1 comprises: - a heavy chain CDR3 sequence that is identical to the heavy chain CDR3 sequence of the variable region selected from MF7689 and MF7703; or - a heavy chain CDR3 sequence that deviates from the heavy chain CDR3 sequence of MF7689 or MF7703 by no more than three, preferably by no more than two, more preferably by no more than one amino acid.

27. The antibody or variant thereof according to any one of claims 1 to 26, comprising a variable domain that can bind to the extracellular portion of PD-L1, and wherein the variable domain that can bind to the extracellular portion of PD-L1 comprises: - a heavy chain CDR1, CDR2 and CDR3 sequence that is identical to the heavy chain CDR1, CDR2 and CDR3 sequence of the variable region selected from MF7689 and MF7703; or - A heavy chain CDR1 and CDR2 and CDR3 sequence that deviates from the heavy chain CDR1 and CDR2 and CDR3 sequence of MF7689 or MF7703 by no more than three, preferably by no more than two, more preferably by no more than one amino acid.

28. The antibody or variant thereof according to any one of claims 1 to 27, comprising a variable domain that can bind to the extracellular portion of PD-L1, and wherein the variable domain that can bind to the extracellular portion of PD-L1 comprises: - a heavy chain variable region comprising the amino acid sequence of MF7689 or MF7703, or - A heavy chain variable region having a sequence that is at least 80% identical to the amino acid sequence of MF7689 or MF7703.

29. The antibody or variant thereof of any one of claims 1 to 28, comprising a light chain variable region having the CDR1, CDR2 and CDR3 sequences of the light chain variable region as shown in Figure 1B, or having CDR1, CDR2 and CDR3 sequences that deviate from the light chain variable region CDR1, CDR2 and CDR3 sequences as shown in Figure 1B by no more than three, preferably no more than two, and more preferably no more than one amino acid.

30. The antibody or variant thereof of any one of claims 1 to 29, comprising a light chain variable region having a sequence that is at least 80% identical to the amino acid sequence as shown in Figure 1B.

31. The antibody or variant thereof of any one of claims 1 to 30, wherein the antigen binding site of the antibody or variant thereof consists of one immunoglobulin variable domain that can bind to the extracellular portion of PD-1 and one immunoglobulin variable domain that can bind to the extracellular portion of PD-L1 or PD-L2.

32. The antibody of any one of claims 1 to 31, wherein the antibody is a full length bispecific antibody.

33. The antibody of any one of claims 1 to 32, wherein the antibody is an IgG.

34. A method of inhibiting a biological activity in a first or second cell mediated by the binding of a CD28 family member (first membrane protein) on the first cell to a B7 family member (second membrane protein) on the second cell, wherein the first and second membrane proteins are binding partners (i.e., a ligand and receptor pair), the method comprising: - providing an antibody or a variant of the antibody that maintains the binding specificity of the antibody to a system comprising the first and second cells, wherein the antibody or the variant of the antibody comprises a variable domain that can bind to the extracellular part of the first membrane protein and a variable domain that can bind to the extracellular part of the second membrane protein; as well as - incubating the system in the absence of the antibody or variant thereof under conditions that allow the first or second cell to express the biological activity mediated by the binding of the first membrane protein to the second membrane protein; - wherein binding of the variable domain that can bind to the extracellular part of the first membrane protein blocks binding of the first membrane protein to the second membrane protein, and / or binding of the variable domain that can bind to the extracellular part of the second membrane protein blocks binding of the first membrane protein to the second membrane protein.

35. The method of claim 34, wherein binding of the antibody or variant thereof reduces inhibitory activity on binding of the receptor-ligand pair in the first cell.

36. The method of claim 34 or claim 35, wherein the first membrane protein is programmed cell death 1 protein (PD-1); cytotoxic T lymphocyte-associated protein 4 (CTLA-4); B and T lymphocyte attenuator (BTLA); or transmembrane and immunoglobulin domain-containing protein 2 (TMIGD2).

37. The method of any one of claims 34 to 36, wherein the second membrane protein is programmed cell death 1 ligand 1 protein (PD-L1); programmed cell death 1 ligand 2 protein (PD-L2); CD80; CD86; B7-H4; TNFRSF14; or B7-H7.

38. The method of any one of claims 34 to 37, wherein the antibody or variant thereof is the antibody or variant thereof of any one of claims 1 to 33.

39. A method for producing the antibody or variant thereof according to any one of claims 1 to 31 from a single cell, wherein the antibody or variant thereof comprises two CH3 domains capable of forming an interface, the method comprising providing: - a cell, the cell having a) a first nucleic acid molecule encoding an IgG heavy chain, the IgG heavy chain specifically recognizing the extracellular portion of a protein of the CD28 family, preferably PD-1, and comprising a first CH3 domain, and b) a second nucleic acid sequence encoding an IgG heavy chain, the IgG heavy chain specifically recognizing the extracellular portion of a protein of the B7 family, preferably PD-L1, and comprising a second CH3 domain, wherein the nucleic acid sequence has a mutation for preferential pairing of the first and second CH3 domains, the method further comprising the following steps: The cells are cultured and the nucleic acid sequence is allowed to express and the antibody or variant thereof is harvested from the culture.

40. The method of claim 39, wherein the cells have a third nucleic acid sequence encoding a common light chain, preferably a rearranged germline human kappa light chain IgVκ1-39*01 / IGJκ1*01.

41. The method of claim 39 or 40, wherein the first CH3 domain comprises the amino acid substitutions L351K and T366K (numbered according to EU numbering), and wherein the second CH3 domain comprises the amino acid substitutions L351D and L368E (numbered according to EU numbering), the method further comprising the steps of: The cells are cultured and the nucleic acid sequence is allowed to express and the antibody or variant thereof is harvested from the culture.

42. The antibody or variant thereof as claimed in any one of claims 1 to 33 for use as a medicament.

43. The antibody or variant thereof of any one of claims 1 to 33 for use in a method of treating cancer or a pathogen infection, preferably a viral or parasitic infection.

44. A composition or kit comprising at least one antibody or variant thereof according to any one of claims 1 to 33.

45. A pharmaceutical composition comprising at least one antibody or variant thereof according to any one of claims 1 to 33 and a pharmaceutically acceptable carrier, diluent or excipient.

46. ​​A nucleic acid molecule having a length of at least 15 nucleotides encoding at least one CDR region of the antibody or variant of any one of claims 1 to 33.

47. A nucleic acid molecule according to claim 46, encoding at least the heavy chain variable region of the antibody or variant according to any one of claims 1 to 33.

48. A nucleic acid molecule according to claim 46 or 47, encoding a heavy chain variable region as shown in Figures 3 and / or 13.

49. A nucleic acid molecule encoding the antibody or variant of any one of claims 1 to 33.

50. A vector comprising a nucleic acid molecule according to any one of claims 46 to 49.

51. An isolated or recombinant cell or non-human animal comprising a nucleic acid molecule according to any one of claims 46 to 49 or a vector according to claim 50.

52. A method for treating cancer or pathogen infection, comprising administering to a subject in need thereof a therapeutically effective amount of an antibody or variant according to any one of claims 1 to 33, or a composition according to claim 44 or 45, or a nucleic acid molecule according to any one of claims 46 to 49, or a vector according to claim 50.

Images