EA047726B1 - MULTISPECIFIC PD-1 BINDING MOLECULES AND METHODS OF THEIR APPLICATION - Google Patents

Description

Cross-reference to related applications

This patent application claims priority to U.S. Patent Application Nos. 62/198,867 (filed July 30, 2015; pending), 62/239,559 (filed October 9, 2015; pending), 62/255,140 (filed November 13, 2015; pending), and 62/322,974 (filed April 15, 2016; pending), the entire contents of which are incorporated herein by reference.

Link to sequence listing

The present invention includes one or more sequence listings in accordance with 37 CFR 1.821 et seq., which are disclosed on a machine-readable medium (file entitled: 1301_0122PCT_Sequence_Listing_ST25.txt, created July 1, 2016 and has a size of 282,789 bytes), which file is incorporated herein by reference in its entirety.

Field of technology

The present invention relates to PD-1-binding molecules that comprise a PD-1-binding domain of selected anti-PD-1 antibodies capable of binding to cynomolgus macaque PD-1 and human PD1: anti-PD-1 monoclonal antibody 1 (MAB 1), anti-PD-1 MAT 2, anti-PD-1 MAT 3, anti-PD-1 MAT 4, anti-PD-1 MAT 5, anti-PD-1 MAT 6, anti-PD-1 MAT 7, anti-PD-1 MAT 8, anti-PD-1 MAT 9, anti-PD-1 MAT 10, anti-PD-1 MAT 11, anti-PD-1 MAT 12, anti-PD-1 MAT 13, anti-PD-1 MAT 14, or anti-PD-1 MAT 15. In particular, the present invention provides PD-1-binding molecules that are humanized or chimeric versions of said antibodies or comprise PD-1-binding fragments of said PD-1 antibodies (in particular, immunoconjugates, diabodies, BiTEs, bispecific antibodies, etc.). In particular, the present invention provides PD-1-binding molecules that are additionally capable of binding to an epitope of a molecule involved in the regulation of an immune checkpoint that is present on the surface of an immune cell. The present invention also provides methods of using said PD-1-binding molecules to detect PD-1 or to stimulate an immune response. The present invention also provides methods of combination therapy in which a PD-1-binding molecule that comprises one or more PD-1-binding domains of said selected PD-1 antibodies is administered in combination with one or more additional molecules that effectively stimulate an immune response and/or in combination with one or more additional molecules that specifically bind to a cancer antigen.

State of the art

I. Cell-mediated immune responses.

The immune system of humans and other mammals provides protection against infections and diseases. This protection is mediated by both humoral and cell-mediated immune responses. The humoral response results in the production of antibodies and other biological molecules that are able to recognize and neutralize foreign targets (antigens). In contrast, the cell-mediated immune response involves the activation of macrophages, natural killer (NK) cells, and antigen-specific cytotoxic T lymphocytes by T cells, as well as the release of various cytokines in response to antigen recognition (Dong, C. et al. (2003) Immune Regulation by Novel Costimulatory Molecules, Immunolog. Res. 28(1):39–48). The ability of T cells to optimally mediate an immune response to antigen requires two distinct signaling interactions (Viglietta, V. et al. (2007) Modulating Co-Stimulation, Neurotherapeutics 4:666-675; Korman, A.J. et al. (2007) Checkpoint Blockade in Cancer Immunotherapy, Adv. Immunol. 90:297-339). First, antigen that has been displayed on the surface of antigen-presenting cells (APCs) must be presented to antigen-specific naive CD4+ T cells. The presentation process involves signaling via the T cell receptor (TCR), directing the T cell to initiate an immune response that is specific for the presented antigen. Second, a series of costimulatory and inhibitory signals mediated by interactions between APCs and various T cell surface molecules trigger first T cell activation and proliferation and, ultimately, their inhibition. The first signal therefore confers specificity to the immune response, whereas the second signal serves to determine the nature, magnitude, and duration of the response.

The immune system is tightly controlled by costimulatory and coinhibitory ligands and receptors. These molecules provide a second signal for T cell activation and ensure a balanced set of positive and negative signals to maximize immune responses to infection while minimizing autoimmune responses (Wang, L. et al. (March 7, 2011) VISTA, A Novel Mouse Ig Superfamily Ligand That Negatively Regulates T-Cell Responses, J. Exp. Med. 10.1084/jem.20100619:1-16; Lepenies, B. et al. (2008) The Role Of Negative Costimulators During Parasitic Infections, Endocrine, Metabolic & Immune Disorders - Drug Targets 8:279-288). Of particular importance is the binding of the ligands B7.1 (CD80) and B7.2 (CD86) of the antigen-presenting cell, as well as CD28 and CTLA-4 receptors of the CD4+ T lymphocyte (Sharpe, A.H. et al. (2002) The B7-CD28 Superfamily, Nature Rev. Immunol. 2:116-126; Dong, C et al. (2003) Immune Regulation by Novel Costimulatory

- 1 047726

Molecules, Immunolog. Res. 28(1):39–48; Lindley, P.S. et al. (2009) The Clinical Utility Of Inhibiting CD28Mediated Costimulation, Immunol. Rev. 229:307–321). Binding of B7.1 or B7.2 to CD28 stimulates T cell activation; binding of B7.1 or B7.2 to CTLA-4 inhibits such activation (Dong, S. et al. (2003) Immune Regulation by Novel Costimulatory Molecules, Immunolog. Res. 28(1):39-48; Lindley, P.S. et al. (2009) The Clinical Utility Of Inhibiting CD28-Mediated Costimulation, Immunol. Rev. 229:3 07-321; Greenwald, R. J. et al. (2005) The B7 Family Rev. 23:515-548). CD28 is constitutively expressed on the surface of T cells (Gross, J., et al. (1992) Identification And Distribution Of The Costimulatory Receptor CD28 In The Mouse, J. Immunol. 149:380-388), whereas CTLA-4 expression is rapidly upregulated following T cell activation (Linsley, P. et al. (1996) Intracellular Trafficking Of CTLA4 And Focal Localization Towards Sites Of TCR Engagement, Immunity 4:535-543). Because CTLA-4 is a higher affinity receptor (Sharpe, A.H. et al. (2002) The B7-CD28 Superfamily, Nature Rev. Immunol. 2:116-126), binding first initiates T cell proliferation (via CD28) and then inhibits it (via nascent CTLA-4 expression), thereby attenuating the effect if proliferation is no longer required.

Further studies of CD28 receptor ligands led to the identification and characterization of a set of related B7 molecules (the B7 superfamily) (Coyle, A. J. et al. (2001) The Expanding B7 Superfamily: Increasing Complexity In Costimulatory Signals Regulating T-Cell Function, Nature Immunol. 2(3):203–209; Sharpe, A. H. et al. (2002) The B7-CD28 Superfamily, Nature Rev. Immunol. 2:116–126; Greenwald, R. J. et al. (2005) The B7 Family Revisited, Ann. Rev. Immunol. 23:515–548; Collins, M. et al. (2005) The B7 Family Of Immune-Regulatory Ligands, Genome Biol. 6:223.1–223.7; Loke, P. et al. (2004) Emerging Mechanisms Of Immune Regulation: The Extended B7 Family And Regulatory T-Cells. Arthritis Res. Ther. 6:208-214; Korman, A.J. et al. (2007) Checkpoint Blockade in Cancer Immunotherapy, Adv. Immunol. 90:297-339; Flies, D.B. et al. (2007) The New B7s: Playing a Pivotal Role in Tumor Immunity, J. Immunother. 30(3):251-260; Agarwal, A. et al. (2008) The Role Of Positive Costimulatory Molecules In Transplantation And Tolerance, Curr. Opin. Organ Transplant. 13:366-372; Lenschow, D.J. et al. (1996) CD28/B7 System of T-Cell Costimulation, Ann. Rev. Immunol. 14:233-258; Wang, S. et al. (2004) Co-Signaling Molecules Of The B7-CD28 Family In Positive And Negative Regulation Of T Lymphocyte Responses, Microbes Infect. 6:759-766). Currently, there are several known members of the family: B7.1 (CD80), B7.2 (CD86), inducible costimulatory ligand (ICOS-L), programmed death ligand-1 (PD-L1; B7-H1), programmed death ligand-2 (PD-L2, B7-DC), B7-H3, B7-H4, and B7-H6 (Collins, M. et al. (2005) The B7 Family Of Immune-Regulatory Ligands, Genome Biol. 6:223.1–223.7; Flajnik, M. F. et al. (2012) Evolution Of The B7 Family: Co-Evolution Of B7H6 And Nkp30, Identification Of A New B7 Family Member, B7H7, And Of B7's Historical Relationship With The MHC, Immunogenetics epub doi.org/10.1007/s00251-012-0616-2).

II. Programmed death protein-1 (PD-1).

Programmed death protein-1 (PD-1, also known as CD279) is a 31 kDa type I membrane protein that is a member of the expanded CD28/CTLA-4 family of T cell regulators that generally negatively regulates immune responses (Ishida, Y. et al. (1992) Induced Expression Of PD-1, A Novel Member Of The Immunoglobulin Gene Superfamily, Upon Programmed Cell Death, EMBO J. 11:3887–3895; U.S. Patent Application Publication Nos. 2007/0202100; 2008/0311117; 2009/00110667, U.S. Patent Nos. 6,808,710, 7,101,500, 7,488,802, 7,635,757, 7722868, PCT publication WO 01/14557).

PD-1 is expressed on activated T cells, B cells, and monocytes (Agata, Y. et al. (1996) Expression Of The PD-1 Antigen On The Surface Of Stimulated Mouse T And B Lymphocytes, Int. Immunol. 8(5):765–772; Yamazaki, T. et al. (2002) Expression Of Programmed Death 1 Ligands By Murine T-Cells And APC, J. Immunol. 169:5538–5545) and at low levels on natural killer T cells (NK cells) (Nishimura, H. et al. (2000) Facilitation Of Beta Selection And Modification Of Positive Selection In The Thymus Of PD-1-Deficient Mice, J. Exp. Med. 191:891–898; Martin-Orozco, N. et al. (2007) Inhibitory Costimulation And Anti-Tumor Immunity, Semin. Cancer Biol. 17(4):288-298). The extracellular region of PD-1 consists of a single immunoglobulin (Ig) variable (V) domain that is 23% identical to the equivalent domain in CTLA-4 (Martin-Orozco, N. et al. (2007) Inhibitory Costimulation And Anti-Tumor Immunity, Semin. Cancer Biol. 17(4):288-298). Following the extracellular Ig variable domain is a transmembrane region and an intracellular tail. The intracellular tail contains two phosphorylation sites located in the immunoreceptor tyrosine-based inhibitory motif and the immunoreceptor tyrosine-based switch motif, suggesting that PD-1 negatively regulates TCR signaling (Ishida, Y. et al. (1992) Induced Expression Of PD-1, A Novel Member Of The Immunoglobulin Gene Superfamily, Upon Programmed Cell Death, EMBO J. 11:3887–3895; Blank, C. et al. (2006) Contribution Of The PD-L1/PD-1 Pathway To T-Cell Exhaustion: An Update On Implications For Chronic Infections And Tumor Evasion Cancer, Immunol. Immunother. 56(5):739–745). PD-1 inhibits the immune system by binding to B7-H1 and B7-DC (Flies, D.B. et al. (2007) The New B7s: Playing a Pivotal Role in Tumor Immunity, J. Immunother. 30(3):251–260; U.S. Patent Nos. 6,803,192, 7,794,710, U.S. Patent Application Publications Nos. 2005/0059051, 2009/0055944, 2009/0274666, 2009/0313687, PCT Publication Nos. WO 01/39722, WO 02/086083).

- 2 047726

B7-H1 and B7-DC are widely expressed on the surfaces of human and mouse tissues, such as the heart, placenta, muscle, fetal liver, spleen, lymph nodes, and thymus, as well as mouse liver, lung, kidney, pancreatic islet cells, and small intestine (Martin-Orozco, N. et al. (2007) Inhibitory Costimulation And Anti-Tumor Immunity, Semin. Cancer Biol. 17(4):288-298). In humans, B7-H1 protein expression has been detected in human endothelial cells (Chen, Y. et al. (2005) Expression of B7-H1 in Inflammatory Renal Tubular Epithelial Cells, Nephron. Exp. Nephrol. 102:e81-e92; de Haij, S. et al. (2005) Renal Tubular Epithelial Cells Modulate T-Cell Responses Via ICOS-L And B7-H1 Kidney Int. 68:2091-2102; Mazanet, M.M. et al. (2002) B7-H1 Is Expressed By Human Endothelial Cells And Suppresses T-Cell Cytokine Synthesis, J. Immunol. 169:3581-3588), myocardium (Brown, J.A. et al. (2003) Blockade Of Programmed Death-1 Ligands On Dendritic Cells Enhances T-Cell Activation And Cytokine Production, J. Immunol. 170:1257-1266), syncytiotrophoblasts (Petroff, M.G. et al. (2002) 57 Family Molecules: Novel Immunomodulators At The Maternal-Fetal Interface, Placenta 23:S95-S101). The molecules are also expressed by resident macrophages of some tissues, by macrophages that have been activated by interferon (IFN)--- or tumor necrosis factor (TNF)-a (Latchman, Y. et al. (2001) PD-L2 Is A Second Ligand For PD-1 And Inhibits T-Cell Activation, Nat. Immunol 2:261-268), and in tumors (Dong, H. (2003) B7-H1 Pathway And Its Role In The Evasion Of Tumor Immunity, J. Mol. Med. 81:281-287).

The interaction of B7-H1 and PD-1 has been found to send a significant negative costimulatory signal to T and B cells (Martin-Orozco, N. et al. (2007) Inhibitory Costimulation And AntiTumor Immunity, Semin. Cancer Biol. 17(4):288-298) and functions as a cell death inducer (Ishida, Y. et al. (1992) Induced Expression Of PD-1, A Novel Member Of The Immunoglobulin Gene Superfamily, Upon Programmed Cell Death, EMBO J. 11:3887-3895; Subudhi, SK et al. (2005) The Balance Of Immune Responses: Costimulation Verse Coinhibition, J. Molec. Med. 83:193-202). More specifically, it was found that interaction of the PD-1 receptor at low concentrations with the B7-H1 ligand results in the transmission of an inhibitory signal that potently inhibits the proliferation of antigen-specific CD8 ⁺ T cells; at higher concentrations, interactions with PD-1 do not inhibit T cell proliferation but significantly reduce the production of many cytokines (Sharpe, AH et al. (2002) The B7-CD28 Superfamily, Nature Rev. Immunol. 2:116-126). T cell proliferation and cytokine production by resting and previously activated CD4 and CD8 T cells, and even naive cord blood T cells, have been found to be inhibited by soluble B7-H1-Fc fusion proteins (Freeman, GJ et al. (2000) Engagement Of The PD-1 Immunoinhibitory Receptor By A Novel B7 Family Member Leads To Negative Regulation Of Lymphocyte Activation, J. Exp. Med. 192:1-9; Latchman, Y. et al. (2001) PD-L2 Is A Second Ligand For PD-1 And Inhibits T-Cell Activation, Nature Immunol. 2:261-268; Carter, L. et al. (2002) PD-1:PD-L Inhibitory Pathway Affects Both CD4(+) and CD8(+) T-cells And Is Overcome By IL-2, Eur. J. Immunol. 32(3):634-643; Sharpe, A.H. et al. (2002) The B7-CD28 Superfamily, Nature Rev. Immunol. 2:116-126).

The role of B7-H1 and PD-1 in inhibiting T cell activation and proliferation suggests that these biological molecules may serve as therapeutic targets for the treatment of inflammation and cancer. In this regard, the use of anti-PD-1 antibodies for the treatment of infections and tumors, as well as activating modulation of the adaptive immune response has been proposed (see US Patent Application Publications Nos. 2010/0040614; 2010/0028330; 2004/0241745; 2008/0311117; 2009/0217401, US Patents Nos. 7521051, 7563869, 7595048, PCT Publications WO 2004/056875, WO 2008/083174). Antibodies capable of specifically binding to PD-1 have been described by Agata, T. et al. (1996) Expression Of The PD-1 Antigen On The Surface Of Stimulated Mouse T And In Lymphocytes, Int. Immunol. 8(5):765-772; and Berger, R. et al. (2008) Phase I Safety And Pharmacokinetic Study Of CT-011, A Humanized Antibody Interacting With PD-1, In Patients With Advanced Hematologic Malignancies, Clin. Cancer Res. 14(10):3044-3051 (see also U.S. Patent Nos. 8,008,449 and 8,552,154, U.S. Patent Publication Nos. 2007/0166281; 2012/0114648; 2012/0114649; 2013/0017199; 2013/0230514 and 2014/0044738; and PCT Patent Publication Nos. WO 2003/099196, WO 2004/004771, WO 2004/056875, WO 2004/072286, WO 2006/121168, WO 2007/005874, WO 2008/083174, WO 2009/014708, WO 2009/073533; WO 2012/135408, WO 2012/145549 and WO 2013/014668).

However, despite all previous achievements, there remains a need for improved compositions that can more actively direct the body's immune system to attack cancer cells or pathogen-infected cells, especially at lower therapeutic concentrations. Although the adaptive immune system can be a potent defense mechanism against cancer and disease, it is often hampered by immune suppression mechanisms in the tumor microenvironment, such as PD-1 expression. In addition, coinhibitory molecules expressed by tumor cells, immune cells, and stromal cells in the tumor environment can preferentially attenuate T cell responses to cancer cells. Therefore, there remains a need for effective PD-1-binding molecules. In particular, there is a need for effective PD-1 binding molecules that have a desirable binding kinetic profile and are capable of acting as antagonists of the PD-1/PD-L1 axis by blocking the PD-1/PD-L1 interaction, which can provide an improved therapeutic effect in patients suffering from cancer or other diseases and conditions. The present invention is directed to these and other purposes.

- 3 047726

Brief description of the invention

The present invention relates to PD-1-binding molecules comprising a PD-1-binding domain from selected anti-PD-1 antibodies capable of binding to cynomolgus macaque PD-1 and human PD1: anti-PD-1 MAT 1, anti-PD-1 MAT 2, anti-PD-1 MAT 3, anti-PD-1 MAT 4, anti-PD-1 MAT 5, anti-PD-1 MAT 6, anti-PD-1 MAT 7, anti-PD-1 MAT 8, anti-PD-1 MAT 9, anti-PD-1 MAT 10, anti-PD-1 MAT 11, anti-PD-1 MAT 12, anti-PD-1 MAT 13, anti-PD-1 MAT 14, or anti-PD-1 MAT 15. The present invention provides, in particular, PD-1-binding molecules that are humanized or chimeric versions of said antibodies or that comprise PD-1-binding fragments of said antibodies to PD-1 (in particular, immunoconjugates, diabodies, BiTEs, bispecific antibodies, etc.). In particular, the present invention provides PD-1-binding molecules that are additionally capable of binding to an epitope of a molecule involved in the regulation of an immune checkpoint that is present on the surface of an immune cell. The present invention also provides methods of using said PD-1-binding molecules to detect PD-1 or to stimulate an immune response. The present invention also provides methods of combination therapy, wherein a PD-1 binding molecule comprising one or more PD-1 binding domains of said selected anti-PD-1 antibodies is administered in combination with one or more additional molecules that effectively stimulate an immune response and/or in combination with one or more additional molecules that specifically bind to a cancer antigen. In particular, the present invention provides a human PD-1 binding molecule that comprises three heavy chain complementarity determining region (CDR) domains (H) CDRH1, CDRH2 and CDRH3 and three light chain (L) CDR domains (L) CDR _L 1, CDR _L 2 and CDR _L 3, wherein:

(A) (1) the CDRH1 domain, the CDRH2 domain, and the CDRH3 domain are the CDR domains of the heavy chain of MAT α-PD-1 1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 71, SEQ ID NO: 72, and SEQ ID NO: 73; and (2) the CDRL1 domain, the CDRL2 domain, and the CDRL3 domain are the CDR domains of the light chain of MAT α-PD-1 1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 76, SEQ ID NO: 77, and SEQ ID NO: 78; or (B) (1) the CDR _H 1 domain, the CDR _H 2 domain, and the CDR _H 3 domain are the CDR domains of the heavy chain of MAT α-PD-1 2 and, respectively, comprise the amino acid sequences of SEQ ID NO: 85, SEQ ID NO: 86, and SEQ ID NO: 87; and (2) the CDR _L 1 domain, the CDR _L 2 domain, and the CDR _L 3 domain are the light chain CDR domains of MAT 2 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 90, SEQ ID NO: 91, and SEQ ID NO: 92; or (C) (1) the CDRH1 domain, the CDRH2 domain, and the CDRH3 domain are the heavy chain CDR domains of MAT 3 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 99, SEQ ID NO: 100, and SEQ ID NO: 101; and (2) the CDRL1 domain, the CDRL2 domain, and the CDRL3 domain are the light chain CDR domains of MAT 3 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 104, SEQ ID NO: 105, and SEQ ID NO: 106; or (D) (1) the CDR _H 1 domain, the CDR _H 2 domain, and the CDR _H 3 domain are the CDR domains of the heavy chain of MAT 4 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 109, SEQ ID NO: 110, and SEQ ID NO: 111; and (2) the CDRL1 domain, the CDRL2 domain, and the CDRL3 domain are the CDR domains of the light chain of MAT 4 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 114, SEQ ID NO: 115, and SEQ ID NO: 116; or (E) (1) the CDRH1 domain, the CDRH2 domain, and the CDRH3 domain are the CDR domains of the heavy chain of MAT 5 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 119, SEQ ID NO: 120, and SEQ ID NO: 121; and (2) the CDR _L 1 domain, the CDR _L 2 domain and the CDR _L 3 domain are the light chain CDR domains of MAT 5 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 124, SEQ ID NO: 125 and SEQ ID NO: 126; or (F) (1) the CDRH1 domain, the CDRH2 domain and the CDRH3 domain are the heavy chain CDR domains of MAT 6 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 129, SEQ ID NO: 130 and SEQ ID NO: 131; and (2) the CDRL1 domain, the CDRL2 domain and the CDRL3 domain are the light chain CDR domains of MAT 6 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 134, SEQ ID NO: 135 and SEQ ID NO: 136; or (G) (1) the CDRH1 domain, the CDRH2 domain, and the CDRH3 domain are the heavy chain CDR domains of MAT 7 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 139, SEQ ID NO: 140, and SEQ ID NO: 141; and (2) the CDR _L 1 domain, the CDR _L 2 domain, and the CDR _L 3 domain are the light chain CDR domains of MAT 7 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 144, SEQ ID NO: 145, and SEQ ID NO: 146;

- 4 047726

ID NO: 145 and SEQ ID NO: 146; or (H) (1) the CDRH1 domain, the CDRH2 domain and the CDRH3 domain are the CDR domains of the heavy chain of MAT 8 to PD-1 and, respectively, comprise the amino acid sequences of: SEQ ID NO: 161, SEQ ID NO: 162 and SEQ ID NO: 163; and (2) the CDRL1 domain, the CDRL2 domain and the CDRL3 domain are the CDR domains of the light chain of MAT 8 to PD-1 and, respectively, comprise the amino acid sequences of: SEQ ID NO: 166, SEQ ID NO: 167 and SEQ ID NO: 168; or (I) (1) the CDR _H 1 domain, the CDR _H 2 domain and the CDR _H 3 domain are the CDR domains of the heavy chain of MAT 9 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 171, SEQ ID NO: 172 and SEQ ID NO: 173; and (2) the CDRL1 domain, the CDRL2 domain and the CDRL3 domain are the CDR domains of the light chain of MAT 9 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 176, SEQ ID NO: 177 and SEQ ID NO: 178; or (J) (1) the CDRH1 domain, the CDRH2 domain and the CDRH3 domain are the CDR domains of the heavy chain of MAT 10 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 192, SEQ ID NO: 193 and SEQ ID NO: 194; and (2) the CDR _L 1 domain, the CDR _L 2 domain and the CDR _L 3 domain are the light chain CDR domains of MAT 10 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 197, SEQ ID NO: 198 and SEQ ID NO: 199; or (K) (1) the CDRH1 domain, the CDRH2 domain and the CDRH3 domain are the heavy chain CDR domains of MAT 11 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 202, SEQ ID NO: 203 and SEQ ID NO: 204; and (2) the CDRL1 domain, the CDRL2 domain and the CDRL3 domain are the light chain CDR domains of MAT 11 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 207, SEQ ID NO: 208 and SEQ ID NO: 209; or (L) (1) the CDRH1 domain, the CDRH2 domain and the CDRH3 domain are the CDR domains of the heavy chain of MAT 12 to PD-1 and, respectively, comprise the amino acid sequences of: SEQ ID NO: 212, SEQ ID NO: 213 and SEQ ID NO: 214; and (2) the CDR _L 1 domain, the CDR _L 2 domain and the CDR _L 3 domain are the CDR domains of the light chain of MAT 12 to PD-1 and, respectively, comprise the amino acid sequences of: SEQ ID NO: 217, SEQ ID NO: 218 and SEQ ID NO: 219 or (M) (1) the CDRH1 domain, the CDRH2 domain and the CDRH3 domain are the CDR domains of the heavy chain of MAT 13 to PD-1 and, respectively, comprise the amino acid sequences of: SEQ ID NO: 222, SEQ ID NO: 223 and SEQ ID NO: 224; and (2) the CDRL1 domain, the CDRL2 domain and the CDRL3 domain are the light chain CDR domains of MAT 13 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 227, SEQ ID NO: 228 and SEQ ID NO: 229; or (N) (1) the CDR _H 1 domain, the CDR _H 2 domain and the CDR _H 3 domain are the heavy chain CDR domains of MAT 14 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 232, SEQ ID NO: 233 and SEQ ID NO: 234; and (2) the CDR _L 1 domain, the CDR _L 2 domain and the CDR _L 3 domain are the light chain CDR domains of MAT 14 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 237, SEQ ID NO: 238 and SEQ ID NO: 239; or (O) (1) the CDRH1 domain, the CDRH2 domain, and the CDRH3 domain are the CDR domains of the heavy chain of MAT 15 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 242, SEQ ID NO: 243, and SEQ ID NO: 244; and (2) the CDRL1 domain, the CDRL2 domain, and the CDRL3 domain are the CDR domains of the light chain of MAT 15 to PD-1 and, respectively, comprise the amino acid sequences of SEQ ID NO: 247, SEQ ID NO: 248, and SEQ ID NO: 249; or (P) (1) the CDRH1 domain, the CDRH2 domain and the CDRH3 domain are the heavy chain CDR domains of MAT to human PD-1 (hPD-1) 7(1.2) and, respectively, comprise the amino acid sequences of SEQ ID NO: 139, SEQ ID NO: 140 and SEQ ID NO: 141; and (2) the CDR _L 1 domain, the CDR _L 2 domain and the CDR _L 3 domain are the light chain CDR domains of MAT to hPD-1 7(1.2) and, respectively, comprise the amino acid sequences of SEQ ID NO: 157, SEQ ID NO: 145 and SEQ ID NO: 146; or (Q) (1) the CDRH1 domain, the CDRH2 domain, and the CDRH3 domain are the heavy chain CDR domains of MAT to hPD-1 7 (1.3) and, respectively, comprise the amino acid sequences of SEQ ID NO: 139, SEQ ID NO: 140, and SEQ ID NO: 141; and (2) the CDRL1 domain, the CDRL2 domain, and the CDRL3 domain are the light chain CDR domains of MAT to hPD-1 7 (1.3) and, respectively, comprise the amino acid sequences of SEQ ID NO: 157, SEQ ID NO: 158, and SEQ ID NO: 145; or (R) (1) the CDR _H 1 domain, the CDR _H 2 domain, and the CDR _H 3 domain are the heavy chain CDR domains of MAT to hPD-1 7 (1.3)

- 5,047,726 pi mAb to hPD-1 9 (2.2) and, respectively, comprise the amino acid sequences: SEQ ID NO: 183, SEQ ID NO: 172 and SEQ ID NO: 173; and (2) the CDRL1 domain, the CDRL2 domain and the CDRL3 domain are the CDR domains of the light chain of mAb to hPD-1 9 (2.2) and, respectively, comprise the amino acid sequences: SEQ ID NO: 188, SEQ ID NO: 189 and SEQ ID NO: 178. According to the present invention, embodiments of all of these molecules binding human PD-1 are also provided, wherein said molecule is an antibody, and in particular said molecule is a chimeric antibody or a humanized antibody.

The present invention also provides embodiments of said human PD-1 binding molecules, wherein the heavy chain variable domain comprises the amino acid sequence set forth in SEQ ID NO: 79, SEQ ID NO: 93, SEQ ID NO: 147, SEQ ID NO: 149, SEQ ID NO: 179, SEQ ID NO: 181, or SEQ ID NO: 250.

The present invention also provides embodiments of said human PD-1 binding molecules, wherein the light chain variable domain comprises the amino acid sequence set forth in SEQ ID NO: 81, SEQ ID NO: 95, SEQ ID NO: 151, SEQ ID NO: 153, SEQ ID NO: 155, SEQ ID NO: 184, SEQ ID NO: 186, or SEQ ID NO: 251.

The present invention also provides an embodiment wherein the human PD-1 binding molecule is a bispecific binding molecule capable of simultaneously binding to human PD-1 and to a second epitope, and in particular provides an embodiment wherein the second epitope is an epitope of a molecule involved in the regulation of an immune checkpoint present on the surface of an immune cell (in particular where the second epitope is an epitope of B7-H3, B7-H4, BTLA, CD40, CD40L, CD47, CD70, CD80, CD86, CD94, CD137, CD137L, CD226, CTLA-4, galectin-9, GITR, GITRL, HHLA2, ICOS, ICOSL, KIR, LAG-3, LIGHT, MHC class I or II, NKG2a, NKG2d, OX40, OX40L, PD1H, PD-1, PD-L1, PD-L2, PVR, SIRPa, TCR, TIGIT, TIM-3, or VISTA, and most particularly if the second epitope is an epitope of CD137, CTLA-4, LAG-3, OX40, TIGIT, or TIM-3).

The present invention also provides embodiments in which the human PD-1 binding molecule is a bispecific molecule comprising a LAG-3 epitope binding site, in particular in which the LAG-3 epitope binding site comprises:

(A) (1) a CDRH1 domain, a CDRH2 domain, and a CDRH3 domain of the heavy chain variable region of MAT anti-LAG-3 1 comprising the amino acid sequences of SEQ ID NO: 42, SEQ ID NO: 43, and SEQ ID NO: 44, respectively; and (2) a CDRL1 domain, a CDRL2 domain, and a CDRL3 domain of the light chain variable region of MAT anti-LAG-3 1 comprising the amino acid sequences of SEQ ID NO: 46, SEQ ID NO: 47, and SEQ ID NO: 48, respectively; or (B) (1) a CDRH1 domain, a CDRH2 domain, and a CDRH3 domain of the human anti-LAG-3 (hLAG-3) MAT VH1 heavy chain variable region 1 comprising the amino acid sequences of SEQ ID NO: 42, SEQ ID NO: 43, and SEQ ID NO: 44, respectively; and (2) a CDR _L 1 domain, a CDR _L 2 domain, and a CDR _L 3 domain of the VL4 light chain variable region of MAT κ hLAG-3 1 comprising the amino acid sequences of SEQ ID NO: 55, SEQ ID NO: 47, and SEQ ID NO: 48, respectively; or (C) (1) a CDRH1 domain, a CDRH2 domain, and a CDRH3 domain of the heavy chain variable region of MAT κ LAG-3 6 comprising the amino acid sequences of SEQ ID NO: 57, SEQ ID NO: 58, and SEQ ID NO: 59, respectively; and (2) a CDRL1 domain, a CDRL2 domain, and a CDRL3 domain of the light chain variable region of MAT κ LAG-3 6 comprising the amino acid sequences of SEQ ID NO: 61, SEQ ID NO: 62, and SEQ ID NO: 63, respectively; or (D) (1) a CDRH1 domain, a CDRH2 domain, and a CDRH3 domain of the VH1 heavy chain variable region of MAT to hLAG-3 6 comprising the amino acid sequences of SEQ ID NO: 57, SEQ ID NO: 58, and SEQ ID NO: 59, respectively; and (2) a CDR _L 1 domain, a CDR _L 2 domain, and a CDR _L 3 domain of the light chain variable region of MAT to LAG-3 6 comprising the amino acid sequences of SEQ ID NO: 298, SEQ ID NO: 62, and SEQ ID NO: 63, respectively.

The present invention also provides embodiments of said human PD-1 binding molecules, wherein said molecule is a diabody, and in particular said diabody is a covalently linked complex that comprises two or three or four or five polypeptide chains. The present invention also provides embodiments of said human PD-1 binding molecules, wherein said molecule is a trivalent binding molecule, and in particular said trivalent binding molecule is a covalently linked complex that comprises three, four, five or more than five polypeptide chains. The present invention further provides embodiments of said human PD-1 binding molecules, wherein said molecule comprises

- 6 047726 Fc region. According to the present invention, embodiments of said molecules binding human PD-1 are further provided, wherein said molecule comprises an albumin binding domain, and in particular a deimmunized albumin binding domain.

The present invention also provides embodiments of all of the above human PD-1 binding molecules, wherein the above molecule comprises an Fc region, and wherein the Fc region is a variant Fc region that comprises one or more amino acid modifications that reduce the affinity of the variant Fc region for FcyR and/or increase the serum half-life, and more particularly wherein the above modifications comprise at least one amino acid substitution selected from the group consisting of:

(1) L234A; L235A;

(2) L234A and L235A;

(3) M252Y; M252Y and S254T;

(4) M252Y and T256E;

(5) M252Y, S254T and T256E; or (6) K288D and Н435К;

wherein the numbering is an index according to the EU numbering system described in Kabat. The present invention also provides embodiments in which any of the above-described PD-1 binding molecules is used to stimulate a T-cell mediated immune response. The present invention further provides embodiments in which any of the above-described PD-1 binding molecules is used to treat a disease or condition associated with a suppressed immune system, in particular cancer or an infection.

According to the present invention, in particular, a variant of use in the treatment or diagnosis, or prognosis of cancer is proposed, wherein the cancer is characterized by the presence of a cancer cell selected from the group consisting of a cell: an adrenal tumor, AIDS-associated cancer, alveolar soft tissue sarcoma, astrocytic tumor, bladder cancer, bone cancer, brain and spinal cord cancer, metastatic brain tumor, breast cancer, carotid artery tumor, cervical cancer, chondrosarcoma, chordoma, chromophobe renal cell sarcoma, clear cell sarcoma, colon cancer, colorectal cancer, cutaneous benign fibrous histiocytoma, desmoplastic small round cell tumor, ependymoma, Ewing's tumor, extraskeletal myxoid chondrosarcoma, imperfect bone fibrogenesis, fibrous dysplasia of bone, gallbladder or bile duct cancer, gastric cancer, gestational trophoblastic disease, germ cell tumor, head and neck cancer, hepatocellular carcinoma, islet cell tumor, Kaposi's sarcoma, renal cell cancer, leukemia, lipoma/benign lipomatous tumor, liposarcoma/malignant lipomatous tumor, liver cancer, lymphoma, lung cancer, medulloblastoma, melanoma, meningioma, multiple endocrine neoplasia, multiple myeloma, myelodysplastic syndrome, neuroblastoma, neuroendocrine tumor, ovarian cancer, pancreatic cancer, papillary thyroid carcinoma, parathyroid tumor, pediatric cancer, peripheral nerve sheath tumor, pheochromocytoma, pituitary tumor, prostate cancer, posterior uveal melanoma, rare hematological disorder, metastatic cancer kidney cancer, rhabdoid tumor, rhabdomyosarcoma, sarcoma, skin cancer, soft tissue sarcoma, squamous cell sarcoma, gastric cancer, synovial sarcoma, testicular cancer, thymic carcinoma, thymoma, metastatic thyroid cancer, and uterine cancer.

According to the present invention, a method is provided for use in the treatment or diagnosis, or prognosis of cancer, wherein the cancer is colorectal cancer, hepatocellular carcinoma, glioma, renal cancer, breast cancer, multiple myeloma, bladder cancer, neuroblastoma; sarcoma, non-Hodgkin's lymphoma, non-small cell lung cancer, ovarian cancer, pancreatic cancer, rectal cancer, acute myeloid leukemia (AML), chronic myelogenous leukemia (CML), B-lineage acute lymphoblastic leukemia (B-ALL), chronic lymphocytic leukemia (CLL), hairy cell leukemia (HCL), blastic plasmacytoid dendritic cell neoplasm (BPDCN), non-Hodgkin's lymphomas (NHL) including mantle cell leukemia (MCL) and small lymphocytic lymphoma (SLL), Hodgkin's lymphoma, systemic mastocytosis, or Burkitt's lymphoma.

The present invention also provides embodiments in which any of the above-described PD-1 binding molecules comprises a detectable label and is used to detect PD-1.

Brief description of the drawings

Fig. 1 is a schematic representation of a typical covalently linked diabody containing two epitope-binding sites consisting of two polypeptide chains, each containing a heterodimer-promoting domain with an E-helix or a K-helix. A cysteine residue may be present in the linker and/or in the heterodimer-promoting domain, as shown in Fig. 3B. The VL and VH domains, which recognize the same epitope, are shown using the same shading or filling.

Fig. 2 shows a schematic representation of a typical covalently linked dia molecule.

- 7 047726 body containing two epitope-binding sites consisting of two polypeptide chains, each containing a CH2 and a CH3 domain, such that the corresponding chains form all or part of an Fc region. The VL and VH domains that recognize the same epitope are shown using the same shading or filling.

Figs. 3A-3C are diagrams illustrating an exemplary tetravalent diabody comprising four epitope-binding sites consisting of two pairs of polypeptide chains (i.e., a total of four polypeptide chains). One polypeptide of each pair comprises a CH2 and a CH3 domain, such that the corresponding chains form all or part of an Fc region. The VL and VH domains that recognize the same epitope are shown using the same shading or filling. The two pairs of polypeptide chains may be the same. In embodiments of the present invention in which the VL and VH domains recognize different epitopes (as shown in Figs. 3A-3C), the resulting molecule has four epitope-binding sites and is bispecific and bivalent with respect to each epitope bound. In embodiments of the invention in which the VL and VH domains recognize the same epitope (e.g., the same VL domain CDRs and the same VH domain CDRs are used in both chains), the resulting molecule has four epitope binding sites and is monospecific and tetravalent for one epitope. Alternatively, the two pairs of polypeptides may be different. In embodiments of the invention in which the VL and VH domains of each pair of polypeptides recognize different epitopes (as shown in Figs. 3A-3C), the resulting molecule has four epitope binding sites and is tetraspecific and monovalent for each epitope bound. Fig. 3A shows a diabody comprising an Fc region that comprises a cysteine residue-containing heterodimer promoting domain peptide. Fig. 3B shows a diabody containing an Fc region that contains heterodimer-promoting domains with an E-helix and a K-helix containing a cysteine residue and a linker (optionally containing a cysteine residue). Fig. 3C shows a diabody containing an Fc region that contains CH1 and CL domains.

Figures 4A and 4B are schematic diagrams of a typical covalently linked diabody molecule containing two epitope-binding sites composed of three polypeptide chains. The two polypeptide chains contain CH2 and CH3 domains, such that the corresponding chains form all or part of an Fc region. The polypeptide chains containing the VL and VH domain additionally contain a domain that facilitates heterodimer formation. VL and VH domains that recognize the same epitope are shown using the same shading or filling.

Figure 5 shows a schematic of a typical covalently linked diabody molecule containing four epitope-binding sites composed of five polypeptide chains. Two of the polypeptide chains contain a CH2 and CH3 domain, such that the corresponding chains form an Fc region that includes all or part of the Fc region. The polypeptide chains containing the linked VL and VH domains additionally contain a domain that facilitates heterodimer formation. VL and VH domains that recognize the same epitope are shown using the same shading or filling.

Figs. 6A-6F are schematic representations of exemplary trivalent binding molecules comprising an Fc region containing three epitope-binding sites. Figs. 6A and 6B, respectively, are schematic representations of the domains of trivalent binding molecules comprising two diabody-specific binding domains and a Fab-type binding domain having different domain orientations in which the diabody-specific binding domains are N-terminal or C-terminal to the Fc region. The molecules of Figs. 6A and 6B contain four chains. Figs. 6C and 6D, respectively, are schematic representations of the domains of trivalent binding molecules comprising two diabody-specific binding domains N-terminal to the Fc region and a Fab-type binding domain in which the light chain and the heavy chain are linked by a polypeptide spacer or an scFv-type binding domain. The trivalent binding molecules in Figs. 6E and 6F, respectively, schematically illustrate the domains of trivalent binding molecules containing two diabody-specific binding domains C-terminal to the Fc region and a Fab-type binding domain or an scFv-type binding domain in which the diabody-specific binding domains reside. The trivalent binding molecules in Figs. 6C-6F contain three chains. The VL and VH domains, which recognize the same epitope, are shown using the same shading or filling.

Figures 7A-7D show that anti-PD-1 MAbs 1-15 bind to human PD-1. Binding curves for binding to shPD-1-His are shown in Figure 7A (anti-PD-1 MAb 1, anti-PD-1 MAb 2, anti-PD-1 MAb 4, and anti-PD-1 MAb 9), Figure 7B (anti-PD-1 MAb 5, anti-PD-1 MAb 6, and anti-PD-1 MAb 7), and Figure 7C (anti-PD-1 MAb 3, anti-PD-1 MAb 8, anti-PD-1 MAb 10, anti-PD-1 MAb 11, anti-PD-1 MAb 12, anti-PD-1 MAb 13, anti-PD-1 MAb 14, and anti-PD-1 MAb 15). Binding curves for binding to human shPD-1-Fc are shown in Fig. 7D (MAT 3 to PD-1, MAT 8 to PD-1, MAT 10 to PD-1, MAT 11 to PD-1, MAT 12 to PD-1, MAT 13 to PD-1, MAT 14 to PD-1, and MAT 15 to PD-1).

Figures 8A–8C show that anti-PD-1 MAbs 1–15 bind to cynomolgus macaque PD-1. Binding curves for binding to scynoPD-1-hFc are shown in Figure 8A (anti-PD-1 MAb 1, anti-PD-1 MAb 2, anti-PD-1 MAb 4, anti-PD-1 MAb 5, anti-PD-1 MAb 6, anti-PD-1 MAb 7, anti-PD-1 MAb 8, anti-PD-1 MAb 9, anti-PD-1 MAb 10, anti-PD-1 MAb 11, anti-PD-1 MAb 12, anti-PD-1 MAb 13, anti-PD-1 MAb 14, anti-PD-1 MAb 15, anti-PD-1 MAb 16, anti-PD-1 MAb 17, anti-PD-1 MAb 18, anti-PD-1 MAb 19, anti-PD

- 8 047726

PD-1, MAT 5 to PD-1, MAT 6 to PD-1, MAT 7 to PD-1), Fig. 8B (MAT 9 to PD-1) and Fig. 8C (MAT 3 to PD-1, MAT 8 to PD-1, MAT 10 to PD-1, MAT 11 to PD-1, MAT 12 to PD-1, MAT 13 to PD-1, MAT 14 to PD-1 and MAT 15 to PD-1). Figs. 9A-9D show the results of assessing the ability of MATs to PD-1 1-15 to block the binding of human PD-L1 to human PD-1. The inhibition curves are shown in Fig. 9A (MAT to PD-1 I, MAT 2 to PD-1, MAT 3 to PD-1, MAT 15 to PD-1 and MAT to PD-1 A), Fig. 9B (MAT 4 to PD-1), Fig. 9C (MAT 5 to PD-1, MAT 6 to PD-1, MAT 7 to PD-1 and MAT to PD-1 A) and Fig. 9D (MAT 3 to PD-1, MAT 8 to PD-1, MAT 10 to PD-1, MAT 11 to PD-1, MAT 12 to PD-1, MAT 13 to PD-1, MAT 14 to PD-1, MAT 15 to PD-1 and MAT to PD-1 A).

Figures 10A-10B show the results of the tissue specificity analysis of hPD-1 mAb 7. Figure 10A shows the histological staining of normal colon (panels i and vii), liver (panels ii and viii), lung (panels iii and ix), pancreas (panels iv and x), kidney (panels v and xi), and heart (panels vi and xii) tissue. Figure 10A, panels i-vi, show the results of tissue incubated with labeled hPD-1 mAb 7 (0.313 μg/mL). Figure 10A, panels viixii, show the results of tissue incubated with labeled isotype control mAb (0.314 μg/mL). 10B shows the results of histological staining of skin (panels i and iv), tonsils (panels ii and v), and PD-1-expressing NSO cells (panels iii and vi). Fig. 10B, panels i-iii, show the results of tissue incubated with labeled MAb 7 to PD-1 (0.313 μg/mL).

Figure 11 shows the binding profiles of humanized anti-PD-1 antibodies, anti-hPD-1 MAb 2, anti-hPD-1 MAb 7 (1.1), anti-hPD-1 MAb 7 (1.2), anti-hPD-1 MAb 9 (1.1), and anti-PD-1 control antibodies, anti-PD-1 MAb A and anti-PD-1 MAb B, containing IgG1 (AA) or IgG4 (P), for binding to PD-1 on the cell surface.

Figures 12A-12B show the results of evaluating the ability of humanized anti-PD antibodies, anti-hPD-1 MAb 2, anti-hPD-1 MAb 7 (1.1), anti-hPD-1 MAb 7 (1.2), anti-hPD-1 MAb 9 (1.1), reference anti-PD1 antibodies, anti-PD-1 MAb A, and anti-PD-1 MAb B containing IgG1 (AA) or IgG4 (P), to block the binding of soluble human PD-L1 (Figure 12A) and soluble human PD-L2 (Figure 12B) to human PD-1 on the cell surface.

Figure 13 shows the results of evaluating the ability of humanized anti-PD antibodies, anti-hPD-1 MAb 2, anti-hPD-1 MAb 7(1.1), anti-hPD-1 MAb 7(1.2), anti-hPD-1 MAb 9(1.1), and reference anti-PD-1 antibodies, anti-PD-1 MAb A and anti-PD-1 MAb B, containing IgG1 (AA) or IgG4 (P), to act as antagonists of the PD1/PD-L1 axis by blocking the PD-1/PD-L1 interaction and preventing the suppression of T cell responses in a quantitative assay using the Jurkat-luc-NFAT/CHO-PD-L1 luciferase reporter.

Figure 14 presents results showing that anti-PD-1 MAb 2, anti-PD-1 MAb 7, anti-PD-1 MAb 9, and anti-PD-1 MAb 15 were able to stimulate cytokine production to levels comparable to or higher than those of the reference anti-PD-1 antibodies (anti-PD-1 MAb A and anti-PD-1 MAb B), and that treatment with anti-PD-1 MAb 2, anti-PD-1 MAb 7, anti-PD-1 MAb 9, and anti-PD-1 MAb 15 in combination with anti-LAG-3 MAb 1 provided the greatest increase in cytokine release. We present IFN-γ secretion profiles from peripheral blood mononuclear cells (PBMCs) stimulated with staphylococcal enterotoxin B (SEB), treated with anti-PD-1 antibodies and anti-LAG-3 antibodies individually and in combination.

Figures 15A-15B show the results of the assessment of the ability of humanized anti-PD antibodies, anti-hPD-1 mAb 2, anti-hPD-1 mAb 7(1.2), anti-hPD-1 mAb 9(1.1), and the reference anti-PD-1 antibodies, anti-PD-1 mAb A and anti-PD-1 mAb B, containing IgG1 (AA) or IgG4 (P), to stimulate cytokine production. The secretion profiles of IFN-γ (Figure 15A) and TNF-α (Figure 15B) from SEB-stimulated PBMCs treated with anti-PD-1 antibodies are shown.

Figures 16A-16B show results demonstrating that the PD-1xLAG-3 bispecific diabody constructs, DART A, DART D, DART E, DART F, DART G, and DART H, were able to stimulate cytokine production to levels comparable to or greater than those observed with the combination of anti-PD-1 mAb+anti-LAG-3 mAb (anti-PD-1 mAb A+anti-LAG-3 mAb A), and that the PD-1xLAG-3 bispecific diabody constructs, DART A, DART D, DART E, DART F, and DART G, provided the greatest enhancement of cytokine release. IFN-γ secretion profiles from PBMCs stimulated with a low concentration of SEB (0.2 ng/ml) treated with PD-1xLAG-3 bispecific diabodies or with antibodies to PD-1 and LAG-3 alone and in combination are shown in graphical form. The results obtained using PBMCs from two representative donors are shown in Fig. 16A and Fig. 16B.

Figures 17A–17B show results demonstrating that the PD-1xLAG-3 bispecific diabodies constructs, DART A, DART B, and DART C, are able to stimulate cytokine production to levels exceeding those observed with the combination of anti-PD-1 mAb+anti-LAG-3 mAb (anti-PD-1 mAb A+anti-LAG-3 mAb A). IFN-γ secretion profiles from PBMCs from two representative donors stimulated with high concentrations of SEB (85 ng/mL) treated with PD-1xLAG-3 bispecific diabodies or anti-PD-1 and anti-LAG-3 antibodies alone and in combination

- 9 047726 tions are presented in graphical form. The results obtained using PBMCs from two typical donors are shown in Fig. 17A and Fig. 17B.

Figures 18A-18B show results demonstrating that the PD-1xLAG-3 bispecific diabodies constructs, DART A, DART B, and DART C, are able to stimulate cytokine production to levels exceeding those observed with the combination of anti-PD-1 mAb+anti-LAG-3 mAb (anti-PD-1 mAb A+anti-LAG-3 mAb A). IFN-γ secretion profiles from PBMCs from two representative donors stimulated with a medium concentration of SEB (0.5 ng/mL) that were treated with the PD-1xLAG-3 bispecific diabodies or with anti-PD-1 and anti-LAG-3 antibodies, either alone or in combination, are shown in graphical form. The results obtained using PBMCs from two representative donors are shown in Figure 18A and Figure 18B.

Figure 19 shows the results demonstrating that the PD-1xLAG-3 bispecific diabodies constructs, DART D and DART H, were able to stimulate cytokine production to levels comparable to or greater than those observed with the combination of anti-PD1 mAb+anti-LAG-3 mAb (anti-PD-1 mAb A+anti-LAG-3 mAb A), and that DART D provided the greatest increase in cytokine release. The IL-2 secretion profiles from typical donor PBMCs stimulated with a high concentration of SEB (85 ng/mL) that were treated with the PD-1xLAG-3 bispecific diabodies or with anti-PD-1 antibodies and anti-LAG-3 antibodies, alone and in combination, are shown in graphical form. 20 presents results indicating that the PD-1xLAG-3 bispecific diabodies constructs, DART B and DART I, are capable of stimulating cytokine production to levels exceeding those observed with the administration of a combination of anti-PD-1 MAb + anti-LAG-3 MAb (anti-PD-1 MAb A + anti-LAG-3 MAb A, anti-hPD-1 MAb 7(1.2) + anti-hLAG-3 MAb 1(1.4), anti-hPD-1 MAb 7(1.2) + anti-hLAG-3 MAb 6(1.1)). IFN-γ secretion profiles from typical donor PBMCs stimulated with a moderate concentration of SEB (0.5 ng/mL) that were treated with PD-1xLAG-3 bispecific diabodies or anti-PD-1 and anti-LAG-3 antibodies, individually and in combination, are shown in graphical form.

Figures 21A-21D show results demonstrating that the PD-1xLAG-3 bispecific diabody, DART I, is able to stimulate cytokine production to levels exceeding those observed with the combination of anti-PD-1 mAb+anti-LAG-3 mAb (anti-PD-1 mAbA+anti-LAG3 mAbA). The secretion profiles of IFN-γ (Figures 21A and 21C) and IL-2 (Figures 21B and 21D) from tetanus toxoid-stimulated (5 μg/ml) memory CD4 T cells from two representative donors treated with the PD-1xLAG-3 bispecific diabody, DART-I, anti-PD-1 antibody, and anti-LAG-3 antibody in combination, or an isotype control antibody are shown in graphical form. Results obtained at day 7 using memory CD4 T cells from two typical donors are shown in Fig. 21AB and Figs. 21C–D. Figure 22 shows results demonstrating that the pharmacokinetic parameters of the PD-1xLAG-3 bispecific molecule, DART I, are comparable to those of the anti-PD-1 antibody, anti-PD-1 A IgG4 mAb (P), in cynomolgus monkeys. Lines indicate mean serum concentrations of DART I (solid line) and anti-PD-1 A mAb (dashed line). Individual values for male (filled symbols) and female (open symbols) monkeys are plotted for DART I (triangles) and anti-PD-1 A mAb (circles).

Figures 23A-23C show serum antibody concentrations and the percentage of bound PD1 on CD4+ or CD8+ T cells as a function of time in animals after treatment with different anti-PD-1 antibodies. The percentage of bound PD-1 on CD4+ or CD8+ T cells after treatment with anti-PD-1 mAbs is plotted on the right OY axes; symbols represent the percentage (%) of bound PD-1 on T cells for each individual animal, and dashed lines represent mean values. Serum mAb concentrations are plotted on the left OY axes; symbols represent serum levels for each individual animal, and solid lines represent nonlinear fits of the data. Each panel shows data for animals (n=1/sex/group) that received 10 mg/kg anti-hPD-1 mAb 7(1.2) IgG4 (P) (Fig. 23A), anti-PD-1 mAb A IgG4 (P) (Fig. 23B), or anti-PD-1 mAb B IgG4 (P) (Fig. 23B) by intravenous (i.v.) infusion on day 1.

Detailed description of the invention

According to the present invention, PD-1-binding molecules are provided that comprise a PD-1-binding domain of selected anti-PD-1 antibodies capable of binding to cynomolgus macaque PD-1 and human PD-1: anti-PD-1 MAT 1, anti-PD-1 MAT 2, anti-PD-1 MAT 3, anti-PD-1 MAT 4, anti-PD-1 MAT 5, anti-PD-1 MAT 6, anti-PD-1 MAT 7, anti-PD-1 MAT 8, anti-PD-1 MAT 9, anti-PD-1 MAT 10, anti-PD-1 MAT 11, anti-PD-1 MAT 12, anti-PD-1 MAT 13, anti-PD-1 MAT 14, or anti-PD-1 MAT 15. In particular, the present invention provides PD-1-binding molecules that are humanized or chimeric versions of said antibodies, or that comprise PD-1-binding fragments of said PD-1 antibodies (in particular, immunoconjugates, diabodies (including, but not limited to, DART-A, DART-B, DART-C, DART-D, DART-E, DART-F, DART-G, DART-H, DART-I and DART-J), BiTEs, bispecific antibodies, etc.). The present invention provides PD-1-binding molecules that complement

- 10 047726 are particularly capable of binding to an epitope of a molecule involved in the regulation of an immune response checkpoint, which is present on the surface of an immune cell. According to the present invention, there are also provided methods of using said PD-1-binding molecules for detecting PD-1 or for stimulating an immune response. According to the present invention, there are also provided methods of combination therapy, in which a PD-1-binding molecule, which comprises one or more PD-1-binding domains of said selected antibodies to PD-1, is administered in combination with one or more additional molecules that effectively stimulate an immune response, and/or in combination with one or more additional molecules that specifically bind to a cancer antigen.

I. Antibodies and their binding domains.

The antibodies of the present invention are immunoglobulin molecules capable of specifically binding to a target such as a carbohydrate, a polynucleotide, a lipid, a polypeptide, etc., via at least one antigen recognition site located in the variable domain of the immunoglobulin molecule. In the present invention, the terms antibody and antibodies refer to monoclonal antibodies, multispecific antibodies, human antibodies, humanized antibodies, synthetic antibodies, chimeric antibodies, polyclonal antibodies, camelid antibodies, single-chain Fvs (scFv), single-chain antibodies, Fab fragments, F(ab') fragments, disulfide-linked bispecific Fvs (sdFv), intracellular antibodies, and epitope-binding fragments of any of the foregoing. In particular, antibodies include immunoglobulin molecules and immunologically active fragments of immunoglobulin molecules, i.e. molecules that contain an antigen-binding site. Immunoglobulin molecules may be of any type (e.g., IgG, IgE, IgM, IgD, IgA, and IgY), class (e.g., IgG1, IgG2, IgG3, IgG4, IgA1, and IgA1), or subclass. In addition to the known diagnostic uses of antibodies, antibodies have been shown to be useful as therapeutic agents. Antibodies are capable of immunologically specifically binding to a polypeptide or protein or nonprotein molecule due to the presence of a particular domain or fragment or conformation (epitope) on such molecule. The epitope-containing molecule may have immunogenic activity such that it elicits the production of antibodies in an animal, such molecules being called antigens. Over the past few decades, there has been a resurgence of interest in the therapeutic potential of antibodies, and antibodies have become one of the leading classes of biotechnological drugs (Chan, C.E. et al. (2009) The Use Of Antibodies In The Treatment Of Infectious Diseases, Singapore Med. J. 50(7):663-666). More than 200 antibody-based drugs have been approved for use or are in development.

The term monoclonal antibody refers to a homogeneous population of antibodies, wherein a monoclonal antibody is composed of amino acids (natural and non-natural) that participate in the selective binding of an antigen. Monoclonal antibodies are highly specific and are directed to a single epitope (or antigenic site). The term monoclonal antibody includes not only intact monoclonal antibodies and full-length monoclonal antibodies, but also fragments thereof (such as Fab, Fab', F(ab') ₂ Fv), single-chain fragments (scFv), mutated variants thereof, fusion proteins containing part of an antibody, humanized monoclonal antibodies, chimeric monoclonal antibodies, and any other modified configuration of the immunoglobulin molecule that contains an antigen recognition site having the desired specificity and ability to bind to an antigen. The term is not limiting as to the source of the antibody or the method of its production (e.g., by hybridoma, phage selection, recombinant expression, transgenic animals, etc.). The term includes full-length immunoglobulins as well as fragments, etc., described above under the definition of antibody. Methods for producing monoclonal antibodies are known in the art. One method that can be used is the method described by Kohler, G. et al. (1975) Continuous Cultures Of Fused Cells Secreting Antibody Of Predefined Specificity, Nature 256:495-497, or a modification thereof. Typically, monoclonal antibodies are produced in mice, rats, or rabbits. The antibodies are produced by immunizing the animal with an immunogenic amount of cells, cell extracts, or protein preparations that contain the desired epitope. The immunogen may be, but is not limited to, primary cells, cultured cell lines, cancer cells, proteins, peptides, nucleic acids, or tissue. Cells used for immunization may be cultured for a period of time (e.g., at least 24 hours) prior to use as an immunogen. Cells may be used as immunogens alone or in combination with a non-denaturing adjuvant such as Ribi (see, e.g., Jennings, VM (1995) Review of Selected Adjuvants Used in Antibody Production, ILAR J. 37(3):119-125). In general, cells should be maintained in an intact state and preferably viable when used as immunogens. Intact cells may provide more efficient antigen detection than damaged cells in immunized animals. The use of denaturing or harsh adjuvants, such as Freud's adjuvant, may cause cell damage and is therefore not recommended. The immunogen can be administered multiple times at periodic intervals, such as once every two weeks or every

- 11 047 726 directly, or can be administered in a manner that maintains the viability of the animal (e.g., in a tissue recombinant). Alternatively, existing monoclonal antibodies and any other equivalent antibodies that are immunospecific for a desired pathogenic epitope can be sequenced and produced using any recombinant techniques known in the art. In one embodiment of the present invention, a suitable antibody is sequenced and the polynucleotide sequence is then cloned into a vector for expression or propagation. The sequence encoding the antibody of interest can be maintained in the vector in a host cell, and the host cell can then be propagated and frozen for later use. The polynucleotide sequence of suitable antibodies can be used for genetic manipulation to obtain monospecific or polyspecific (e.g. bispecific, trispecific and tetraspecific) molecules according to the present invention, as well as an affinity-optimized antibody, a chimeric antibody, a humanized antibody and/or an antibody containing fragments of canine antibodies in order to improve the affinity or other characteristics of the antibody. The general principle of humanization of an antibody includes maintaining the basic sequence of the antigen-binding portion of the antibody and replacing the remaining portion of the antibody from a non-human species with human antibody sequences.

Natural antibodies (such as IgG antibodies) consist of two light chains linked to two heavy chains. Each light chain contains a variable domain (VL) and a constant domain (CL). Each heavy chain contains a variable domain (VH), three constant domains (CH1, CH2, and CH3), and a hinge domain located between the CH1 and CH2 domains. The basic structural unit of natural immunoglobulins (e.g., IgG) is thus a tetramer containing two light chains and two heavy chains, usually expressed as a glycoprotein of approximately 150,000 Da. The amino-terminal (N-terminal) portion of each chain contains a variable domain of 100 to 110 or more amino acids primarily responsible for antigen recognition. The carboxyl-terminal (C-terminal) portion of each chain defines the constant region, with light chains containing one constant domain and heavy chains typically containing three constant domains and a hinge domain. Therefore, the structure of the light chains of the IgG molecule is nVL-CL-c, and the structure of the heavy chains of IgG is n-VH-CH1-H-CH2-CH3-C (where H represents the hinge domain and n and c represent the N-terminus and C-terminus of the polypeptide, respectively). The variable domains of the IgG molecule are composed of hypervariable regions (CDRs), which contain the residues that make contact with the epitope, and segments other than the CDRs, called framework segments (FRs), which generally maintain the structure and determine the arrangement of the CDR loops to enable such contact (although some residues of the framework regions may also bind the antigen). Accordingly, the VL and VH domains have the structure n-FR1-CDR1-FR2-CDR2-FR3-CDR3FR4-C. The polypeptides that are (or can perform the function of) the first, second, and third CDRs of the light chain of an antibody are designated, respectively, as the CDRL1 domain, the CDRL2 domain, and the CDRL3 domain. Similarly, the polypeptides that are (or can perform the function of) the first, second, and third CDRs of the heavy chain of an antibody are designated, respectively, as the CDRH1 domain, the CDRH2 domain, and the CDRH3 domain. Accordingly, the terms CDRL1 domain, CDRL2 domain, CDRL3 domain, CDRH1 domain, CDRH2 domain and CDRH3 domain refer to polypeptides that, when incorporated into a protein, confer on said protein the ability to bind to a particular epitope, regardless of whether said protein is an antibody comprising light and heavy chains, or a diabody or a single-chain binding molecule (e.g., scFv, BiTe, etc.), or is another type of protein. Accordingly, in the present application, the term epitope-binding fragment means a fragment of an antibody capable of immunospecifically binding to an epitope, and the term epitope-binding site refers to that portion of the molecule that contains the epitope-binding fragment that is responsible for binding the epitope. The epitope-binding site may comprise 1, 2, 3, 4, 5 or all 6 CDR domains of said antibody and, although capable of immunospecifically binding to said epitope, may exhibit immunospecificity, affinity or selectivity for an epitope that is different from the epitope of said antibody. Preferably, however, the epitope-binding fragment comprises all 6 CDR domains of said antibody. The epitope-binding fragment of an antibody may be a single polypeptide chain (e.g., scFv) or may comprise two or more polypeptide chains, each having an amino terminus and a carboxy terminus (e.g., diabody, Fab fragment, F(ab') ₂ fragment, etc.).

The present invention provides, in particular, single-chain variable domain fragments (scFv) of the anti-PD-1 antibodies of the present invention and multispecific binding molecules comprising the same. The single-chain variable domain fragments are prepared by joining the variable domains of a light chain and/or a heavy chain using a short linking peptide. Bird et al. (1988) (Single-Chain Antigen-Binding Proteins, Science 242:423-426) describe an example of linking peptides that form a bond of approximately 3.5 nm in length between the carboxy terminus of one variable domain and the amino terminus of another variable domain. Linkers having other sequences have been designed and used (Bird et al.

- 12 047726 (1988) Single-Chain Antigen-Binding Proteins, Science 242:423-426). In turn, the linkers can be modified to impart additional functions, such as drug attachment or attachment to solid supports. Single-chain variants can be produced by recombinant or synthetic methods. To produce scFv by synthetic methods, an automated synthesizer can be used. To produce scFv by recombinant methods, a suitable plasmid containing a polynucleotide that encodes an scFv can be introduced into a suitable host cell, for example, a eukaryotic cell such as yeast, plant, insect or mammalian cells, or a prokaryotic cell such as E. coli. Polynucleotides that encode the scFv of interest can be obtained by standard manipulations such as polynucleotide ligation. The resulting scFv can be isolated using standard protein purification methods known in the art.

The present invention also specifically includes humanized variants of the PD-1 antibodies of the present invention and multispecific binding molecules comprising them. The term humanized antibody refers to a chimeric molecule, typically produced using recombinant techniques, comprising an antigen-binding site of an immunoglobulin from a non-human species and the remainder of the immunoglobulin structure of the molecule, which is based on the structure and/or sequence of a human immunoglobulin. The human PD-1 antibodies of the present invention include humanized, chimeric variants and variants comprising fragments of canine antibodies, antibodies MAT to PD-1 1, MAT 2 to PD-1, MAT 3 to PD-1, MAT 4 to PD-1, MAT 5 to PD-1, MAT 6 to PD-1, MAT 7 to PD-1, MAT 8 to PD-1, MAT 9 to PD-1, MAT 10 to PD-1, MAT 11 to PD-1, MAT 12 to PD-1, MAT 13 to PD-1, MAT 14 to PD-1 or MAT 15 to PD-1. The polynucleotide sequence of the variable domains of said antibodies can be used for genetic manipulation to obtain said derivatives and improve the affinity or other characteristics of said antibodies. The general principle of humanization of an antibody involves maintaining the core sequence of the antigen-binding portion of the antibody while replacing the remainder of the antibody from a non-human species with human antibody sequences. There are four general steps in humanizing a monoclonal antibody. These steps include: (1) determining the nucleotide and predicted amino acid sequence of the original light and heavy chain variable domains of the antibody, (2) designing the humanized or canine-derived antibody, i.e., deciding which framework region of the antibody will be used during the humanization or canine-derived antibody incorporation process, (3) the actual humanization or canine-derived antibody incorporation methodologies/techniques, and (4) transfection and expression of the humanized antibody. See, e.g., U.S. Pat. Nos. 4,816,567; 5,807,715; and 6,331,415.

The antigen-binding site may comprise the entire variable domain fused to the constant domain, or only the complementarity determining regions (CDRs) of the variable domain grafted onto the appropriate framework regions. The antigen-binding sites may be wild-type or may be modified by one or more amino acid substitutions. This eliminates the constant region as an immunogen in humans, but retains the possibility of an immune response to the foreign variable domain (LoBuglio, A. F. et al. (1989) Mouse/Human Chimeric Monoclonal Antibody In Man: Kinetics And Immune Response, Proc. Natl. Acad. Sci. (U.S.A.) 86:4220-4224). Another approach focuses not only on providing constant regions of human origin, but also on modifying the variable domains to change them as closely as possible to the human form. The variable domains of the heavy and light chains are known to contain three complementarity determining regions (CDRs), which vary depending on the antigens in question and determine the binding ability, flanked by four framework regions (FRs), which are relatively conserved in a given species and which are thought to provide a framework for the CDRs. When producing antibodies from a non-human species against a given antigen, the variable domains can be reshaped or humanized by grafting CDRs derived from antibodies from non-human species onto FRs present in the human antibody to be modified. The application of this approach to various antibodies has been described in

- 13 047726

Sato, K. et al. (1993) Cancer Res 53:851–856. Riechmann, L. et al. (1988) “Reshaping Human Antibodies for Therapy,” Nature 332:323–327; Verhoeyen, M. et al. (1988) “Reshaping Human Antibodies: Grafting An Antilysozyme Activity,” Science 239:1534–1536; Kettleborough, C. A. et al. (1991) "Humanization Of A Mouse Monoclonal Antibody By CDRGrafting: The Importance Of Framework Residues On Loop Conformation", Protein Engineering 4:773-3783; Maeda, H. et al. (1991) “Construction Of Reshaped Human Antibodies With HTV-Neutralizing Activity,” Human Antibodies Hybridoma 2:124–134; Gorman, S. D. et al. (1991) "Reshaping A Therapeutic CD4 Antibody", Proc. Natl. Acad. Sci. (U.S.A.) 88:41814185; Tempest, P.R. et al. (1991) “Reshaping A Human Monoclonal Antibody To Inhibit Human Respiratory Syncytial Virus Infection in Vivo,” Bio/Technology 9:266–271; Co, M. S. et al. (1991) "Humanized Antibodies For Antiviral Therapy", Proc. Natl. Acad. Sci. (U.S.A.) 88:28692873; Carter, P. et al. (1992) "Humanization Of An Anti-pl 8 5her 2 Antibody For Human Cancer Therapy", Proc. Natl. Acad. Sci. (U.S.A.) 89:4285-4289; and Co., M.S. et al. (1992) "Chimeric And Humanized Antibodies With Specificity For The CD33 Antigen", J. Immunol. 148:11491154.

In some embodiments of the present invention, humanized antibodies retain all CDR sequences (e.g., a humanized murine antibody that contains all six CDRs from murine antibodies). In other embodiments of the present invention, humanized antibodies contain one or more CDRs (one, two, three, four, five, or six) whose sequences differ from those of the parent antibody. A number of humanized antibody molecules containing an antigen-binding site derived from an immunoglobulin from a non-human species have been described, including chimeric antibodies containing a rodent or modified rodent variable domain and their associated complementarity determining regions (CDRs) fused to human constant domains (see, e.g., Winter et al. (1991) Man-made Antibodies, Nature 349:293–299; Lobuglio et al. (1989) Mouse/Human Chimeric Monoclonal Antibody In Man: Kinetics And Immune Response, Proc. Natl. Acad. Sci. (U.S.A.) 86:4220–4224 (1989), Shaw et al. (1987) Characterization Of A Mouse/Human Chimeric Monoclonal Antibody (17-1A) To A Colon Cancer Tumor-Associated Antigen, J. Immunol. 138:4534-4538, and Brown et al. (1987) Tumor-Specific Genetically Engineered Murine/Human Chimeric Monoclonal Antibody, Cancer Res. 47:3577-3583). Other references describe rodent CDRs grafted into a human scaffold framework (FR) region prior to hybridization with the corresponding human constant domains (see, e.g., Riechmann, L. et al. (1988) Reshaping Human Antibodies for Therapy, Nature 332:323-327; Verhoeyen, M. et al. (1988) Reshaping Human Antibodies: Grafting An Antilysozyme Activity, Science 239:1534-1536; and Jones et al. (1986) Replacing The Complementarity-Determining Regions In A Human Antibody With Those From A Mouse, Nature 321:522-525). Another reference describes rodent CDRs supported by rodent framework regions after recombination of surface residues. See, for example, European Patent Publication No. 519,596. These humanized molecules are intended to minimize the undesirable immunological response to rodent anti-human antibody molecules, which limits the duration and effectiveness of therapeutic use of these fragments in human recipients. Other methods of humanizing antibodies that may also be used are disclosed in Daugherty et al. (1991) Polymerase Chain Reaction Facilitates The Cloning, CDR-Grafting, And Rapid Expression Of A Murine Monoclonal Antibody Directed Against The CD18 Component Of Leukocyte Integrins, Nucl. Acids Res. 19:2471-2476, and in U.S. Patents 6,180,377; 6,054,297; and 5,866,692.

II. Fcy receptors (FcyR).

The CH2 and CH3 domains of the two heavy chains interact to form an Fc region, which is a domain recognized by cellular Fc receptors, including, but not limited to, Fc-gamma receptors (FcyRs). In the present invention, the term Fc region is used to define the C-terminal region of the IgG heavy chain. The amino acid sequence of the CH2-CH3 domain of an exemplary human IgG1 is the sequence shown below (SEQ ID NO: 1):

- 14 047726

231 240 250 260 270 280

APELLGGPSV FLFPPKPKDT LMISRTPEVT CVWDVSHED PEVKFNWYVD

290 300 310 320 330

GVEVHNAKTK PREEQYNSTY RWSVLTVLH QDWLNGKEYK CKVSNKALPA

340 350 360 370 380

PIEKTISKAK GQPREPQVYT LPPSREEMTK NQVSLTCLVK GFYPSDIAVE

390 400 410 420 430

WESNGQPENN YKTTPPVLDS DGSFFLYSKL TVDKSRWQQG NVFSCSVMHE

440 447

ALHNHYTQKS LSLSPGX numbered using the EU index as described in Kabat, where X is lysine (K) or absent.

The amino acid sequence of the CH2-CH3 domain of an exemplary human IgG2 is as follows (SEQ ID NO: 2):

231 240 250 260 270 280

APPVA-GPSV FLFPPKPKDT LMISRTPEVT CVWDVSHED PEVQFNWYVD 290 300 310 320 330 GVEVHNAKTK PREEQFNSTF RWSVLTWH QDWLNGKEYK CKVSNKGLPA 340 350 360 370 380 PIEKTISKTK GQPREPQVYT LPPSREEMTK NQVSLTCLVK GFYPSDISVE 390 400 410 420 430 WESNGQPENN YKTTPPMLDS DGSFFLYSKL TVDKSRWQQG NVFSCSVMHE

440 447

ALHNHYTQKS LSLSPGX numbered using the EU index as described in Kabat, where X is lysine (K) or absent.

The amino acid sequence of the CH2-CH3 domain of an exemplary human IgG3 is as follows (SEQ ID NO: 3):

231 240 250 260 270 280

APELLGGPSV FLFPPKPKDT LMISRTPEVT CVWDVSHED PEVQFKWYVD

290 300 310 320 330

GVEVHNAKTK PREEQYNSTF RWSVLTVLH QDWLNGKEYK CKVSNKALPA

340 350 360 370 380

PIEKTISKTK GQPREPQVYT LPPSREEMTK NQVSLTCLVK GFYPSDIAVE

390 400 410 420 430

WESSGQPENN YNTTPPMLDS DGSFFLYSKL TVDKSRWQQG NIFCSVMHE

440 447

ALHNRFTQKS LSLSPGX numbered using the EU index as described in Kabat, where X is lysine (K) or absent.

The amino acid sequence of the CH2-CH3 domain of an exemplary human IgG4 is as follows (SEQ ID NO: 4):

- 15 047726

231 240 250 260 270 280

APEFLGGPSV FLFPPKPKDT

LMISRTPEVT CVWDVSQED

PEVQFNWYVD

290 300 310 320 330

GVEVHNAKTK PREEQFNSTY

RWSVLTVLH QDWLNGKEYK

CKVSNKGLPS

340 350 360 370 380

SIEKTISKAK GQPREPQVYT

LPPSQEEMTK NQVSLTCLVK

GFYPSDIAVE

390 400 410 420 430

WESNGQPENN YKTTPPVLDS

DGSFFLYSRL TVDKSRWQEG

NVFSCSVMHE

440 447

ALHNHYTQKS LSLSLGX numbered using the EU index as described in Kabat, where X is lysine (K) or absent.

As used herein, the numbering of residues in the constant region of the IgG heavy chain corresponds to the EU index as set forth in Kabat et al., Sequences of Proteins of Immunological Interest, ^5th Ed. Public Health Service, NH1, MD (1991) (Kabat), incorporated herein by reference in its entirety. The term EU index as set forth in Kabat refers to the numbering according to the EU system of the human IgG1 antibody. Amino acids from the variable domains of mature heavy and light chains of immunoglobulins are designated according to the position of the amino acid in the chain. Kabat et al. describe numerous amino acid sequences for antibodies, identify an amino acid consensus sequence for each subgroup, and assign a residue number to each amino acid and identify the CDRs as defined in Kabat et al. (It should be understood that CDRH1 defined in Chothia, C. & Lesk, A. M. ((1987) Canonical structures for the hypervariable regions of immunoglobulins, J. Mol. Biol. 196:901-917) begins five residues earlier.) The Kabat numbering scheme can be extended to antibodies not included in its compendium by matching the antibody in question to one of the consensus sequences in Kabat with respect to conserved amino acids. This method of assigning residue numbers has become standard in the art and allows for the ready identification of amino acids at equivalent positions in different antibodies, including chimeric or humanized variants. For example, the amino acid at position 50 of the light chain of a human antibody occupies a position equivalent to that of the amino acid at position 50 of the light chain of a mouse antibody.

Polymorphisms have been observed at a number of different positions in the constant regions of the antibody (e.g., at CH1 positions including but not limited to positions 192, 193 and 214, at Fc positions including but not limited to positions 270, 272, 312, 315, 356 and 358, according to the EU numbering system set out in Kabat), and accordingly, there may be minor differences between the presented sequence and sequences known in the prior art. Polymorphic forms of human immunoglobulins are well known. Currently, 18 GM allotypes are known: G1m (1, 2, 3, 17) or G1m (a, x, f, z), G2m (23) or G2m (n), G3m (5, 6, 10, 11, 13, 14, 15, 16, 21, 24, 26, 27, 28) or G3m (b1, c3, , b0, b3, b4, s, t, g1, c5, u, v, g5) (Lefranc, et al., The Human IgG Subclasses: Molecular Analysis Of Structure, Function And Regulation. Pergamon, Oxford, p. 4378 (1990); Lefranc, G. et al., 1979, Hum. Genet.: 50, 199-211). The present invention specifically contemplates that the antibodies of the present invention may comprise any allotype, isoallotype, or haplotype of any immunoglobulin gene, and are not limited to the allotype, isoallotype, or haplotype of the sequences provided herein. In addition, in some expression systems, the C-terminal amino acid residue (shown in bold) of the CH3 domain may be removed after translation. Accordingly, the C-terminal residue of the CH3 domain is an optional amino acid residue in the PD-1 binding molecules of the present invention. Specifically included within the scope of the present invention are PD-1 binding molecules lacking the C-terminal residue of the CH3 domain. Specifically included within the scope of the present invention are constructs comprising the C-terminal lysine residue of the CH3 domain.

Activating and inhibitory signals are transmitted through ligation of the Fc region to the cellular Fc-gamma receptor (FcyR). The ability of this ligation to perform diametrically opposed functions is due to structural differences between the different FcyRs. Two distinct domains within the cytoplasmic signaling domains of the receptor, called immunoreceptor tyrosine-based activation motifs (ITAMs) and immunoreceptor tyrosine-based inhibitory motifs (ITIMs), are responsible for the distinct responses. Recruitment of different cytoplasmic enzymes to these structures determines the outcome of FcyR-mediated cellular responses.

- 16 047726 items. ITAM-containing FcyR complexes include FcyRI, FcyRIIA, FcyRIIIA, whereas ITIM-containing complexes include only FcyRIIB. Human neutrophils express the FcyRIIA gene. Clustering of FcyRIIA by immune complexes or specific cross-linking of antibodies serves to aggregate ITAMs together with receptor-associated kinases, which promote ITAM phosphorylation. ITAM phosphorylation serves as a docking site for Syk kinase, whose activation leads to activation of subsequent substrates (e.g., PI3K). Cellular activation leads to the release of proinflammatory mediators. The FcyRIIB gene is expressed in B lymphocytes; the extracellular domain of FcyRIIB is 96% identical to FcyRIIA and binds to IgG complexes in a similar manner. The presence of ITIM in the cytoplasmic domain of FcyRIIB defines this inhibitory subclass of FcyR. The molecular basis for this inhibition has recently been elucidated. When co-ligated with the activating FcyR, the ITIM of FcyRIIB becomes phosphorylated and recruits the SH2 domain of inositol polyphosphate 5'-phosphatase (SHIP), which hydrolyzes phosphoinositol messengers that are released upon tyrosine kinase activation mediated by ITAM-containing FcyR, subsequently leading to the prevention of intracellular Ca ²⁺ influx. Accordingly, cross-linking of FcyRIIB attenuates the activating response to FcyR ligation and inhibits cellular susceptibility. As a consequence, B-cell activation, B-cell proliferation, and antibody secretion are prevented.

III. Bispecific antibodies, polyspecific diabodies and DART™ diabodies.

The ability of an antibody to bind to an epitope of an antigen depends on the presence and amino acid sequence of the VL and VH domains of the antibody. The interaction of the antibody light chain and the antibody heavy chain, and in particular the interaction of their VL and VH domains, forms one of the two epitope binding sites of a natural antibody. Natural antibodies are capable of binding to only one type of epitope (i.e., they are monospecific), although they can bind to multiple copies of a given type of epitope (i.e., they exhibit bivalentity or polyvalency). The binding domains of the present invention bind to epitopes in an immunospecific manner. In the present invention, an antibody, diabody or other epitope-binding molecule is said to immunospecifically bind to a region of another molecule (i.e., an epitope) if it reacts or binds more frequently, more rapidly, with a longer duration and/or with a higher affinity to a given epitope compared to other epitopes. For example, an antibody that immunospecifically binds to a viral epitope is an antibody that binds to that viral epitope with greater affinity, avidity, more readily, and/or for a longer duration than it immunospecifically binds to other viral epitopes or non-viral epitopes. It should also be understood from reading this definition that, for example, an antibody (or fragment or epitope) that immunospecifically binds to a first target may, but need not, bind specifically or preferentially to a second target. Accordingly, immunospecific binding does not necessarily require (although it may include) exclusive binding. Typically, but not necessarily, reference to binding means specific binding. Two molecules are said to be capable of binding to each other in a physiospecific manner if such binding exhibits the specificity with which receptors bind to their respective ligands.

The functionality of antibodies can be enhanced by producing polyspecific antibody-based molecules that can simultaneously bind to two separate and different antigens (or different epitopes of the same antigen) and/or by producing an antibody-based molecule that has a higher valency (i.e., more than two binding sites) for the same epitope and/or antigen.

In order to provide molecules with greater functionality than natural antibodies, a wide range of recombinant bispecific antibody formats have been developed (see, e.g., PCT Publications WO 2008/003116, WO 2009/132876, WO 2008/003103, WO 2007/146968, WO 2009/018386, WO 2012/009544, WO 2013/070565), most of which use linker peptides to fusion an additional epitope-binding moiety (e.g., scFv, VL, VH, etc.) to or within the core portion of the antibody (IgA, IgD, IgE, IgG, or IgM) or to fusion multiple epitope-binding moieties (e.g., two Fab or scFv moieties). Alternative formats use linker peptides to hybridize an epitope binding moiety (e.g. scFv, VL, VH, etc.) to a dimerization domain such as a CH2-CH3 domain or alternative polypeptides (WO 2005/070966, WO 2006/107786A WO 2006/107617A, WO 2007/046893). Typically, these approaches include compromise options and acceptable options. For example, PCT publications WO 2013/174873, WO 2011/133886 and WO 2010/136172 disclose that the use of linkers may cause problems in a therapeutic setting and describe a trispecific antibody in which the CL and CH1 domains have been moved from their respective natural positions and the VL and VH domains have been diversified (WO 2008/027236, WO 2010/108127) to confer the ability to bind to more than one antigen. Accordingly, in the molecules disclosed in these documents, the binding specificity is altered in favor of the ability to bind to additional types of antigens. PCT publications WO 2013/163427 and WO 2013/119903 disclose a modification

- 17 047726 of the CH2 domain to comprise a hybrid protein adduct containing the binding domain. The document notes that the CH2 domain likely plays only a minimal role as a mediator of effector function. PCT publications WO 2010/028797, WO2010028796 and WO 2010/028795 disclose recombinant antibodies whose Fc regions have been replaced with additional VL and VH domains so as to yield trivalent binding molecules. PCT publications WO 2003/025018 and WO2003012069 disclose recombinant diabodies whose individual chains contain scFv domains. PCT Publication WO 2013/006544 discloses polyvalent Fab molecules that are synthesized as a single polypeptide chain and then subjected to proteolysis to form heterodimeric structures. Accordingly, in the molecules disclosed in these documents, all or some of the functional capabilities to act as mediators of effector function are altered in favor of the ability to bind to additional types of antigens. PCT publications WO 2014/022540, WO 2013/003652, WO 2012/162583, WO 2012/156430, WO 2011/086091, WO 2008/024188, WO 2007/024715, WO 2007/075270, WO 1998/002463, WO 1992/022583 and WO 1991/003493 disclose methods for adding additional binding domains or functional groups to an antibody or antibody portion (e.g., adding a diabody to the light chain of an antibody, or adding additional VL and VH domains to the light and heavy chains of an antibody, or adding a heterologous fusion protein, or linking multiple Fab domains to each other). Accordingly, in the molecules disclosed in the said documents, the native structure of the antibodies is modified in favor of the ability to bind to additional types of antigens.

The art also mentions the possibility of producing diabodies that differ from said natural antibodies in their ability to bind to two or more different types of epitopes (i.e. exhibit bispecificity or polyspecificity in addition to the bivalent or polyvalent form) (see, for example,

Holliger et al. (1993) "'Diabodies': Small Bivalent And Bispecific Antibody

Fragments", Proc. Natl. Acad. Sci. (U.S.A.) 90:6444-6448; US 2004/0058400 (Hollinger et al.

US 2004/0220388/WO 02/02781 (Mertens et al. Alt et aL (1999) FEBS Lett. 454(l-2):90-94; Lu, D. et al. (2005) "L Fully Human Recombinant IgG- Like Bispecific Antibody To Both The Epidermal Growth Factor Receptor And The Insulin-Like Growth Factor Receptor For Enhanced Antitumor Activity", J. Biol. Chem. 280(20): 19665-19672; WO 02/02781 (Mertens et al. Olafsen, T. et al. (2004) "Covalent Di sulfide-Linked Anti-CEA Diabody Allows Site-Specific Conjugation and Radiolabeling for Tumor Targeting Applications", Protein Sei.

17(l):21-27; Wu, A. et al. (2001) "Multimerization Of A Chimeric Anti-CD20 Single Chain FvFv Fusion Protein Is Mediated Through Variable Domain Exchange", Protein Engineering

14(2): 1025-1033; Asano et al. (2004) "L Diabody For Cancer Immunotherapy And Its Functional Enhancement By Fusion Of Human Fc Domain", Abstract 3P-683, J. Biochem.

76(8):992; Takemura, S. et al. (2000) "Construction Of A Diabody (Small Recombinant Bispecific Antibody) Using A Refolding System", Protein Eng. 13(8):583-588; Baeuerle, P.A. et al. (2009) "Bispecific T-Cell Engaging Antibodies For Cancer Therapy", Cancer Res. 69(12):4941–4944).

The design of diabodies is based on the production of an antibody derivative known as a single-chain variable fragment (scFv). Suitable molecules are prepared by linking the variable domains of the light chain and/or heavy chain using a short linking peptide. Bird et al. (1988) (Single-Chain Antigen-Binding Proteins, Science 242:423-426) describe an example of linking peptides that form a bridge of approximately 3.5 nm in length between the carboxy terminus of one variable domain and the amino terminus of the other variable domain. Linkers having other sequences have been designed and used (Bird et al. (1988) Single-Chain Antigen-Binding Proteins, Science 242:423-426). In turn, the linkers can be modified to impart additional functions, such as drug attachment or attachment to solid supports. Single-chain variants can be produced by recombinant or synthetic methods. An automated synthesizer can be used to produce scFv by synthetic methods. To produce scFv by recombinant methods, a suitable plasmid containing a polynucleotide that encodes an scFv can be introduced into a suitable host cell, for example, a eukaryotic cell, such as yeast, plant, insect or mammalian cells, or a prokaryotic cell, such as an E. coli cell. Polynucleotides that encode the scFv of interest can be obtained by standard manipulations, such as ligation of polynucleotides. The resulting scFv can be isolated using standard protein purification methods known in the art. The provision of non-monospecific diabodies represents a significant advantage over antibodies, including, but not limited to,

- 18 047726 the ability to ligate and colocalize cells that express different epitopes. Accordingly, bispecific diabodies have a wide range of applications, including therapy and immunodiagnostics. Bispecificity provides greater flexibility in designing and modifying diabodies for a variety of applications by providing increased avidity for multimeric antigens, cross-linking of different antigens, and targeted targeting to specific cell types based on the presence of both target antigens. Due to their increased valency, low off-rates, and rapid clearance from the circulation (for small diabodies of about or less than ~50 kDa), diabody molecules known in the art have also found particular application in the field of tumor imaging (Fitzgerald et al. (1997) Improved Tumour Targeting By Disulphide Stabilized Diabodies Expressed In Pichia pastoris, Protein Eng. 10:1221). The bispecificity of diabodies has led to their use for the coligation of various cells, for example, the cross-linking of cytotoxic T cells to tumor cells (Staerz et al. (1985) Hybrid Antibodies Can Target Sites For Attack By T Cells, Nature 314:628-631, and Holliger et al. (1996) Specific Killing Of Lymphoma Cells By Cytotoxic T-Cells Mediated By A Bispecific Diabody, Protein Eng. 9:299-305; Marvin et al. (2005) Recombinant Approaches To IgG-Like Bispecific Antibodies, Acta Pharmacol. Sin. 26:649-658). Alternatively, or additionally, bispecific diabodies can be used to coligate receptors on the surface of different cells or on the same cell. Co-ligation of different cells and/or receptors can be used to modulate effector functions and/or signaling involving immune cells. Polyspecific molecules (e.g. bispecific diabodies) containing epitope binding sites can be directed to the surface determinant of any immune cell, such as B7-H3 (CD276), B7-H4 (VTCN1), BTLA (CD272), CD3, CD8, CD16, CD27, CD32, CD40, CD40L, CD47, CD64, CD70 (CD27L), CD80 (B7-1), CD86 (B7-2), CD94 (KLRD1), CD137 (4-1BB), CD137L (41BBL), CD226, CTLA-4 (CD152), galectin-9, GITR, GITRL, HHLA2, ICOS (CD278), ICOSL (CD275), killer cell activation receptor (KIR), LAG-3 (CD223), LIGHT (TNFSF14, CD258), MHC class I or II, NKG2a, NKG2d, OX40 (CD134), OX40L (CD134L), PD1H, PD-1 (CD279), PD-L1 (B7-H1, CD274), PDL2 (B7-CD, CD273), PVR (NECL5, CD155), SIRPa, TCR, TIGIT, TIM-3 (HAVCR2), and/or VISTA (PD-1H), which are expressed on T lymphocytes, natural killer (NK) cells, antigen-presenting cells, or other mononuclear cells. In particular, epitope binding sites directed to a cell surface receptor (or its ligand) that is involved in regulating an immune checkpoint response can be used to generate bispecific or polyspecific binding molecules that antagonize or block inhibitory signaling by immune checkpoint molecules and thereby stimulate, activate, or enhance immune responses in a subject. Molecules involved in regulating immune checkpoints include, but are not limited to, B7-H3, B7-H4, BTLA, CD40, CD40L, CD47, CD70, CD80, CD86, CD94, CD137, CD137L, CD226, CTLA-4, galectin-9, GITR, GITRL, HHLA2, ICOS, ICOSL, KIR, LAG-3, LIGHT, MHC class I or II, NKG2a, NKG2d, OX40, OX40L, PD1H, PD-1, PD-L1, PD-L2, PVR, SIRPa, TCR, TIGIT, TIM-3, and/or VISTA.

However, the above advantages are associated with significant costs. The formation of these non-monospecific diabodies requires the successful assembly of two or more separate and distinct polypeptides (i.e., this formation process requires that the diabodies be formed by heterodimerization of different types of polypeptide chains). This is in contrast to the situation with monospecific diabodies, which are formed by homodimerization of identical polypeptide chains. Since at least two different polypeptides (i.e., two polypeptide molecules) must be provided to form a non-monospecific diabody and since homodimerization of such polypeptides results in inactive molecules (Takemura, S. et al. (2000) Construction Of A Diabody (Small Recombinant Bispecific Antibody) Using A Refolding System, Protein Eng. 13(8):583-588), the production of said polypeptides must be carried out in a way that prevents covalent association between polypeptides of the same type (i.e., to prevent homodimerization) (Takemura, S. et al. (2000) Construction Of A Diabody (Small Recombinant Bispecific Antibody) Using A Refolding System, Protein Eng. 13(8):583-588). In this regard, non-covalent linking of such polypeptides has been proposed in the art (see, for example, Olafsen et al. (2004) Covalent Disulfide-Linked Anti-CEA Diabody Allows Site-Specific Conjugation And Radiolabeling For Tumor Targeting Applications, Prot. Engr. Des. Sel. 17:21-27; Asano et al. (2004) A Diabody For Cancer Immunotherapy And Its Functional Enhancement By Fusion Of Human Fc Domain, Abstract 3P-683, J. Biochem. 76(8):992; Takemura, S. et al. (2000) Construction Of A Diabody (Small Recombinant Bispecific Antibody) Using A Refolding System, Protein Eng. 13(8):583-588; Lu, D. et al. (2005) A Fully Human Recombinant IgG-Like Bispecific Antibody To Both The Epidermal Growth Factor Receptor And The Insulin-Like Growth Factor Receptor For Enhanced Antitumor Activity, J. Biol. Chem. 280(20): 19665-19672).

However, it has been established in the art that bispecific diabodies consisting of non-covalently linked polypeptides are unstable and easily dissociate into non-functional monomers (see, for example, Lu, D. et al. (2005) A Fully Human Recombinant IgG-Like Bispecific Antibody To Both The Epidermal Growth Factor Receptor And The Insulin-Like Growth Factor Receptor For En

- 19 047726 hanced Antitumor Activity, J. Biol. Chem. 280(20): 19665-19672).

In light of this problem, the art has developed stable, covalently linked heterodimeric non-monospecific diabodies termed DART™ diabodies (Dual Affinity ReTargeting Reagents); see, e.g., U.S. Patent Publication Nos. 2013-0295121; 2010-0174053; and 2009-0060910; European Patent Publications EP 2714079; EP 2601216; EP 2376109; EP 2158221; and PCT Publications WO 2012/162068; WO 2012/018687; WO 2010/080538; and Sloan, D.D. et al. (2015) Targeting HIV Reservoir in Infected CD4 T Cells by DualAffinity Re-targeting Molecules (DARTs) that Bind HIV Envelope and Recruit Cytotoxic T Cells, PLoS Pathog. 11(11):e1005233. doi: 10.1371/journal.ppat. 1005233; Al Hussaini, M. et al. (2015) Targeting CD123 In AML Using A T-Cell Directed Dual-Affinity Re-Targeting (DART™) Platform, Blood pii: blood-2014-05575704; Chichili, G.R. et al. (2015) A CD3xCD123 Bispecific DART For Redirecting Host T Cells To Myelogenous Leukemia: Preclinical Activity And Safety In Nonhuman Primates, Sci. Transl. Med. 7(289):289ra82; Moore, P.A. et al. (2011) Application Of Dual Affinity Retargeting Molecules To Achieve Optimally Redirected T-Cell Killing Of B-Cell Lymphoma, Blood 117(17):4542-4551; Veri, M.C. et al. (2010) Therapeutic Control Of In Cell Activation Via Recruitment Of Fcgamma Receptor IIb (CD32B) Inhibitory Function With A Novel Bispecific Antibody Scaffold, Arthritis Rheum. 62(7): 1933-1943; Johnson, S. et al. (2010) Effector Cell Recruitment With Novel Fv-Based Dual-Affinity Re-Targeting Protein Leads To Potent Tumor Cytolysis And in vivo B-Cell Depletion, J. Mol. Biol. 399(3):436–449). These diabodies comprise two or more covalently linked polypeptides and involve modification of one or more cysteine residues in each of the polypeptide molecules used, which allows for the formation of disulfide bonds and thereby covalently linking the two polypeptide chains. For example, it has been shown that addition of a cysteine residue to the C-terminus of these constructs provides for the formation of a disulfide bond between the polypeptide chains, stabilizing the resulting heterodimer without adversely affecting the binding properties of the divalent molecule.

Each of the two polypeptides of the simplest bispecific DART diabody consists of three domains. The first polypeptide comprises (in the direction from the N-terminus to the C-terminus): (i) a first domain that contains a binding region from the variable domain of the light chain of a first immunoglobulin (VL1), (ii) a second domain that contains a binding region from the variable domain of the heavy chain of a second immunoglobulin (VH2), and (iii) a third domain that contains a cysteine residue (or a cysteine-containing domain) and a heterodimer-promoting domain that serves to promote heterodimerization with the second diabody polypeptide and to covalently link the first and second diabody polypeptides to each other. The second polypeptide comprises (in the direction from the N-terminus to the C-terminus): (i) a first domain that comprises a binding region from the variable domain of the light chain of a second immunoglobulin (VL2), (ii) a second domain that comprises a binding region from the variable domain of the heavy chain of a first immunoglobulin (VH1), and (iii) a third domain that comprises a cysteine residue (or a cysteine-containing domain) and a complementary heterodimer-promoting domain that combines with the heterodimer-promoting domain of the first polypeptide chain to promote heterodimerization with the first polypeptide chain. The cysteine residue (or cysteine-containing domain) of the third domain of the second polypeptide chain promotes the formation of a covalent bond between the second polypeptide chain and the first polypeptide chain of the diabody. The said molecules are stable, effective and have the ability to simultaneously bind to two or more antigens. According to one embodiment of the present invention, the third domains of the first and second polypeptides comprise a cysteine residue that serves to link the polypeptides via a disulfide bond. Fig. 1 is a diagram of said diabody that utilizes E-helix/K-helix heterodimer-promoting domains and a cysteine-containing linker to form a covalent bond. As shown in Fig. 2 and Fig. 3A-3C, one or both polypeptides may further comprise a CH2-CH3 domain sequence such that the junction of the two polypeptides of the diabody forms an Fc region that is capable of binding to an Fc receptor of cells (such as B lymphocytes, dendritic cells, natural killer cells, macrophages, neutrophils, eosinophils, basophils, and mast cells). As described in more detail below, the CH2 and/or CH3 domains of said polypeptide chains need not have identical sequences, and are preferably modified to facilitate joining of the two polypeptide chains.

Numerous variations of these molecules have been described (see, for example, U.S. Patent Publication Nos. 2015/0175697; 2014/0255407; 2014/0099318, 2013/0295121, 2010/0174053, and 2009/0060910; European Patent Publications EP 2714079, EP 2601216, EP 2376109, EP 2158221; and PCT Publications WO 2012/162068, WO 2012/018687, WO 2010/080538). These DART™ diabodies containing an Fc region may comprise two pairs of polypeptide chains. The first polypeptide chain comprises (in the direction from the N-terminus to the C-terminus): (i) a first domain that comprises a binding region from the variable domain of the light chain of a first immunoglobulin (VL1), (ii) a second domain that comprises a binding region from the variable domain of the heavy chain of a second immunoglobulin (VH2), (iii) a third to

- 20 047 726 me, which contains a cysteine residue (or a cysteine-containing domain) and promotes heterodimerization with a second diabody polypeptide and the formation of a covalent bond between the first and second diabody polypeptides, and (iv) a CH2-CH3 domain. The second polypeptide comprises (in the direction from the N-terminus to the C-terminus): (i) a first domain that contains a binding region from the variable domain of the light chain of a second immunoglobulin (VL2), (ii) a second domain that contains a binding region from the variable domain of the heavy chain of a first immunoglobulin (VH1), and (iii) a third domain that contains a cysteine residue (or a cysteine-containing domain) and a heterodimer formation promoting domain that promotes heterodimerization with the first polypeptide chain. In this case, the first two polypeptides are associated with each other to form an Fc region. In Fig. 3A-3C are schematic diagrams of three embodiments of said diabodies that utilize different domains that promote heterodimer formation. Other DART™ diabodies that contain an Fc region may contain three polypeptide chains. The first polypeptide of said DART™ diabodies contains three domains: (i) a VL1 containing domain, (ii) a VH2 containing domain, and (iii) a domain that contains the sequence CH2CH3. The second polypeptide of said DART™ diabodies contains: (i) a VL2 containing domain, (ii) a VH1 containing domain, and (iii) a domain that promotes heterodimerization and the formation of a covalent bond with the first polypeptide chain of the diabody. The third polypeptide of said DART™ diabodies contains the sequence CH2-CH3. Accordingly, the first and second polypeptide chains of the said DART™ diabodies associate to form a VL1/VH1 binding site that is capable of binding to an epitope, as well as a VL2/VH2 binding site that is capable of binding to a second epitope. Such more complex DART molecules also contain cysteine-containing domains that participate in the formation of a covalently linked complex. Accordingly, the first and second polypeptides are linked to each other via a disulfide bond between cysteine residues in their respective third domains. Notably, the first and third polypeptide chains are linked to each other to form an Fc region that is stabilized by a disulfide bond. Figs. 4A-4B show schematics of the said diabodies containing three polypeptide chains.

Other DART™ diabodies containing an Fc region may contain five polypeptide chains, which may contain binding regions from the light chain and heavy chain variable domains of up to three different immunoglobulins (referred to as VL1/VH1, VL2/VH2, and VL3/VH3). For example, the first polypeptide chain of such diabodies may contain: (i) a VH1 containing domain, (ii) a CH1 containing domain, and (iii) a CH2-CH3 containing domain. The second and fifth polypeptide chains of such diabodies may contain: (i) a VL1 containing domain, and (ii) a CL containing domain. The third polypeptide chain of said diabodies may comprise: (i) a domain containing VH1, (ii) a domain containing CH1, (iii) a domain containing the sequence CH2-CH3, (iv) a domain containing VL2, (v) a domain containing VH3, and (vi) a domain promoting the formation of a heterodimer, wherein the domains promoting the formation of a heterodimer stimulate the dimerization of the third chain with the fourth chain. The fourth polypeptide of said diabodies may comprise: (i) a domain containing VL3, (ii) a domain containing VH2, and (iii) a domain that stimulates heterodimerization and the formation of a covalent bond with the third polypeptide chain of the diabody. In this case, the first and third polypeptides are connected to each other to form an Fc region. Such more complex DART™ molecules also contain cysteine-containing domains that participate in the formation of a covalently linked complex such that each polypeptide chain is linked to at least one additional polypeptide chain via a disulfide bond involving cysteine residues. Preferably, said domains are ordered in an N-terminal to C-terminal direction. Figure 5 shows schematics of said diabodies containing five polypeptide chains.

Alternative constructs are known in the art for applications in which a tetravalent molecule is preferred but an Fc region is not required, including, but not limited to, tetravalent tandem antibodies, also referred to as TandAbs (see, e.g., U.S. Patent Publication Nos. 2005-0079170, 2007-0031436, 2010-0099853, 2011-020667, 20130189263; European Patent Publication Nos. EP 1078004, EP 2371866, EP 2361936, and EP 1293514; PCT Publication Nos. WO 1999/057150, WO 2003/025018, and WO 2013/013700), which are formed by homodimerization of two identical chains, each containing a VH1, VL2, VH2 and VL2 domain.

Recently, trivalent structures comprising two diabody-type binding domains and one non-diabody-type domain and an Fc region have been described (see, e.g., PCT Publication Nos. PCT/US15/33076, entitled Trispecific Binding Molecules and Methods of Using Them, filed May 29, 2015, and PCT/US15/33081, entitled Trispecific Binding Molecules That Specifically Bind to Multiple Cancer Antigens and Methods of Using Them, filed May 29, 2015). These trivalent molecules can be used to prepare monospecific, bispecific, or trispecific molecules. Figures 6A-6F are schematic diagrams of these trivalent molecules comprising 3 or 4 polypeptide chains.

IV. Human PD-1 binding molecules according to the present invention.

Preferred PD-1 binding molecules of the present invention include an

- 21 047 726 titos, diabodies, BiTEs, etc. and are capable of binding to a continuous or discontinuous (e.g., conformational) portion (epitope) of human PD-1 (CD279). The human PD-1 binding molecules of the present invention are preferably also capable of binding to PD-1 molecules of one or more non-human species, in particular to primate PD-1 (and in particular to primate species such as cynomolgus macaques). An exemplary PD-1 polypeptide (sequence in NCBI NP_005009.2; including a signal sequence of 20 amino acid residues (shown underlined) and 268 amino acid residues of the mature protein) comprises the amino acid sequence shown below (SEQ ID NO: 68):

MQIPQAPWPV VWAVLQLGWR PGWFLDSPDR PWNPPTFSPA LLWTEGDNA

TFTCSFSNTS ESFVLNWYRM SPSNQTDKLA AFPEDRSQPG QDCRFRVTQL

PNGRDFHMSV VRARRNDSGT YLCGAISLAP KAQIKESLRA ELRVTERRAE VPTAHPSPSP RPAGQFQTLV VGWGGLLGS LVLLVWVLAV ICSRAARGTI GARRTGQPLK EDPSAVPVFS VDYGELDFQW REKTPEPPVP CVPEQTEYAT

According to some embodiments of the present invention, human PD-1 binding molecules of the present invention are characterized by any one or more of the following criteria:

(1) specifically binds to human PD-1 endogenously expressed on the surface of stimulated human T cells;

(2) specifically binds to human PD-1 with an equilibrium binding constant (K _D ) of 40 nM or less;

(3) specifically binds to human PD-1 with an equilibrium binding constant (K _D ) of 5 nM or less;

(4) binds specifically to human PD-1 with a forward reaction velocity (ka) of 1.5x104 M ^-1 min ^-1 or more;

(5) specifically binds to human PD-1 with a forward reaction velocity (ka) of 90.0x104 M ^-1 min ^-1 or more;

(6) binds specifically to human PD-1 with a kd of 7x10 ^-4 min ^-1 or less;

(7) binds specifically to human PD-1 with a kd of 2x10 ^-4 min ^-1 or less;

(8) binds specifically to non-human primate PD-1 (e.g., cynomolgus macaque PD-1);

(9) inhibits (i.e. blocks or interferes with) the binding/inhibitory activity of PD-1 ligand (PD-L1/PD-L2) to PD-1;

(10) stimulates an immune response; and/or (11) synergizes with an antibody to human LAG-3 to stimulate an antigen-specific T cell response.

In the present invention, the term antigen-specific T cell response refers to T cell responses that result from stimulation of a T cell with an antigen for which the T cell is specific. Non-limiting examples of T cell responses upon antigen-specific stimulation include proliferation and cytokine production (e.g., TNF-α, IFN-γ production). The ability of a molecule to stimulate an antigen-specific T cell response can be determined, for example, using the Staphylococcus aureus enterotoxin type B (SEB)-stimulated PBMC assay described herein.

Preferred human PD-1 binding molecules according to the present invention comprise the VH and/or VL domains of the mouse anti-human PD-1 monoclonal antibodies MAb anti-PD-1 1, MAb anti-PD-1 2, MAb anti-PD-1 3, MAb anti-PD-1 4, MAb anti-PD-1 5, MAb anti-PD-1 6, MAb anti-PD-1 7, MAb anti-PD-1 8, MAb anti-PD-1 9, MAb anti-PD-1 10, MAb anti-PD-1 11, MAb anti-PD-1 12, MAb anti-PD-1 13, MAb anti-PD-1 14 or MAb anti-PD-1 15, and more preferably comprise 1, 2 or all 3 CDR _H of the VH domain and/or 1, 2 or all 3 CDR _L of the VL domain of said anti-PD-1 monoclonal antibodies human. Said preferred molecules binding to human PD-1 include bispecific (or polyspecific) antibodies, chimeric or humanized antibodies, BiTe, diabodies, etc., and said binding molecules containing variant Fc regions.

According to the present invention, in particular, PD-1-binding molecules comprising a PD-1-binding domain are provided, which comprise:

(A) (1) three CDRH domains of VH MAT to PD-1 1;

(2) three CDRl domains of VL MAT to PD-1 1;

(3) three CDRH domains of VH MAT to PD-1 1 and three CDRL domains of VL MAT to PD-1 1;

(4) VH domain from VH1 MAT to hPD-1 1;

(5) VL domain of VL1 MAT to hPD-1 1;

- 22 047726 (6) VH and VL domains of MAT to hPD-1 1;

(B) (1) three CDRH domains of VH MAT 2 to PD-1;

(2) three CdR _l domains of VL MAT 2 to PD-1;

(3) three CDRH domains of VH MAT 2 to PD-1 and three CDR _L domains of VL MAT 2 to PD-1;

(4) VH domain from VH1 MAT to hPD-1 2;

(5) VL domain of VL1 MAT to hPD-1 2;

(6) VH and VL domains of MAT to hPD-1 2;

(C) (1) three CDRH domains of VH MAT 3 to PD-1;

(2) three CDRL domains of VL MAT 3 to PD-1;

(3) three CDRH domains of VH MAT 3 to PD-1 and three CDRL domains of VL MAT 3 to PD-1;

(D) (1) three CDRH domains of VH MAT 4 to PD-1;

(2) three CDRL domains of MAT 4 VL to PD-1;

(3) three CDRH domains of VH MAT 4 to PD-1 and three CDRL domains of VL MAT 4 to PD-1;

(E) (1) three CDRH domains of VH MAT 5 to PD-1;

(2) three CDRs of _{the L} domain of VL MAT 5 to PD-1;

(3) three CDRH domains of MAT 5 VH to PD-1 and three CDRL domains of MAT 5 VL to PD-1;

(F) (1) three CDRH domains of VH MAT 6 to PD-1;

(2) three CDRs of _{the L} domain of VL MAT 6 to PD-1;

(3) three CDRH domains of MAT 6 VH to PD-1 and three CDRL domains of MAT 6 VL to PD-1;

(G) (1) three CDRH domains of MAT 7 VH to PD-1;

(2) three CDRL domains of VL MAT 7 to PD-1 or VL2 MAT to hPD-1 7 or VL3 MAT to hPD-1 7;

(3) three CDRH domains of VH MAT 7 to PD-1 and three CDR _L domains of VL MAT 7 to PD-1 or VL2 MAT to hPD1 7, VL3 MAT to hPD-1 7;

(4) the VH domain of VH1 MAT to hPD-1 7 or VH2 MAT to hPD-1 7;

(5) the VL domain of VL1 MAT to hPD-1 7 or VL2 MAT to hPD-1 7, or VL3 MAT to hPD-1 7;

(6) VH and VL domains of MAT to hPD-1 7(1.1) or MAT to hPD-1 7(1.2), or MAT to hPD-1 7(1.3), or MAT to hPD-1 7(2.1), or MAT to hPD-1 7(2.2), or MAT to hPD-1 7(2.3);

(H) (1) three CDRH domains of VH MAT 8 to PD-1;

(2) three CDRs of _{the L} domain of VL MAT 8 to PD-1;

(3) three CDRH domains of MAT 8 VH to PD-1 and three CDR _L domains of MAT 8 VL to PD-1;

(I) (1) three CDRH domains of VH MAT 9 to PD-1 or VH2 MAT to hPD-1 9;

(2) three CDRL domains of VL MAT 9 to PD-1 or VL2 MAT to hPD-1 9;

(3) three CDRH domains of VH MAT 9 to PD-1 or VH2 MAT to hPD-1 9 and three CDR _L domains of VL MAT 9 to PD-1 or VL2 MAT to hPD-1 9;

(4) the VH domain of VH1 MAT to hPD-1 9 or VH2 MAT to hPD-1 9;

(5) the VL domain of VL1 MAT to hPD-1 9 or VL2 MAT to hPD-1 9;

(6) VH and VL domains of MAT to hPD-1 9(1.1) or MAT to hPD-1 9(1.2), or MAT to hPD-1 9(2.1), or MAT to hPD-1 9(2.2);

(J) (1) three CDRH domains of VH MAT 10 to PD-1;

(2) three CDRs of _{the L} domain of VL MAT 10 to PD-1;

(3) three CDRH domains of VH MAT 10 to PD-1 and three CDR _L domains of VL MAT 10 to PD-1;

(K) (1) three CDRH domains of VH MAT 11 to PD-1;

(2) three CDRL domains of VL MAT 11 to PD-1;

(3) three CDRH domains of VH MAT 11 to PD-1 and three CDR _L domains of VL MAT 11 to PD-1;

(L) (1) three CDRH domains of VH MAT 12 to PD-1;

(2) three CDRL domains of VL MAT 12 to PD-1;

(3) three CDRH domains of VH MAT 12 to PD-1 and three CDR _L domains of VL MAT 12 to PD-1;

(M) (1) three CDRH domains of VH MAT 13 to PD-1;

(2) three CDRs of _{the L} domain of VL MAT 13 to PD-1;

(3) three CDRH domains of VH MAT 13 to PD-1 and three CDR _L domains of VL MAT 13 to PD-1;

(N) (1) three CDRH domains of VH MAT 14 to PD-1;

(2) three CDRs of _{the L} domain of VL MAT 14 to PD-1;

(3) three CDRH domains of VH MAT 14 to PD-1 and three CDR _L domains of VL MAT 14 to PD-1;

(O) (1) three CDRH domains of VH MAT 15 to PD-1;

(2) three CDRL domains of VL MAT 15 to PD-1;

(3) three CDRH domains of VH MAT 15 to PD-1 and three CDR _L domains of VL MAT 15 to PD-1;

(4) VH domain from VH1 MAT to hPD-1 15;

(5) VL domain of VL1 MAT to hPD-1 15;

(6) the VH and VL domains of anti-hPD-1 MAT 15; or which binds to or competes for binding to the same epitope as anti-PD-1 MAT 1, anti-PD-1 MAT 2, anti-PD-1 MAT 3, anti-PD-1 MAT 4, anti-PD-1 MAT 5, anti-PD-1 MAT 6, anti-PD-1 MAT 7, anti-PD-1 MAT 8, anti-PD-1 MAT 9, anti-PD-1 MAT 10, anti-PD-1 MAT 11, anti-PD-1 MAT 12, anti-PD-1 MAT 13, anti-PD-1 MAT 14

- 23 047726 or MAT 15 to PD-1.

A. Monoclonal antibody to human PD-1 1.

1. Mouse monoclonal antibody to human PD-1 1.

The amino acid sequence of the VH domain of MAT to PD-1 1 (SEQ ID NO: 69) is shown below (CDR _H residues are underlined).

DVQLQESGPG RVKPSQSLSL TCTVTGFSIT NDYAWNWIRQ FPGNKLEWMG

HITYSGSTSY NPSLKSRISI TRDTSKNHFF LQLSSVTPED TATYYCARDY

GSGYPYTLDY WGQGTSVTVS S

CDRhI MAT to PD-1 1 (SEQ GO NO: 71): ND YAWN

CDR _H 2 MAT to PD-1 1 (SEQ ID NO: 72): HITYSGSTSYNPSLKS

CDR _H 3 MAT to PD-1 1 (SEQ ID NO: 73): DYGSGYPYTLDY

An exemplary polynucleotide that encodes the VH domain of MAT to PD-1 1 comprises the sequence set forth in SEQ ID NO: 70 (nucleotides encoding CDRH residues are underlined):

cagatccagt gatgtgcagc ttcaggagtc gggacctggc cgggtgaaac cttctcagtc tctgtccctc acctgcactg tcactggctt ctcaatcacc aatgattatg cctggaactg gatccgacag tttccaggaa acaaactgga gtggatgggc cacataacct acagtggcag cactagctac a acccatctc tcaaaagtcg aatctctatc actcgggaca catccaagaa ccacttcttc ctgcagttga gttctgtgac tcctgaggac acagccacat attactgtgc a a g agattac ggtagtggct acccctatac tttggactac tggggtcaag gtacctcagt caccgtctcc tcc

The amino acid sequence of the VL domain of MAT to PD-1 1 (SEQ ID NO: 74) is shown below (CDR _l residues are underlined):

QIVLTQSPAL MSASPGEKVT MTCSATSIVS YVYWYQQKPG SSPQPWIYLT

SNLASGVPAR FSGSGSGTSY SLTISSMEAE DAATYYCQQW SDNPYTFGGG TKLEIK

CDRlI MAT to PD-1 1 (SEQ ID NO: 76): SATSIVSYVY

CDR _l 2 MAT to PD-1 1 (SEQ ID NO: 77): LTSNLAS

CDR _l 3 MAT to PD-1 1 (SEQ ID NO: 78): QQWSDNPYT

An exemplary polynucleotide that encodes the VL domain of MAT to PD-1 1 comprises the sequence set forth in SEQ ID NO: 75 (nucleotides encoding CDR _L residues are underlined):

caaattgttc tcacccagtc tccagcactc atgtctgcat ctccagggga gaaggtcacc atgacctgca gtgccacctc aattgtaagt tacgtttact ggtaccagca gaagcctgga tcctcccccc aaccctggat ttatctcaca tccaacctgg cttctggagt ccctgctcgc ttcagtggca g tgggtctgg gacctcttac tctctcacaa tcagcagcat ggaggctgaa gatgctgcca cttattactg ccagcagtgg agtgataacc cgtacacgtt cggagggggg accaagctgg aaataaaa

2. Humanization of the monoclonal antibody to PD-1, mAb to PD-1 1, to obtain mAb to hPD-1 1.

The above-described murine anti-human PD-1 antibody, anti-PD-1 mAb 1, was humanized and further deimmunized after identification of antigenic epitopes to demonstrate the feasibility of humanizing an anti-human PD-1 antibody to reduce its antigenicity when administered to a human recipient. The humanization resulted in one humanized VH domain, designated herein as VH1 anti-hPD-1 mAb 1, and one humanized VL domain, designated herein as VL1 anti-hPD-1 mAb 1. Accordingly, an antibody comprising humanized VL domains paired with a humanized VH domain is referred to as anti-hPD-1 mAb 1.

The amino acid sequence of the VH domain of VH1 MAT to hPD-1 1 (SEQ ID NO: 79) is shown below (CDRH residues are underlined):

DVQLQESGPG LVKPSQTLSL TCTVSGFSIS NDYAWNWIRQ PPGKGLEWIG HITYSGSTSY NPSLKSRLTI TRDTSKNQFV LTMTNMDPVD TATYYCARDY GSGYPYTLDY WGQGTTVTVS S

An exemplary polynucleotide that encodes VH1 MAT to hPD-1 1 contains the sequence,

- 24 047726 given in SEQ ID NO: 80 (nucleotides encoding CDRH residues are underlined): gacgtacagc tccaggaaag tggcccaggt ctggtgaagc catcccagac actgagcctg acttgcaccg tgagtggctt ctccatctca aatgactacg cctggaattg gattaggcag cctcccggta aagggctgga gtggatcggc cacatcacat acagcggctc cacatcatat aatcccagtc tgaagagсcg tcttaccatt actcgcgaca ctagtaagaa ccagtttgtt ctgaccatga ccaacatgga ccctgtggat actgcaacat actattgtgc tcgagattat ggttctggtt acccttatac actcgactac tggggacagg gaaccactgt gaccgtgagc tcc

The amino acid sequence of the VL domain from the VL1 MAT to hPD-1 1 (SEQ ID NO: 81) is shown below (CDRH residues are underlined): EIVLTQSPAT LSVSPGEKVT ITCSATSIVS YVYWYQQKPG QAPQPLIYLT SNLASGIPAR FSGSGSGTDF TLTISSLEAE DAATYYCQQW SDNPYTFGGG TKVEIK

An exemplary polynucleotide that encodes the VL1 mAb to hPD-1 1 comprises the sequence set forth in SEQ ID NO: 82 (nucleotides encoding CDRH residues are underlined): gaaatcgttc tgacccagag cccagcaacc ctgtctgtct cccccggaga aaaggtcacc attacttgct ctgctacttc tatcgtgtcc tacgtgtact ggtatcagca gaagcccggt caggctcccc agccattgat atatctgacc agcaacctgg cttctggtat cccagctcgt ttttccggta gcgggtccgg gactgatttc actttgacta tcagctctct ggaggcagaa gacgccgcca cctattattg tcaacagtgg tcagacaatc catacacttt tggcggtggc accaaagtcg aaataaag

B. Monoclonal antibody to human PD-1 2.

1. Mouse monoclonal antibody to human PD-1 2.

The amino acid sequence of the VH domain of MAT 2 to PD-1 (SEQ ID NO: 83) is shown below (CDRH residues are underlined).

DVQLVESGGG LVQPGGSRKL SCAASGFVFS SFGMHWVRQA PEKGLEWVAY ISSGSMSISY ADTVKGRFTV TRDNAKNTLF LQMTSLRSED TAIYYCASLS DYFDYWGQGT TLTVSS

CDRhI MAT 2 to PD-1 (SEQ GO NO: 85): SFGMH

CDR _H 2 MAT 2 to PD-1 (SEQ ID NO: 86): YISSGSMSISYADTVKG

CDR _H 3 MAT 2 to PD-1 (SEQ GO NO: 87): LSDYFDY

An exemplary polynucleotide that encodes the VH domain of MAT 2 to PD-1 comprises the sequence set forth in SEQ ID NO: 84 (nucleotides encoding CDRH residues are underlined): gatgtgcagc tcgtggagtc tgggggaggc ttagtgcagc ctggagggtc ccggaaactc tcctgtgcag cctctggatt cgttttcagt agctttggaa tgcactgggt tcgtcaggct ccagagaagg ggctggagtg ggtcgcatac atcagtagtg gcagtatgag catttcctat gcagacacag tgaagggccg attcaccgtc accagagaca atgccaagaa caccctgttc ctgcaaatga ccagtctaag gtctgaggac acggccattt attactgtgc atccctgagt gactactttg actactgggg ccaaggcacc actctcacag tctcctcc

The amino acid sequence of the VL domain of MAT 2 to PD-1 (SEQ ID NO: 88) is shown below (CDR _l residues are underlined):

DWMSQTPLS LPVSLGDQAS ISCRSSQSLV HSTGNTYLHW YLQKPGQSPK LLIYRVSNRF SGVPDRFSGS GSGTDFTLKI SRVEAEDLGV FFCSQTTHVP WTFGGGTKLE IK

CDRlI MAT 2 to PD-1 (SEQ ID NO: 90): RSSQSLVHSTGNTYLH

CDR _l 2 MAT 2 to PD-1 (SEQ ID NO: 91): RVSNRFS

CDR _l 3 MAT 2 to PD-1 (SEQ ID NO: 92): SQTTHVPWT

An exemplary polynucleotide that encodes the VL domain of MAT 2 to PD-1 comprises the sequence set forth in SEQ ID NO: 89 (nucleotides encoding CDR _l residues are underlined):

- 25 047726 gatgttgtga tgtcccaaac tccactctcc ctgcctgtca gtcttggaga tcaagcctcc atctcttgca gatetagtea gagccttgtt cacagtactg

Igaaacaccta tttacattgg tacctgcaga agccaggcca gtctccaaag ctcctgatct acagggtttc taaccgattt tctggggtcc ccgacaggtt cagtggcagt ggatcaggga cagatttcac actcaagatc agtagagtgg aggetgagga tctgggagtt tttttctgct ctcaaactac acatgt tccg tggacgttcg gtggaggcac caagctggaa atcaaa

2. Humanization of the anti-human PD-1 monoclonal antibody, anti-PD-1 mAb 2, to produce anti-hPD-1 mAb 2.

The above-described murine monoclonal antibody to human PD-1, anti-PD-1 Mab 2, was humanized and further deimmunized after identification of antigenic epitopes to demonstrate the feasibility of humanizing an anti-human PD-1 antibody to reduce its antigenicity when administered to a human recipient. The humanization resulted in one humanized VH domain, designated herein as VH1 anti-hPD-1 Mab 2, and one humanized VL domain, designated herein as VL1 anti-hPD-1 Mab 1. Accordingly, any antibody comprising humanized VL domains paired with a humanized VH domain is referred to as an anti-hPD-1 Mab 2.

The amino acid sequence of the VH domain of VH1 MAT to hPD-1 2 (SEQ ID NO: 93) is shown below (CDRH residues are underlined):

EVQLVESGGG LVQPGGSLRL SCAASGFVFS SFGMHWVRQA PGKGLEWVAY ISSGSMSISY ADTVKGRFTI SRDNAKNTLY LQMNSLRTED TALYYCASLS DYFDYWGQGT TVTVSS

An exemplary polynucleotide that encodes a VH1 mAb to hPD-1 2 comprises the sequence set forth in SEQ ID NO: 94 (nucleotides encoding CDRH residues are underlined): gaagtgeaat tggttgagag tggtggtggc etggtgeage caggtggaag tctgcggttg tcctgtgcag caagcggatt tgtgttcagc tcttttggga tgcattgggt gcgccaggct cccggcaagg gtctcgagtg ggtagcatac atctccagcg ggtccatgtc tattagttat gccgacacag tgaaaggcag gtttactatc tcccgtgaca atgeaaaaa cacactgtac etgeaaatga atageetgeg caccgaggac accgccttgt actactgcgc ttccctgtct gattaetteg actactgggg tcagggcaca actgtgacag tttcttcc

The amino acid sequence of the VL domain of VL1 MAT to hPD-1 2 (SEQ ID NO: 95) is shown below (CDRH residues are underlined):

DWMTQSPLS LPVTLGQPAS ISCRSSQSLV HSTGNTYLHW YLQKPGQSPQ

LLIYRVSNRF SGVPDRFSGS GSGTDFTLKI SRVEAEDVGV YYCSQTTHVP WTFGQGTKLE IK

An exemplary polynucleotide that encodes the VL1 MAT to hPD-1 2 comprises the sequence set forth in SEQ ID NO: 96 (nucleotides encoding CDRH residues are underlined): gacgttgtga tgacacagtc accactgagt ctgccagtta ccctgggcca gccagccagt atttcttgtc ggagttcaca gagtctggta cattccacag

Igaaatacata tctccattgg tacctgcaaa aaccagggca gagcccccag etgetgattt atagagtgte taategattt tctggcgtgc cagatcggtt cagcggcagc gggtctggca ctgatttcac aetgaaaate tctagggtgg aggcagagga cgtaggcgtt tactactgta gtcagaccac ccatgtacc c tggacttttg gccaaggtac taagctggaa atcaag

C. Murine monoclonal antibody to human PD-1 3.

The amino acid sequence of the VH domain of MAT 3 to PD-1 (SEQ ID NO: 97) is shown below (CDRH residues are underlined).

QVQLQQSGAE LVRPGASVTL SCKASGYTFT DYVMHWVKQT PVHGLEWIGT

IDPETGGTAY NQKFKGKAIL TADKSSNTAY MELRSLTSED SAVYYFTREK

ITTIVEGTYW YFDVWGTGTT VTVSS

CDRhI MAT 3 to PD-1 (SEQ ID NO: 99): DYVMH

CDR _H 2 MAT 3 to PD-1 (SEQ ID NO: 100): TIDPETGGTAYNQKFKG

CDR _H 3 MAT 3 to PD-1 (SEQ ID NO: 101): EKITTIVEGTYWYFDV

An exemplary polynucleotide that encodes the VH domain of MAT 3 to PD-1 contains the sequence

- 26 047726 the sequence shown in SEQ ID NO: 98 (nucleotides encoding CDRH residues are underlined):

caggttcaac tgcaacagtc tggggctgag ctggtgaggc ctggggcttc agtgacgctg tcctgcaagg cttcgggcta cacatttact gactatgtaa tgcactgggt gaagcagaca cctgtgcatg gcctggaatg gattggaact attgatcctg aaactggtgg tactgcctac aatcagaag t tcaagggcaa ggccatactg actgcagaca agtcctccaa cacagcctac atggagctcc gcagcctgac atctgaggac tctgccgtct attactttac aagagagaag attactacga tagtagaggg gacatactgg tacttcgatg tctggggс a c agggaccacg gtcaccgtct cctca

The amino acid sequence of the VL domain of MAT 3 to PD-1 (SEQ ID NO: 102) is shown below (CDR _L residues are underlined):

DVLLTQTPLS LPVSLGDQAS ISCRSSQNIV HSNGDTYLEW YLQKPGQSPK LLIYKVSNRF SGVPDRFSGS GSGTDFTLKI SRVEAEDLGV YYCFQGSHLP YTFGGGTKLE IK

CDRlI MAT 3 to PD-1 (SEQ Ш NO: 104): RSSQNIVHSNGDTYLE

CDR _l 2 MAT 3 to PD-1 (SEQ Ш NO: 105): KVSNRFS

CDR _l 3 MAT 3 to PD-1 (SEQ Ш NO: 106): FQGSHLPYT

An exemplary polynucleotide that encodes the VL domain of MAT 3 to PD-1 comprises the sequence set forth in SEQ ID NO: 103 (nucleotides encoding CDR _L residues are underlined): gatgttttgc tgacccaaac tccactctcc ctgcctgtca gtcttggaga tcaagcctcc atctcttgca gatetagtea gaacattgta catagtaatg gagacaccta tttggaatgg tacctgcaga aaccaggcca gtctccaaag ctcctgatct ataaagtttc caaccgattt tctggggtcc cagacaggtt cagtggcagt gggtcaggga cagattttac actcaaaatc agcagagtgg aggetgagga tctgggagtt tattaetget ttcaaggttc acatcttccg tacacgttcg gaggggggac caagctggaa ataaaa

D. Murine monoclonal antibody to human PD-1 4.

The amino acid sequence of the VH domain of MAT 4 to PD-1 (SEQ ID NO: 107) is shown below (CDRH residues are underlined).

DVQLVESGGG LVQPGGSRKL SCAASGFVFS SFGMHWVRQA PEKGLEWVAY ISSGSMSISY ADTVKGRFTV TRDNAKNTLF LQMTSLRSED TAIYYCASLT DYFDYWGQGT TLTVSS

CDRhI MAT 4 to PD-1 (SEQ GO NO: 109): SFGMH

CDR _H 2 MAT 4 to PD-1 (SEQ ID NO: 110): YISSGSMSISYADTVKG

CDR _H 3 MAT 4 to PD-1 (SEQ GO NO: 111): LTDYFDY

An exemplary polynucleotide that encodes the VH domain of MAT 4 to PD-1 comprises the sequence set forth in SEQ ID NO: 108 (nucleotides encoding CDRH residues are underlined): gatgtgeage tegtggagte tgggggaggc ttagtgeage ctggagggtc ccggaaactc tcctgtgcag cctctggatt cgttttcagt agctttggaa tgcactgggt tegteagget ccagagaagg ggctggagtg ggtcgcatat attagtagtg gcagtatgag tatttcctat gcagacacag tgaagggccg attcaccgtc accagagaca atgccaagaa caccctgttc etgeaaatga ccagtctaag gtctgaggac acggccattt attactgtgc atccctgact gactactttg actactgggg ccaaggcacc actctcacag tctcctca

The amino acid sequence of the VL domain of MAT 4 to PD-1 (SEQ ID NO: 112) is shown below (CDR _L residues are underlined):

- 27 047726

DWMSQTPLS LPVSLGDQAS ISCRSSQSLV HSTGNTYFHW YLQKPGQSPK LLIYRVSNRF SGVPDRFSGS GSGTDFTLKI SRVEAEDLGV YFCSQTTHVP WTFGGGTKLE IK

CDRlI MAT 4 to PD-1 (SEQ ID NO: 114): RSSQSLVHSTGNTYFH

CDR _l 2 MAT 4 to PD-1 (SEQ ID NO: 115): RVSNRFS

CDR _l 3 MAT 4 to PD-1 (SEQ ID NO: 116): SQTTHVPWT

An exemplary polynucleotide that encodes the VL domain of MAT 4 to PD-1 comprises the sequence set forth in SEQ ID NO: 113 (nucleotides encoding CDR _L residues are underlined): gatgttgtga tgtcccaaac tccactctcc ctgcctgtca gtcttggaga tcaagcctcc atctcctgca gatetagtea gagccttgtt cacagtactg gaaacaccta tttccattgg tacctgcaga agccaggcca gtctccaaag ctcctgatct acagggtttc taaccgattt tctggggtcc ccgacaggtt cagtggcagt ggatcaggga cagatttcac actcaagatc agcagagtgg aggetgagga tctgggagtt tatttetget ctcaaactac acatgttccg tggacgttcg gtggaggcac caagctggaa atcaaa

E. Murine monoclonal antibody to human PD-1 5.

The amino acid sequence of the VH domain of MAT 5 to PD-1 (SEQ ID NO: 117) is shown below (CDR _H residues are underlined).

QVQLQQPGVE LVRPGASVKL SCKASGYSFT AYWMNWMKQR PGQGLEWIGV IHPSDSETWL NQKFKDKATL TVDKSSSTAY MQLISPTSED SAVYYCAREH YGSSPFAYWG QGTLVTVSA

CDRhI MAT 5 to PD-1 (SEQ ID NO: 119): AYWMN

CDR _H 2 MAT 5 to PD-1 (SEQ ID NO: 120): VIHPSDSETWLNQKFKD

CDR _H 3 MAT 5 to PD-1 (SEQ ID NO: 121): EHYGSSPFAY

An exemplary polynucleotide that encodes the VH domain of MAT 5 to PD-1 comprises the sequence set forth in SEQ ID NO: 118 (nucleotides encoding CDRH residues are underlined): caggtccaac tgcagcagcc tggggttgaa etggtgagge etggagette agtgaagctg teetgeaagg ettetggeta ctccttcacc gcctactgga tgaactggat gaaacagagg cctggacaag geettgagtg gattggcgtg attcatcctt ccgatagtga aacttggtta aatcagaagt tcaaggac a a ggccacattg actgtagaca aatcctccag cacagcctac atgcaactca tcagcccgac atctgaggac tetgeggtet attactgtgc aagagagcac taeggtagta gcccgtttgc ttactggggc caagggactc tggtcactgt ctctgca

The amino acid sequence of the VL domain of MAT 5 to PD-1 (SEQ ID NO: 122) is shown below (CDR _L residues are underlined): DIVLTQSPAS LAVSLGQRAT ISCRANESVD NYGMSFMNWF QQKPGQPPKL LIYAASNQGS GVPARFSGSG SGTDFSLNIH PMEEDDTAMY FCQQSKEVPY TFGGGTKLEI K

CDRlI MAT 5 to PD-1 (SEQ ID NO: 124): RANESVDNYGMSFMN

CDR _l 2 MAT 5 to PD-1 (SEQ ID NO: 125): AASNQGS

CDR _l 3 MAT 5 to PD-1 (SEQ ID NO: 126): QQSKEVPYT

An exemplary polynucleotide that encodes the VL domain of MAT 5 to PD-1 comprises the sequence set forth in SEQ ID NO: 123 (nucleotides encoding CDR _L residues are underlined):

- 28 047726 gacattgtgc tgacccaatc tccagcttct ttggctgtgt ctctagggca gagggccacc atctcctgca gagccaacga aagtgttgat aattatggca tgagttttat gaactggttc caacagaaac caggacagcc acccaaactc ctcatctatg ctgcatccaa ccaaggatcc gg ggtccctg ccaggtttag tggcagtggg tctgggacag atttcagcct caacatccat cctatggagg aggatgatac tgcaatgtat ttctgtcagc aaagtaagga ggttccgtac acgttcggag gggggaccaa gctggaaata aaa

F. Mouse monoclonal antibody to human PD-1 6.

The amino acid sequence of the VH domain of MAT 6 to PD-1 (SEQ ID NO: 127) is shown below (CDR _H residues are underlined).

EVKLVESGGG LVNPGGSLKL SCAASGFTFS SYGMSWVRQT PEKRLEWVAT ISGGGSDTYY PDSVKGRFTI SRDNAKNNLY LQMSSLRSED TALYYCARQK ATTWFAYWGQ GTLVTVST

CDRhI MAT 6 to PD-1 (SEQ GO NO: 129): SYGMS CDR _H 2 MAT 6 to PD-1 (SEQ ID NO: 130): TISGGGSDTYYPDSVKG CDR _H 3 MAT 6 to PD-1 (SEQ ID NO: 131): QKATTWFAY

An exemplary polynucleotide that encodes the VH domain of MAT 6 to PD-1 comprises the sequence set forth in SEQ ID NO: 128 (nucleotides encoding CDRH residues are underlined):

gaaatcgtac tcacccagtc acctgcaacc ctttctctga gccccggtga acgtgccact ctcagctgca gagcaagtga gagtgtggac aattacggca tgtccttcat gaactggttt cagcagaagc ctgggcagcc acctaagctg ctcatccacg ccgcctctaa ccgcggatct ggggt gcctt cacgtttttc tggatcagga agtggcactg acttcaccct tacaatcagc tctctggagc cagaggactt tgccgtctat ttctgccagc aatctaaaga ggtgccctat acttttggtg gcgggaccaa ggttgagatc aaa

The amino acid sequence of the VL domain of MAT 6 to PD-1 (SEQ ID NO: 132) is shown below (CDR _L residues are underlined):

DIVLTQSPAS LAVSLGQRAT ISCRASESVD NYGISFMNWF QQKPGQPPKL

LIYPASNQGS GVPARFSGSG SGTDFSLNIH PMEEDDAAMY FCQQSKEVPW TFGGGTKLEI K

CDRlI MAT 6 to PD-1 (SEQ ID NO: 134): RASESVDNYGISFMN

CDR _l 2 MAT 6 to PD-1 (SEQ ID NO: 135): PASNQGS

CDR _l 3 MAT 6 to PD-1 (SEQ Ш NO: 136): QQSKEVPWT

An exemplary polynucleotide that encodes the VL domain of MAT 6 to PD-1 comprises the sequence set forth in SEQ ID NO: 133 (nucleotides encoding CDR _L residues are underlined):

gacattgtgc tgacccaatc tccagcttct ttggctgtgt ctctagggca gagggccacc atctcctgca gagccagcga aagtgttgat aattatggca ttagttttat gaactggttc caacagaaac caggacagcc acccaaactc ctcatctatc ctgcatccaa ccaaggatcc ggggtccctg ccaggtttag tggcagtggg tctgggacag acttcagcct caacatccat cctatggagg aggatgatgc tgcaatgtat ttctgtcagc aaagtaagga ggttccgtgg acgttcggtg gaggcaccaa gctggaaatc aaa

G. Monoclonal antibody to human PD-1 7.

1. Murine monoclonal antibody to human PD-1 7.

The amino acid sequence of the VH domain of MAT 7 to PD-1 (SEQ ID NO: 137) is shown below (CDRH residues are underlined).

- 29 047726

QVQLQQPGAE LVRPGASVKL SCKASGYSFT SYWMNWVKQR PGQGLEWIGV IHPSDSETWL DQKFKDKATL TVDKSSTTAY MQLISPTSED SAVYYCAREH YGTSPFAYWG QGTLVTVSS

CDRhI MAT 7 to PD-1 (SEQ ID NO: 139): SYWMN

CDR _H 2 MAT 7 to PD-1 (SEQ ID NO: 140): VIHPSDSETWLDQKFKD

CDR _H 3 MAT 7 to PD-1 (SEQ ID NO: 141): EHYGTSPFAY

An exemplary polynucleotide that encodes the VH domain of MAT 7 to PD-1 comprises the sequence set forth in SEQ ID NO: 138 (nucleotides encoding CDR _H residues are underlined): gaggtccaac tgcagcagcc tggggctgaa ctggtgaggc ctggagcttc agtgaagctg tcctgcaagg cttctggcta ctccttcacc agctactgga tgaactgggt gaagcagagg cctggacaag gccttgagtg gattggcgtg attcatcctt ccgatagtga aacttggtta gatcagaagt tcaaggacaa ggccacattg actgtagaca aatcctccac cacagcctac atgcaactca tcagcccgac atctgaggac tctgcggtct attactgtgc aagggagcac tacggtacta gcccgtttgc ttactggggc caagggactc tggtcactgt gtcttcc

The amino acid sequence of the VL domain of MAT 7 to PD-1 (SEQ ID NO: 142) is shown below (CDR _L residues are underlined): DIVLTQSPAS LAVSLGQRAT ISCRANESVD NYGMSFMNWF QQKPGQPPKL LIHAASNQGS GVPARFSGSG FGTDFSLNIH PMEEDDAAMY FCQQSKEVPY TFGGGTKLEI K

CDRlI MAT 7 to PD-1 (SEQ ID NO: 144): RANESVDNYGMSFMN

CDR _l 2 MAT 7 to PD-1 (SEQ ID NO: 145): AASNQGS

CDR _l 3 MAT 7 to PD-1 (SEQ ID NO: 146): QQSKEVPYT

An exemplary polynucleotide that encodes the VL domain of MAT 7 to PD-1 comprises the sequence set forth in SEQ ID NO: 143 (nucleotides encoding CDR _L residues are underlined):

gacattgtgc tgacccaatc tccagcttct ttggctgtgt ctctagggca gagggccacc atctcctgca gagccaacga aagtgttgat aattatggca tgagttttat gaactggttc caacagaaac caggacagcc acccaaactc ctcatccatg ctgcatccaa ccaaggatcc ggggtccctg ccaggtttag tggcagtggg tttgggacag acttcagcct caacatccat cctatggagg aggatgatgc tgcaatgtat ttctgtcagc aaagtaagga ggttccgtac acgttcggag gggggaccaa gctggaaata aaa

2. Humanization of monoclonal antibody to human PD-1 7 to obtain mAb to hPD-1 7.

The above-described murine monoclonal antibody to human PD-1, anti-PD-1 MAb 7, was humanized and further deimmunized after identification of antigenic epitopes to demonstrate the feasibility of humanizing an anti-human PD-1 antibody to reduce its antigenicity when administered to a human recipient. The humanization resulted in two humanized VH domains, designated herein as anti-hPD-1 MAb 7 VH1 and anti-hPD-1 MAb 7 VH2, and three humanized VL domains, designated herein as anti-hPD-1 MAb 7 VL1, anti-hPD-1 MAb 7 VL2, and anti-hPD-1 MAb 7 VL3. Any of the humanized VL domains may be paired with any of the humanized VH domains. Accordingly, any antibody comprising one of the humanized VL domains paired with a humanized VH domain is generally referred to as an anti-hPD-17 mAb, and specific combinations of humanized VH/VL domains are referred to by reference to specific VH/VL domains, for example, a humanized antibody comprising a VH1 anti-hPD-17 mAb and a VL2 anti-hPD-11 mAb is specifically referred to as an anti-hPD-17 mAb(1.2).

The amino acid sequence of the VH domain of the VH1 MAT to hPD-1 7 (SEQ ID NO: 147) is shown below (CDRH residues are underlined):

QVQLVQSGAE VKKPGASVKV SCKASGYSFT SYWMNWVRQA PGQGLEWIGV

IHPSDSETWL DQKFKDRVTI TVDKSTSTAY MELSSLRSED TAVYYCAREH

YGTSPFAYWG QGTLVTVSS

An exemplary polynucleotide that encodes VH1 MAT to hPD-1 7 contains the sequence,

- 30 047726 given in SEQ ID NO: 148 (nucleotides encoding CDRH residues are underlined): caagttcaat tggtacagag cggggcagag gtgaagaaac ccggcgccag tgttaaggtg tcctgcaaag ccagcggtta cagctttaca agctattgga tgaattgggt gcgtcaagca ccaggcagg gtctggaatg gattggggtg atacatcctt ctgacagcga aacatggttg gaccagaaat ttaaagatcg tgtgacaatt acagtcgata agtccacaag cactgcttac atggaactct ccagcttgcg gtccgaggac accgctgtgt attattgcgc cagagagcac tacggcacat caccttttgc atactggggc cagggaactc tcgtaaccgt atcctcc

The amino acid sequence of the VH domain of the VH2 MAT to hPD-1 7 (SEQ ID NO: 149) is shown below (CDRH residues are underlined):

QVQLVQSGAE VKKPGASVKV SCKASGYSFT SYWMNWVRQA PGQGLEWAGV

IHPSDSETWL DQKFKDRVTI TVDKSTSTAY MELSSLRSED TAVYYCAREH YGTSPFAYWG QGTLVTVSS

An exemplary polynucleotide that encodes a VH2 mAb to hPD-17 comprises the sequence set forth in SEQ ID NO: 150 (nucleotides encoding CDRH residues are underlined): caagttcaat tggtacagag cggggcagag gtgaagaaac ccggcgccag tgttaaggtg tcctgcaaag ccagcggtta cagctttaca agctattgga tgaattgggt gcgtcaagca ccagggcagg gtctggaatg ggctggggtg atacatcctt ctgacagcga aacatggttg gaccagaaat ttaaagatcg tgtgacaatt acagtcgata agtccacaag cactgcttac atggaactct ccagcttgcg gtccgaggac accgctgtgt attattgcgc cagagagcac tacggcacat caccttttgc atactggggc cagggaactc tcgtaaccgt atcctcc

The amino acid sequence of the VL domain of the hPD-1 MAT VL1 7 (SEQ ID NO: 151) is shown below (CDRH residues are underlined):

EIVLTQSPAT LLSLSPGERAT LSCRANESVD NYGMSFMNWF QQKPGQPPKL

LIHAASNQGS GVPSRFSGSG SGTDFTLTIS SLEPEDFAVY FCQQSKEVPY TFGGGTKVEI K

An exemplary polynucleotide that encodes the hPD-1 7 MAT VL1 comprises the sequence set forth in SEQ ID NO: 152 (nucleotides encoding CDRH residues are underlined): gaaatcgtac tcacccagtc acctgcaacc ctttctctga gccccggtga acgtgccact ctcagctgca gagcaaatga gagtgtggac aattacggca tgtccttcat gaactggttt cagcagaagc ctgggcagcc acctaagctg ctcatccacg ccgcctctaa ccagggatct ggggtgcctt cacgtttttc tggatcagga agtggcactg acttcaccct tacaatcagc tctctggagc cagaggactt tgccgtctat ttctgccagc aatctaaaga ggtgccctat acttttggtg gcgggaccaa ggttgagatc aaa

The amino acid sequence of the VL domain of the VL2 MAT to hPD-1 7 (SEQ ID NO: 153) is shown below (CDRH residues are underlined):

EIVLTQSPAT LSLSPGERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL

LIHAASNQGS GVPSRFSGSG SGTDFTLTIS SLEPEDFAVY FCQQSKEVPY TFGGGTKVEI K

An exemplary polynucleotide that encodes the VL2 mAb to hPD-17 has the sequence set forth in SEQ ID NO: 154 (nucleotides encoding CDRH residues are underlined): gaaatcgtac tcacccagtc acctgcaacc ctttctctga gccccggtga acgtgccact ctcagctgca gagcaagtga gagtgtggac aattacggca tgtccttcat gaactggttt cagcagaagc ctgggcagcc acctaagctg ctcatccacg ccgcctctaa ccagggatct ggggtgcctt cacgtttttc tggatcagga agtggcactg acttcaccct tacaatcagc tctctggagc cagaggactt tgccgtctat ttctgccagc aatctaaaga ggtgccctat acttttggtg gcgggaccaa ggttgagatc aaa

The amino acid sequence of the VL domain from VL3 MAT to hPD-1 7 (SEQ ID NO: 155) is shown below (CDR _H residues are underlined):

- 31 047726

EIVLTQSPAT LSLSPGERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL LIHAASNRGS GVPSRFSGSG SGTDFTLTIS SLEPEDFAVY FCQQSKEVPY TFGGGTKVEI K

An exemplary polynucleotide that encodes the VL3 mAb to hPD-17 comprises the sequence set forth in SEQ ID NO: 156 (nucleotides encoding CDRH residues are underlined): gaaatcgtac tcacccagtc acctgcaacc ctttctctga gccccggtga acgtgccact ctcagctgca gagcaagtga gagtgtggac aattacggca tgtccttcat gaactggttt cagcagaagc ctgggcagcc acctaagctg ctcatccacg ccgcctctaa ccgcggatct ggggtgcctt cacgtttttc tggatcagga agtggcactg acttcaccct tacaatcagc tctctggagc cagaggactt tgccgtctat ttctgccagc aatctaaaga ggtgccctat acttttggtg gcgggaccaa ggttgagatc aaa

CDR _l 1 of the VL domain from VL2 of MAT to hPD-1 7 and VL3 of MAT to hPD-1 7 contains a substitution of asparagine for the amino acid serine and contains the amino acid sequence: ((SEQ ID NO:

157), the substituted serine is underlined). According to the present invention, it is envisaged that a similar substitution can be included in any of the CDR _L 1 domains of MAT 7 to PD-1 described above. In addition, CDR _L 2 of the VL VL3 domain of MAT 7 to hPD-1 comprises a substitution of glutamine with the amino acid arginine and comprises the amino acid sequence: ^LLXX 5 ^XX ((SEQ ID NO: 158), the substituted arginine is underlined). According to the present invention, it is envisaged that a similar substitution can be included in any of the CDR _L 2 domains of MAT 7 to PD-1 described above.

H. Murine monoclonal antibody to human PD-1 8.

The amino acid sequence of the VH domain of MAT 8 to PD-1 (SEQ ID NO: 159) is shown below (CDRH residues are underlined).

EGQLQQSGPE LVKPGASVKI SCKASGYTFT DYYMNWVKQN HGKSLEWIGD

INPKNGDTHY NQKFKGEATL TVDKSSTTAY MELRSLTSED SAVYYCASDF DYWGQGTTLT VSS

CDRhI MAT 8 to PD-1 (SEQ ID NO: 161): DYYMN

CDR _H 2 MAT 8 to PD-1 (SEQ ID NO: 162): DINPKNGDTHYNQKFKG

CDRh3 MAT 8 to PD-1 (SEQ ID NO: 163): DFDY

An exemplary polynucleotide that encodes the VH domain of MAT 8 to PD-1 comprises the sequence set forth in SEQ ID NO: 160 (nucleotides encoding CDRH residues are underlined): gagggccagc tgcaacaatc tggacctgag ctggtgaagc ctggggcttc agtgaagata tcctgtaagg cttctggata cacgttcact gactactaca tgaactgggt gaagcagaac catggaaaga gccttgagtg gattggagat attaatccta aaaatggtga cactcactac aaccagaagt tcaagggcga ggccacattg actgtagaca agtcctccac cacagcctac atggagctcc gcagcctgac atctgaggac tctgcagtct attactgtgc gagcgatttt gactactggg gccaaggcac cactctcaca gtctcctcc

The amino acid sequence of the VL domain of MAT 8 to PD-1 (SEQ ID NO: 164) is shown below (CDRL residues are underlined):

DWMTQTPLS LPVGLGDQAS ISCRSSQTLV YSNGNTYLNW FLQKPGQSPK

LLIYKVSNRF SGVPDRFSGS GSGTDFTLKI SRVEAEDLGV YFCSQSTHVP FTFGSGTKLE IK

CDRlI MAT 8 to PD-1 (SEQ ID NO: 166):

CDR _l 2 MAT 8 to PD-1 (SEQ ID NO: 167):

CDR _l 3 MAT 8 to PD-1 (SEQ ID NO: 168):

RSSQTLVYSNGNTYLN

KVSNRFS

SQSTHVPFT

An exemplary polynucleotide that encodes the VL domain of MAT 8 to PD-1 comprises the sequence set forth in SEQ ID NO: 165 (nucleotides encoding CDRL residues are underlined):

- 32 047726 gatgttgtga tgacccaaac tccactctcc ctgcctgtcg gtcttggaga tcaagcctcc atctcttgca gatetagtea gacccttgta tatagtaatg gaaacaccta tttaaattgg ttcctgcaga agccaggcca gtctccaaag ctcctgatct acaaagtttc caaccgattt tctggggtcc cagacaggtt cagtggcagt ggatcaggga cagatttcac actcaagatc agcagagtgg aggetgagga tctgggagtt tatttetget ctcaaagtac acatgttcca ttcacgttcg gctcggggac aaagttggaa ataaaa

I. Monoclonal antibody to human PD-1 9.

1. Murine monoclonal antibody to human PD-1 9.

The amino acid sequence of the VH domain of MAT 9 to PD-1 (SEQ ID NO: 169) is shown below (CDR _H residues are underlined).

EVMLVESGGG LVKPGGSLKL SCAASGFTFS SYLVSWVRQT PEKRLEWVAT

ISGGGGNTYY SDSVKGRFTI SRDNAKNTLY LQISSLRSED TALYYCARYG

FDGAWFAYWGQGTLVTVSS

CDRhI MAT 9 to PD-1 (SEQ GO NO: 171): SYLVS

CDR _H 2 MAT 9 kPD-1 (SEQ ID NO: 172): TISGGGGNTYYSDSVKG

CDRh3 MAT 9 to PD-1 (SEQ ID NO: 173): YGFDGAWFAY

An exemplary polynucleotide that encodes the VH domain of MAT 9 to PD-1 comprises the sequence set forth in SEQ ID NO: 170 (nucleotides encoding CDRH residues are underlined):

gaagtgatgc tggtggagtc tgggggaggc ttagtgaagc ctggagggtc cctgaaactc tcctgtgcag cctctggatt cactttcagt agttatcttg tgtcttgggt tcgccagact ccggagaaga ggctggagtg ggtcgcaacc attagtggtg gtggtggtaa cacctactat tcagaca gtg tgaagggtcg attcaccatc tccagagaca atgccaagaa caccctgtac etgeaaatea gcagtctgag gtctgaggac aeggeettgt attactgtgc aaggtatggt ttcgacggcg cctggtttgc ttactggggc caagggactc tggtcactgt ctcttcc

The amino acid sequence of the VL domain of MAT 9 to PD-1 (SEQ ID NO: 174) is shown below (CDR _L residues are underlined):

DIQMTQSPAS LSASVGDIVT ITCRASENIY SYLAWYQQKQ EKSPQLLVYN

AKTLAAGVPS RFSGSGSGTQ FSLTINSLQP EDFGNYYCQH HYAVPWTFGG GTRLEIT

CDRlI MAT 9 to PD-1 (SEQ ID NO: 176): RASENIYSYLA

CDR _l 2 MAT 9 to PD-1 (SEQ ID NO: 177): NAKTLAA

CDR _l 3 MAT 9 to PD-1 (SEQ ID NO: 178): QHHYAVPWT

An exemplary polynucleotide that encodes the VL domain of MAT 9 to PD-1 comprises the sequence set forth in SEQ ID NO: 175 (nucleotides encoding CDR _L residues are underlined):

gacatccaga tgactcagtc tccagcctcc etatetgeat ctgtgggaga tattgtcacc atcacatgtc gagcaagtga gaatatttac agttatttag catggtatca gcagaaacag gaaaaatctc ctcagctcct ggtetataat gcaaaaacct tggcagcagg tgtgccatca aggttcagtg tggatc aggcacacag ttttctctga ccatcaacag cctgcagcct gaagattttg ggaattatta ctgtcagcat cattatgctg ttccgtggac gtteggtgga ggcaccagac tggaaatcac a

2. Humanization of monoclonal antibody to human PD-1 9 to obtain mAb to hPD-1 9.

The above-described murine monoclonal antibody to human PD-1 9 was humanized and further deimmunized after identification of antigenic epitopes to demonstrate the feasibility of humanizing an antibody to human PD-1 to reduce its antigenicity when administered to a human recipient. The humanization resulted in two humanized VH domains, designated herein as VH1 MAT to hPD-1 9 and VH2 MAT to hPD-1 9, and two humanized VL domains, designated herein as VL1 MAT to hPD-1 9 and VL2 MAT to hPD-1 9. Any of the humanized VL domains can be paired with humanized VH domains. Accordingly, any antibody comprising one of the humanized VL domains paired with a humanized

- 33 047726 humanized VH domain is generally referred to as an anti-hPD-19 MAb, and specific combinations of humanized VH/VL domains are referred to by reference to specific VH/VL domains, for example, a humanized antibody comprising VH1 of an anti-hPD-19 MAb and VL2 of an anti-hPD-19 MAb is specifically referred to as an anti-hPD-19 MAb(1.2).

The amino acid sequence of the VH domain of the VH1 MAT to hPD-1 9 (SEQ ID NO: 179) is shown below (CDR _H residues are underlined):

EVQLVESGGG LVRPGGSLKL SCAASGFTFS SYLVSWVRQA PGKGLEWVAT

ISGGGGNTYY SDSVKGRFTI SRDNAKNSLY LQMNSLRAED TATYYCARYG FDGAWFAYWG QGTLVTVSS

An exemplary polynucleotide that encodes the VH1 mAb to hPD-1 9 comprises the sequence set forth in SEQ ID NO: 180 (nucleotides encoding CDRH residues are underlined): gaggtgcagc tggtggaaag tgggggcggc ctggtgcgac ccgggggaag tctgaaactg tcctgtgcag catcaggatt tactttttca tcttatctcg tgtcttgggt aagacaagca cccggaaaag gcttggaatg ggtggccact atctccggtg gaggtggcaa cacctactat agcgacagtg tcaagggaag atttaccatc agtcgcgaca acgctaagaa tagcctgtac ctccagatga actccctgcg cgccgaggac accgccacct attactgtgc acgctatgga tttgacggcg catggtttgc ctactgggga cagggcacat tggtaaccgt tagctcc

The amino acid sequence of the VH domain of the VH2 MAT to hPD-1 9 (SEQ ID NO: 181) is shown below (CDRH residues are underlined):

EVQLVESGGG LARPGGSLKL SCAASGFTFS SYLVGWVRQA PGKGLEWTAT

ISGGGGNTYY SDSVKGRFTI SRDNAKNSLY LQMNSARAED TATYYCARYG FDGAWFAYWG QGTLVTVSS

An exemplary polynucleotide that encodes a VH2 mAb to hPD-19 comprises the sequence set forth in SEQ ID NO: 182 (nucleotides encoding CDRH residues are underlined): gaggtgcagc tggtggaaag tgggggcggc ctggcgcgac ccgggggaag tctgaaactg tcctgtgcag catcaggatt tactttttca tcttatctcg tgggctgggt aagacaagca cccggaaaag gcttggaatg gacggccact atctccggtg gaggtggcaa cacctactat agcgacagtg tcaagggaag atttaccatc agtcgcgaca acgctaagaa tagcctgtac ctccagatga actccgcacg cgccgaggac accgccacct attactgtgc acgctatgga tttgacggcg catggtttgc ctactgggga cagggcacat tggtaaccgt tagctcc

CDRH1 of the VH domain of VH2 of MAT to hPD-1 9 comprises a substitution of serine with the amino acid glycine and comprises the amino acid sequence: ^sylv - ((SEQ ID NO: 183), the substituted glycine is underlined). According to the present invention, it is envisaged that a similar substitution may be included in any of the CDRH1 domains of MAT 9 to PD-1 described above.

The amino acid sequence of the VL domain of the hPD-1 mAb VL1 9 (SEQ ID NO: 184) is shown below (CDRH residues are underlined):

DIQMTQSPSS LSASVGDRVT ITCRASENIY SYLAWYQQKP GKAPKLLIYN

AKTLAAGVPS RFSGSGSGTD FTLTISSLQP EDFATYYCQH HYAVPWTFGQ GTKLEIK

An exemplary polynucleotide that encodes the VL1 mAb to hPD-19 comprises the sequence set forth in SEQ ID NO: 185 (nucleotides encoding CDRH residues are underlined): gacattcaga tgactcagtc tcccagcagt ctgtccgcat ccgtggggga tcgggtcacc atcacctgcc gtgcctcaga aaacatctat tcatacctcg cctggtatca acagaaacct ggtaaagccc caaaattgct catttacaac gccaagaccc tcgcagctgg cgtgccaagt aggttctcag gcagcggctc agggacagat ttcaccctca ccatatcctc actgcagccc gaggattttg ccacttacta ctgccagcat cattacgcag tgccctggac cttcggacaa ggcactaagc tcgagatcaa a

The amino acid sequence of the VL domain of the VL2 MAT to hPD-1 9 (SEQ ID NO: 186) is shown below (CDRH residues are underlined):

DIQMTQSPSS LSASVGDRVT ITCRASENIY NYLAWYQQKP GKAPKLLIYD

AKTLAAGVPS RFSGSGSGTD FTLTISSLQP EDFATYYCQH HYAVPWTFGQ GTKLEIK

An exemplary polynucleotide that encodes the VL2 MAT to hPD-1 9 comprises the sequence set forth in SEQ ID NO: 187 (nucleotides encoding CDRH residues are underlined):

- 34 047726 gacattcaga tgactcagtc tcccagcagt ctgtccgcat ccgtggggga tcgggtcacc atcacctgcc gtgcctcaga aaacatctat aactacctcg cctggtatca acagaaacct ggtaaagccc caaaattgct catttacgac gccaagaccc tcgcagctgg cgtgcca agt aggttctcag gcagcggctc agggacagat ttcaccctca ccatatcctc actgcagccc gaggattttg ccacttacta ctgccagcat cattacgcag tgccctggac cttcggacaa ggcactaagc tcgagatcaa a

CDRL1 of the VL domain of the VL2 MAT to hPD-1 9 contains a substitution of serine by the amino acid asparagine and contains the amino acid sequence: kaseniynyla (SEQ ID NO: 188), the substituted asparagine is underlined).

According to the present invention, it is envisaged that the same substitution may be included in any of the CDRL1 domains of MAT 9 to PD-1 described above. CDRL2 of the VL domain of VL2 MAT 9 to hPD-1 comprises a substitution of asparagine with the amino acid aspartate and comprises the amino acid sequence: daktlaa ((SEQ ID\() 189), the substituted aspartate is underlined). According to the present invention, it is envisaged that the same substitution may be included in any of the CDRL2 domains of MAT 7 to PD-1 described above.

J. Murine monoclonal antibody to human PD-1 10.

The amino acid sequence of the VH domain of MAT 10 to PD-1 (SEQ ID NO: 190) is shown below (CDR _H residues are underlined).

EVILVESGGG LVKPGGSLKL SCAASGFTFS NYLMSWVRQT PEKRLEWVAS ISGGGSNIYY PDSVKGRFTI SRDNAKNTLY LQMNSLRSED TALYYCARQE LAFDYWGQGT TLTVSS

CDRhI MAT 10 to PD-1 (SEQ Ш NO: 192): NYLMS

CDR _H 2 MAT 10 kPD-1 (SEQ Ш NO: 193): SISGGGSNIYYPDSVKG

CDR _H 3 MAT 10 to PD-1 (SEQ Ш NO: 194): QELAFDY

An exemplary polynucleotide that encodes the VH domain of MAT 10 to PD-1 comprises the sequence set forth in SEQ ID NO: 191 (nucleotides encoding CDRH residues are underlined): gaagtgatac tggtggagtc tgggggaggc ttagtgaagc ctggagggtc cctgaaactc tcctgtgcag cctctggatt cactttcagt aactatctca tgtcttgggt tcgccagact ccggagaaga ggctggagtg ggtcgcaagt attagtggtg gtggtagtaa tatctactat ccagacagtg tgaagggtcg attcaccata tccagggaca atgccaagaa caccctgtac ctgcaaatga acagtctgag gtctgaggac acggccttgt attactgtgc aagacaagaa ctggcttttg actactgggg ccaaggcacc actctcacag tctcctcc

The amino acid sequence of the VL domain of MAT 10 to PD-1 (SEQ ID NO: 195) is shown below (CDR _L residues are underlined):

DIQMTQTTSS LSASLGDRVT ISCRTSQDIS NFLNWYQQKP DGTIKLLIYY

TSRLHSGVPS RFSGSGSGTD YSLTISNLEQ EDIATYFCQQ GSTLPWTFGG GTKLEII

CDR _l 1 MAT 10 to PD-1 (SEQ ID NO: 197): RTSQDISNFLN

CDR _l 2 MAT 10 to PD-1 (SEQ Ш NO: 198): YTSRLHS

CDR _l 3 MAT 10 to PD-1 (SEQ Ш NO: 199): QQGSTLPWT

An exemplary polynucleotide that encodes the VL domain of MAT 10 to PD-1 comprises the sequence set forth in SEQ ID NO: 196 (nucleotides encoding CDR _L residues are underlined):

gatatccaga tgacacagac tacatcctcc ctgtctgcct ctctgggaga cagagtcacc atcagttgca ggacaagtca ggacattagc aattttttaa actggtatca gcagaaacca gatggaacta ttaaactcct gatctactac acatcaagat tacactcagg agtcccatca aggttcagtg gcagtgggtc t ggaacagat tattctctca ccattagcaa cctggagcaa gaagatattg ccacttactt ttgccaacag ggtagtacgc ttccgtggac gttcggtgga ggcaccaagc tggaaatcat a

- 35 047726

K. Murine monoclonal antibody to human PD-1 11.

The amino acid sequence of the VH domain of MAT 11 to PD-1 (SEQ ID NO: 200) is shown below (CDR _H residues are underlined).

EVQLQQSGTV LARPGASVKM SCKTSGYTFT GYWMHWVKQR PGQGLKWMGA IYPGNSDTHY NQKFKGKAKL TAVTSASTAY MELSSLTNED SAIYYCTTGT YSYFDVWGTG TTVTVSS

CDRhI MAT 11 to PD-1 (SEQ Ш NO: 202): GYWMH

CDR _H 2 MAT 11 to PD-1 (SEQ Ш NO: 203): AI YPGNSDTHYNQKFKG

CDR _H 3 MAT 11 to PD-1 (SEQ Ш NO: 204): GTYS YFDV

An exemplary polynucleotide that encodes the VH domain of MAT 11 to PD-1 comprises the sequence set forth in SEQ ID NO: 201 (nucleotides encoding CDRH residues are underlined): gaggttcagc tccagcagtc tgggactgtg ctggcaaggc ctggggcttc agtgaagatg tcctgcaaga cttctggcta cacatttacc ggctactgga tgcactgggt aaaacagagg cctggacagg gtctgaaatg gatgggggct atttatcctg gaaatagtga tactcactac aaccagaagt tcaagggcaa ggccaaactg actgcagtca catccgccag cactgcctac atggagctca gcagcctgac aaatgaggac tctgcgatct attactgtac tactgggacc tactcgtact tcgatgtctg gggcacaggg accacggtca ccgtctcctc a

The amino acid sequence of the VL domain of MAT 11 to PD-1 (SEQ ID NO: 205) is shown below (CDR _L residues are underlined):

DILLTQSPAI LSVSPGERVS FSCRASQSIG TSIHWYQHRT NGSPRLLIKY

ASESISGIPS RFSGSGSGTD FTLSINSVES EDIADYYCQQ SNSWLTFGAG TKLELK

CDRlI MAT AND to PD-1 (SEQ Ш NO: 207): RASQSIGTSIH

CDR _l 2 MAT AND to PD-1 (SEQ Ш NO: 208): YASES IS

CDR _l 3 MAT 11 to PD-1 (SEQ Ш NO: 209): QQSNSWLT

An exemplary polynucleotide that encodes the VL domain of MAT 11 to PD-1 comprises the sequence set forth in SEQ ID NO: 206 (nucleotides encoding CDR _L residues are underlined): gacatcttgc tgactcagtc tccagccatc ctgtctgtga gtccaggaga aagagtcagt ttctcctgca gggccagtca gagcattggc acaagcatac actggtatca gcacagaaca aatggttctc caaggcttct cataaagtat gcttctgagt ctatctctgg gatcccttcc aggtttagtg gcagtggatc agggactgat tttactctta gcatcaacag tgtggagtct gaagatattg cagattatta ctgtcaacaa agtaatagct ggctcacgtt cggtgctggg accaagctgg agctgaaa

L. Murine monoclonal antibody to human PD-1 12.

The amino acid sequence of the VH domain of MAT 12 to PD-1 (SEQ ID NO: 210) is shown below (CDRH residues are underlined).

QGHLQQSGAE LVRPGASVTL SCKASGFTFT DYEMHWVKQT PVHGLEWIGT

IDPETGGTAY NQKFKGKAIL TVDKSSTTTY MELRSLTSED SAVFYCSRER ITTWEGAYW YFDVWGTGTT VTVSS

CDRhI MAT 12 to PD-1 (SEQ ID NO: 212): DYEMH

CDR _H 2 MAT 12 to PD-1 (SEQ Ш NO: 213): TIDPETGGTAYNQKFKG

CDRh3 MAT 12 to PD-1 (SEQ Ш NO: 214): ERITTVVEGAYWYFDV

An exemplary polynucleotide that encodes the VH domain of MAT 12 to PD-1 comprises the sequence set forth in SEQ ID NO: 211 (nucleotides encoding CDRH residues are underlined):

- 36 047726 cagggtcacc tgcagcagtc tggggctgag ctggtgaggc ctggggcttc agtgacgctg tcctgcaagg cttcgggctt cacatttact gactatgaga tgcactgggt gaaacagaca cctgtgcatg gcctggaatg gattgggact attgatcctg aaactggtgg tactgcctac aatcagaagt tcaagggcaa ggccatactg acagtagaca aatcttccac tacaacctac atggagctcc gcagcctgac atctgaggac tctgccgtct tttattgttc aagagagagg attactacgg ttgttgaggg ggcatactgg tacttcgatg tctggggс a c agggaccacg gtcaccgtct cctca

The amino acid sequence of the VL domain of MAT 4 to PD-1 (SEQ ID NO: 215) is shown below (CDR _L residues are underlined):

DVLMTQTPLS LPVSLGDQAS ISCRSSQNIV HSNGNTYLEW YLQKPGQSPK LLICKVSTRF SGVPDRFSGS GSGTDFTLKI SRVEAEDLGV YYCFQGSHVP IYTFGGGTKLE IK

CDRlI MAT 12 to PD-1 (SEQ Ш NO: 217): RSSQNIVHSNGNTYLE

CDR _l 2 MAT 12 to PD-1 (SEQ ID NO: 218): KVSTRFS

CDR _l 3 MAT 12 to PD-1 (SEQ TO NO: 219): FQGSHVPYT

An exemplary polynucleotide that encodes the VL domain of MAT 12 to PD-1 comprises the sequence set forth in SEQ ID NO: 216 (nucleotides encoding CDR _L residues are underlined):

gatgttttga tgacccagac tccactctcc ctgcctgtca gtcttggaga tcaagcctcc atctcttgca gatetagtea gaacattgta catagtaatg gaaacaccta tttagaatgg tacctgcaga aaccaggcca gtctccaaag ctcctgatct gcaaagtttc cacccgattt tctggggtcc cagacaggtt ca gtggcagt ggatcaggga cagatttcac actcaagatc agcagagtgg aggetgagga tctgggagtt tattattget ttcaaggttc acatgttccg tacacgttcg gaggggggac caagctggaa ataaaa

M. Murine monoclonal antibody to human PD-1 13.

The amino acid sequence of the VH domain of MAT 13 to PD-1 (SEQ ID NO: 220) is shown below (CDR _H residues are underlined).

EVMLVESGGG LVKPGGSLKL SCAASGFTFS SHTMSWVRQT PEKRLEWVAT ISGGGSNIYY PDSVKGRFTI SRDNAKNTLY LQMSSLRSED TALYYCARQA YYGNYWYFDV WGTGTTVTVS S

CDRhI MAT 13 to PD-1 (SEQ ID NO: 222): SHTMS

CDR _H 2 MAT 13 to PD-1 (SEQ Ш NO: 223): TISGGGSNIYYPDSVKG

CDRh3 MAT 13 to PD-1 (SEQ Ш NO: 224): QAYYGNYWYFDV

An exemplary polynucleotide that encodes the VH domain of MAT 13 to PD-1 comprises the sequence set forth in SEQ ID NO: 221 (nucleotides encoding CDRH residues are underlined): gaagtgatgc tggtggagtc tgggggaggc ttagtgaagc ctggagggtc cctgaaactc tcctgtgcag cctctggatt cactttcagt agccatacca tgtcttgggt tcgccagact ccggagaaga ggctggagtg ggtcgcaacc attagtggtg gtggttctaa tatctactat ccagacagtg tgaagggtcg attcaccatc tccagagaca atgccaagaa caccctgtac etgeaaatga gcagtctgag gtctgaggac aeggeettgt attactgtgc aagacaagct tactacggta attactggta ettegatgte tggggcacag ggaccacggt caccgtctcc tcc

The amino acid sequence of the VL domain of MAT 13 to PD-1 (SEQ ID NO: 225) is shown below (CDR _L residues are underlined):

- 37 047726

DIQMTQSPAT QSASLGESVT ITCLASQTIG TWLAWYQQKP GKSPQLLIYA

ATSLADGVPS RFSGSGSGTK FSFKISSLQA EDFVSYYCQQ LDSIPWTFGG GTKLEIK

CDRlI MAT 13 to PD-1 (SEQ Ш NO: 227): LASQTIGTWLA

CDR _l 2 MAT 13 to PD-1 (SEQ ID NO: 228): AATSLAD

CDR _l 3 MAT 13 to PD-1 (SEQ Ш NO: 229): QQLDSIPWT

An exemplary polynucleotide that encodes the VL domain of MAT 13 to PD-1 comprises the sequence set forth in SEQ ID NO: 226 (nucleotides encoding CDR _L residues are underlined): gacattcaga tgacccagtc tcctgccacc cagtctgcat ctctgggaga aagtgtcacc atcacgtgcc tggcaagtca gaccattggt acatggttag catggtatca gcagaaacca gggaaatctc ctcagctcct gatttatgct gcaaccagct tggcagatgg ggtcccatca aggttcagtg gtagtggatc tggcacaaaa ttttctttca agatcagcag cctacaggct gaagattttg taagttatta ctgtcaacaa cttgacagta ttccgtggac gttcggtgga ggcaccaagc tggaaatcaa a

N. Murine monoclonal antibody to human PD-1 14.

The amino acid sequence of the VH domain of MAT 14 to PD-1 (SEQ ID NO: 230) is shown below (CDR _H residues are underlined).

QVQLQQPGAE LVKPGASVKM SCKASGYNFI SYWITWVKQR PGQGLQWIGN IYPGTDGTTY NEKFKSKATL TVDTSSSTAY MHLSRLTSED SAVYYCATGL HWYFDVWGTG TTVTVSS

CDRhI MAT 14 kPD-1 (SEQ ID NO: 232): SYWIT

CDR _H 2 MAT 14 to PD-1 (SEQ Ш NO: 233): NIYPGTDGTTYNEKFKS

CDR _H 3 MAT 14 to PD-1 (SEQ Ш NO: 234): GLHWYFDV

An exemplary polynucleotide that encodes the VH domain of MAT 14 to PD-1 comprises the sequence set forth in SEQ ID NO: 231 (nucleotides encoding CDRH residues are underlined): caggtccaac tgcagcagcc tggggctgag cttgtgaagc ctggggcttc agtgaagatg tcctgcaagg cttctggcta caacttcatc agctactgga taacctgggt gaaacagagg cctggacaag gccttcagtg gattggaaat atttatcctg gtactgatgg tactacctac aatgagaagt tcaagagcaa ggccacactg actgtagaca catcctccag cacagcctac atgcacctca gtcgcctgac atctgaggac tctgcggtct attactgtgc aactgggcta cactggtact tcgatgtctg gggcacaggg accacggtca ccgtctcctc c

The amino acid sequence of the VL domain of MAT 14 to PD-1 (SEQ ID NO: 235) is shown below (CDR _L residues are underlined):

DIVMTQSQKF MSTSVGDRVS VTCKASQSVG TNVAWYQQKP GQSPKALIYS

ASSRFSGVPD RFTGSGSGTD FTLTISNVQS EDLAEYFCQQ YNSYPYTFGG GTKLEIK

CDRlI MAT 14 to PD-1 (SEQ Ш NO: 237): KASQSVGTNVA

CDR _l 2 MAT 14 to PD-1 (SEQ Ш NO: 238): SASSRFS

CDR _l 3 MAT 14 to PD-1 (SEQ Ш NO: 239): QQYNSYPYT

An exemplary polynucleotide that encodes the VL domain of MAT 14 to PD-1 comprises the sequence set forth in SEQ ID NO: 236 (nucleotides encoding CDR _L residues are underlined): gacattgtga tgacccagtc tcaaaaattc atgtccacat cagtaggaga cagggtcagt gtcacctgca aggccagtca gagtgtgggt actaatgtag cctggtatca acagaagccc ggtcaatctc ctaaagcact gatttactcg gcatcctccc gattcagtgg cgtccctgat cgcttcacag gcagtggatc tgggacagat ttcactctca ccatcagtaa tgtgcagtct gaagacttgg cagagtattt ctgtcagcaa tataacagct atccgtacac gttcggaggg gggaccaagc tggaaataaa a

- 38 047726

O. Monoclonal antibody to human PD-1 15.

1. Murine monoclonal antibody to human PD-1 15.

The amino acid sequence of the VH domain of MAT 15 to PD-1 (SEQ ID NO: 240) is shown below (CDR _H residues are underlined).

EVMLVESGGG LVKPGGSLKL SCAASGFIFS SYLISWVRQT PEKRLEWVAA ISGGGADTYY ADSVKGRFTI SRDNAKNTLY LQMSSLRSED TALYYCTRRG TYAMDYWGQG TSVTVSS

CDRhI MAT 15 kPD-1 (SEQ Ш NO: 242): SYLIS

CDR _H 2 MAT 15 kPD-1 (SEQ Ш NO: 243): AISGGGADTYYADSVKG

CDR _H 3 MAT 15 to PD-1 (SEQ Ш NO: 244): RGTYAMDY

An exemplary polynucleotide that encodes the VH domain of MAT 15 to PD-1 comprises the sequence set forth in SEQ ID NO: 241 (nucleotides encoding CDRH residues are underlined): gaagtgatgc tggtggagtc tgggggaggc ttagtgaagc ctggagggtc cctgaaactc tcctgtgcag cctctggatt cattttcagt agctatctca tctcttgggt tcgccagact ccggagaaga ggctggagtg ggtcgctgcc attagtggtg gtggtgctga cacctactat gccgacagtg tgaagggtcg attcaccatc tccagagaca atgccaagaa caccctgtat ctgcaaatga gcagtctgag gtctgaggac acggccttat attactgtac aagacgaggg acctatgcta tggactactg gggtcaagga acctcagtca ccgtctcctc c

The amino acid sequence of the VL domain of MAT 15 to PD-1 (SEQ ID NO: 245) is shown below (CDR _L residues are underlined):

DIQMTQSPAS QSASLGESVT ITCLASQTIG TWLAWYQQKP GKSPQLLIYA

ATSLADGVPS RFSGSGSGTK FSFKISSLQA EDFVNYYCQQ LYSIPWTFGG GTKLEIK

CDRlI MAT 15 kPD-1 (SEQ Ш NO: 247): LASQTIGTWLA

CDR _l 2 MAT 15 to PD-1 (SEQ ID NO: 248): AATSLAD

CDR _l 3 MAT 15 kPD-1 (SEQ GO NO: 249): QQLYSIPWT

An exemplary polynucleotide that encodes the VL domain of MAT 15 to PD-1 comprises the sequence set forth in SEQ ID NO: 246 (nucleotides encoding CDR _L residues are underlined):

gacattcaga tgacccagtc tcccgcctcc cagtctgcat ctctggggaga aagtgtcacc atcacatgcc tggcaagtca gaccattggt acatggttag catggtatca gcagaaacca gggaaatctc ctcagctcct gatttatgct gcaaccagct tggcagatgg ggtcccatca aggttcagtg gtagtggatc tgg cacaaaa ttttctttca agatcagcag cctacaggct gaagattttg taaattatta ctgtcaacaa ctttacagta ttccgtggac gttcggtgga ggcaccaagc tggaaatcaa a

2. Humanization of monoclonal antibody to human PD-1 15 to obtain mAb to hPD-1 15.

The above-described murine monoclonal antibody to human PD-1 was humanized and further deimmunized after identification of antigenic epitopes to demonstrate the feasibility of humanizing an antibody to human PD-1 to reduce its antigenicity when administered to a human recipient. The humanization resulted in one humanized VH domain, designated herein as VH1 mAb to hPD-1 2, and one humanized VL domain, designated herein as VL1 mAb to hPD-1 1. An antibody comprising a humanized VL domain paired with a humanized VH domain is referred to as mAb to hPD-1 15.

The amino acid sequence of the VH domain of the VH1 MAT to hPD-1 15 (SEQ ID NO: 250) is shown below (CDRH residues are underlined):

EVQLVESGGG LVRPGGSLRL SCAASGFTFS SYLISWVRQA PGKGLEWVAA

ISGGGADTYY ADSVKGRFTI SRDNAKNSLY LQMNSLRAED TATYYCARRG

TYAMDYWGQG TLVTVSS

An exemplary polynucleotide that encodes VH1 MAT to hPD-1 15 comprises the sequence set forth in SEQ ID NO: 251 (nucleotides encoding CDRH residues are underlined):

- 39 047726 gaagtgcaac tggttgaaag tggcggcggg ctggtgcggc caggtggttc actcagactg tcttgtgcag cttcaggctt tacattctcc tcttatctta tctcttgggt gcgccaagcc ccaggtaagg gccttgaatg ggtcgccgcc attagtgggg gtggtgcc ga tacatattat gccgacagcg tcaagggacg tttcaccatc agcagggaca acgccaagaa tagcctttac ctgcagatga actcacttag agctgaagac accgctactt attactgtgc ccggcgcggg acttacgcta tggactattg gggccagggc accttggtca ctgtctcatc c

The amino acid sequence of the VH domain from VL1 MAT to hPD-1 15 (SEQ ID NO: 252) is shown below (CDR _H residues are underlined):

DIQMTQSPSS LSASVGDRVT ITCLASQTIG TWLAWYQQKP GKAPKLLIYA

ATSLADGVPS RFSGSGSGTD FTFTISSLQP EDFATYYCQQ LYSIPWTFGQ GTKLEIK

An exemplary polynucleotide that encodes the VL1 mAb to hPD-1 15 has the sequence set forth in SEQ ID NO: 253 (nucleotides encoding CDRH residues are underlined): gatatccaga tgacccagtc tcccagctct ctcagtgcaa gcgtaggcga ccgtgtgacc atcacctgtc tggccagtca gaccattgga acctggctcg cctggtatca gcagaaacct ggcaaggccc ctaagctgct gatttacgcc gccacctccc tcgcagatgg agtgccctcc cgatttagcg ggtccgggtc cggcaccgac ttcacattca caatcagcag cctccagccc gaggatttcg ctacatacta ctgtcaacag ctctactcca ttccatggac ctttggtcag ggtactaaac tggagatcaa a

V. Monoclonal antibodies to human PD-1 1-15 and their derivatives containing a modified Fc region.

During normal immune system function, interactions of antibody–antigen complexes with immune cells result in a wide range of responses, ranging from effector functions such as antibody-dependent cytotoxicity, mast cell degranulation, and phagocytosis to immunomodulatory signals such as regulation of lymphocyte proliferation and antibody secretion. All of these interactions are initiated by binding of the Fc region of antibodies or immune complexes to specialized cell surface receptors on hematopoietic cells. The diversity of cellular responses elicited by antibodies and immune complexes is due to the structural heterogeneity of the three Fc receptors: FcyRI (CD64), FcyRII (CD32), and FcyRIII (CD16). FcyRI (CD64), FcyRIIA (CD32A), and FcyRIII (CD16) are activating (i.e., immune-enhancing) receptors; FcγRIIB (CD32B) is an inhibitory (i.e., weakening the immune system) receptor. In addition, interaction with the neonatal Fc receptor (FcRn) mediates the recycling of IgG molecules from the endosome to the cell surface and release into the blood. The amino acid sequence of exemplary wild-type IgG1 (SEQ ID NO: 1), IgG2 (SEQ ID NO: 2), IgG3 (SEQ ID NO: 3), and IgG4 (SEQ ID NO: 4) is shown below.

Modification of the Fc region typically results in an altered phenotype, such as altered serum half-life, altered stability, altered susceptibility to cellular enzymes, or altered effector function. It may be desirable to modify an antibody or other binding molecule of the present invention with respect to effector function, for example, in order to increase the efficacy of said molecule in the treatment of cancer. Reduction or elimination of effector function is desirable in some cases, for example, in the case of antibodies whose mechanism of action includes blocking or antagonizing, but not lysing, cells bearing the target antigen. Increased effector function is typically desirable when it is directed to unwanted cells, such as tumor and foreign cells in which FcyRs are expressed at low levels, for example, tumor-specific B cells with low levels of FcyRIIB (e.g., non-Hodgkin's lymphoma, CLL, and Burkitt's lymphoma). In these embodiments, molecules of the present invention that have been rendered or have altered effector functional activity may be used to treat and/or prevent a disease, disorder or infection in which increased efficacy of effector functional activity is desired.

In some embodiments, the PD-1 binding molecules of the invention comprise an Fc region that comprises one or more modifications (e.g., substitutions, deletions, or insertions) in the amino acid sequence of the wild-type Fc region (e.g., SEQ ID NO: 1) that reduce the affinity and avidity of the Fc region, and thus the molecule of the invention, for one or more FcyR receptors. In other embodiments, the molecules of the invention

- 40 047 726 comprise an Fc region that comprises one or more amino acid modifications of a wild-type Fc region that increase the affinity and avidity of the Fc region, and thus the molecule of the present invention, for one or more FcyR receptors. According to other embodiments of the present invention, the molecules comprise a variant Fc region, wherein the variant confers or mediates increased antibody-dependent cellular cytotoxicity (ADCC) activity and/or increased binding to FcyRIIA compared to a molecule that lacks an Fc region or that comprises a wild-type Fc region. In other embodiments of the present invention, the molecules comprise a variant Fc region, wherein the variant confers or mediates decreased ADCC activity (or other effector function) and/or increased binding to FcyRIIB compared to a molecule that lacks an Fc region or that comprises a wild-type Fc region. In some embodiments, the present invention includes within its scope PD-1 binding molecules comprising a variant Fc region, wherein said variant Fc region exhibits no detectable binding to any FcyR, compared to a comparable molecule comprising a wild-type Fc region. In other embodiments, the present invention includes within its scope PD-1 binding molecules comprising a variant Fc region, wherein said variant Fc region binds to only one FcyR, preferably one of FcyRIIA, FcyRIIB, or FcyRIIIA. Any suitable increase in affinity and/or avidity is preferably assessed by measuring in vitro the extent of detectable binding to FcyR or FcYR-related activity in cells that express low levels of FcyR, where binding activity of the parent molecule (without the modified Fc region) is not detectable in the cells, or in cells that express target antigens other than Fcy receptors at a density of 30,000 to 20,000 molecules/cell, at a density of 20,000 to 10,000 molecules/cell, at a density of 10,000 to 5,000 molecules/cell, at a density of 5,000 to 1,000 molecules/cell, at a density of 1,000 to 200 molecules/cell, or at a density of 200 molecules/cell or less (but not less than 10, 50, 100, or 150 molecules/cell). The PD-1 binding molecules of the present invention may comprise a variant Fc region having an altered affinity for the activating and/or inhibitory receptor Fcy. In one embodiment of the present invention, a PD-1 binding molecule comprises a variant Fc region that has an increased affinity for FcyRIIB and a decreased affinity for FcyRIIIA and/or FcyRIIA compared to a comparable molecule comprising a wild-type Fc region. In another embodiment of the present invention, a PD-1 binding molecule of the present invention comprises a variant Fc region that has a decreased affinity for FcyRIIB and an increased affinity for FcyRIIIA and/or FcyRIIA compared to a comparable molecule comprising a wild-type Fc region. In another embodiment of the present invention, the PD-1 binding molecules of the present invention comprise a variant Fc region that has a reduced affinity for FcyRIIB and a reduced affinity for FcyRIIIA and/or FcyRIIA relative to a comparable molecule comprising a wild-type Fc region. In another embodiment of the present invention, the PD-1 binding molecules of the present invention comprise a variant Fc region that has an unchanged affinity for FcyRIIB and a reduced (or increased) affinity for FcyRIIIA and/or FcyRIIA relative to a comparable molecule comprising a wild-type Fcy region.

In some embodiments of the present invention, the PD-1 binding molecules of the present invention comprise a variant Fc region having an altered affinity for FcγRIIIA and/or FcγRIIA, such that the immunoglobulin has an enhanced effector function. Non-limiting examples of cellular effector functions include antibody-dependent cellular cytotoxicity, antibody-dependent phagocytosis, phagocytosis, opsonization, opsonophagocytosis, cell binding, rosetting, binding to complement component G1q, and complement-dependent cellular cytotoxicity. In a preferred embodiment of the present invention, the affinity or effector function is altered by at least 2-fold, preferably at least 4-fold, at least 5-fold, at least 6-fold, at least 7-fold, at least 8-fold, at least 9-fold, at least 10-fold, at least 50-fold, or at least 100-fold compared to a comparable molecule comprising a wild-type Fc region. According to other embodiments of the present invention, the variant Fc region immunospecifically binds to one or more FcRs with an affinity that is at least 65%, preferably at least 70%, 75%, 80%, 85%, 90%, 95%, 100%, 125%, 150%, 175%, 200%, 225% or 250% greater than the affinity of a molecule comprising a wild-type Fc region. Such measurements may be quantitative in vivo or in vitro assays, and are preferably quantitative in vitro assays such as enzyme-linked immunosorbent assay (ELISA) or surface plasmon resonance.

According to various embodiments of the present invention, the PD-1 binding molecules of the present invention comprise a variant Fc region, wherein said variant is

- 41 047726 an agonist of at least one FcyR receptor activity or an antagonist of at least one FcyR receptor activity. According to a preferred embodiment of the present invention, the molecules comprise a variant that is an antagonist of one or more FcyRIIB activities, such as B cell-mediated signaling, B cell activation, B cell proliferation, antibody production, intracellular calcium influx into B cells, cell cycle progression, FcεRIIB-mediated inhibition of FcεRI signaling, FcyRIIB phosphorylation, SHIP recruitment, SHIP phosphorylation and binding to Shc, or the activity of one or more downstream molecules (e.g., MAP kinase, JNK, p38, or Akt) in the FcyRIIB signaling pathway. According to another embodiment of the present invention, the PD-1 binding molecules of the present invention comprise a variant that is an agonist of one or more activities of FcεRI, such as mast cell activation, calcium mobilization, degranulation, cytokine production, or serotonin release.

In some embodiments, the molecules comprise an Fc region comprising regions from two or more IgG isotypes (e.g., IgG1, IgG2, IgG3, and IgG4). In the present invention, an Fc region is considered to be of a particular IgG isotype if its amino acid sequence is most homologous to that isotype relative to other IgG isotypes. Different IgG isotypes exhibit different physical and functional properties, including serum half-life, complement binding, binding affinity for FcyR, and effector function activity (e.g., ADCC, CDC, etc.), due to differences in the amino acid sequences of the hinge region and/or the Fc region, e.g., as described in Flesch and Neppert (1999) J. Clin. Lab. Anal. 14:141-156; Chappel et al. (1993) J. Biol. Chem. 33:25124-25131; Chappel et al. (1991) Proc. Natl. Acad. Sci. (U.S.A.) 88:9036-9040; or Bruggemann et al. (1987) J. Exp. Med 166:1351-1361. The type of Fcy region variant may be used alone or in combination with an amino acid modification to affect Fc-mediated effector function and/or binding activity. When used in combination, the amino acid modification and the hinge/IgG Fc region may exhibit similar functionality (e.g., increased affinity for FcyRIIA) and may act additively or, more preferably, synergistically to modify effector functionality in a molecule of the present invention, as compared to a molecule of the present invention that comprises a wild-type Fc region. In other embodiments of the present invention, the amino acid modification and the Fc region of an IgG may exhibit opposite functionality (e.g., increasing and decreasing affinity for FcγRIIA, respectively) and may selectively regulate or decrease specific functionality in a molecule of the present invention, compared to a molecule of the present invention that does not contain an Fc region or contains a wild-type Fc region of the same isotype.

In a preferred specific embodiment, the PD-1 binding molecules of the present invention comprise a variant Fc region, wherein said variant Fc region comprises at least one amino acid modification, relative to a wild-type Fc region, such that said molecule has an altered affinity for FcR, with the proviso that said variant Fc region does not comprise a substitution at positions that directly contact FcγR, based on crystallographic and structural analysis of Fc-FcR interactions, such as those described by Sondermann et al. (2000) Nature 406:267-73. Exemplary positions in the Fc region that directly contact FcγR are amino acid residues 234-239, amino acid residues 265-269 (B/C loop), amino acid residues 297-299 (C'/E loop), and amino acid residues 327-332 (F/G loop). In some embodiments of the present invention, the molecules of the present invention comprise variants of the Fc region comprising a modification of at least one residue that does not directly contact FcyR, based on structural and crystallographic analysis, such as being outside the Fc-FcyR binding site.

Fc region variants are well known in the art, and any known Fc region variant can be used according to the present invention to confer or modify the effector function exhibited by a molecule of the present invention comprising an Fc region (or a portion thereof), as determined by functional assays, such as NK-dependent or macrophage-dependent assays. For example, Fc region variants identified as effector function altering variants are disclosed in PCT publications WO 04/063351; WO 06/088494; WO 07/024249; WO 06/113665; WO 07/021841; WO 07/106707; and WO 2008/140603, and any suitable variant disclosed in these publications can be used in the molecules of the present invention.

According to some embodiments of the present invention, the PD-1 binding molecules of the present invention comprise a variant Fc region comprising one or more amino acid modifications in one or more regions, the modification(s) of which alters (relative to a wild-type Fc region) the affinity ratio of the variant Fc region in favor of the active

- 42 047726 of a activating FcyR (such as FcyRIIA or FcyRIIIA) relative to an inhibitory FcyR (such as FcYRIIB):

Affinity ratio = Change in affinity of Fc variant, relative to wild type, in favor of activating FcYR / Change in affinity of Fc variant, relative to wild type, in favor of inhibitory FcyR

Particularly preferred are PD-1 binding molecules of the present invention that comprise a variant Fc region (compared to a wild-type Fc region) in which the variant Fc region has an affinity ratio of greater than 1. These molecules are particularly useful in providing therapeutic or prophylactic treatment for, or amelioration of, a disease, disorder or infection where increased efficiency of FcyR-mediated cellular effector function (e.g., ADCC) is desired, such as in cancer or an infectious disease. In contrast, a variant Fc region having an affinity ratio of less than 1 causes a decrease in the efficiency of cellular effector function. Table 1 lists exemplary single, double, triple, quadruple and quintuple mutations, regardless of whether they have an affinity ratio of greater than or less than 1.

Table 1

Typical single and multiple mutations listed according to the ratio of affinity values

Single Double Triple Quadruple Fivefold Ratio of affinity values >1 F243L F243L and R292P F243L, P247L and N421K L234F, F243L, R292P and Y300L L235V, F243L, R292P, Y300LHP396L D270E F243L and Y300L F243L, R292P and Y300L L235I, F243L, R292P and Y300L L235P, F243L, R292P, Y300LhP396L R292G F243L and P396L F243L, R292P and V305I L235Q, F243L, R292P and Y300L F243L, R292P, V305I, Y300LhP396L R292P D270E and P396L F243L, R292P and P396L F243L, P247L, D270EhN421K R292P and Y300L F243L, Y300L and P396L F243L, R255L, D270E and P396L R292P and P247L, D270E and F243L, D270E,

- 43 047726

V305I N421K G316DhR416G R292P and P396L R255L, D270E and P396L F243L, D270E, K392T and P396L Y300L and P396L D270E, G316DH R416G F243L, D270E, P396LhQ419H P396L and Q419H D270E, K392T and P396L F243L, R292P, Y300L, and P396L D270E, P396L and Q419H F243L, R292P, V305I and P396L V284M, R292L and K370N P247L, D270E, Y300LhN421K R292P, Y300L and P396L R255L, D270E, R292G and P396L R255L, D270E, Y300LhP396L D270E, G316D, P396LhR416G Ratio of affinity values <1 Y300L F243L and P396L F243L, R292P and V305I P396L P247L and N421K R255L and P396L R292P and V305I K392T and P396L P396L and Q419H

According to a specific embodiment of the present invention, the Fc region variants comprise any amino acid modifications (e.g., substitutions) at any of positions 235, 240, 241, 243, 244, 247, 262, 263, 269, 298, 328, or 330, and preferably one or more of the following residues: A240, I240, L241, L243, H244, N298, I328, or V330. According to another embodiment of the present invention, the Fc region variants comprise any amino acid modifications (e.g., substitutions) at any of positions 268, 269, 270, 272, 276, 278, 283, 285, 286, 289, 292, 293, 301, 303, 305, 307, 309, 331, 333, 334, 335, 337, 338, 340, 360, 373, 376, 416, 419, 430, 434, 435, 437, 438, or 439 and preferably one or more of the following residues: H280, Q280, Y280, G290, S290, T290, Y290, N294, K295, P296, D298, N298, P298, V298, I300 or L300. According to a preferred embodiment of the present invention, Fc region variants that bind FcyR with altered affinity comprise any amino acid modifications (e.g., substitutions) at any of positions 255, 256, 258, 267, 268, 269, 270, 272, 276, 278, 280, 283, 285, 286, 289, 290, 292, 293, 294, 295, 296, 298, 300, 301, 303, 305, 307, 309, 312, 320, 322, 326, 329, 330, 332, 331, 333, 334, 335, 337, 338, 339, 340, 359, 360, 373, 376, 416, 419, 430, 434, 435, 437, 438 or 439. Preferably, the Fc region variant comprises any of the following residues: A256, N268, Q272, D286, Q286, S286, A290, S290, A298, M301, A312, E320, M320, Q320, R320, E322, A326, D326, E326, N326, S326, K330, T339, A333, A334, E334, H334, L334, M334, Q334, V334, K335, Q335, A359, A360 or A430.

According to another embodiment of the present invention, Fc region variants that bind FcyR (via their Fc region) with reduced affinity comprise any amino acid modifications (e.g., substitutions) at any of positions 252, 254, 265, 268, 269, 270, 278, 289, 292, 293, 294, 295, 296, 298, 300, 301, 303, 322, 324, 327, 329, 333, 335, 338, 340, 373, 376, 382, 388, 389, 414, 416, 419, 434, 435, 437, 438, or 439.

According to another embodiment of the present invention, Fc region variants that are associated

- 44 047726 bind FcγR (via its Fc region) with increased affinity, comprise any amino acid modifications (e.g., substitutions) at any of positions 280, 283, 285, 286, 290, 294, 295, 298, 300, 301, 305, 307, 309, 312, 315, 331, 333, 334, 337, 340, 360, 378, 398, or 430. According to another embodiment of the present invention, Fc region variants that bind FcγRIIA with increased affinity comprise any of the following residues: A255, A256, A258, A267, A268, N268, A272, Q272, A276, A280, A283, A285, A286, D286, Q286, S286, A290, S290, M301, E320, M320, Q320, R320, E322, A326, D326, E326, S326, K330, A331, Q335, A337 or A43 0.

Preferred embodiments comprise one or more modifications at any of positions 228, 230, 231, 232, 233, 234, 235, 239, 240, 241, 243, 244, 245, 247, 262, 263, 264, 265, 266, 271, 273, 275, 281, 284, 291, 296, 297, 298, 299, 302, 304, 305, 313, 323, 325, 326, 328, 330 or 332.

Particularly preferred embodiments comprise one or more modifications selected from groups A-AI: ______________________________________________________________________

A 228E, 228K, 228Y or 228G; IN 230A, 23 0E, 23 0Y or 23 0G; WITH 231E, 231K, 231Y, 231P or 231G; D 232E, 232K, 232Y, 232G; E 233D; F 2341 or 234F; G 235D, 235Q, 235Р, 2351 or 235V; N 239D, 239E, 239N or 239Q; I 240A, 2401, 240M or 240T; J 243R, 243, 243Y, 243L, 243Q, 243W, 243H or 2431; TO 244N; L 245A; M 247G, 247V or 247L; N 262A, 262E, 2621, 262T, 262E or 262F; ABOUT 263A, 2631, 263M or 263T; R 264F, 264E, 264R, 2641, 264A, 264T or 264W; Q 265F, 265Y, 265H, 2651, 265L, 265T, 265V, 265N or 265Q; R 266A, 2661, 266M or 266T; S 271D, 271E, 271N, 271Q, 271K, 271R, 271S, 271T, 271H, 271A, 271V, 271L, 2711, 271F, 271M, 271Y, 271W or 271G; T 2731; And 275L or 275W; V 281D, 281К, 281Y or 281Р; W 284E, 284N, 284T, 284L, 284Y or 284M; X 291D, 291E, 291Q, 291T, 291H, 2911 or 291G; Y 299A, 299D, 299E, 299F, 299G, 299Н, 2991, 299К, 299L, 299М, 299N, 299Р, 299Q,

- 45 047726

299R, 299S, 299V, 299W or 299Y; Z 3021; AA 304D, 304N, 304T, 304H or 304L AB 3051; AS 313F; AD 3231; AE 325A, 325D, 325E, 325G, 325H, 3251, 325L, 325K, 325R, 325S, 325F, 325M, 325T, 325V, 325Y, 325W or 325P; AF 328D, 328Q, 328K, 328R, 328S, 328T, 328V, 3281, 328Y, 328W, 328P, 328G, 328A, 328E, 328F, 328N, 328M or 328N; AG 330L, 330Y, 3301 or 330V; AN 332A, 332D, 332E, 332N, 332N, 332Q, 332T, 332K, 332R, 332S, 332V, 332L, 332F, 332M, 332W, 332P, 332G or 332Y; And AI 336E, 336K or 336Y

Even more preferred embodiments contain one or more modifications selected from groups 1-105: __________________________________________________________________________

Group Option Group Option 1 A330L/I332E 54 S239D/D265L/N297D/I332E 2 D265F/N297E/I332E 55 S239D/D265T/N297D/I332E 3 D265Y/N297D/I332E 56 S239D/D265V/N297D/I332E 4 D265Y/N297D/T299L/I332E 57 S239D/D265Y/N297D/I332E 5 F241E/F243Q/V262T/V264F 58 S239D/I332D 6 F241E/F243 Q/V262T/V264E/I332E 59 S239D/I332E 7 F241E/F243R/V262E/V264R 60 S239D/I332E/A330I 8 F241E/F243R/V262E/V264R/I332E 61 S239D/I332N 9 F241E/F243Y/V262T/V264R 62 S239D/I332Q 10 F241E/F243Y/V262T/V264R/I332E 63 S239D/N297D/I332E AND F241L/F243L/V262W264I 64 S23 9D/N297D/I3 3 2E/AZ 3 0 Y 12 F241L/V262I 65 S23 9D/N297D/I3 3 2E/AZ 3 0 Y/F241S/F24 3H/V262T/V264T

- 46 047726

13 F241R/F243Q/V262T/V264R 66 S239D/N297D/I332E/K326E 14 F241R/F243 Q/V262T/V264R/I332E 67 S23 9D/N297D/I3 3 2E/L23 5D 15 F241W/F243W/V262A/V264 A 68 S239D/S298A/I332E 16 F241Y/F243Y/V262T/V264T 69 S23 9D/V264RAZ 3 0L/I3 3 2E 17 F241Y/F243Y/V262T/V264T/N297D/I332E 70 S239D/V264VI332E 18 F243L/V262VV264W 71 S23 9D/V264VS298A/I3 32E 19 P243L/V264I 72 S239E/D265N 20 L328D/I332E 73 S239E/D265Q 21 L328E/I332E 74 S239E/I332D 22 L328H/I332E 75 S239E/I332E 23 L328VI332E 76 S239E/I332N 24 L328M/I332E 77 S239E/I332Q 25 L328N/I332E 78 S239E/N297D/I332E 26 L328Q/I332E 79 S239E/V264VA330Y/I332 E 27 L328T/I332E 80 S239E/V264VI332 E 28 L328V/I332E 81 S23 9E/V2641/S298А/AZ 3 0 Y/I3 3 2E 29 N297D/A330Y/I332E 82 S239N/A330L/I332E 30 N297D/I332E 83 S239N/A330Y/I332E 31 N297D/I3 3 2E/S23 9D/A3 3 0L 84 S239N/I332D 32 N297D/S298A/A330Y/I 332E 85 S239N/I332E 33 N297D/T299L/I332E 86 S239N/I332N 34 N297D/T299F/I332E/N297D/ T299H/I332E 87 S239N/I332Q 35 N297D/T299VI332E 88 S239N1S298A/I332E 36 N297D/T299L/I332E 89 S239Q/I332D 37 N297D/T299V/I332E 90 S239Q/I332E 38 N297E/I332E 91 S239Q/I332N 39 N297S/I332E 92 S239Q/I332Q 40 P230A/E233D/I332E 93 S239Q/V264VI332E 41 P244H/P245A/P247V 94 S298A/I332E 42 S239D/A330L/I332E 95 V264E/N297D/I332E 43 S239D/A330Y/I332E 96 V264VA330L/I332E 44 S23 9D/A3 3 0Y/I3 3 2E/KZ 26E 97 V264VA330Y/I332E 45 S23 9D/A3 3 0Y/I3 3 2E/KZ 26T 98 V264VI332E 46 S23 9D/A3 3 0Y/I3 3 2E/L2341 99 V264VS298A/I332E 47 S23 9D/A3 3 0Y/I3 3 2E/L23 5D 100 Y296D/N297D/I332E 48 S23 9D/A3 3 0Y/I3 3 2E/V240I 101 Y296E/N297D/I332 E 49 S23 9D/A3 3 0Y/I3 3 2E/V264T 102 Y296H/N297D/I332E 50 S23 9D/A3 3 0Y/I3 3 2E/V266I 103 Y296N/N297D/I332E 51 S23 9D/D265F/N297D/I3 3 2E 104 Y296Q/N297VI332E 52 S239D/D265H/N297D/I332E 105 Y296I7N297D/I332E 53 S239D/D265I/N297D/I332E

In one embodiment of the present invention, a PD-1 binding molecule of the present invention will comprise a variant Fc region comprising at least one modification in the Fc region. In some embodiments of the present invention

- 47 047726 the Fc region variant comprises at least one substitution selected from the group consisting of L235V, F243L, R292P, Y300L, V305I and P396L.

According to a specific embodiment of the present invention, the Fc region variant comprises:

(A) at least one substitution selected from the group consisting of F243L, R292P, Y300L, V305I and P396L;

(B) at least two replacements selected from the group consisting of:

(1) F243L and P396L;

(2) F243L and R292P; and (3) R292P and V305I;

(C) at least three replacements selected from the group consisting of:

(1) F243L, R292P and Y300L;

(2) F243L, R292P and V305I;

(3) F243L, R292P and P396L; and (4) R292P, V305I and P396L;

(D) at least four replacements selected from the group consisting of:

(1) F243L, R292P, Y300L and P396L; and (2) F243L, R292P, V305I and P396L; or (E) at least five substitutions selected from the group consisting of:

(1) F243L, R292P, Y300L, V305I and P396L; and (2) L235V, F243L, R292P, Y300L and P396L.

In another specific embodiment of the present invention, the Fc region variant comprises the substitutions:

(A) F243L, R292P and Y300L;

(b) L235V, F243L, R292P, Y300L and P396L; or (C) F243L, R292P, Y300L, V305I and P396L.

In one embodiment of the present invention, a PD-1 binding molecule of the present invention comprises a variant Fc region that exhibits reduced (or substantially absent) binding to FcγRIA (CD64), FcγRIIA (CD32A), FcγRIIB (CD32B), FcγRIIIA (CD16a), or FcγRIIIB (CD16b) (compared to the binding exhibited by the Fc region of wild-type IgG1 (SEQ ID NO: 1)). In one embodiment of the present invention, a PD-1 binding molecule of the present invention will comprise a variant Fc region that exhibits reduced (or substantially absent) binding to FcγR (e.g., FcγRIIIA) and reduced (or substantially absent) ADCC effector function. According to some embodiments of the present invention, the variant Fc region comprises at least one substitution selected from the group consisting of L234A, L235A, D265A, N297Q, and N297G. According to a specific embodiment of the present invention, the variant Fc region comprises the substitution L234A; L235A; L234A and L235A; D265A; N297Q, or N297G.

Preferred IgG1 sequences for the CH2 and CH3 domains of PD-1 binding molecules according to the present invention will contain the substitutions L234A/L235A (SEQ ID NO: 5):

APEAAGGPSV FLFPPKPKDT LMISRTPEVT CVWDVSHED PEVKFNWYVD

GVEVHNAKTK PREEQYNSTY RWSVLTVLH QDWLNGKEYK CKVSNKALPA PIEKTISKAK GQPREPQVYT LPPSREEMTK NQVSLTCLVK GFYPSDIAVE WESNGQPENN YKTTPPVLDS DGSFFLYSKL TVDKSRWQQG NVFSCSVMHE ALHNHYTQKS LSLSPGX where X is lysine (K) or absent.

According to another embodiment of the present invention, a PD-1 binding molecule of the present invention comprises an Fc region that inherently exhibits reduced (or substantially absent) binding to FcγRIIIA (CD16a) and/or reduced effector function (compared to the binding exhibited by wild-type IgG1 Fc (SEQ ID NO: 1)). According to a specific embodiment of the present invention, a PD-1 binding molecule of the present invention does not comprise an IgG2 Fc region (SEQ ID NO: 2) or an IgG4 Fc region, (SEQ ID NO: 4). In the case of using the Fc region of IgG4, it is also within the scope of the present invention to introduce a stabilizing mutation, such as a S228P substitution of the IgG4 hinge region (see, for example, SEQ ID NO: 13: eskygppcppcp, (l _u et al. (2008) The Effect Of A Point Mutation On The Stability Of Igg4 As Monitored By Analytical Ultracentrifugation, J. Pharmaceutical Sciences 97:960-969), to reduce the frequency of strand exchange. Other stabilizing mutations known in the art can be introduced into the Fc region of IgG4 (Peters, P et al. (2012) Engineering an Improved IgG4 Molecule with Reduced Disulfide Bond Heterogeneity and Increased Fab Domain Thermal Stability, J. Biol. Chem., 287:24525-24533; PCT Patent Publication WO 2008/145142).

According to other embodiments, the present invention includes within its scope the use of

- 48 047726 any Fc region variant known in the art, such as those disclosed in Jefferis, B.J. et al. (2002) “Interaction Sites On Human

IgG-Fc For FcgammaR: Current Models,” Immunol. Lett. 82:57-65; Presta, L.G. et al. (2002) "Engineering Therapeutic Antibodies For Improved Function", Biochem. Soc. Trans. 30:48790; Idusogie, E.E. et al. (2001) "Engineered Antibodies With Increased Activity To Recruit Complement", J. Immunol. 166:2571-75; Shields, R.L. et al. (2001) "High Resolution Mapping Of The Binding Site On Human IgGl For Fc Gamma RI, Fc Gamma RII, Fc Gamma RIII, And FcRn And Design Of IgGl Variants With Improved Binding To The Fc gamma R", J. Biol.

Chern. 276:6591-6604; Idusogie, E.E. et al. (2000) "Mapping Of The Clq Binding Site On Rituxan, A Chimeric Antibody With A Human IgG Fc", J. Immunol. 164:4178-84; Reddy, M.P. et al. (2000) "Elimination Of Fc Receptor-Dependent Effector Functions Of A Modified IgG4 Monoclonal Antibody To Human CD4", J. Immunol. 164:1925-1933; Xu, D. et al. (2000) “In vitro Characterization of Five Humanized OKT3 Effector Function Variant Antibodies,” Cell.

Immunol. 200:16-26; Armour, K.L. et al. (1999) "Recombinant human IgG Molecules Lacking

Fcgamma Receptor I Binding And Monocyte Triggering Activities,” Eur. J. Immunol. 29:261324; Jefferis, R. et al. (1996) "Modulation Of Fc(Gamma)R And Human Complement Activation By IgG3-Core Oligosaccharide Interactions", Immunol. Lett. 54:101-04; Lund, J. et al. (1996) "Multiple Interactions Of IgG With Its Core Oligosaccharide Can Modulate Recognition By Complement And Human Fc Gamma Receptor I And Influence The Synthesis Of Its Oligosaccharide Chains", J. Immunol. 157:4963-4969; Hutchins et al. (1995) "Improved Biodistribution, Tumor Targeting, And Reduced Immunogenicity In Mice With A Gamma 4 Variant Of Campath-1H", Proc. Natl. Acad. Sci. (U.S.A.) 92:11980-84; Jefferis, R. et al. (1995) "Recognition Sites On Human IgG For Fc Gamma Receptors: The Role Of Glycosylation", Immunol. Lett. 44:111-17; Lund, J. et al. (1995) "Oligosaccharide-Protein Interactions In IgG Can Modulate Recognition By Fc Gamma Receptors", FASEB J. 9:115-19; Alegre, M. L. et al.

(1994) “A Non-Activating “Humanized” Anti-CD3 Monoclonal Antibody Retains Immunosuppressive Properties in Vivo”, Transplantation 57:1537-1543; Lund et al. (1992) "Multiple Binding Sites On The CH2 Domain Of IgG For Mouse Fc Gamma Rll", Mol.

Immunol. 29:53-59; Lund et al. (1991) "Human Fc Gamma RI And Fc Gamma RII Interact With Distinct But Overlapping Sites On Human IgG", J. Immunol. 147:2657-2662; Duncan, A.R. et al. (1988) "Localization Of The Binding Site For The Human High-Affinity Fc Receptor On IgG", Nature 332:563-564;

U.S. Patents 5,624,821; 5,885,573; 6,194,551; 7,276,586; and 7,317,091; and PCT Publications WO 00/42072 and PCT WO 99/58572.

According to some embodiments of the present invention, the molecules of the present invention further comprise one or more glycosylation sites such that one or more carbohydrate moieties are covalently attached to the molecule. Preferably, the molecules of the present invention comprising one or more glycosylation sites and/or one or more modifications in the Fc region provide or have an enhanced effector function mediated by the antibody, such as enhanced ADCC activity, compared to an unmodified antibody. According to some embodiments of the present invention are provided molecules comprising one or more amino acid modifications that are known to interact directly or indirectly with the carbohydrate moiety of the Fc region, including, but not limited to, amino acids at positions 241, 243, 244, 245, 245, 249, 256, 258, 260, 262, 264, 265, 296, 299, and 301. Amino acids that interact directly or indirectly with the carbohydrate moiety of the Fc region are known in the art, see, for example, Jefferis et al., 1995 Immunology Letters, 44: 111-7, which is incorporated herein by reference in its entirety.

According to another embodiment of the present invention, there are provided molecules that have been modified by introducing one or more glycosylation sites into one or more sites of the molecules, preferably without altering the functionality of the molecules, such as the binding activity to a target antigen or FcyR. The glycosylation sites may be introduced into the variable and/or constant region of the molecules of the present invention. As used herein, the term glycosylation sites includes any particular amino acid sequence in an antibody to which an oligosaccharide (i.e., carbohydrates containing

- 49 047726 containing two or more simple sugars linked together). Oligosaccharide side chains are typically attached to the antibody backbone via N- or O-linkages. N-linked glycosylation refers to the attachment of an oligosaccharide moiety to the side chain of an asparagine residue. O-linked glycosylation refers to the attachment of an oligosaccharide moiety to a hydroxyamino acid, such as serine, threonine. The molecules of the present invention may comprise one or more glycosylation sites, including N-linked and O-linked glycosylation sites. Any glycosylation site for N-linked or O-linked glycosylation known in the art may be used in accordance with the present invention. An exemplary N-linked glycosylation site that can be used in accordance with the methods of the present invention is the amino acid sequence: Asn-X-Thr/Ser, wherein X can be any amino acid and Thr/Ser is threonine or serine. A suitable site or sites can be introduced into a molecule of the present invention using methods well known in the art to which the present invention pertains (see, for example, in vitro Mutagenesis, Recombinant DNA: A Short Course, J. D. Watson, et al. W. H. Freeman and Company, New York, 1983, chapter 8, pp. 106-116, which is incorporated herein by reference in its entirety. An exemplary method for introducing a glycosylation site into a molecule of the present invention can comprise: modifying or mutating the amino acid sequence of the molecule so as to obtain the desired sequence Asn-X-Thr/Ser.

In some embodiments, the present invention provides methods of modifying the carbohydrate content of a molecule of the present invention by adding or removing a glycosylation site. Methods of modifying the carbohydrate content of antibodies (and molecules comprising antibody domains, e.g., an Fc region) are well known in the art and are included within the scope of the present invention, see, e.g., U.S. Patent No. 6,218,149; EP 0 359 096 B1; U.S. Patent Publication No. US 2002/0028486; WO 03/035835; U.S. Patent Publication No. 2003/0115614; U.S. Patent No. 6,218,149; U.S. Patent No. 6,472,511; all of which are incorporated herein by reference in their entireties. In other embodiments, the present invention provides methods of modifying the carbohydrate content of a molecule of the present invention by removing one or more endogenous carbohydrate moieties from the molecule. In a specific embodiment, the present invention provides for shifting the glycosylation site of the Fc region of an antibody by modifying positions adjacent to 297. In a specific embodiment, the present invention provides for modifying position 296 such that position 296, but not position 297, is glycosylated.

Effector function can also be modified by techniques such as the introduction of one or more cysteine residues into the Fc region, thereby allowing for the formation of an interchain disulfide bond in said region to produce a homodimeric antibody that may have improved internalization capacity and/or enhanced complement-dependent cytotoxicity and ADCC (Caron, P.C. et al. (1992) Engineered Humanized Dimeric Forms Of IgG Are More Effective Antibodies, J. Exp. Med. 176:1191-1195; Shopes, B. (1992) A Genetically Engineered Human IgG Mutant With Enhanced Cytolytic Activity, J. Immunol. 148(9):2918-2922. Homodimeric antibodies with enhanced antitumor activity can also be produced using heterobifunctional crosslinkers as described in Wolff, E.A. et al. (1992) Engineered Humanized Dimeric Forms Of IgG Are More Effective Antibodies, J. Exp. Med. 176:1191-1195; Shopes, B. (1992) A Genetically Engineered Human IgG Mutant With Enhanced Cytolytic Activity, J. Immunol. 148(9):2918-2922. Homodimeric antibodies with enhanced antitumor activity can also be produced using the heterobifunctional crosslinkers described in Wolff, E.A. et al. (1992) Engineered Humanized Dimeric Forms Of IgG Are More Effective Antibodies, J. Exp. Med. 176:1191-1195; Shopes, B. (1992) A Genetically Engineered Human IgG Mutant With Enhanced Cytolytic Activity, J. Immunol. 148(9):2918-2922. al. (1993) Monoclonal Antibody Homodimers: Enhanced Antitumor Activity In Nude Mice, Cancer Research 53:2560-2565. In another embodiment, the antibody can be modified to include two Fc regions and, accordingly, can have enhanced complement-mediated cytotoxicity and ADCC activity (Stevenson, G.T. et al. (1989) A Chimeric Antibody With Dual Fc Regions (bisFabFc) Prepared By Manipulations At The IgG Hinge, Anti-Cancer Drug Design 3:219-230)

The serum half-life of the molecules of the present invention comprising Fc regions can be increased by increasing the binding affinity of the Fc region for FcRn. As used herein, the term half-life refers to a pharmacokinetic property of a molecule that is a measure of the average time the molecules exist after administration. The half-life can be expressed as the time required to eliminate fifty percent (50%) of a known amount of the molecule from the body of a subject (e.g., a human patient or other mammal) or a specific compartment of the body, such as when measured in blood serum, i.e., the half-life in the bloodstream, or in other tissues. In general, an increase in the half-life results in an increase in the mean residence time (MRT) in the bloodstream for the administered molecule.

In some embodiments of the present invention, PD-1 binding molecules of the present invention comprise a variant Fc region, wherein said variant Fc region comprises at least one amino acid modification, relative to a wild-type Fc region, such that said molecule has an increased half-life (relative to a wild-type Fc region). In some embodiments of the present invention, PD-1 binding molecules of the present invention comprise a variant Fc region, wherein said variant

- 50 047726 Fc region comprises an amino acid substitution that increases half-life at one or more positions selected from the group consisting of 238, 250, 252, 254, 256, 257, 256, 265, 272, 286, 288, 303, 305, 307, 308, 309, 311, 312, 317, 340, 356, 360, 362, 376, 378, 380, 382, 413, 424, 428, 433, 434, 435 and 436. Numerous specific mutations capable of increasing the half-life of a molecule comprising an Fc region are known in the art and include, for example, M252Y, S254T, T256E, and combinations thereof. For example, see the mutations described in U.S. Patent Nos. 6,277,375, 7,083,784; 7,217,797, 8,088,376; U.S. Patent Application Publications 2002/0147311; 2007/0148164; and International Publications WO 98/23289; WO 2009/058492; and WO 2010/033279, which are incorporated herein by reference in their entireties. Fc region-containing molecules with an increased half-life also include those that comprise substitutions in two or more Fc region residues 250, 252, 254, 256, 257, 288, 307, 308, 309, 311, 378, 428, 433, 434, 435 and 436. In particular, said molecules comprise two or more substitutions selected from: T250Q, M252Y, S254T, T256E, K288D, T307Q, V308P, A378V, M428L, N434A, H435K and Y436I. According to a specific embodiment of the present invention, the Fc region variant comprises substitutions:

(A) M252Y, S254T and T256E;

(B) M252Y and S254T;

(C) M252Y and T256E;

(D) T250Q and M428L;

(E) T307Q and N434A;

(F) A378V and N434A;

(G) N434A and Y436I;

(H) V308P and N34A; or (I) K288D and H435K.

The present invention further includes within its scope variants of Fc regions that comprise:

(A) one or more mutations that alter effector function and/or FcyR; and (B) one or more mutations that increase serum half-life.

VI. Bispecific molecules binding to human PD-1.

According to one embodiment, the present invention provides bispecific binding molecules that are capable of binding to a first epitope and a second epitope, wherein said first epitope is an epitope of human PD-1, and said second epitope is similar to or different from an epitope of PD-1, or is an epitope of another molecule that is present on the surface of an immune cell (such as a T lymphocyte) and is involved in regulating an immune checkpoint response. According to one embodiment of the present invention, the second epitope is an epitope of B7-H3, B7-H4, BTLA, CD3, CD8, CD16, CD27, CD32, CD40, CD40L, CD47, CD64, CD70, CD80, CD86, CD94, CD137, CD137L, CD226, CTLA-4, galectin-9, GITR, GITRL, HHLA2, ICOS, ICOSL, KIR, LAG-3, LIGHT, MHC class I or II, NKG2a, NKG2d, OX40, OX40L, PD1H, PD-1, PD-L1, PD-L2, PVR, SIRPa, TCR, TIGIT, TIM-3 or VISTA. According to one embodiment of the present invention, the second epitope is not an epitope of PD-1. According to a specific embodiment of the present invention, the second epitope is CD137, CTLA-4, LAG-3, OX40, TIGIT or TIM-3. According to some embodiments of the present invention, the bispecific molecule comprises more than two epitope binding sites. These bispecific molecules can bind two or more different epitopes of LAG-3 and at least one epitope of a molecule that is different from LAG-3.

The present invention includes bispecific antibodies capable of simultaneously binding to PD-1 and a second epitope (e.g., B7-H3, B7-H4, BTLA, CD40, CD80, CD86, CD137, CTLA-4, ICOS, KIR, LAG-3, MHC class I or II, OX40, PD-L1, TCR, TIM-3, etc.). According to some embodiments of the present invention, a bispecific antibody capable of simultaneously binding to PD-1 and a second epitope is produced using any of the methods described in PCT publications WO 1998/002463, WO 2005/070966, WO 2006/107786 WO 2007/024715, WO 2007/075270, WO 2006/107617, WO 2007/046893, WO 2007/146968, WO 2008/003103, WO 2008/003116, WO 2008/027236, WO 2008/024188, WO 2009/132876, WO 2009/018386, WO 2010/028797, WO2010028796, WO 2010/028795, WO 2010/108127, WO 2010/136172, WO 2011/086091, WO 2011/133886, WO 2012/009544, WO 2013/003652, WO 2013/070565, WO 2012/162583, WO 2012/156430, WO 2013/174873 and WO 2014/022540, each of which is incorporated herein by reference in its entirety.

A. Bispecific diabodies lacking the Fc region.

According to one embodiment, the present invention provides bispecific diabodies that comprise, and most preferably consist of, a first polypeptide chain and a second polypeptide chain, the sequences of which provide for the covalent binding of said polypeptide chains to each other to form a covalently linked diabody that is capable of simultaneously binding to a first epitope and a second epitope, wherein said

- 51 047726 epitopes are not identical to each other. Accordingly, said bispecific diabodies comprise VL1/VH1 domains which are capable of binding to a first epitope and VL2VVH2 domains which are capable of binding to a second epitope. The designation VL1 and VH1 means, respectively, a variable domain of a light chain and a variable domain of a heavy chain which bind to the first epitope of said bispecific diabody. Likewise, the designation VL2 and VH2 means, respectively, a variable domain of a light chain and a variable domain of a heavy chain which bind to the second epitope of said bispecific diabody. Regardless of whether a particular epitope is designated as the first and second epitope; such designation is relevant only to the presence and orientation of the domains of the polypeptide chains of the binding molecules according to the present invention. According to one embodiment of the present invention, one of said epitopes is a PD-1 epitope and the other of said epitopes is not a PD-1 epitope (e.g., an epitope of B7-H3, B7-H4, BTLA, CD40, CD80, CD86, CD137, CTLA-4, ICOS, KIR, LAG-3, MHC class I or II, OX40, PD-L1, TCR, TIM-3, etc.).

The VL domain of the first polypeptide chain interacts with the VH domain of the second polypeptide chain to form a first functional antigen-binding site that is specific for a first antigen (i.e., PD-1 or an antigen that contains a second epitope). Similarly, the VL domain of the second polypeptide chain interacts with the VH domain of the first polypeptide chain to form a second functional antigen-binding site that is specific for a second antigen (i.e., an antigen that contains a second epitope, or PD-1). Consequently, the selection of the VL and VH domains of the first and second polypeptide chains is coordinated such that the two polypeptide chains of the diabody collectively comprise VL and VH domains capable of binding to both the PD-1 epitope and a second epitope (i.e., they comprise VL _PD-1 /VH _PD-1 and VL2/VH2, where PD-1 is the first epitope, or VL1/VH1 and VL _PD-1 /VH _PD-1 , where PD-1 is the second epitope).

The first polypeptide chain of an embodiment of said bispecific diabodies comprises, in the direction from the N-terminus to the C-terminus, the N-terminus, the VL1 domain of the monoclonal antibody capable of binding to a first or second epitope (i.e. VL _PD-1 or VL _epitope 2), a first intermediate spacer peptide (linker 1), the VH2 domain of the monoclonal antibody capable of binding to a second epitope (if said first polypeptide chain comprises VL _PD-1 ) or the first epitope (if said first polypeptide chain comprises VL _epitope 2), a second intermediate spacer peptide (linker 2), optionally comprising a cysteine residue, a domain facilitating the formation of a heterodimer, and a C-terminus (Fig. 1).

The second polypeptide chain of said embodiment of the bispecific diabodies comprises, in the direction from the N-terminus to the C-terminus, the N-terminus, a VL2 domain of a monoclonal antibody capable of binding to PD-1 or a second epitope (i.e. VL _PD-1 or VL _{epitope 2} , and wherein said VL domain is not suitable for inclusion in the first polypeptide chain of the diabody), an intermediate linker peptide (linker 1), a VH1 domain of a monoclonal antibody capable of binding to a second epitope (if said second polypeptide chain comprises VL _PD-1 ) or PD-1 (if said second polypeptide chain comprises VL _{epitope 2} ), a second intermediate spacer peptide (linker 2), optionally comprising a cysteine residue, a domain that promotes heterodimer formation, and a C-terminus (Fig. 1).

Most preferably, the length of the intermediate linker peptide (e.g., linker 1), which separates said VL and VH domains, is selected so as to substantially or completely prevent the VL and VH domains of the polypeptide chain from binding to each other. Accordingly, the VL and VH domains of the first polypeptide chain are substantially or completely unable to bind to each other. Likewise, the VL and VH domains of the second polypeptide chain are substantially or completely unable to bind to each other. A preferred intermediate spacer peptide (linker 1) comprises the sequence presented in ( ^SE Q ^{w N0:14} ) ^: gggsgggg.

The length and composition of the second intermediate linker peptide (linker 2) are selected based on the selection of heterodimer-promoting domains. Typically, the second intermediate linker peptide (linker 2) will contain 3-20 amino acid residues. In particular, if the heterodimer-promoting domains do not contain a cysteine residue, a second intermediate linker peptide containing cysteine (linker 2) is used. The second intermediate linker peptide containing cysteine (linker 2) will contain 1, 2, 3 or more cysteine residues. A preferred cysteine-containing spacer peptide (linker 2) comprises the sequence given in SEQ ID NO: 15: GGCGGG.

In another embodiment, linker 2 does not contain cysteine (e.g.,

GGG, GGGS (SEQ ID NO: 29),

LGGGSG (SEQ GO NO: 261), GGGSGGGSGGG (SEQ GO NO: 262), ASTKG (SEQ GO NO: 30),

LEPKSS (SEQ ID NO: 33), APSSS (SEQ ID NO: 34), etc.), and the cysteine-containing heterodimer-promoting domain described below is used. Optionally, a cysteine-containing linker 2 and a cysteine-containing heterodimer-promoting domain may be used. The heterodimer-promoting domains

- 52 047726 heterodimers may be (SEQ ID NO: 16) or vepksc (SEQ ID NO: 17) or AEPKSC (SEQ ID NO: 18) on one polypeptide chain, and GFNRGEC (SEQ ID NO: 19) or FNRGEC (SEQ ID NO: 20) on the other polypeptide chain (US2007/0004909).

More preferably, however, the heterodimer-promoting domains of said diabodies are formed by one, two, three or four tandemly repeating helical domains of opposite charge that comprise a sequence of at least six, at least seven or at least eight amino acid residues such that the heterodimer-promoting domain has an overall charge (Apostolovic, B. et al. (2008) “pH-Sensitivity of the

E3/K3 Heterodimeric Coil,” Biomacromolecules 9:3173–3180; Arndt, K.M. et al. (2001) “Helix-stabilized Fv (hsFv) Antibody Fragments: Substituting the Constant Domains of a Fab Fragment for a Heterodimeric Coiled-coil Domain,” J. Molec. Biol. 312:221-228; Arndt, K.M. et al. (2002) “Comparison of in vivo Selection and Rational Design of Heterodimeric Coils,” Structure 10:1235–1248; Boucher, C. et al. (2010) “Protein Detection By Western Blot Via Coiled-Coil Interactions”, Analytical Biochemistry 399:138-140; Cachia, P.J. et al. (2004) "Synthetic Peptide Vaccine Development: Measurement Of Polyclonal Antibody Affinity And Cross-Reactivity Using A New Peptide Capture And Release System For Surface Plasmon Resonance Spectroscopy", J. Mol. Recognize 17:540-557; De Crescenzo, G.D. et al. (2003) “Real-Time Monitoring of the Interactions of Two-Stranded De Novo Designed Coiled-Coils: Effect of Chain Length on the Kinetic and Thermodynamic Constants of Binding,” Biochemistry 42:1754–1763; Fernandez-Rodriquez, J. et al. (2012) “Induced Heterodimerization And Purification Of Two Target Proteins By A Synthetic Coiled-Coil Tag,” Protein Science 21:511519; Ghosh, T.S. et al. (2009) "End-To-End And End-To-Middle Interhelical Interactions: New Classes Of Interacting Helix Pairs In Protein Structures", Neta. Crystallographica D65:10321041; Grigoryan, G. et al. (2008) “Structural Specificity in Coiled-Coil Interactions”, Curr. Opin. Struc. Biol. 18:477-483; Litowski, J.R. et al. (2002) "Designing Heterodimeric TwoStranded a-Helical Coiled-Coils: The Effects of Hydrophobicity And a-Helical Propensity On Protein Folding, Stability, And Specificity", J. Biol. Chern. 277:37272-37279; Steinkruger, J.D. et al. (2012) “The d'—d—d' Vertical Triad is Less Discriminating Than the a'—a—a' Vertical Triad in the Antiparallel Coiled-coil Dimer Motif,” J. Amer. Chern. Soc. 134(5):2626-2633; Straussman, R. et al. (2007) "Kinking the Coiled Coil - Negatively Charged Residues at the Coiled-coil Interface", J. Molec. Biol. 366:1232-1242; Tripet, B. et al. (2002) "Kinetic Analysis of the Interactions between Troponin C and the C-terminal Troponin I Regulatory Region and Validation of a New Peptide Delivery/Capture System used for Surface Plasmon Resonance", J.

Molec. Biol. 323:345-362; Woolfson, D.N. (2005) "The Design Of Coiled-Coil Structures And Assemblies", Adv. Prot. Chem. 70:79-112; Zeng, Y. et al. (2008) "A Ligand-Pseudoreceptor System Based On De Novo Designed Peptides For The Generation Of Adenoviral Vectors With Altered Tropism", J. Gene Med. 10:355-367).

Said repeating helical domains may be exact repeats or may comprise substitutions. For example, a helical domain from a heterodimer-promoting domain of a first polypeptide chain may comprise a sequence of eight amino acid residues selected to impart a negative charge to said heterodimer-promoting domain, and a helical domain from a heterodimer-promoting domain of a second polypeptide chain may comprise a sequence of eight amino acid residues selected to impart a positive charge to said heterodimer-promoting domain. Regardless of which helix is provided for the first or second polypeptide chains, it is envisaged that a helix of the opposite charge is used for the other polypeptide chain. The positively charged amino acid may be lysine, arginine, histidine, etc. and/or the negatively charged amino acid may be glutamic acid, aspartic acid, etc. The positively charged amino acid is preferably lysine and/or the negatively charged amino acid is preferably glutamic acid. It is possible to use only one heterodimer formation promoting domain (since said domain will inhibit homodimerization and thus promote heterodimerization), however, it is preferable that the first and second polypeptide chains of the diabodies according to

- 53 047726 of the present invention contained domains that promote the formation of a heterodimer.

According to a preferred embodiment of the present invention, one of the heterodimer formation promoting domains will comprise four tandem helical E-helix domains ( ^{SEQ ID NO: 21:} evaalek-evaalek-evaalek-evaalek) in which the glutamate residues will impart a negative charge at pH 7, while the other heterodimer formation promoting domains will comprise four tandem K-helix domains (SEQ ID NO: 22: kvaalke-kvaalke-kvaalke-kvaalke) in which the lysine residues will impart a positive charge at pH 7. The presence of these charged domains promotes the association of the first and second polypeptides and, accordingly, promotes the formation of the heterodimer. Particularly preferred is a heterodimer formation promoting domain in which one of the four tandem helical E-helix domains comprising the sequence given in SEQ ID NO: 21 has been modified to include a cysteine residue: EVAACEK-EVAALEK-EVAALEK-EVAALEK (SEQ ID NO: 23).

Similarly, particularly preferred is a heterodimer formation promoting domain in which one of the four tandem helical K-helix domains comprising the sequence set forth in SEQ ID NO: 22 has been modified to include a cysteine residue: kvaacke-kvaalke-kvaalke-kvaalke (SEQ ID NO: 24).

As disclosed in WO 2012/018687, in order to improve the pharmacokinetic properties of the diabodies in vivo, the diabody can be modified to contain a polypeptide portion of a serum binding protein at one or more ends of the diabody. Most preferably, said polypeptide portion of a serum binding protein will be inserted at the C-terminus of the diabody. Albumin is the most abundant protein in blood plasma and has a half-life of 19 days in humans. Albumin contains several small molecule binding sites that allow it to non-covalently bind to other proteins and thereby prolong their half-lives in serum. The albumin-binding domain 3 (ABD3) of the G protein of Streptococcus strain G148 consists of 46 amino acid residues that form a stable three-helical cord and has broad albumin-binding specificity (Johansson, M.U. et al. (2002) Structure, Specificity, And Mode Of Interaction For Bacterial Albumin-Binding Modules, J. Biol. Chem. 277(10):8114-8120). Accordingly, a particularly preferred polypeptide portion of a serum binding protein for improving the pharmacokinetic properties of the diabody in vivo is the albumin binding domain (ABD) of streptococcal protein G, and more preferably the albumin binding domain 3 (ABD3) of protein G of Streptococcus dysgalactiae strain G148 (SEQ ID NO: 25): laeakvlanr eldkygvsdy yknlidnaks

AEGVKALIDE ILAALP.

As disclosed in WO 2012/162068 (which is incorporated herein by reference), deimmunized variants comprising the sequence of SEQ ID NO: 25 are capable of reducing or eliminating binding to MHC class II. Based on the results of the combined mutations, the following substitution combinations are considered as preferred substitutions for obtaining said deimmunized ABD: 66D/70S+71A; 66S/70S+71A; 66S/70S+79A; 64A/65A/71A; 64A/65A/71A+66S; 64A/65A/71A+66D; 64A/65A/71A+66E; 64A/65A/79A+66S; 64A/65A/79A+66D; 64A/65A/79A+66E. ABD variants contain modifications L64A, I65A and D79A or modifications N66S, T70S and D79A. Deimmunized ABD variants containing the amino acid sequence:

LAEAKVLANR ELDKYGVSDY YKNLID ₆₆ NAKS ₇₀ A ₇₁ EGVKALIDE ILAALP (seq id ^. 26), or the amino acid sequence:

LAEAKVLANR ELDKYGVSDY YKNLIS ₆ 6NAKS ₇₀ VEGVKALIA79E ilaalp (seq id no: 27), or amino acid sequence:

LAEAKVLANR ELDKYGVSDY YKNLIS ₆₆ NAKS ₇₀ VEGVKALIA ₇₉ E ILAALP (seq id ^, 28), are particularly preferred because said deimmunized ABDs exhibit substantially wild-type binding while providing attenuated binding to MHC class II. Accordingly, the first polypeptide chain of said ABD-containing diabody comprises a peptide linker, preferably positioned C-terminally relative to the E-helix (or K-helix)-containing domain of said polypeptide chain, such that said linker is located between the E-helix (or K-helix)-containing domain and the ABD (which is preferably a deimmunized ABD). A preferred sequence for said linker peptide is the sequence set forth in SEQ ID NO: 29: gggs.

B. Bispecific diabodies containing Fc regions.

In one embodiment, the present invention provides bispecific diabodies comprising an Fc region capable of simultaneously binding to PD-1 and a second epitope (e.g., B7-H3, B7-H4, BTLA, CD40, CD80, CD86, CD137, CTLA-4, ICOS, KIR, LAG-3, MHC class I or II, OX40, PD-1, PD-L1, TCR, TIM-3, etc.). The addition of a CH2-CH3 domain of IgG to one or both polypeptide chains of the antibody such that the joining of the diabody chains results in the formation of an Fc region, increases the biological half-life and/or alters the valency of the diabody. Inclusion of the CH2-CH3 domains of IgG into both polypeptides of the diabody will ensure the formation of a two-chain bispecific diabody containing the Fc region (Fig. 2).

In another embodiment, inclusion of the CH2-CH3 domains of IgG in only one of the diabody polypeptides would result in the formation of a more complex four-chain bispecific diabody containing the Fc region (Fig. 3A-3C). 3C shows an exemplary four-chain diabody comprising a light chain constant domain (CL) and a heavy chain constant domain CH1, however, fragments of these domains, as well as other polypeptides, can be used as an alternative (see, for example, Figs. 3A and 3B, U.S. Patent Publications 2013-0295121; 2010-0174053; and 2009-0060910; European Patent Publication EP 2714079; EP 2601216; EP 2376109; EP 2158221; and PCT Publications WO 2012/162068; WO 2012/018687; WO 2010/080538). Therefore, for example, instead of the CH1 domain, a peptide comprising the amino acid sequence gvepksc (SEQ ID NO: 16), vepksc (SEQ ID NO: 17) or aepksc (SEQ ID NO: 18) derived from the hinge domain of human IgG can be used, and instead of the CL domain, the 6 C-terminal amino acids of the human kappa light chain GFNRGEC (SEQ ID NO: 19) _or FNRGEC (SEQ ID NO: 20) can _be used.

A typical four-chain diabody comprising the peptide is shown in Fig. 3A. Alternatively or in addition, a peptide comprising tandem helical domains of opposite charge, such as E-helix helical domains (SEQ ID NO: 21: EVAALEK-EVAALEK-EVAALEK-EVAALEK or SEQ ID NO:

23: EVAAGEK-EVAALEK-EVAALEK-EVAALEK);

and K-helix domains (SEQ ID NO: 22:

KVikALKE-KVikALKE-KVikALKE-KVikALKE or SEQ GO NO: 24: KVAACKE-KVAALKEKVAALKE - KVAALKE).

A typical helical domain that is part of a four-stranded diabody is shown in Fig. 3B.

The Fc region-containing diabody molecules of the present invention typically contain additional intervening linker peptides (linkers). Typically, the additional linkers will contain 3-20 amino acid residues. Additional or alternative linkers that may be used in the Fc region-containing antibody molecules of the present invention include:

GGGS (SEQ GO NO: 29), LGGGSG (SEQ GO NO: 261),

GGGSGGGSGGG (SEQ ID NO: 262), ASTKG (SEQ ID NO: 30), DKTHTCPPCP (SEQ ID NO: 31), EPKSCDKTHTCPPCP (SEQ ID NO: 32), LEPKSS (SEQ ID NO: 33), APSSS (SEQ ID NO: 34), and APSSSPME (SEQ ID NO: 35), LEPKSADKTHTCPPC SEQ ID NO: 36), GGC and GGG. SEQ ID NO: 33 can be used instead of GGG or GGC to facilitate cloning. In addition, the amino acids GGG or the sequence presented in SEQ ID NO: 33 can be immediately followed by the sequence presented in SEQ ID NO: 31 to form alternative linkers: GGGDKTHTCPPCP (SEQ ID NO: 263); and LEPKSSDKTHTCPPCP (SEQ ID NO: 37).

The Fc region-containing diabody molecule of the present invention may comprise an IgG hinge region in addition to or in place of a linker. An exemplary hinge region comprises:

EPKSCDKTHTCPPCP (SEQ ID NO: 32) from IgGl,

ERKCCVECPPCP (SEQ ID NO: 11) from IgG2, ESKYGPPCPSCP (SEQ ID NO: 12) from IgG4 and

ESKYGPPCPPCP (SEQ ID NO: 13) and a variant IgG4 hinge region containing a stabilizing substitution to reduce strand exchange.

As shown in Fig. 3A-3C, the diabodies of the present invention may comprise four different chains. The first and third polypeptide chains of said diabody comprise three domains: (i) a VL1 containing domain, (ii) a VH2 containing domain, (iii) a heterodimer forming domain, and (iv) a CH2-CH3 containing domain. The second and fourth polypeptide chains comprise: (i) a VL2 containing domain, (ii) a VH1 containing domain, and (iii) a heterodimer forming domain, wherein the heterodimer forming domains facilitate dimerization of the first/third polypeptide chains with the second/fourth polypeptide chains. The VL and/or VH domains of the third and fourth polypeptide chains and the VL and/or VH domains of the first and second polypeptide chains may be the same or different to provide tetravalent binding that is monospecific, bispecific or tetraspecific. The designation VL3 and VH3 denotes, respectively, the variable domain of the light chain and the variable domain of the heavy chain that bind to the third epitope of the said diabody. Similarly,

- 55 047726 the designation VL4 and VH4 means, respectively, the variable domain of the light chain and the variable domain of the heavy chain, which bind to the fourth epitope of the said diabody. The general structure of the polypeptide chains of typical four-chain diabodies containing the Fc region, according to the present invention is presented in Table 2:

Table 2

Bispecific diabody 2 chain NH2-VL2-VH1-HPD-COOH 1 chain NH ₂ -VL 1 - VH2-HPD-CH2-CH3 -COOH 1 chain NH2-VL 1 - VH2-HPD-CH2-CH3 -COOH 2 chain _NH2 -VL2-VH1-HPD-COOH Tetraspecific diabody 2 chain _NH2 -VL2-VH1-HPD-COOH 1 chain NH ₂ -VL 1 - VH2-HPD-CH2-CH3 -COOH 3 chain NH ₂ -VL3-VH4-HPD-CH2-CH3-COOH 4 chain _NH2 -VL4-VH3-HPD-COOH

HPD=heterodimer promoting domain.

According to a specific embodiment of the present invention, the diabodies of the present invention are bispecific, tetravalent (i.e., contain four epitope-binding sites), Fc-containing diabodies (Figs. 3A-3C) that are comprised of four complete polypeptide chains. The bispecific tetravalent Fc region-containing diabodies of the present invention comprise two epitope-binding sites immunologically specific for PD1 (which may be capable of binding to the same PD-1 epitope or to different PD-1 epitopes) and two epitope-binding sites specific for a second epitope (e.g., B7H3, B7-H4, BTLA, CD40, CD80, CD86, CD137, CTLA-4, ICOS, KIR, LAG-3, MHC class I or II, OX40, PD-L1, TCR, TIM-3, etc.).

According to another embodiment of the present invention, bispecific diabodies comprising Fc regions may comprise three polypeptide chains. The first polypeptide of said diabody comprises three domains: (i) a domain comprising VL1, (ii) a domain comprising VH2, and (iii) a domain comprising the sequence CH2-CH3. The second polypeptide of said diabodies comprises: (i) a domain comprising VL2, (ii) a domain comprising VH1, and (iii) a domain that promotes heterodimerization and the formation of a covalent bond with the first polypeptide chain of the diabody. The third polypeptide of said diabodies comprises the sequence CH2-CH3. Accordingly, the first and second polypeptide chains of said diabodies bind to each other to form a VL1/VH1 binding site that is capable of binding to the first epitope, as well as a VL2/VH2 binding site that is capable of binding to the second epitope. The first and second polypeptides are linked to each other by a disulfide bond involving cysteine residues in their respective third domains. Notably, the first and third polypeptide chains are linked to each other to form an Fc region that is stabilized by a disulfide bond. These diabodies have increased potency. Figures 4A and 4B show the structures of these diabodies. These bispecific diabodies containing the Fc region can have one of two orientations (Table 3).

Table 3

First orientation 3 chain NH2-CH2-CH3-COOH 1 chain NH2-VL 1 - VH2-HPD-CH2-CH3 -COOH 2 chain NH2-VL2-VH1-HPD-COOH Second orientation 3 chain NH2-CH2-CH3-COOH 1 chain NH2-CH2-CH3 - VL 1 - VH2-HPD-COOH 2 chain NH2-VL2-VH1-HPD-COOH

HPD=heterodimer promoting domain.

In a specific embodiment of the present invention, the diabodies of the present invention are bispecific, bivalent (i.e., contain two epitope-binding sites), Fc-containing diabodies (Figs. 4A-4B) that are comprised of three complete polypeptide chains. The bispecific, bivalent Fc-containing diabodies of the present invention comprise one epitope-binding site immunologically specific for PD-1 and one epitope-binding site specific for a second epitope (e.g., B7-H3, B7-H4, BTLA, CD40, CD80, CD86, CD137, CTLA-4, ICOS, KIR, LAG-3, MHC class I or II, OX40, PD-L1, TCR, TIM-3, etc.).

According to another embodiment of the present invention, bispecific diabodies containing

- 56 047726 containing Fc regions may comprise a total of five polypeptide chains. According to a specific embodiment of the present invention, two of said five polypeptide chains comprise the same amino acid sequence. The first polypeptide chain of said diabodies comprises: (i) a domain comprising VH1, (ii) a domain comprising CH1, and (iii) a domain comprising the sequence CH2-CH3. The first polypeptide chain may be a heavy chain of an antibody that comprises VH1 and a heavy chain constant region. The second and fifth polypeptide chains of said diabodies comprise: (i) a domain comprising VL1, and (ii) a domain comprising CL. The second and/or fifth polypeptide chains of said diabodies may be light chains of an antibody that comprises VL1 complementary to VH1 of the first/third polypeptide chain. The first, second and/or fifth polypeptide chains may be isolated from natural antibodies. In another embodiment, they can be constructed recombinantly. The third polypeptide chain of said diabodies comprises: (i) a domain containing VH1, (ii) a domain containing CH1, (iii) a domain containing the sequence CH2-CH3, (iv) a domain containing VL2, (v) a domain containing VH3, and (vi) a domain facilitating the formation of a heterodimer, wherein the domains facilitating the formation of a heterodimer facilitate the dimerization of the third chain with the fourth chain. The fourth polypeptide of said diabodies comprises: (i) a domain containing VL3, (ii) a domain containing VH2, and (iii) a domain that facilitates heterodimerization and the formation of a covalent bond with the third polypeptide chain of the diabody.

Accordingly, the first and second, as well as the third and fifth, polypeptide chains of said diabodies bind to form two VL1/VH1 binding sites capable of binding to the first epitope. The third and fourth polypeptide chains of said diabodies bind to form a VL2/VH2 binding site, which is capable of binding to the second epitope, as well as a VL3/VH3 binding site, which is capable of binding to the third epitope. The first and third polypeptides bind to each other via a disulfide bond involving cysteine residues in their respective constant regions. It is noteworthy that the first and third polypeptide chains bind to each other to form an Fc region. Said diabodies have increased efficiency. Fig. 5 shows the structure of said diabodies. It should be understood that the VL1/VH1, VL2/VH2 and VL3/VH3 domains may be the same or different so as to provide binding that is monospecific, bispecific or trispecific. However, as indicated herein, said domains are preferably selected so as to bind to PD-1 and a second epitope (e.g., B7-H3, B7-H4, BTLA, CD40, CD80, CD86, CD137, CTLA-4, ICOS, KIR, LAG-3, MHC class I or II, OX40, PD-L1, TCR, TIM-3, etc.).

The VL and VH domains of the polypeptide chains are selected to form VL/VH binding sites specific for a desired epitope. The VL/VH binding sites formed upon binding of the polypeptide chains may be the same or different to provide tetravalent binding that is monospecific, bispecific, trispecific, or tetraspecific. In particular, the VL and VH domains may be selected such that the bispecific antibody may comprise two binding sites for a first epitope and two binding sites for a second epitope, or three binding sites for the first epitope and one binding site for the second epitope, or two binding sites for the first epitope, one binding site for the second epitope, and one binding site for a third epitope (as shown in Fig. 5). The general structure of the polypeptide chains of typical five-chain diabodies containing an Fc region is presented in Table 4.

Table 4

Bispecific diabody (2x2) 2 chain NH2-VL1-CL-COOH 1 chain NH2-VH1-CH1-CH2-CH3-COOH 3 chain NH2-VH1 -CHI -CH2-CH3-VL2-VH2-HPD-COOH 5 chain NH2-VL1-CL-COOH 4 chain NH2-VL2-VH2-HPD-COOH Bispecific diabody (3x1) 2 chain NH2-VL1-CL-COOH 1 chain NH2-VH1-CH1-CH2-CH3-COOH 3 chain NH2-VH1 -CHI -CH2-CHZ -VL1-VH2-HPD-COOH 5 chain NH2-VL1-CL-COOH 4 chain NH2-VL2-VH1-HPD-COOH Trispecific diabody (2x1x1) 2 chain NH2-VL1-CL-COOH 1 chain NH2-VH1-CH1-CH2-CH3-COOH 3 chain NH2-VH 1 -CH 1-CH2-CHZ - VL2-VH3 -HPD-COOH 5 chain NH2-VL1-CL-COOH 4 chain NH2-VL3-VH2-HPD-COOH

HPD=heterodimer promoting domain.

- 57 047726

According to a specific embodiment of the present invention, the diabodies of the present invention are bispecific tetravalent (i.e., contain four epitope-binding sites) diabodies containing an Fc region that consist of five complete polypeptide chains containing two binding sites for a first epitope and two binding sites for a second epitope. According to one embodiment of the present invention, the bispecific tetravalent diabodies comprising an Fc region according to the present invention comprise two epitope-binding sites immunologically specific for PD-1 (which may be capable of binding to the same epitope of PD-1 or to different epitopes of PD-1) and two epitope-binding sites specific for a second epitope (e.g., B7-H3, B7-H4, BTLA, CD40, CD80, CD86, CD137, CTLA-4, ICOS, KIR, LAG-3 MHC class I or II, OX40, PD-L1, TCR, TIM-3, etc.). According to another embodiment of the present invention, the bispecific tetravalent diabodies comprising an Fc region according to the present invention comprise three epitope-binding sites immunologically specific for PD-1, which may be capable of binding to the same epitope of PD-1 or to different epitopes of PD-1), and one epitope-binding site specific for a second epitope (e.g., B7-H3, B7-H4, BTLA, CD40, CD80, CD86, CD137, CTLA-4, ICOS, KIR, LAG-3 MHC class I or II, OX40, PD-L1, TCR, TIM-3, etc.). According to another embodiment of the present invention, the bispecific tetravalent diabodies comprising an Fc region according to the present invention comprise one epitope-binding site immunologically specific for PD-1 and three epitope-binding sites specific for a second epitope (e.g., B7-H3, B7-H4, BTLA, CD40, CD80, CD86, CD137, CTLA-4, ICOS, KIR, LAG-3 MHC class I or II, OX40, PD-L1, TCR, TIM-3, etc.).

C. Bispecific trivalent binding molecules containing Fc regions.

According to another embodiment, according to the present invention, bispecific trivalent binding molecules are provided, comprising an Fc region, capable of simultaneously binding to a first epitope, a second epitope and a third epitope, wherein at least one of said epitopes is not identical to the other. Accordingly, said bispecific diabodies comprise VL1/VH1 domains, which are capable of binding to the first epitope, VL2/VH2 domains, which are capable of binding to the second epitope, and VL3/VH3 domains, which are capable of binding to the third epitope. According to one embodiment of the present invention, one or two of said epitopes are a PD-1 epitope and the other (or others) of said epitopes are not a PD-1 epitope (e.g., an epitope of B7-H3, B7-H4, BTLA, CD40, CD80, CD86, CD137, CTLA4, ICOS, KIR, LAG-3, MHC class I or II, OX40, PD-1, PD-L1, TCR, TIM-3, etc.). The bispecific trivalent binding molecules comprise three epitope-binding sites, two of which are diabody-specific binding domains that provide binding site A and binding site B, and one of which is a non-diabody-specific binding domain that provides binding site C (see, for example, Figs. 6A-6F and PCT application PCT/US15/33081; and PCT/US15/33076).

Typically, the trivalent binding molecules of the present invention will comprise four different polypeptide chains (see Figs. 6A-6B), however, the molecules may comprise fewer or more polypeptide chains, such as by hybridizing said polypeptide chains to each other (e.g., via a peptide bond) or by dividing said polypeptide chains to form additional polypeptide chains or by linking fewer or additional polypeptide chains via disulfide bonds. Figs. 6B-6F illustrate this aspect of the present invention by schematically representing said molecules comprising three polypeptide chains. As shown in Figs. 6A-6F, the trivalent binding molecules of the present invention may have alternative orientations in which the binding domains characteristic of diabodies are N-terminal (Figs. 6A, 6C and 6D) or C-terminal (Figs. 6B, 6E and 6F) relative to the Fc region. According to some embodiments of the present invention, the first polypeptide chain of the trivalent binding molecules of the present invention comprises: (i) a VL1 containing domain, (ii) a VH2 containing domain, (iii) a heterodimer formation promoting domain, and (iv) a CH2-CH3 containing domain. The VL1 and VL2 domains will be located in an N-terminal or C-terminal direction relative to the CH2-CH3 containing domain, as shown in Table 5 (Figs. 6A and 6B). The second polypeptide chain of the embodiments comprises: (i) a VL2 containing domain, (ii) a VH1 containing domain, and (iii) a heterodimer formation promoting domain. The third polypeptide chain of the embodiments comprises: (i) a VH3 containing domain, (ii) a CH1 containing domain, and (iii) a CH2-CH3 containing domain. The third polypeptide chain may be a heavy chain of an antibody that comprises VH3 and a heavy chain constant region. The fourth polypeptide of these embodiments comprises: (i) a VL3-containing domain, and (ii) a CL-containing domain. The fourth polypeptide chain may be a light chain of an antibody that comprises a VL3 complementary to the VH3 of the third polypeptide chain. The third or fourth polypeptide chains may be isolated from natural antibodies. Alternatively, they may be engineered using

- 58 047726 recombinant, synthetic or other methods.

The light chain variable domains of the first and second polypeptide chains are separated from the heavy chain variable domains of said polypeptide chains by an intervening spacer linker that is insufficient in length to connect their VL1/VH2 (or VL2/VH1) domains to form an epitope-binding site capable of binding to either of the first or second epitopes. A preferred intervening spacer peptide (linker 1) for this purpose comprises the sequence presented in (SEQ ID NO: 14): GGGSGGGG. Other domains of the trivalent binding molecules may be separated by one or more intervening spacer peptides, optionally comprising a cysteine residue. Exemplary linkers useful for preparing trivalent binding molecules are provided herein and are also provided in PCT application: PCT/US15/33081; and PCT/US15/33076. Accordingly, the first and second polypeptide chains of said trivalent binding molecules are joined to form a VL1/VH1 binding site capable of binding to a first epitope, as well as a VL2/VH2 binding site that is capable of binding to a second epitope. The third and fourth polypeptide chains of said trivalent binding molecules are joined to form a VL3/VH3 binding site that is capable of binding to a third epitope. It should be understood that the VL1/VH1, VL2/VH2 and VL3/VH3 domains can be the same or different so as to provide binding that is monospecific, bispecific or trispecific.

As described above, the trivalent binding molecules of the present invention may comprise three polypeptides. Trivalent binding molecules comprising three polypeptide chains may be prepared by joining domains of a fourth polypeptide to the N-terminus of a VH3-containing domain of a third polypeptide. In another embodiment, a third polypeptide chain of the trivalent binding molecule of the present invention is used, comprising the following three domains: (i) a VL3-containing domain, (ii) a VH3-containing domain, and (iii) a CH2-CH3-containing domain, wherein VL3 and VH3 are separated from each other by an intervening spacer peptide, which is sufficiently long (containing at least 9 or more amino acid residues) to allow the said domains to be joined to form an epitope-binding site. It should be understood that the VL1/VH1, VL2/VH2 and VL3/VH3 domains may be the same or different so as to provide binding that is monospecific, bispecific or trispecific. However, herein, said domains are preferably selected so as to bind to PD-1 and a second epitope (or a second and third epitope) (preferably, said epitopes are epitopes of B7-H3, B7-H4, BTLA, CD40, CD80, CD86, CD137, CTLA-4, ICOS, KIR, LAG-3 MHC class I or II, OX40, PD-L1, TCR, TIM-3, etc.). In particular, the VL and VH domains can be selected such that the trivalent binding molecule comprises two binding sites for a first epitope and one binding site for a second epitope, or one binding site for a first epitope and two binding sites for a second epitope, or one binding site for a first epitope, one binding site for a second epitope and one binding site for a third epitope. The general structure of the polypeptide chains of typical trivalent binding molecules according to the present invention is shown in Fig. 6A-6F and in Table 5.

Table 5

Four chains 1 orientation 2 chain NH2-VL2-VH1-HPD-COOH 1 chain NH2-VL 1 - VH2-HPD-CH2-CH3 -COOH 3 chain NH2-VH3 -CH 1-CH2-CH3 -COOH 4 chain NH2-VL3-CL-COOH Four chains 2 orientation 2 chain NH2-VL2-VH1-HPD-COOH 1 chain NH2 -CH2-CHZ— VL1-VH2-HPD COOH 3 chain NH2-VH3 -CH 1-CH2-CH3 -COOH 4 chain NH2-VL3-CL-COOH Three chains 1 orientation 2 chain NH2-VL2-VH1-HPD-COOH 1 chain NH2-VL 1 - VH2-HPD-CH2-CH3 -COOH 3 chain NH2-VL3 - VH3 -HPD-CH2-CH3 -COOH Three chains 2 orientation 2 chain NH2-VL2-VH1-HPD-COOH 1 chain NH2 -CH2-CH3-VL1-VH2-HPD COOH 3 chain NH2-VL3 - VH3 -HPD-CH2-CH3 -COOH

HPD=heterodimer promoting domain.

In one embodiment, the present invention provides bispecific trivalent binding molecules that comprise two epitope-binding sites for PD1 and one epitope-binding site for a second epitope present on a molecule other than PD1 (e.g., B7-H3, B7-H4, BTLA, CD40, CD80, CD86, CD137, CTLA-4, ICOS, KIR, LAG-3, MHC class I or II, OX40, PD-L1, TCR, TIM-3, etc.). The two epitope-binding sites for PD-1 may bind to the same epitope or different epitopes. According to another embodiment, the present invention provides bispecific trivalent binding molecules that comprise one epitope-binding site for PD-1 and two epitope-binding sites that bind to a second antigen present on a molecule other than PD-1 (e.g., B7-H3, B7H4, BTLA, CD40, CD80, CD86, CD137, CTLA-4, ICOS, KIR, LAG-3, MHC class I or II, OX40, PD-L1, TCR, TIM-3, etc.). The two epitope-binding sites for the second antigen may bind to the same epitope or different epitopes of the antigen (e.g., the same or different epitopes of LAG-3). As noted above, the bispecific trivalent binding molecules may comprise three or four polypeptide chains.

VII. Constant domains and Fc regions.

The present invention provides antibody constant domains that can be used in the production of PD-1 binding molecules (e.g., antibodies, diabodies, trivalent binding molecules, etc.) according to the present invention.

A preferred CL domain is the CL domain of the human IgG kappa light chain. The amino acid sequence of an exemplary CL domain of the human kappa light chain is the sequence shown below (SEQ ID NO: 8):

RTVAAPSVFI FPPSDEQLKS GTASWCLLN NFYPREAKVQ WKVDNALQSG

NSQESVTEQD SKDSTYSLSS TLTLSKADYE KHKVYACEVT HQGLSSPVTK SFNRGEC

In another embodiment, an exemplary CL domain is a CL domain of a human IgG lambda chain. The amino acid sequence of an exemplary CL domain of a human lambda chain is the sequence shown below (SEQ ID NO: 9):

QPKAAPSVTL FPPSSEELQA NKATLVCLIS DFYPGAVTVA WKADSSPVKA

GVETTPSKQS NNKYAASSYL SLTPEQWKSH RSYSCQVTHE GSTVEKTVAP

TECS

According to the present invention, the PD-1 binding molecules of the present invention may comprise an Fc region. The Fc region of said molecules of the present invention may be of any isotype (e.g., IgG1, IgG2, IgG3, or IgG4). The PD-1 binding molecules of the present invention may further comprise a CH1 domain and/or a hinge region. If a CH1 domain and/or a hinge region are present in said molecules, they may be of any isotype (e.g., IgG1, IgG2, IgG3, or IgG4), and preferably have the same isotype as the desired Fc region.

An exemplary CH1 domain is the CH1 domain of human IgG1.

The amino acid sequence of an exemplary CH1 domain of human IgG1 is as follows (SEQ ID NO: 10):

ASTKGPSVFP LAPSSKSTSG GTAALGCLVK DYFPEPVTVS WNSGALTSGV

HTFPAVLQSS GLYSLSSWT VPSSSLGTQT YICNVNHKPS NTKVDKRV

An exemplary CH1 domain is the CH1 domain of human IgG2.

The amino acid sequence of an exemplary CH1 domain of human IgG2 is as follows (SEQ ID NO: 257):

ASTKGPSVFP LAPCSRSTSE STAALGCLVK DYFPEPVTVS WNSGALTSGV

HTFPAVLQSS GLYSLSSWT VPSSNFGTQT YTCNVDHKPS NTKVDKTV

An exemplary CH1 domain is the CH1 domain of human IgG4.

The amino acid sequence of an exemplary CH1 domain of human IgG4 is as follows (SEQ ID NO: 254):

I ASTKGPSVFP LAPCSRSTSE STAALGCLVK DYFPEPVTVS WNSGALTSGV

HTFPAVLQSS GLYSLSSWT VPSSSLGTKT YTCNVDHKPS NTKVDKRV

One exemplary hinge region is a human IgG1 hinge region. The amino acid sequence of an exemplary human IgG1 hinge region is as follows (SEQ ID NO:32): epkscdkthtcppcp.

Another exemplary hinge region is a human IgG2 hinge region. The amino acid sequence of an exemplary human IgG2 hinge region is as follows (SEQ ID NO: 11): erkccvecppcp.

Another exemplary hinge region is a human IgG4 hinge region. The amino acid sequence of an exemplary human IgG4 hinge region is the sequence shown below (SEQ ID NO: 12): eskygppcpscp.

As described herein, the hinge region of IgG4 may contain a stabilizing

- 60 047726 mutation such as S228P substitution. The amino acid sequence of a typical stabilized hinge region of IgG4 is the sequence shown below (SEQ ID NO: 13): ESKYGPPCPPCP.

The Fc region of the Fc region-containing molecules (e.g., antibodies, diabodies, and trivalent molecules) of the present invention may be a complete Fc region (e.g., a complete Fc region of an IgG) or only a fragment of an Fc region. Optionally, the Fc region of the Fc region-containing molecules of the present invention does not have a C-terminal lysine amino acid residue. In particular, the Fc region of the Fc region-containing molecules of the present invention may be a modified version of the Fc region. Although the Fc region of the bispecific molecules comprising the Fc regions of the present invention may be capable of binding to one or more Fc receptors (e.g., FcyR), more preferably, said variant Fc region has altered binding characteristics to FcyRIA (CD64), FcyRIIA (CD32A), FcyRIIB (CD32B), FcyRIIIA (CD16a) or FcyRIIIB (CD16b) (relative to the binding exhibited by the wild-type Fc region) or will have a substantially reduced or absent ability to bind to an inhibitory receptor. Accordingly, the Fc region of the Fc region-containing molecules of the present invention may comprise some or all of the CH2 domains and/or some or all of the CH3 domains of a full Fc region, or may comprise a CH2 sequence variant and/or a CH3 sequence variant (which may comprise, for example, one or more insertions and/or one or more deletions with respect to the CH2 or CH3 domains of a full Fc region). Said Fc regions may contain regions of a polypeptide other than the Fc region, or may contain portions of Fc regions that are not complete in nature, or may contain non-natural orientations of the CH2 and/or CH3 domains (such as, for example, two CH2 domains or two CH3 domains, or, in the N-terminal to C-terminal direction, a CH3 domain connected to a CH2 domain, etc.).

Modifications of the Fc region identified as altering effector function are known in the art, including modifications that increase binding to activating receptors (e.g., FcγRIIA (CD16A)) and decrease binding to inhibitory receptors (e.g., FcγRIIB (CD32B)) (see, e.g., Stavenhagen, J.B. et al. (2007) Fc Optimization Of Therapeutic Antibodies Enhances Their Ability To Kill Tumor Cells in Vitro And Controls Tumor Expansion in Vivo Via Low-Affinity Activating Fcgamma Receptors, Cancer Res. 57(18):8882-8890). Exemplary embodiments of human IgG1 Fc regions with reduced binding to CD32B and/or increased binding to CD16A comprise the substitutions F243L, R292P, Y300L, V305I, or P296L. These amino acid substitutions may be present in the human IgG1 Fc region in any combination or additional combination. According to one embodiment of the present invention, the human IgG1 Fc region variant comprises the substitution F243L, R292P and Y300L. According to another embodiment of the present invention, the human IgG1 Fc region variant comprises the substitution F243L, R292P, Y300L, V305I and P296L.

In particular, it is preferred that the Fc regions of the polypeptide chains of the Fc region-containing molecules of the present invention exhibit reduced (or substantially absent) binding to FcγRIA (CD64), FcγRIIA (CD32A), FcγRIIB (CD32B), FcγRIIIA (CD16a) or FcγRIIIB (CD16b) (relative to the binding exhibited by the wild-type IgG1 Fc region (SEQ ID NO: 1). Variants of the Fc regions and mutated forms capable of mediating such altered binding are described above. According to a specific embodiment of the present invention, the Fc region-containing molecules of the present invention comprise an IgG Fc region that exhibits reduced ADCC effector function. According to a preferred embodiment of the present invention, the CH2-CH3 domain of the first and/or third polypeptide chains of said Fc region-containing molecules comprises any 1, 2 or 3 of the substitutions: L234A, L235A, N297Q and N297G. In another embodiment of the present invention, a variant human IgG Fc region comprises a N297Q substitution, an N297G substitution, L234A and L235A substitutions, or a D265A substitution, since these mutations abolish binding to FcRs. In another embodiment, a CH2-CH3 domain of an Fc region is used that inherently exhibits reduced (or substantially absent) binding to FcγRIIIA (CD16a) and/or reduced effector function (relative to the binding exhibited by the wild-type IgG1 Fc region (SEQ ID NO: 1)). In a specific embodiment of the present invention, molecules comprising Fc regions of the present invention comprise an IgG2 Fc region (SEQ ID NO: 2) or an IgG4 Fc region (SEQ ID NO: 4). When using the Fc region of IgG4, it is within the scope of the present invention to introduce a stabilizing mutation, such as the S228P substitution in the hinge region described above (see, for example, SEQ ID NO: 13). Since the N297G, N297Q, L234A, L235A and D265A substitutions eliminate effector function, in cases where effector function is desired, these substitutions will preferably not be used.

In particular, it is preferred that the Fc regions of the polypeptide chains of the Fc region-containing molecules of the present invention exhibit an increased serum half-life (relative to the half-life exhibited by the corresponding wild-type Fc). Fc region variants and mutated forms having an increased serum half-life,

- 61 047726 are described above. According to a preferred embodiment of the present invention, the CH2CH3 domain of the first and/or third polypeptide chains of said Fc region-containing molecules comprises any 1, 2 or 3 of the substitutions: M252Y, S254T and T256E. Also included within the scope of the present invention are Fc region-containing molecules according to the present invention comprising Fc region variants comprising:

(A) one or more mutations that alter effector function and/or FcyR; and (B) one or more mutations that increase serum half-life.

A preferred IgG1 sequence for the CH2 and CH3 domains of the Fc region-containing molecules of the present invention will comprise the substitutions L234A/L235A/M252Y/S254T/T256E (SEQ ID NO: 258):

APEAAGGPSV FLFPPKPKDT LYITREPEVT CVWDVSHED PEVKFNWYVD

GVEVHNAKTK PREEQYNSTY RWSVLTVLH QDWLNGKEYK CKVSNKALPA

PIEKTISKAK GQPREPQVYT LPPSREEMTK NQVSLTCLVK GFYPSDIAVE

WESNGQPENN YKTTPPVLDS DGSFFLYSKL TVDKSRWQQG NVFSCSVMHE ALHNHYTQKS LSLSPGX where X is lysine (K) or absent.

A preferred IgG4 sequence for the CH2 and CH3 domains of the Fc region-containing molecules of the present invention will comprise the substitutions M252Y/S254T/T256E (SEQ ID NO: 259):

APEFLGGPSV FLFPPKPKDT LYITREPEVT CVWDVSQED PEVQFNWYVD

GVEVHNAKTK PREEQFNSTY RWSVLTVLH QDWLNGKEYK CKVSNKGLPS

SIEKTISKAK GQPREPQVYT LPPSQEEMTK NQVSLTCLVK GFYPSDIAVE

WESNGQPENN YKTTPPVLDS DGSFFLYSRL TVDKSRWQEG NVFSCSVMHE ALHNHYTQKS LSLSLGX where X is lysine (K) or absent.

For diabodies and trivalent binding molecules whose first and third polypeptide chains are not identical, it is desirable to reduce or prevent homodimerization between the CH2-CH3 domains of the first two polypeptide chains or between the CH2-CH3 domains of the third two polypeptide chains. The CH2 and/or CH3 domains of said polypeptide chains need not have identical sequences and are advantageously modified to promote complex formation between the two polypeptide chains. For example, an amino acid substitution (preferably a substitution with an amino acid containing a bulky side group forming a knob, such as tryptophan) can be introduced into a CH2 or CH3 domain such that steric influence will prevent interaction with a domain mutated in a similar manner and will result in pairing of the mutated domain with a domain into which a complementary or promoting mutation, i.e. a hole, has been introduced (e.g., a glycine substitution). Such groups of mutations can be introduced into any pair of polypeptides containing the CH2-CH3 domains that form the Fc region. Methods for modifying proteins to promote heterodimerization over homodimerization are well known in the art, particularly with respect to the modification of immunoglobulin-like molecules, and are included within the scope of the present invention (see, for example, Ridgway et al. (1996) 'Knobs-Into-Holes' Engineering Of Antibody CHS Domains For Heavy Chain Heterodimerization, Protein Engr. 9:617-621, Atwell et al. (1997) Stable Heterodimers From Remodeling The Domain Interface Of A Homodimer Using A Phage Display Library, J. Mol. Biol. 270: 26-35, and Xie et al. (2005) A New Format Of Bispecific Antibody: Highly Efficient Heterodimerization, Expression And Tumor Cell Lysis, J. Immunol. Methods 296:95-101; each of which is incorporated herein by reference in its entirety). Preferably, the knob is introduced into the CH2-CH3 domains of the first polypeptide chain and the hole is introduced into the CH2-CH3 domains of the third polypeptide chain of diabodies comprising three polypeptide chains. Accordingly, the knob will prevent homodimerization of the first polypeptide chain via its CH2 and/or CH3 domains. Since the third polypeptide chain preferably contains a substitution providing a hole, it will heterodimerize with the first polypeptide chain and also homodimerize with itself. This strategy can be used for antibodies and trivalent binding molecules comprising three, four or five chains described above, in which the knob is introduced into the CH2-CH3 domains of the first polypeptide chain and the hole is introduced into the CH2-CH3 domains of the third polypeptide chain.

A preferred knob is created by modifying the Fc region of IgG to include the modification T366W. A preferred hole is created by modifying the Fc region of IgG to include the modification T366S, L368A, and Y407V. To facilitate purification of the hole-bearing third polypeptide chain homodimer from the final bispecific heterodimeric molecule comprising the Fc region, the protein A binding site of the hole-bearing CH2 and CH3 domains of the third polypeptide chain is preferably mutated by substituting an amino acid at position 435 (H435R). Accordingly, the hole-bearing third polypeptide chain homodimer will not bind to protein A, whereas the bis

- 62 047726 the specific heterodimer will retain its ability to bind to protein A via the protein A binding site in the first polypeptide chain. In another embodiment, the third polypeptide chain bearing the hole may contain amino acid substitutions at positions 434 and 435 (N434A/N435K).

A preferred IgG1 amino acid sequence for the CH2 and CH3 domains of the first polypeptide chain of the Fc region-containing molecule of the present invention will comprise the knob-bearing sequence shown below (SEQ ID NO: 6):

APEAAGGPSV FLFPPKPKDT LMISRTPEVT CVWDVSHED PEVKFNWYVD

GVEVHNAKTK PREEQYNSTY RWSVLTVLH QDWLNGKEYK CKVSNKALPA

PIEKTISKAK GQPREPQVYT LPPSREEMTK NQVSLWCLVK GFYPSDIAVE

WESNGQPENN YKTTPPVLDS DGSFFLYSKL TVDKSRWQQG NVFSCSVMHE ALHNHYTQKS LSLSPGX where X is lysine (K) or absent.

A preferred IgG1 amino acid sequence for the CH2 and CH3 domains of the second polypeptide chain of the Fc region-containing molecule of the present invention comprising two polypeptide chains (or the third polypeptide chain of the Fc region-containing molecule comprising three, four or five polypeptide chains) will comprise the cavity-bearing sequence presented below (SEQ ID NO: 7):

APEAAGGPSV FLFPPKPKDT LMISRTPEVT CVWDVSHED PEVKFNWYVD

GVEVHNAKTK PREEQYNSTY RWSVLTVLH QDWLNGKEYK CKVSNKALPA

PIEKTISKAK GQPREPQVYT LPPSREEMTK NQVSLSCAVK GFYPSDIAVE

WESNGQPENN YKTTPPVLDS DGSFFLVSKL TVDKSRWQQG NVFSCSVMHE

ALHNRYTQKS LSLSPGX where X is lysine (K) or absent.

As should be noted, the CH2-CH3 domains of the sequences set forth in SEQ ID NO: 6 and SEQ ID NO: 7 comprise a substitution at position 234 with an alanine residue and at position 235 with an alanine residue and, accordingly, form an Fc region that exhibits reduced (or substantially absent) binding to FcγRIA (CD64), FcγRIIA (CD32A), FcγRIIB (CD32B), FcγRIIIA (CD16a), or FcγRIIIB (CD16b) (relative to the binding exhibited by the wild-type Fc region (SEQ ID NO: 1). Also included within the scope of the present invention are said CH2-CH3 domains that comprise alternative and/or additional substitutions that modify the effector function and/or binding activity of the Fc region with FyR. Also included within the scope of the present invention are said CH2-CH3 domains that further comprise one or more amino acid substitutions, increasing the half-life. In particular, the scope of the present invention includes said CH2-CH3 cavity-bearing and knob-bearing domains, which further comprise M252Y/S254T/T256E.

Preferably, the first polypeptide chain will comprise a CH2-CH3 knob-bearing sequence, such as that shown in SEQ ID NO: 6. However, as will be noted, a CH2-CH3 hole-bearing domain (e.g., SEQ ID NO: 7) may be used in the first polypeptide chain, in which case a CH2-CH3 knob-bearing domain (e.g., SEQ ID NO: 6) may be used in a second polypeptide chain of an Fc region-containing molecule of the present invention comprising two polypeptide chains (or in a third polypeptide chain of an Fc region-containing molecule comprising three, four or five polypeptide chains).

As described in detail above, included within the scope of the present invention are molecules comprising an Fc region (e.g., antibodies and diabodies comprising an Fc region) comprising wild-type CH2 and CH3 domains or comprising CH2 and CH3 domains with the combinations of substitutions described above. An exemplary amino acid sequence of an IgG1 CH2-CH3 domain comprising such variants is the sequence set forth below (SEQ ID NO: 260):

APEX1X2GGPSV FLFPPKPKDT LX3IX4RX5PEVT CVWDVSHED PEVKFNWYVD

GVEVHNAKTK PREEQYNSTY RWSVLTVLH QDWLNGKEYK CKVSNKALPA

PIEKTISKAK GQPREPQVYT LPPSREEMTK NQVSLX ₆ CX ₇ VK GFYPSDIAVE

WESNGQPENN YKTTPPVLDS DGSFFLXgSKL TVDKSRWQQG NVFSCSVMHE

ALHX9X10YTQKS LSLSPGX11 in which:

(a) X1 and _X2 are both L (wild type) or both are A (reduced binding to FcyR);

(b) X ₃ , X ₄ and X ₅ , respectively, represent M, S and T (wild type) or Y, T and E (extended half-life), (c) X ₆ , X ₇ and X ₈ , respectively, represent T, L and Y (wild type) or W, L and Y (protrusion), or S, A and V (valley);

- 63 047726 (d) X ₉ and X ₁₀ , respectively, are N and H (wild type), or are N and R (no binding to protein A), or A and K (no binding to protein A); and (e) X11 is K or is absent.

In other embodiments, the present invention includes within its scope PD-1 binding molecules comprising CH2 and/or CH3 domains that have been modified to promote heterodimerization over homodimerization using mutations known in the art, such as those disclosed in PCT publications WO 2007/110205; WO 2011/143545; WO 2012/058768; WO 2013/06867, all of which are incorporated herein by reference in their entireties.

VIII. Bispecific binding molecules PD-1xLAG-3.

The present invention provides in particular PD-1xLAG-3 bispecific binding molecules (e.g. bispecific antibodies, bispecific diabodies, etc.) comprising an epitope-binding fragment of an antibody to PD-1, and preferably one of the novel human PD-1 antibodies provided herein, and an epitope-binding fragment of an antibody to human LAG-3, preferably one of the novel human LAG-3 antibodies provided herein. Preferred PD-1xLAG-3 bispecific binding molecules according to the present invention comprise epitope-binding fragments of antibodies that allow them to coordinately bind to two different epitopes: a PD-1 epitope and a LAG-3 epitope, so as to attenuate the inhibitory activity of said molecules. In the present invention, such attenuation refers to a decrease by at least 20%, a decrease by at least 50%, a decrease by at least 80%, or a decrease by at least 90% of the detectable inhibitory activity of PD-1 and/or LAG-3, or to a complete elimination of the detectable inhibitory activity of PD-1 and/or LAG-3. The selection of epitope-binding fragments (e.g., VL and VH domains) of the anti-human PD-1 antibody and the anti-LAG-3 antibody is coordinated such that the polypeptide chains that form said PD-1xLAG-3 bispecific binding molecules are joined to form at least one functional antigen-binding site that is specific for a first antigen (i.e., PD-1 or LAG-3) and at least one functional antigen-binding site that is specific for a second antigen (i.e., PD-1 or LAG-3, depending on the identity of the first antigen).

According to a specific embodiment of the present invention, the PD-1xLAG-3 bispecific binding molecule of the present invention is a bispecific antibody that preferably comprises two, three, four or five polypeptide chains as described herein. In another specific embodiment, a PD-1xLAG-3 bispecific binding molecule of the present invention is a bispecific antibody that preferably comprises two, three or four polypeptide chains as described herein (see, e.g., WO 2007/024715; WO2007/110205; WO 2009/080251; WO 2009/080254; WO 2009/089004; WO 2011/069104; WO 2011/117329; WO 2011/131746; WO 2011/133886; WO 2011/143545; WO 2012/023053; WO 2013/060867, all of which are fully are incorporated into the present invention by reference).

A. Antibodies to human LAG-3.

Exemplary antibodies that are immunospecific for human LAG-3 are provided below. Additional desirable antibodies can be obtained by isolating hybridomas that secrete antibodies detected using LAG-3 or a peptide fragment thereof, or by screening recombinant antibody libraries for binding to LAG-3 or a peptide fragment thereof. Human LAG-3 (including the 28 amino acid residue signal sequence (underlined) and the 497 amino acid residue mature protein) has the amino acid sequence provided below (SEQ ID NO: 38):

MWEAQFLGLL FLQPLWVAPV KPLQPGAEVP WWAQEGAPA QLPCSPTIPL

QDLSLLRRAG VTWQHQPDSG PPAAAPGHPL APGPHPAAPS SWGPRPRRYT VLSVGPGGLR SGRLPLQPRV QLDERGRQRG DFSLWLRPAR RADAGEYRAA VHLRDRALSC RLRLRLGQAS MTASPPGSLR ASDWVILNCS FSRPDRPASV HWFRNRGQGR VPVRESPHHH LAESFLFLPQ VSPMDSGPWG CILTY RDGFN VSIMYNLTVL GLEPPTPLTV YAGAGSRVGL PCRLPAGVGT RSFLTAKWTP PGGGPDLLVT GDNGDFTLRL EDVSQAQAGT YTCHIHLQEQ QLNATVTLAI ITVTPKSFGS PGSLGKLLCE VTPVSGQERF VWSSLDTPSQ RSFSGPWLEA QEAQLLSQPW QCQLYQGERL LGAAVYFTEL SSPGAQRSGR APGALPAGHL LLFLILGVLS LLLLVTGAFG FHLWRRQWRP RRFSALEQGI HPPQAQSKIE ELEQEPEPEP EPEPEPEPEP EPEQL

- 64 047726

1. Monoclonal antibody to LAG-3 A.

The anti-human LAG-3 antibody, BMS-986016 (25F7; Medarex/BMS), herein designated anti-LAG-3 mAb, and variants thereof have been described (see, e.g., WO 2014/008218). The amino acid sequence of the heavy chain variable domain of the anti-LAG-3 mAb comprises the amino acid sequence set forth below (SEQ ID NO: 39) (CDRs are underlined): QVQLQQWGAG LLKPSETLSL TCAVYGGSFS DYYWNWIRQP PGKGLEWIGE INHNGNTNSN PSLKSRVTLS LDTSKNQFSL KLRSVTAADT AVYYCAFGYS DYEYNWFDPW GQGTLVTVSS

The amino acid sequence of the light chain variable domain of the anti-LAG-3A MAT contains the amino acid sequence shown below (SEQ ID NO: 40) (CDRs are underlined):

EIVLTQSPAT LSLSPGERAT LSCRASQSIS SYLAWYQQKP GQAPRLLIYD

ASNRATGIPA RFSGSGSGTD FTLTISSLEP EDFAVYYCQQ RSNWPLTFGQ

GTNLEIK

Recently, additional murine anti-human LAG-3 antibodies have been identified that have unique binding characteristics (see U.S. Patent Application Ser. No. 62/172,277). Preferred PD-1xLAG-3 bispecific binding molecules of the present invention comprise epitope-binding fragments of an anti-LAG-3 antibody, anti-LAG-3 Mab 1, or anti-LAG-3 Mab 6 that bind to a novel epitope and do not compete with BMS-986016 for binding to LAG-3. Particularly preferred are PD-1xLAG-3 bispecific binding molecules of the present invention that comprise humanized VH and/or VL domains of anti-LAG-3 Mab 1 or anti-LAG-3 Mab 6.

2. MAT to LAG-3 1.

The amino acid sequence of the VH domain of MAT to LAG-3 1 (SEQ ID NO: 41) is shown below (CDR _H residues are underlined).

QIQLVQSGPE LKKPGETVKI SCKASGYTFR NYGMNWVKQA PGKVLKWMGW

INTYTGESTY ADDFEGRFAF SLGTSASTAY LQINILKNED TATYFCARES

LYDYYSMDYW GQGTSVTVSS

CDRhI MAT to LAG-3 1 (SEQ Ш NO: 42): RNYGMN

CDR _H 2 MAT to LAG-3 1 (SEQ Ш NO43): WINTYTGESTYADDFEG

CDR _H 3 MAT to LAG-3 1 (SEQ Ш NO: 44): ESLYDYYSMDY

The amino acid sequence of the VL domain of MAT to LAG-3 1 (SEQ ID NO: 45) is shown below (CDRL residues are underlined):

DVWTQTPLT LSVTIGQPAS ISCKSSQSLL HSDGKTYLNW LLQRPGQSPE

RLIYLVSELD SGVPDRFTGS GSGTDFTLKI SRVEAEDLGV YYCWQGTHFP

YTFGGGTKLE IK

CDRlI MAT to LAG-3 1 (SEQ Ш NO: 46): KSSQSLLHSDGKTYLN

CDR _l 2 MAT to LAG-3 1 (SEQ Ш NO: 47): LVSELDS

CDR _l 3 MAT to LAG-3 1 (SEQ Ш NO: 48): WQGTHFPYT

Two exemplary humanized VH domains of the anti-LAG-3 MAT 1, designated herein as VH1 of the anti-hLAG-3 MAT 1 and VH2 of the anti-hLAG-3 MAT, and four exemplary humanized VL domains of the anti-LAG-3 MAT 1, VL1 of the anti-hLAG-3 MAT 1, VL2 of the anti-hLAG-3 MAT 1, VL3 of the anti-hLAG-3 MAT 1, and VL4 of the anti-hLAG-3 MAT 1, are provided below. Any of the humanized VL domains can be paired with any of the humanized VH domains to produce a LAG-3 binding domain. Accordingly, any antibody comprising one of the humanized VL domains paired with a humanized VH domain is generally referred to as an anti-hLAG-3 mAb 1, and specific combinations of humanized VH/VL domains are referred to by reference to specific VH/VL domains, e.g., a humanized antibody comprising a VH1 anti-hLAG-3 mAb 1 and a VL2 anti-hLAG-3 mAb 1 is specifically referred to as an anti-hLAG-3 mAb 1 (1.2).

The amino acid sequence of the VH domain from VH1 MAT to hLAG-3 1 (SEQ ID NO: 49) is shown below (CDRH residues are underlined):

QVQLVQSGAE VKKPGASVKV SCKASGYTFT NYGMNWVRQA PGQGLEWMGW

INTYTGESTY ADDFEGRFVF SMDTSASTAY LQISSLKAED TAVYYCARES LYDYYSMDYW GQGTTVTVSS

The amino acid sequence of the VH domain from VH2 MAT to hLAG-3 1 (SEQ ID NO: 50) is shown below (CDRH residues are underlined):

- 65 047726

QVQLVQSGAE VKKPGASVKV SCKASGYTFT NYGMNWVRQA PGQGLEWMGW INTYTGESTY ADDFEGRFVF SMDTSASTAY LQISSLKAED TAVYFCARES LYDYYSMDYW GQGTTVTVSS

The amino acid sequence of the VL domain from VL1 MAT to hLAG-3 1 (SEQ ID NO: 51) is shown below (CDR _L residues are underlined):

DIVMTQTPLS LSVTPGQPAS ISCKSSQSLL HSDGKTYLNW LLQKPGQSPE

RLIYLVSELD SGVPDRFSGS GSGTDFTLKI SRVEAEDVGV YYCWQGTHFP

YTFGGGTKVE IK

The amino acid sequence of the VL domain from VL2 MAT to hLAG-3 1 (SEQ ID NO: 52) is shown below (CDRL residues are underlined):

DIVMTQTPLS LSVTPGQPAS ISCKSSQSLL HSDGKTYLNW LLQRPGQSPE

RLIYLVSELD SGVPDRFSGS GSGTDFTLKI SRVEAEDVGV YYCWQGTHFP YTFGGGTKVE IK

The amino acid sequence of the VL domain from VL3 MAT to hLAG-3 1 (SEQ ID NO: 53) is shown below (CDRL residues are underlined):

DIVMTQTPLS LSVTPGQPAS ISCKSSQSLL HSDGKTYLNW LLQKPGQPPE

RLIYLVSELD SGVPDRFSGS GSGTDFTLKI SRVEAEDVGV YYCWQGTHFP

YTFGGGTKVE IK

The amino acid sequence of the VL domain from VL4 MAT to hLAG-3 1 (SEQ ID NO: 54) is shown below (CDRL residues are underlined):

DIVMTQTPLS LSVTPGQPAS ISCKSSQSLL HSDAKTYLNW LLQKPGQPPE

RLIYLVSELD SGVPDRFSGS GSGTDFTLKI SRVEAEDVGV YYCWQGTHFP YTFGGGTKVE IK

CDRL1 of the VL domain from VL4 MAT to hLAG-3 1 contains a substitution of glycine by an alanine residue and contains the amino acid sequence: kssqsllhsdaktyln (SEQ ID NO: 55), the substituted alanine is underlined.

According to the present invention, it is envisaged that a similar substitution may be included in any of the above-described domains of the CDRL1 MAT to LAG-3 1.

3. MAT to LAG-3 6.

The amino acid sequence of the VH domain of MAT to LAG-3 6 (SEQ ID NO: 56) is shown below (CDR _H residues are underlined):

EVLLQQSGPE LVKPGASVKI PCKASGYTFT DYNMDWVKQS HGESLEWIGD INPDNGVTIY NQKFEGKATL TVDKSSSTAY MELRSLTSED TAVYYCAREA DYFYFDYWGQ GTTLTVSS

CDRhI MAT to LAG-3 6 (SEQ Ш NO: 57): DYNMD

CDR _H 2 MAT to LAG-3 6 (SEQ Ш NO: 58): DINPDNGVTIYNQKFEG

CDR _H 3 MAT to LAG-3 6 (SEQ Ш NO: 59): EADYFYFDY

The amino acid sequence of the VL domain of MAT to LAG-3 6 (SEQ ID NO: 60) is shown below (CDRL residues are underlined):

DIVMTQSHRF MSTSVGDRVS ITCKASQDVS SWAWYQQKP GQSPKLLIFS ASYRYTGVPD RFTGSGSGTD FTFTISSVQA ADLAVYYCQQ HYSTPWTFGG GTKLEIK

CDRlI MAT to LAG-3 6 (SEQ Ш NO: 61): KASQDVSSWA

CDR _l 2 MAT to LAG-3 6 (SEQ Ш NO: 62): SASYRYT

CDR _l 3 MAT to LAG-3 6 (SEQ ID NO: 63): HYSTPWT

Two exemplary humanized VH domains of the anti-LAG-3 MAT 6, designated herein as VH1 MAT 6 to hLAG-3 and VH2 MAT 6, and two exemplary humanized VL domains of the anti-LAG-3 MAT 6, designated VL1 MAT 6 to hLAG-3 1 and VL2 MAT 6 to hLAG-3, are shown below. Any of the humanized VL domains can be paired with any of the humanized VH domains to produce a LAG-3 binding domain. Accordingly, any antibody comprising one of the humanized VL domains paired with a humanized VH domain is generally referred to as an anti-hLAG-3 mAb 6, and specific combinations of humanized VH/VL domains are referred to by reference to specific VH/VL domains, e.g., a humanized antibody comprising VH1 of an anti-hLAG-3 mAb 6 and VL2 of an anti-hLAG-3 mAb 6 is specifically referred to as an anti-hLAG-3 mAb 6 (1.2).

- 66 047726

The amino acid sequence of the VH domain from VH1 MAT to hLAG-3 6 (SEQ ID NO: 294) is shown below (CDR _H residues are underlined):

QVQLVQSGAE VKKPGASVKV SCKASGYTFT DYNMDWVRQA PGQGLEWMGD

INPDNGVTIY NQKFEGRVTM TTDTSTSTAY MELRSLRSDD TAVYYCAREA DYFYFDYWGQ GTTLTVSS

The amino acid sequence of the VH domain from VH2 MAT to hLAG-3 6 (SEQ ID NO: 295) is shown below (CDRH residues are underlined):

EVQLVESGGG LVKPGGSLRL SCAASGFTFS DYNMDWVRQA PGKGLEWVSD

INPDNGVTIY NQKFEGRFTI SRDNAKNSLY LQMNSLRAED TAVYYCAREA DYFYFDYWGQ GTTLTVSS

The amino acid sequence of the VL domain from VL1 MAT to hLAG-3 6 (SEQ ID NO: 296) is shown below (CDRL residues are underlined):

DIQMTQSPSS LSASVGDRVT ITCRASQDVS SWAWYQQKP GKAPKLLIYS

ASYRYTGVPS RFSGSGSGTD FTLTISSLQP EDFATYYCQQ HYSTPWTFGG GTKLEIK

The amino acid sequence of the VL domain from VL2 MAT to hLAG-3 6 (SEQ ID NO: 297) is shown below (CDRL residues are underlined):

DIVMTQSPSS LSASVGDRVT ITCRASQDVS SWAWYQQKP GKAPKLLIYS

ASYRYTGVPD RFSGSGSGTD FTFTISSLQP EDIAVYYCQQ HYSTPWTFGG GTKLEIK

CDRL1 of the VL domain from VL1 and VL2 of MAT to hLAG-3 6 contains a substitution of lysine with the amino acid residue arginine, and contains the amino acid sequence: kasqdvsswa (SEQ ID NO: 298), the substituted arginine is underlined.

According to the present invention, it is envisaged that a similar substitution may be included in any of the CDRL1 MAT domains to LAG-3 6 described above.

B. Exemplary four-stranded diabodies containing the Fc region and having an E^-helix.

Four exemplary bispecific four-chain Fc region-containing diabodies of PD-1xLAG-3 having heterodimer-promoting domains with Eβ-helices were obtained (designated DART A, DART B, DART C, and DART I). The structure of these Fc region-containing diabodies is described in detail below. These exemplary PD-1xLAG-3 diabodies are illustrative and in no way limit the scope of the present invention.

1. DART A.

DART A is a bispecific four-chain diabody comprising an Fc region containing two binding sites specific for PD-1, two binding sites specific for LAG-3, a variant of the IgG4 Fc region modified to increase its half-life, and cysteine-containing domains that facilitate heterodimer formation with an E/K helix. The first and third polypeptide chains of DART A comprise, from N-terminus to C-terminus: N-terminus, the VL domain of a monoclonal antibody capable of binding to LAG-3 (VL _lag-3 VL4 MAb to hLAG-3 1) (SEQ ID NO: 54); an intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)); the VH domain of a monoclonal antibody capable of binding to PD-1 (VH _PD-1 VH1 MAb to hPD-1 7) (SEQ ID NO: 147); a cysteine-containing intermediate linker peptide (linker 2: ggcggg (SEQ ID NO: 15)); a cysteine-containing heterodimer formation promoting domain (E-helix) (EVAACEK-EVAALEK-EVAALEK-EVAALEK (SEQ ID NO: 23)); a stabilized hinge region of IgG4 (SEQ ID NO: 13); a variant CH2-CH3 domain of IgG4 containing the M252Y/S254T/T256E substitutions and lacking the C-terminal residue (SEQ ID NO: 259); and a C-terminus.

The amino acid sequence of the first and third polypeptide chains of DART A is a variant of the sequence shown in SEQ ID NO: 267:

DIVMTQTPLS LSVTPGQPAS ISCKSSQSLL HSDXiKTYLNW LLQKPGQPPE

RLIYLVSELD SGVPDRFSGS GSGTDFTLKI SRVEAEDVGV YYCWQGTHFP YTFGGGTKVE IKGGGSGGGG QVQLVQSGAE VKKPGASVKV SCKASGYSFT SYWMNWVRQA PGQGLEWIGV IHPSDSETWL DQKFKDRVTI TVDKSTSTAY MELSSLRSED TAVYYCAREH G QGTLVTVSSG GCGGGEVAAC EKEVAALEKE VAALEKEVAA LEKESKYGPP CPPCPAPEFL GGPSVFLFPP KPKDTLX2IX3R X4PEVTCVWD VSQEDPEVQF NWYVDGVEVH NAKTKPREEQ FNSTYRVVSV LTVLHQDWLN GKEYKCKVSN KGLPSSIEKT ISKAKGQPRE PQVYTLPPSQ EEMTKNQVSL TCLVKGFYPS DIAVEWESNG QPENNYKTTP PVLDSDGSFF LYSRLTVDKS RWQEGNVFSC SVMHEALHNH YTQKSLSLSL G in which X1, _X2 , _X3 , and _X4 are independently chosen, and in which X1 is A or G; X2

- 67 047726 represents Y or M; X3 represents T or S; and _X4 represents E or T.

The amino acid sequences of the first and third polypeptide chains of DART A are the sequence shown in SEQ ID NO: 267, wherein X1 is A; _X2 is Y; X3 is T; and _X4 is E.

The second and fourth polypeptide chains of DART A comprise, in the direction from N-terminus to C-terminus: N-terminus, VL domain of a monoclonal antibody capable of binding to PD-1 (VL _PD-1 VL2 MAT to hPD-1 7) (SEQ ID NO: 153); an intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)); VH domain of a monoclonal antibody capable of binding to LAG-3 (VH _LAG-3 VH1 MAT to hLAG-3 1) (SEQ ID NO: 49); a cysteine-containing intermediate linker peptide (linker 2: ggcggg (SEQ ID NO: 15)); a cysteine-containing domain that facilitates heterodimer formation (K-helix) (KVAACKE-KVAALKE-KVAALKE-KVAALKE (SEQ IDNO: 24); and the C-terminus.

The amino acid sequence of the second and fourth polypeptide chains of DART A is as follows (SEQ ID NO: 268):

EIVLTQSPAT LSLSPGERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL

LIHAASNQGS GVPSRFSGSG SGTDFTLTIS SLEPEDFAVY FCQQSKEVPY TFGGGTKVEI KGGGSGGGGQ VQLVQSGAEV KKPGASVKVS CKASGYTFTN YGMNWVRQAP GQGLEWMGWI NTYTGESTYA DDFEGRFVFS MDTSASTAYL QISSLKAEDT AVYYCARESL YDYYSMDYWG TVTVSSG GCGGGKVAAC KEKVAALKEK VAALKEKVAA LKE

2. DART B.

DART B is identical to DART A except that the first and third polypeptide chains of DART B comprise the VL domain of VL3 MAT to hLAG-3 1 (SEQ ID NO: 53), which contains an amino acid substitution in CDRL1. Accordingly, the first and third polypeptide chains of DART B comprise, from N-terminus to C-terminus: N-terminus; VL domain of a monoclonal antibody capable of binding to LAG-3 (VL _lag-3 VL3 MAT to hLAG-3 1) (SEQ ID NO: 53); an intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)); VH domain of a monoclonal antibody capable of binding to PD-1 (VH _PD-1 VH1 MAT to hPD-1 7) (SEQ ID NO: 147); an intermediate linker peptide (linker 2: ggcggg (SEQ ID NO: 15)); a cysteine-containing heterodimer formation promoting domain (Ehelix) (evaacek-evaalek-evaalek-evaalek (SEQ ID NO: 23)); a stabilized hinge region of IgG4 (SEQ ID NO: 13); a variant CH2-CH3 domain of IgG4 containing the substitutions M252Y/S254T/T256E and lacking the C-terminal residue (SEQ ID NO: 259); and a C-terminus. The amino acid sequence of the first and third polypeptide chains of DART B is the sequence set forth in SEQ ID NO: 267, wherein X1 is G; _X2 is Y; _X3 is T; and _X4 is E.

The amino acid sequence of the second and fourth polypeptide chains of DART B is the sequence shown in SEQ ID NO: 268.

3. DART C.

DART C is identical to DART B except that the first and third polypeptide chains of DART B comprise a wild-type IgG4 CH2-CH3 domain lacking the C-terminal residue (SEQ ID NO: 4). Accordingly, the first and third polypeptide chains of DART C comprise, from N-terminus to C-terminus: N-terminus, a VL domain of a monoclonal antibody capable of binding to LAG-3 (VL _lag-3 VL3 MAT to hLAG-3 1) (SEQ ID NO: 53); an intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)); a VH domain of a monoclonal antibody capable of binding to PD-1 (VH _PD-1 VH1 MAT to hPD-1 7) (SEQ ID NO: 147); an intermediate linker peptide (linker 2: ggcggg (SEQ ID NO: 15)); a cysteine-containing heterodimer-promoting domain (E-helix) (evaacek-evaalek-evaalek-evaalek (SEQ ID NO: 23)); a stabilized hinge region of IgG4 (SEQ ID NO: 13); the CH2-CH3 domain of IgG4 lacking the C-terminal residue (SEQ ID NO: 4); and a C-terminus.

The amino acid sequence of the first and third polypeptide chains of DART C is the sequence set forth in SEQ ID NO: 267, wherein X1 is G; _X2 is M; X3 is S; and _X4 is T.

The amino acid sequence of the second and fourth polypeptide chains of DART C is the sequence shown in SEQ ID NO: 268.

4. DART I.

DART I is a bispecific four-chain diabody comprising an Fc region containing two binding sites specific for PD-1, two binding sites specific for LAG-3, a variant of the IgG4 Fc region modified to increase half-life, and cysteine-containing domains that facilitate heterodimer formation with an E/K helix. The first and third polypeptide chains of DART I comprise, from N-terminus to C-terminus: N-terminus, the VL domain of a monoclonal antibody capable of binding to LAG-3 (VL _lag-3 VL1 MAb to hLAG-3 6) (SEQ ID NO: 296); an intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)); a VH domain

- 68 047726 monoclonal antibody capable of binding to PD-1 (VH _PD-1 VH1 MAb to hPD-1 7) (SEQ ID NO: 147); a cysteine-containing intermediate linker peptide (linker 2: ggcggg (SEQ ID NO: 15)) a cysteine-containing heterodimer-promoting domain (E-helix) (EVAACEK-EVAALEK-EVAALEK-EVAALEK (SEQ ID NO: 23)) a stabilized hinge region of IgG4 (SEQ ID NO: 13); a variant of the CH2-CH3 domain of IgG4 containing the M252Y/S254T/T256E substitutions and lacking the C-terminal residue (SEQ ID NO: 259); and a C-terminus.

The amino acid sequence of the first and third polypeptide chains of DART I is as follows (SEQ ID NO: 290):

DIQMTQSPSS LSASVGDRVT ITCRASQDVS SWAWYQQKP GKAPKLLIYS

ASYRYTGVPS RFSGSGSGTD FTLTISSLQP EDFATYYCQQ HYSTPWTFGG

GTKLEIKGGG SGGGGQVQLV QSGAEVKKPG ASVKVSCKAS GYSFTSYWMN

WVRQAPGQGL EWIGVIHPSD SETWLDQKFK DRVTTITVDKS TSTAYMELSS

LRSEDTAVYY CAREHYGTSP FAYWGQGTLV TVSSGGCGGG EVAACEKEVA

ALEKEVAALE KEVAALEKES KYGPPCPPCP APEFLGGPSV FLFPPKPKDT

LYITREPEVT CVWDVSQED PEVQFNWYVD GVEVHNAKTK PREEQFNSTY

RWSVLTVLH QDWLNGKEYK CKVSNKGLPS SIEKTISKAK GQPREPQVYT

LPPSQEEMTK NQVSLTCLVK GFYPSDIAVE WESNGQPENN YKTTPPVLDS

DGSFFLYSRL TVDKSRWQEG NVFSCSVMHE ALHNHYTQKS LSLSLG

The second and fourth polypeptide chains of DART I comprise, in the direction from N-terminus to C-terminus: N-terminus, the VL domain of a monoclonal antibody capable of binding to PD-1 (VL _PD-1 VL2 MAT to hPD-1 7) (SEQ ID NO: 153); an intermediate linker peptide (linker 1: GGGSGGGG (SEQ ID NO: 14)); the VH domain of a monoclonal antibody capable of binding to LAG-3 (VH _LAG-3 VH1 MAT to hLAG-3 6) (SEQ ID NO: 294); a cysteine-containing intermediate linker peptide (linker 2: GGCGGG (SEQ ID NO: 15)); a cysteine-containing domain that facilitates heterodimer formation (K-helix) (KVAACKE-KVAALKE-KVAALKE-KVAALKE (SEQ IDNO: 24); and the C-terminus.

The amino acid sequence of the second and fourth polypeptide chains of DART I is as follows (SEQ ID NO: 291):

EIVLTQSPAT LSLSPGERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL

LIHAASNQGS GVPSRFSGSG SGTDFTLTIS SLEPEDFAVY FCQQSKEVPY TFGGGTKVEI KGGGSGGGGQ VQLVQSGAEV KKPGASVKVS CKASGYTFTD YNMDWVRQAP GQGLEWMGDI NPDNGVTIYN QKFEGRVTMT TDTSTSTAYM ELRSLRSDDT AVYYCAREAD YFYFDYWGQG T TLTVSSGGC GGGKVAACKE KVAALKEKVA ALKEKVAALK E

C. Exemplary four-chain diabodies containing the Fc region and CL/CH1 domains.

Four exemplary bispecific four-chain Fc region-containing PD-1xLAG-3 diabodies containing CL/CH1 domains were obtained and designated DART D, DART E, DART J, and DART 1. The structure of these Fc region-containing diabodies is described in detail below. These exemplary PD-1xLAG-3 diabodies are illustrative and in no way limit the scope of the present invention.

1. DART D.

DART D is a bispecific four-chain diabody comprising an Fc region containing two binding sites specific for PD-1, two binding sites specific for LAG-3, CL/CH1 domains, and a variant IgG4 Fc region modified to increase half-life. The first and third polypeptide chains of DART D comprise, from N-terminus to C-terminus: N-terminus; the VL domain of a monoclonal antibody capable of binding to PD1 (VL _pd-1 VL2 MAT to hPD-1 7) (SEQ ID NO: 153); an intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)); the VH domain of a monoclonal antibody capable of binding to LAG-3 (VH _LAG-3 VH1 MAT to hLAG-3 1) (SEQ ID NO: 49); an intermediate linker peptide (linker 2: lgggsg (SEQ ID NO: 261)); the CH1 domain of IgG4 (SEQ ID NO: 254); a stabilized hinge region of IgG4 (SEQ ID NO: 13); a variant of the GH2-CH3 domain of IgG4 containing the substitutions M252Y/S254T/T256E and lacking the C-terminal residue (SEQ ID NO: 259); and a C-terminus.

The amino acid sequence of the first and third polypeptide chains of DART D is as follows (SEQ ID NO: 269):

- 69 047726

EIVLTQSPAT LSLSPGERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL

LIHAASNQGS GVPSRFSGSG SGTDFTLTIS SLEPEDFAVY FCQQSKEVPY

TFGGGTKVEI KGGGSGGGGQ VQLVQSGAEV KKPGASVKVS CKASGYTFTN

YGMNWVRQAP GQGLEWMGWI NTYTGESTYA DDFEGRFVFS MDTSASTAYL

QISSLKAEDT AVYYCARESL YDYYSMDYWG QGTTVTVSSL GGGSGASTKG

PSVFPLAPCS RSTSESTAAL GCLVKDYFPE PVTVSWNSGA LTSGVHTFPA

VLQSSGLYSL SSWTVPSSS LGTKTYTCNV DHKPSNTKVD KRVESKYGPP

CPPCPAPEFL GGPSVFLFPP KPKDTLYITR EPEVTCVWD VSQEDPEVQF

NWYVDGVEVH NAKTKPREEQ FNSTYRWSV LTVLHQDWLN GKEYKCKVSN

KGLPSSIEKT ISKAKGQPRE PQVYTLPPSQ EEMTKNQVSL TCLVKGFYPS

DIAVEWESNG QPENNYKTTP PVLDSDGSFF LYSRLTVDKS RWQEGNVFSC

SVMHEALHNH YTQKSLSLSL G

The second and fourth polypeptide chains of DART D comprise, in direction from N-terminus to C-terminus: N-terminus; VL domain of a monoclonal antibody capable of binding to LAG-3 (VL _lag- 3 VL4 MAT to hLAG-3 1) (SEQ ID NO: 54); intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)); VH domain of a monoclonal antibody capable of binding to PD-1 (VH _pd-1 VH1 MAT to hPD-1 7) (SEQ ID NO: 147); intermediate linker peptide (linker 2: lgggsg (SEQ ID NO: 261)); CL domain of kappa chain (SEQ ID NO: 8); and C-terminus.

The amino acid sequence of the second and fourth polypeptide chains of DART D is as follows (SEQ ID NO: 270):

DIVMTQTPLS LSVTPGQPAS ISCKSSQSLL HSDAKTYLNW LLQKPGQPPE RLIYLVSELD SGVPDRFSGS GSGTDFTLKI SRVEAEDVGV YYCWQGTHFP YTFGGGTKVE IKGGGSGGGG QVQLVQSGAE VKKPGASVKV SCKASGYSFT SYWMNWVRQA PGQGLEWIGV SETWL DQKFKDRVTI TVDKSTSTAY MELSSLRSED TAVYYCAREH YGTSPFAYWG QGTLVTVSSL GGGSGRTVAA PSVFIFPPSD EQLKSGTASV VCLLNNFYPR EAKVQWKVDN ALQSGNSQES VTEQDSKDST YSLSSTLTLS KADYEKHKVY ACEVTHQGLS SPVTKSFNRG EC 2. DART E.

DART E is a different bispecific four-chain diabody comprising an Fc region containing two PD-1-specific binding sites, two LAG-3-specific binding sites, CL/CH1 domains, and an IgG4 Fc variant modified to increase half-life. The PD-1 and LAG-3 binding sites of DART E are in the reverse order compared to DART D.

The first and third polypeptide chains of DART E comprise, in the direction from N-terminus to C-terminus: N-terminus, VL domain of a monoclonal antibody capable of binding to LAG-3 (VL _lag-3 VL4 MAT to hLAG-3 1) (SEQ ID NO: 54); an intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)); VH domain of a monoclonal antibody capable of binding to PD-1 (VH _pd-1 VH1 MAT to hPD-1 7) (SEQ ID NO: 147); an intermediate linker peptide (linker 2: lgggsg (SEQ ID NO: 261)); CH1 domain of IgG4 (SEQ ID NO: 254); stabilized hinge region of IgG4 (SEQ ID NO: 13); a variant of the IgG4 CH2CH3 domain containing the M252Y/S254T/T256E substitutions and lacking the C-terminal residue (SEQ ID NO: 259); and the C-terminus.

The amino acid sequence of the first and third polypeptide chains of DART E is as follows (SEQ ID NO: 271):

DIVMTQTPLS LSVTPGQPAS ISCKSSQSLL HSDAKTYLNW LLQKPGQPPE RLIYLVSELD SGVPDRFSGS GSGTDFTLKI SRVEAEDVGV YYCWQGTHFP YTFGGGTKVE IKGGGSGGGG QVQLVQSGAE VKKPGASVKV SCKASGYSFT SYWMNWVRQA PGQGLEWIGV SETWL DQKFKDRVTI TVDKSTSTAY MELSSLRSED TAVYYCAREH YGTSPFAYWG QGTLVTVSSL GGGSGASTKG PSVFPLAPCS RSTSESTAAL GCLVKDYFPE PVTVSWNSGA LTSGVHTFPA VLQSSGLYSL SSWTVPSSS LGTKTYTCNV DHKPSNTKVD KRVESKYGPP CPPCPAPEFL GGPSVFLFPP KPKDTLYITR EPEVTCVWD VSQEDPEVQF NWYVDGVEVH NAKTKPREEQ FNSTYRWSV LTVLHQDWLN GKEYKCKVSN KGLPSSIEKT ISKAKGQPRE PQVYTLPPSQ EEMTKNQVSL TCLVKGFYPS DIAVEWESNG QPENNYKTTP LDSDGSFF LYSRLTVDKS RWQEGNVFSC SVMHEALHNH YTQKSLSLSL G

- 70 047726

The second and fourth polypeptide chains of DART E comprise, in direction from N-terminus to C-terminus: N-terminus; VL domain of a monoclonal antibody capable of binding to PD-1 (VL _PD-1 VL2 MAT to hPD-1 7) (SEQ ID NO: 153); an intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)); VH domain of a monoclonal antibody capable of binding to LAG-3 (VH _LAG-3 VH1 MAT to hLAG-3 1) (SEQ ID NO: 49); intermediate linker peptide (linker 2: lgggsg (SEQ ID NO: 261)); CL domain of kappa chain (SEQ ID NO: 8); and a C-terminus.

The amino acid sequence of the second and fourth polypeptide chains of DART E is as follows (SEQ ID NO: 272): EIVLTQSPAT LSLSPGERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL LIHAASNQGS GVPSRFSGSG SGTDFTLTIS SLEPEDFAVY FCQQSKEVPY TFGGGTKVEI KGGGSGGGGQ VQLVQSGAEV KKPGASVKVS CKASGYTFTN YGMNWVRQAP GQGLEWMGWI NTYTGESTYA DDFEGRFVFS MDTSASTAYL QISSLKAEDT AVYYCARESL YDYYSMDYWG QGTTVTVSSL GGGSGRTVAA PSVFIFPPSD EQLKSGTASV VCLLNNFYPR EAKVQWKVDN ALQSGNSQES VTEQDSKDST YSLSSTLTLS KADYEKHKVY ACEVTHQGLS SPVTKSFNRG EC 3. DART J.

DART J is a bispecific four-chain diabody comprising an Fc region containing two binding sites specific for PD-1, two binding sites specific for LAG-3, CL/CH1 domains, and an IgG4 Fc variant modified to increase half-life. The first and third polypeptide chains of DART J comprise, from N-terminus to C-terminus: N-terminus; the VL domain of a monoclonal antibody capable of binding to LAG-3 (VL _lag-3 VL1 MAT to hLAG-3 6) (SEQ ID NO: 296); an intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)); the VH domain of a monoclonal antibody capable of binding to PD-1 (VH _PD-1 VH1 MAT to hPD-1 7) (SEQ ID NO: 147); an intermediate linker peptide (linker 2: lgggsg (SEQ ID NO: 261)); the CH1 domain of IgG4 (SEQ ID NO: 254); a stabilized hinge region of IgG4 (SEQ ID NO: 13); a variant of the CH2-CH3 domain of IgG4 containing the substitutions M252Y/S254T/T256E and lacking the C-terminal residue (SEQ ID NO: 259); and a C-terminus.

The amino acid sequence of the first and third polypeptide chains of DART J is as follows (SEQ ID NO: 292):

DIQMTQSPSS LSASVGDRVT ITCRASQDVS SWAWYQQKP GKAPKLLIYS ASYRYTGVPS RFSGSGSGTD FTLTISSLQP EDFATYYCQQ HYSTPWTFGG GTKLEIKGGG SGGGGQVQLV QSGAEVKKPG ASVKVSCKAS GYSFTSYWMN WVRQAPGQGL EWIGVIHPSD SETWLDQKF K DRVTITVDKS TSTAYMELSSS LRSEDTAVYY CAREHYGTSP FAYWGQGTLV TVSSLGGGSG ASTKGPSVFP LAPCSRSTSE STAALGCLVK DYFPEPVTVS WNSGALTSGV HTFPAVLQSS GLYSLSSWT VPSSSLGTKT YTCNVDHKPS NTKVDKRVES KYGPPCPPCP APEFLGGPSV FLFPPKPKDT LYITREPEVT CVWDVSQED PEVQFNWYVD GVEVHNAKTK PREEQFNSTY RWSVLTVLH QDWLNGKEYK CKVSNKGLPS SIEKTISKAK GQPREPQVYT LPPSQEEMTK NQVSLTCLVK GFYPSDIAVE WESNGQPENN YKTTPPVLDS SRL TVDKSRWQEG NVFSCSVMHE ALHNHYTQKS LSLSLG

The second and fourth polypeptide chains of DART J comprise, in direction from N-terminus to C-terminus: N-terminus; VL domain of a monoclonal antibody capable of binding to PD-1 (VL _PD-1 VL2 MAb to hPD-1 7) (SEQ ID NO: 153); an intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)); VH domain of a monoclonal antibody capable of binding to LAG-3 (VH _LAG-3 VH1 MAb to hLAG-3 6) (SEQ ID NO: 294); intermediate linker peptide (linker 2: lgggsg (SEQ ID NO: 261)); CL domain of kappa chain (SEQ ID NO: 8); and a C-terminus.

The amino acid sequence of the second and fourth polypeptide chains of DART J is as follows (SEQ ID NO: 293):

EIVLTQSPAT LSLSPGERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL

LIHAASNQGS GVPSRFSGSG SGTDFTLTIS SLEPEDFAVY FCQQSKEVPY

TFGGGTKVEI KGGGSGGGGQ VQLVQSGAEV KKPGASVKVS CKASGYTFTD

YNMDWVRQAP GQGLEWMGDI NPDNGVTIYN QKFEGRVTMT TDTSTSTAYM

ELRSLRSDDT AVYYCAREAD YFYFDYWGQG TTLTVSSLGG GSGRTVAAPS

VFIFPPSDEQ LKSGTASWC LLNNFYPREA KVQWKVDNAL QSGNSQESVT

EQDSKDSTYS LSSTLTLSKA DYEKHKVYAC EVTHQGLSSP VTKSFNRGEC

- 71 047726

4. DART 1.

DART 1 is a bispecific four-chain diabody comprising an Fc region having two binding sites specific for PD-1, two binding sites specific for LAG-3, CL/CH1 domains, and a variant IgG1 Fc region modified to reduce binding to FcyR. The first and third polypeptide chains of DART 1 comprise, from N-terminus to C-terminus: N-terminus; the VL domain of a monoclonal antibody capable of binding to PD-1 (VL _PD-1 VL MAT to PD-1 A) (SEQ ID NO: 65); an intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)) the VH domain of a monoclonal antibody capable of binding to LAG-3 (VH _LAG-3 VH1 MAT to LAG-3 A) (SEQ ID NO: 39); intermediate linker peptide (linker 2: lgggsg (SEQ ID NO: 261)) IgG1 CH1 domain (SEQ ID NO: 10); IgG1 hinge region (SEQ ID NO: 32); a variant IgG1 CH2-CH3 domain containing L234A/L235A substitutions and lacking the C-terminal residue (SeQ ID NO: 5); and the C-terminus.

The amino acid sequence of the first and third polypeptide chains of DART 1 is as follows (SEQ ID NO: 284):

EIVLTQSPAT LSLSPGERAT LSCRASQSIS SYLAWYQQKP GQAPRLLIYD ASNRATGIPA RFSGSGSGTD FTLTISSLEP EDFAVYYCQQ RSNWPLTFGQ GTNLEIKGGG SGGGGQVQLV ESGGGWQPG RSLRLDCKAS GITFSNSGMH WVRQAPGKGL EWVAVIWYDG SKRYYADSVK ISRDNS KNTLFLQMNS LRAEDTAVYY CATNDDYWGQ GTLVTVSSLG GGSGASTKGP SVFPLAPSSK STSGGTAALG CLVKDYFPEP VTVSWNSGAL TSGVHTFPAV LQSSGLYSLS SWTVPSSSL GTQTYICNVN HKPSNTKVDK RVEPKSCDKT HTCPPCPAPE AAGGPSVFLF PPKPKDTLYI TREPEVTCW VDVSHEDPEV KFNWYVDGVE VHNAKTKPRE EQYNSTYRW SVLTVLHQDW LNGKEYKCKV SNKALPAPIE KTISKAKGQP REPQVYTLPP SREEMTKNQV SLTCLVKGFY PSDIAVEWES NGQPENNYKT TPPVLDSDGS FFLYSKLTVD KSRWQQGNVF SCS VMHEALH NHYTQKSLSL SPG

The second and fourth polypeptide chains of DART 1 comprise, in direction from N-terminus to C-terminus: N-terminus; VL domain of a monoclonal antibody capable of binding to LAG-3 (VL _lag-3 VL MAT to LAG3 A) (SEQ ID NO: 40); an intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)); VH domain of a monoclonal antibody capable of binding to PD-1 (VH _PD-1 VH MAT to PD-1 A) (SEQ ID NO: 64); intermediate linker peptide (linker 2: lgggsg (SEQ ID NO: 261)); CL domain of kappa chain (SEQ ID NO: 8); and a C-terminus.

The amino acid sequence of the second and fourth polypeptide chains of DART 1 is as follows (SEQ ID NO: 285):

EIVLTQSPAT LSLSPGERAT LSCRASQQKP GQAPRLLIYD ASNRATGIPA RFSGSGSGTD FTLTISSLEP EDFAVYYCQQ SSNWPRTFGQ GTKVEIKGGG SGGGGQVQLQ QWGAGLLKPS ETLSLTCAVY GGSFSDYYWN WIRQPPGKGL EWIGEINHNG NTNSNPSLKS R VTLSLDTSK NQFSLKLRSV TAADTAVYYC AFGYSDYEYN WFDPWGQGTL VTVSSLGGGS GRTVAAPSVF IFPPSDEQLK SGTASWCLL NNFYPREAKV QWKVDNALQS GNSQESVTEQ DSKDSTYSLS STLTLSKADY EKHKVYACEV THQGLSSPVT KSFNRGEC D. Approximate five-chain diabodies containing the Fc region.

Two exemplary bispecific five-chain diabodies containing the Fc region, PD-1xLAG-3, were obtained, containing CL/CH1 domains and domains promoting heterodimer formation with an E/K helix, and are designated DART F and DART G. The structure of these diabodies containing the Fc region is described in detail below. These exemplary diabodies PD-1xLAG-3 are illustrative and in no way limit the scope of the present invention.

1. DART F.

DART F is a bispecific five-chain diabody comprising an Fc region having three binding sites specific for PD-1, one binding site specific for LAG-3, CL/CH1 domains, a knob/hole variant of the IgG1 Fc region modified to reduce binding to FcyR and increase half-life, and E/K helix heterodimer formation promoting domains. The first polypeptide chain of DART F comprises, from N-terminus to C-terminus: N-terminus; the VH domain of a monoclonal antibody capable of binding to PD-1 (VH _PD-1 VH1 MAb to hPD-1 7) (SEQ ID NO: 147); the CH1 domain of IgG1 (SEQ ID NO: 10); the hinge region of IgG1 (SEQ ID NO: 32); IgG1 CH2-CH3 cavity domain containing the L234A/L235A/M252Y/S254T/T256E/N434A/H435K substitutions and lacking the C-terminal residue (SEQ ID NO: 260, in which X1 is A, X2 is A; _X3 is Y, X4 is

- 72 047726 is T, X5 is E, X6 is S, X7 is A, X8 is V, _X9 is A, _X10 is K and X11 is absent); and the C-terminus.

The amino acid sequence of the first polypeptide chain of DART F is as follows (SEQ ID NO: 273):

QVQLVQSGAE VKKPGASVKV SCKASGYSFT SYWMNWVRQA PGQGLEWIGV

IHPSDSETWL DQKFKDRVTI TVDKSTSTAY MELSSLRSED TAVYYCAREH

YGTSPFAYWG QGTLVTVSSA STKGPSVFPL APSSKSTSGG TAALGCLVKD YFPEPVTVSW NSGALTSGVH TFPAVLQSSG LYSLSSWTV PSSSLGTQTY ICNVNHKPSN TKVDKRVEPK SCDKTHTCPP CPAPEAAGGP SVFLFPPKPK DTLYITREPE VTCVWDVSH EDPEVKFNWY V DGVEVHNAK TKPREEQYNS TYRWSVLTV LHQDWLNGKE YKCKVSNKAL PAPIEKTISK AKGQPREPQV YTLPPSREEM TKNQVSLSCA VKGFYPSDIA VEWESNGQPE NNYKTTPPVL DSDGSFFLVS KLTVDKSRWQ QGNVFSCSVM HEALHAKYTQ KSLSLSPG

The second and fifth polypeptide chains of DART F contain, in the direction from N-terminus to C-terminus: N-terminus; VL domain of a monoclonal antibody capable of binding to PD-1 (VL _PD-1 VL2 MAb to hPD-1 7) (SEQ ID NO: 153), CL domain of the kappa chain (SEQ ID NO: 8) and C-terminus.

The amino acid sequence of the second and fifth polypeptide chains of DART F is the sequence shown below (SEQ ID NO: 274):

EIVLTQSPAT LSLSPGERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL

LIHAASNQGS GVPSRFSGSG SGTDFTLTIS SLEPEDFAVY FCQQSKEVPY TFGGGTKVEI KRTVAAPSVF IFPPSDEQLK SGTASWCLL NNFYPREAKV QWKVDNALQS GNSQESVTEQ DSKDSTYSLS STLTLSKADY EKHKVYACEV THQGLSSPVT KSFNRGEC

The third polypeptide chain of DART F comprises, in the direction from the N-terminus to the C-terminus: the N-terminus; the VH domain of a monoclonal antibody capable of binding to PD-1 (VH _PD-1 VH1 MAb to hPD-1 7) (SEQ ID NO: 147); the CH1 domain of IgG1 (SEQ ID NO: 10); the hinge region of IgG1 (SEQ ID NO: 32); a CH2-CH3 knob-bearing domain of IgG1 containing the substitutions L234A/L235A/M252Y/S254T/T256E and lacking the C-terminal residue (SEQ ID NO: 260, in which X1 is A, X2 is A, _X3 is Y, X4 is T, X5 is E, X6 is W, X7 is L, _X8 is Y, _X9 is H, _X10 is H and X11 is absent); an intermediate linker peptide (gggsgggsggg (SEQ ID NO: 262)); the VL domain of a monoclonal antibody capable of binding to LAG-3 (VL _lag-3 VL4 MAb to hLAG-3 1) (SEQ ID NO: 54); an intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)); a VH domain of a monoclonal antibody capable of binding to PD-1 (VH _PD-1 VH1 MAb to hPD-1 7) (SEQ ID NO: 147); a cysteine-containing intermediate linker peptide (linker 2: ggcggg (SEQ ID NO: 15)); a domain facilitating heterodimer formation (E-helix) (evaalek-evaalek-evaalek-evaalek (SEQ ID NO: 21)); and a C-terminus.

The amino acid sequence of the third polypeptide chain of DART F is as follows (SEQ ID NO: 275):

QVQLVQSGAE VKKPGASVKV SCKASGYSFT SYWMNWVRQA PGQGLEWIGV

IHPSDSETWL DQKFKDRVTI TVDKSTSTAY MELSSLRSED TAVYYCAREH YGTSPFAYWG QGTLVTVSSA STKGPSVFPL APSSKSTSGG TAALGCLVKD YFPEPVTVSW NSGALTSGVH TFPAVLQSSG LYSLSSWTV PSSSLGTQTY ICNVNHKPSN TKVDKRVEPK SCDKTHTCPP EAAGGP SVFLFPPKPK DTLYITREPE VTCVWDVSH EDPEVKFNWY VDGVEVHNAK TKPREEQYNS TYRWSVLTV LHQDWLNGKE YKCKVSNKAL PAPIEKTISK AKGQPREPQV YTLPPSREEM TKNQVSLWCL VKGFYPSDIA VEWESNGQPE NNYKTTPPVL DSDGSFFLYS KLTVDKSRWQ QGNVFSCSVM HEALHNHYTQ KSLSLSPGGG GSGGGSGGGD IVMTQTPLSL SVTPGQPASI SCKSSQSLLH SDAKTYLNWL LQKPGQPPER LIYLVSELDS GVPDRFSSGG SGTDFTLKIS RVEAEDVGVY YCWQGTHFPY TFGGGTKVEI KGGGSGGGGQ VQLVQSGAEV KKPGASVKVS CKASGYSFTS YWMNWVRQAP GQGLEWIGVI HPSDSETWLD QKFKDRVTIT VDKSTSTAYM ELSSLRSEDT AVYYCAREHY GTSPFAYWGQ GTLVTVSSGG CGGGEVAALE KEVAALEKEV AALEKEVAAL EK

The fourth polypeptide chain of DART F contains, in the direction from the N-terminus to the C-terminus: N-terminus; VL domain of a monoclonal antibody capable of binding to PD-1 (VLPD-1 VL2 MAb to hPD-1 7) (SEQ ID NO: 1)

- 73 047726

ID NO: 153); an intermediate linker peptide (linker 1: GGGSGGGG (SEQ ID NO: 14)); the VH domain of a monoclonal antibody capable of binding to LAG-3 (VH _LAG-3 VH1 MAT to hLAG-3 1) (SEQ ID NO: 49); a cysteine-containing intermediate linker peptide (linker 2: ggcggg (SEQ ID NO: 15)); a domain facilitating the formation of a heterodimer (K-helix) (KVAALKE-KVAALKE-KVAALKE-KVAALKE (SEQ ID NO: 22))); and the C-terminus.

The amino acid sequence of the fourth polypeptide chains of DART F is as follows (SEQ ID NO: 276):

EIVLTQSPAT LSLSPGERAT LSCRASESVD NYGMSEMNWF QQKPGQPPKL

LIHAASNQGS GVPSRFSGSG SGTDFTLTIS SLEPEDFAVY FCQQSKEVPY TFGGGTKVEI KGGGSGGGGQ VQLVQSGAEV KKPGASVKVS CKASGYTFTN YGMNWVRQAP GQGLEWMGWI NTYTGESTYA DDFEGRFVFS MDTSASTAYL QISSLKAEDT AVYYCARESL YDYYSMDYWG TVTVSSG GCGGGKVAAL KEKVAALKEK VAALKEKVAA LKE

2. DART G.

DART G is a bispecific five-chain diabody comprising an Fc region having two binding sites specific for PD-1, two binding sites specific for LAG-3, CL/CH1 domains, a knob/hole variant of the IgG1 Fc region modified to decrease binding to FcyR and increase half-life, and E/K helix heterodimer formation promoting domains. The first polypeptide chain of DART G comprises, from N-terminus to C-terminus: N-terminus; the VH domain of a monoclonal antibody capable of binding to LAG-3 (VH _LAG-3 VH1 MAb to hLAG-3 1) (SEQ ID NO: 49); the CH1 domain of IgG1 (SEQ ID NO: 10); the hinge region of IgG1 (SEQ ID NO: 32); a CH2-CH3 hole-bearing domain of IgG1 containing the substitutions L234A/L235A/M252Y/S254T/T256E/N434A/H435K and lacking the C-terminal residue (SEQ ID NO: 260, in which X1 is A, X2 is A; _X3 is Y, X4 is T, X5 is E, _X6 is S, _X7 is A, X8 is V, _X9 is A, _X10 is K and X11 is absent); and a C-terminus.

The amino acid sequence of the first polypeptide chain of DART G is as follows (SEQ ID NO: 277):

QVQLVQSGAE VKKPGASVKV SCKASGYTFT NYGMNWVRQA PGQGLEWMGW

INTYTGESTY ADDFEGRFVF SMDTSASTAY LQISSLKAED TAVYYCARES LYDYYSMDYW GQGTTVTVSS ASTKGPSVFP LAPSSKSTSG GTAALGCLVK DYFPEPVTVS WNSGALTSGV HTFPAVLQSS GLYSLSSWT VPSSSLGTQT YICNVNHKPS NTKVDKRVEP KSCDKTHTCP PCPAPEAAGG PSVFLFPPKP KDTLYITREP EVTCVWDVS HEDPEVKFNW YVDGVEVHNA KTKPREEQYN STYRWSVLT VLHQDWLNGK EYKCKVSNKA LPAPIEKTIS KAKGQPREPQ VYTLPPSREE MTKNQVSLSC AVKGFYPSDI AVEWESNGQP ENNYKTTPPV LDSDGSFFLV SKLTV DKSRW QQGNVFSCSV MHEALHAKYT QKSLSLSPG

The second and fifth polypeptide chains of DART G contain, in the direction from the N-terminus to the C-terminus: the N-terminus; the VL domain of a monoclonal antibody capable of binding to LAG-3 (VL _lag-3 VL4 MAT to hLAG-3 1) (SEQ ID NO: 54), the CL domain of the kappa chain (SEQ ID NO: 8) and the C-terminus.

The amino acid sequence of the second and fifth polypeptide chains of DART G is the sequence shown below (SEQ ID NO: 278):

DIVMTQTPLS LSVTPGQPAS ISCKSSQSLL HSDAKTYLNW LLQKPGQPPE

RLIYLVSELD SGVPDRFSGS GSGTDFTLKI SRVEAEDVGV YYCWQGTHFP YTFGGGTKVE IKRTVAAPSV FIFPPSDEQL KSGTASWCL LNNFYPREAK VQWKVDNALQ SGNSQESVTE QDSKDSTYSL SSTLTLSKAD YEKHKVYACE VTHQGLSSPV TKSFNRGEC

The third polypeptide chain of DART G comprises, in the direction from the N-terminus to the C-terminus: the N-terminus; the VH domain of a monoclonal antibody capable of binding to LAG-3 (VH _LAG-3 VH1 MAT to hLAG-3 1) (SEQ ID NO: 49); the IgG1 CH1 domain (SEQ ID NO: 10); the hinge region of IgG1 (SEQ ID NO: 32); a CH2-CH3 knob-bearing domain of IgG1 containing the L234A/L235A/M252Y/S254T/T256E substitutions and lacking the C-terminal residue (SEQ ID NO: 260, in which X1 is A, X2 is A, _X3 is Y, X4 is T, X5 is E, X6 is W, X7 is L, X8 is Y, _X9 is H, _X10 is H and X11 is absent); an intermediate linker peptide (gggsgggsggg (SEQ ID NO: 262)); the VL domain of a monoclonal antibody capable of binding to PD-1 (VL _PD-1 VL2 MAb to hPD-1 7) (SEQ ID NO: 153); an intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)); VH domain monoclonal anti

- 74 047726 a body capable of binding to PD-1 (VH _PD-1 VH1 MAT to hPD-1 7) (SEQ ID NO: 147); a cysteine-containing intermediate linker peptide (linker 2: ggcggg (SEQ ID NO: 15)); a domain facilitating the formation of a heterodimer (E-helix) (evaalek-evaalek-evaalek-evaalek (SEQ ID NO: 21)); and a C-terminus.

The amino acid sequence of the third polypeptide chain of DART G is as follows (SEQ ID NO: 279):

QVQLVQSGAE VKKPGASVKV SCKASGYTFT NYGMNWVRQA PGQGLEWMGW INTYTGESTY ADDFEGRFVF SMDTSASTAY LQISSLKAED TAVYYCARES LYDYYSMDYW GQGTTVTVSS ASTKGPSVFP LAPSSKSTSG GTAALGCLVK DYFPEPVTVS WNSGALTSGV HTFPAVLQSS GLYSLSSWT VPSSSLGTQT YICNVNHKPS NTKVDKRVEP KSCDKTHTCP PCPAPEAAGG PSVFLFPPKP KDTLYITREP EVTCVWDVS HEDPEVKFNW YVDGVEVHNA KTKPREEQYN STYRWSVLT VLHQDWLNGK EYKCKVSNKA LPAPIEKTIS KAKGQPREPQ VYTLPPSREE SLEPE DFAVY FCQQSKEVPY TFGGGTKVEI KGGGSGGGGQ VQLVQSGAEV KKPGASVKVS CKASGYSFTS YWMNWVRQAP GQGLEWIGVI HPSDSETWLD QKFKDRVTIT VDKSTSTAYM ELSSLRSEDT AVYYCAREHY GTSPFAYWGQ GTLVTVSSGG CGGGEVAALE KEVAALEKEV AALEKEVAAL EK

The fourth polypeptide chain of DART G comprises, in direction from N-terminus to C-terminus: N-terminus; VL domain of a monoclonal antibody capable of binding to PD-1 (VL _PD-1 VL2 MAT to hPD-1 7) (SEQ ID NO: 153); intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)); VH domain of a monoclonal antibody capable of binding to PD-1 (VH _PD-1 VH1 MAT to hPD-1 7) (SEQ ID NO: 147); cysteine-containing intermediate linker peptide (linker 2: ggcggg (SEQ ID NO: 15)); domain facilitating heterodimer formation (K-helix) (KVAALKE-KVAALKE-KVAALKE-KVAALKE (SEQ ID NO: 22)); and C-end.

The amino acid sequence of the fourth polypeptide chains of DART G is as follows (SEQ ID NO: 280):

EIVLTQSPAT LSLSPGERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL LIHAASNQGS GVPSRFSGSG SGTDFTLTIS SLEPEDFAVY FCQQSKEVPY TFGGGTKVEI KGGGSGGGGQ VQLVQSGAEV KKPGASVKVS CKASGYSFTS YWMNWVRQAP GQGLEWIGVI HPSDSETWLD QKFKDR VTIT VDKSTSTAYM ELSSLRSEDT AVYYCAREHY GTSPFAYWGQ GTLVTVSSGG CGGGKVAALK EKVAALKEKV AALKEKVAAL KE

E. Exemplary three-chain diabody containing an Fc region with an E/K helix.

The present invention further provides bispecific three-chain diabodies comprising an Fc region, PD-1xLAG-3, containing domains that facilitate heterodimer formation with an E/K helix. An exemplary bispecific three-chain diabody comprising an Fc region, PD-1xLAG-3, containing domains that facilitate heterodimer formation with an E/K helix, designated as DART H, was obtained. The structure of said diabody comprising an Fc region is described in detail below. Said exemplary diabody PD-1xLAG-3 is illustrative and in no way limits the scope of the present invention. DART H is a bispecific three-chain diabody comprising an Fc region containing one binding site specific for PD-1, one binding site specific for LAG-3, a variant of the IgG1 Fc knob/hole region modified to reduce binding to FcyR, and heterodimer-promoting domains with an E/K helix.

The first polypeptide chain of DART H comprises, in direction from N-terminus to C-terminus: N-terminus; VL domain of monoclonal antibody capable of binding to PD-1 (VL _PD-1 VL2 MAT to hPD-1 7) (SEQ ID NO: 153); intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)); VH domain of monoclonal antibody capable of binding to LAG-3 (VH _LAG-3 VH1 MAT to hLAG-3 1) (SEQ ID NO: 49); cysteine-containing intermediate linker peptide (linker 2: ggcggg (SEQ ID NO: 15)); domain facilitating heterodimer formation (E-helix) (evaalek-evaalek-evaalek-evaalek (SEQ ID NO: 21)); an intermediate linker (spacer-linker 3: gggdkthtcppcp (SEQ ID NO: 263)); a CH2-CH3 knob-bearing domain of IgG1 containing the L234A/L235A substitutions and containing a C-terminal lysine residue (SEQ ID NO: 6); and a C-terminus.

The amino acid sequence of the first polypeptide chain of DART H is

- 75 047726 the sequence given below (SEQ ID NO: 281):

EIVLTQSPAT LSLSPGERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL LIHAASNQGS GVPSRFSGSG SGTDFTLTIS SLEPEDFAVY FCQQSKEVPY ITFGGGTKVEI KGGGSGGGGQ VQLVQSGAEV KKPGASVKVS CKASGYTFTN YGMNWVRQAP GQGLEWMGWI NTYTGESTYA VFS MDTSASTAYL QISSLKAEDT AVYYCARESL YDYYSMDYWG QGTTVTVSSG GCGGGEVAAL EKEVAALEKE VAALEKEVAA LEKGGGDKTH TCPPCPAPEA AGGPSVFLFP PKPKDTLMIS RTPEVTCVW DVSHEDPEVK FNWYVDGVEV HNAKTKPREE QYNSTYRWS VLTVLHQDWL NGKEYKCKVS NKALPAPIEK TISKAKGQPR EPQVYTLPPS REEMTKNQVS LWCLVKGFYP SDIAVEWESN GQPENNYKTT PPVLDSDGSF FLYSKLTVDK SRWQQGNVFS CSVMHEALHN HYTQKSLSLS PGK

The second polypeptide chain of DART H comprises, in the direction from N-terminus to C-terminus: N-terminus; VL domain of a monoclonal antibody capable of binding to LAG-3 (VL _L a _G- 3 VL4 MAT to hLAG-3 1) (SEQ ID NO: 54); intermediate linker peptide (linker 1: gggsgggg (SEQ ID NO: 14)); VH domain of a monoclonal antibody capable of binding to PD-1 (VH _Pd-1 VH1 MAT to hPD-1 7) (SEQ ID NO: 147); cysteine-containing intermediate linker peptide (linker 2: ggcggg (SEQ ID NO: 15)); domain facilitating heterodimer formation (K-helix) (KVAALKE-kvaalke-KVAALKE-KVAALKE (SEQ ID NO: 22)); and C-end.

The amino acid sequence of the second polypeptide chain of DART H is the sequence shown below (SEQ ID NO: 282):

DIVMTQTPLS LSVTPGQPAS ISCKSSQSLL HSDAKTYLNW LLQKPGQPPE RLIYLVSELD SGVPDRFSGS GSGTDFTLKI SRVEAEDVGV YYCWQGTHFP YTFGGGTKVE IKGGGSGGGG QVQLVQSGAE VKKPGASVKV SCKASGYSFT SYWMNWVRQA PGQGLEWIGV SETWL DQKFKDRVTI TVDKSTSTAY MELSSLRSED TAVYYCAREH YGTSPFAYWG QGTLVTVSSG GCGGGKVAAL KEKVAALKEK VAALKEKVAA LKE

The third polypeptide chain of DART H comprises, in the direction from the N-terminus to the C-terminus: the N-terminus; the hinge region (dkthtcppcp (SEQ ID NO: 31), the CH2-CH3 knob-bearing domain of IgG1 containing the L234A/L235A substitutions and containing the C-terminal lysine residue (SEQ ID NO: 7) and the C-terminus.

The amino acid sequence of the third polypeptide chain of DART H is as follows (SEQ ID NO: 283):

DKTHTCPPCP APEAAGGPSV FLFPPKPKDT LMISRTPEVT CVWDVSHED

PEVKFNWYVD GVEVHNAKTK PREEQYNSTY RWSVLTVLH QDWLNGKEYK

CKVSNKALPA PIEKTISKAK GQPREPQVYT LPPSREEMTK NQVSLSCAVK GFYPSDIAVE WESNGQPENN YKTTPPVLDS DGSFFLVSKL TVDKSRWQQG NVFSCSVMHE ALHNRYTQKS LSLSPGK

F. Exemplary bispecific antibody.

An exemplary four-chain bispecific antibody PD-1xLAG-3, designated as BSAB A, was obtained. The structure of said bispecific antibody is described in detail below. Said exemplary bispecific antibody PD-1xLAG-3 is illustrative and does not limit the scope of the present invention in any way.

BSAB A is a bispecific antibody comprising one binding site specific for PD-1, one binding site specific for LAG-3, an IgG1 Fc variant modified to reduce binding to FcyR and to promote complex formation between two different heavy chain polypeptides (see, for example, WO 2011/143545).

The first polypeptide chain of BSAB A comprises, in direction from N-terminus to C-terminus: an N-terminus; a VH domain of a monoclonal antibody capable of binding to PD-1 (VH _Pd-1 VH1 MAb to hPD-1 7) (SEQ ID NO: 147); a CH1 domain of IgG1 (SEQ ID NO: 10); a variant hinge region of IgG1 containing the substitutions D221E/P228E (numbered according to the EU system described in Kabat et al. and highlighted by underlining in SEQ ID NO: 286 below); a variant CH2-CH3 domain of IgG1 containing the substitutions L234A/L235A/L368E (substitutions highlighted by underlining in SEQ ID NO: 286 below) and lacking the C-terminal residue; and a C-terminus.

The amino acid sequence of the first polypeptide chain of BSAB A is the sequence shown below (SEQ ID NO: 286):

- 76 047726

QVQLVQSGAE VKKPGASVKV SCKASGYSFT SYWMNWVRQA PGQGLEWIGV IHPSDSETWL DQKFKDRVTI TVDKSTSTAY MELSSLRSED TAVYYCAREH YGTSPFAYWG QGTLVTVSSA STKGPSVFPL APSSKSTSGG TAALGCLVKD YFPEPVTVSW NSGALTSGVH TFPAVLQSSG LY SLSSWTV PSSSLGTQTY ICNVNHKPSN TKVDKRVEPK SCEKTHTCPE CPAPEAAGGP SVFLFPPKPK DTLMISRTPE VTCVWDVSH EDPEVKFNWY VDGVEVHNAK TKPREEQYNS TYRWSVLTV LHQDWLNGKE YKCKVSNKAL PAPIEKTISK AKGQPREPQV YTLPPSREEM TKNQVSLTCE VKGFYPSDIA VEWESNGQPE NNYKTTPPVL DSDGSFFLYS KLTVDKSRWQ QGNVFSCSVM HEALHNHYTQ KSLSLSPG

The second polypeptide chain of BSAB comprises, in the direction from the N-terminus to the C-terminus: the N-terminus; the VL domain of a monoclonal antibody capable of binding to PD-1 (VL _PD-1 VL2 MAb to hPD-1 7) (SEQ ID NO: 153); the CL domain of the kappa chain (SEQ ID NO: 8) and the C-terminus.

The amino acid sequence of the second polypeptide chain of BSAB is the sequence shown below (SEQ ID NO: 287):

EIVLTQSPAT LSLSPGERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL LIHAASNQGS GVPSRFSGSG SGTDFTLTIS SLEPEDFAVY FCQQSKEVPY TFGGGTKVEI KRTVAAPSVF IFPPSDEQLK SGTASWCLL NNFYPREAKV QWKVDNALQS GNSQESVTEQ DSKDSTYSLS STLT LSKADY EKHKVYACEV THQGLSSPVT KSFNRGEC

The third polypeptide chain of BSAB A comprises, in direction from N-terminus to C-terminus: an N-terminus; a VH domain of a monoclonal antibody capable of binding to LAG-3 (VH _LAG-3 VH1 MAb to hLAG-3 1) (SEQ ID NO: 49); an IgG1 CH1 domain (SEQ ID NO: 10); a variant IgG1 hinge region containing the substitutions D221R/P228R (substitutions are underlined in SEQ ID NO: 288 below); a variant IgG1 CH2CH3 domain containing the substitutions L234A/L235A/L409R (substitutions are underlined in SEQ ID NO: 288 below) and lacking the C-terminal residue; and a C-terminus.

The amino acid sequence of the third polypeptide chain of BSAB A is the sequence shown below (SEQ ID NO: 288):

QVQLVQSGAE VKKPGASVKV SCKASGYTFT NYGMNWVRQA PGQGLEWMGW INTYTGESTY ADDFEGRFVF SMDTSASTAY LQISSLKAED TAVYYCARES LYDYYSMDYW GQGTTVTVSS ASTKGPSVFP LAPSSKSTSG GTAALGCLVK DYFPEPVTVS WNSGALTSGV HTFPAVLQSS GLYSLSSWT VPSSSLGTQT YICNVNHKPS NTKVDKRVEP KSCRKTHTCP RCPAPEAAGG PSVFLFPPKP KDTLMISRTP EVTCVWDVS HEDPEVKFNW YVDGVEVHNA KTKPREEQYN STYRWSVLT VLHQDWLNGK EYKCKVSNKA LPAPIEKTIS KAKGQPREPQ VYTLPPSREE MTKNQVSLTC LVKGFYPSDI AVEWESNGQP ENNYKTTPPV LDSDGSFFLY SRLTVDKSRW QQGNVFSCSV MHEALHNHYT QKSLSLSPG

The fourth polypeptide chain of BSAB A contains, in the direction from the N-terminus to the C-terminus: the N-terminus; the VL domain of a monoclonal antibody capable of binding to LAG-3 (VL _lag-3 VL4 MAT to hLAG-3 1) (SEQ ID NO: 54); the CL domain of the kappa chain (SEQ ID NO: 8) and the C-terminus.

The amino acid sequence of the fourth polypeptide chain of BSAB A is the sequence shown below (SEQ ID NO: 289):

DIVMTQTPLS LSVTPGQPAS ISCKSSQSLL HSDAKTYLNW LLQKPGQPPE RLIYLVSELD SGVPDRFSGS GSGTDFTLKI SRVEAEDVGV YYCWQGTHFP YTFGGGTKVE IKRTVAAPSV FIFPPSDEQL KSGTASWCL LNNFYPREAK VQWKVDNALQ SGNSQESVTE SSTLTLSKAD YEKHKVYACE VTHQGLSSPV TKSFNRGEC IX. Reference antibodies.

A. Reference antibodies to human PD-1.

In order to evaluate and characterize the novel human PD-1 binding molecules of the present invention, the following reference antibodies were used: nivolumab (also known as 5C4, BMS-936558, ONO-4538, MDX-1106 and marketed as OPDIVO™ by Bristol-Myers Squibb), a human IgG4 antibody, referred to herein as anti-PD-1A mAb, and pembrolizumab (formerly known as lambrolizumab, also known as MK-3475, SCH-900475, and marketed as KEYTRUDA™ by Merck), a humanized IgG4 antibody, referred to herein as anti-PD-III mAb.

1. Nivolumab (mAb to PD-1A).

- 77 047726

The amino acid sequence of the heavy chain variable domain of the anti-PD-1A mAb contains the amino acid sequence shown below (SEQ ID NO: 64) (CDRH residues are underlined):

QVQLVESGGG WQPGRSLRL DCKASGITFS NSGMHWVRQA PGKGLEWVAV

IWYDGSKRYY ADSVKGRFTI SRDNSKNTLF LQMNSLRAED TAVYYCATND

DYWGQGTLVT VSS

The amino acid sequence of the light chain variable domain of the anti-PD-1A mAb contains the amino acid sequence shown below (SEQ ID NO: 65) (CDRL residues are underlined):

EIVLTQSPAT LSLSPGERAT LSCRASQSVS SYLAWYQQKP GQAPRLLIYD

ASNRATGIPA RFSGSGSGTD FTLTISSLEP EDFAVYYCQQ SSNWPRTFGQ

GTKVEIK

2. Pembrolizumab (mAb to PD-1B).

The amino acid sequence of the heavy chain variable domain of the anti-PD-1B mAb contains the amino acid sequence shown below (SEQ ID NO: 66) (CDRH residues are underlined):

QVQLVQSGVE VKKPGASVKV SCKASGYTFT NYYMYWVRQA PGQGLEWMGG

INPSNGGTNF NEKFKNRVTL TTDSSTTTAY MELKSLQFDD TAVYYCARRD

YRFDMGFDYW GQGTTVTVSS

The amino acid sequence of the light chain variable domain of the anti-PD-1B mAb contains the amino acid sequence shown below (SEQ ID NO: 67) (CDRL residues are underlined):

EIVLTQSPAT LSLSPGERAT LSCRASKGVS TSGYSYLHWY QQKPGQAPRL

LIYIASYLES GVPARFSGSG SGTDFTLTIS SLEPEDFAVY YCQHSRDLPL TFGGGTKVEIK

X. Methods of obtaining.

Polynucleotide to human PD-1 and other agonists, antagonists and modulators of PD-1 can be created from polynucleotides and/or sequences of antibodies to PD-1 1-15 by methods known in the art, for example, synthetic or recombinant methods. One method for producing said peptide agonists, antagonists and modulators includes chemical synthesis of the polypeptide followed by treatment under oxidative conditions suitable for obtaining the native conformation, i.e., correct disulfide bonds. The method can be carried out using methodologies well known to those skilled in the art (see, for example, Kelley, R. F. et al. (1990) In: Genetic Engineering Principles and Methods, Setlow, J.K. Ed., Plenum Press, N.Y., vol. 12, pp 1-19; Stewart, J.M. et al. (1984) Solid Phase Peptide Synthesis, Pierce Chemical Co., Rockford, IL; see also U.S. Patents 4,105,603; 3,972,859; 3,842,067; and 3,862,925). The polypeptides of the present invention can be readily prepared using solid phase peptide synthesis (Merrifield, B. (1986) Solid Phase Synthesis, Science 232(4748):341-347; Houghten, R.A. (1985) General Method For The Rapid Solid-Phase Synthesis Of Large Numbers Of Peptides: Specificity Of Antigen-Antibody Interaction At The Level Of Individual Amino Acids, Proc. Natl. Acad. Sci. (U.S.A.) 82(15):5131-5135; Ganesan, A. (2006) Solid-Phase Synthesis In The Twenty-First Century, Mini Rev. Med. Chem. 6(1):3-10).

According to another embodiment, fully human antibodies comprising one or more CDRs from anti-PD-1 MAT 1, anti-PD-1 MAT 2, anti-PD-1 MAT 3, anti-PD-1 MAT 4, anti-PD-1 MAT 5, anti-PD-1 MAT 6, anti-PD-1 MAT 7, anti-PD-1 MAT 8, anti-PD-1 MAT 9, anti-PD-1 MAT 10, anti-PD-1 MAT 11, anti-PD-1 MAT 12, anti-PD-1 MAT 13, anti-PD-1 MAT 14, or anti-PD-1 MAT 15, or those that compete with anti-PD-1 MAT 1, anti-PD-1 MAT 2, anti-PD-1 MAT 3, anti-PD-1 MAT 4, anti-PD-1 MAT 5, anti-PD-1 MAT 6, 7 to PD-1, MAT 8 to PD-1, MAT 9 to PD-1, MAT 10 to PD-1, MAT 11 to PD-1, MAT 12 to PD-1, MAT 13 to PD-1, MAT 14 to PD-1 or MAT 15 to PD-1 for binding to human PD-1 or a soluble form thereof can be produced using commercially available mouse strains that have been modified to express specific human immunoglobulin proteins. Transgenic animals that are designed to produce a more desirable (e.g., fully human antibodies) or a more robust immune response can also be used to produce humanized or human antibodies. Examples of suitable technology include XENOMOUSE™ (Abgenix, Inc., Fremont, CA, USA), HuMAb-Mouse™ and TC MOUSE™ (both technologies from Medarex, Inc., Princeton, NJ, USA).

In another embodiment, the antibodies can be produced recombinantly and expressed using any method known in the art. The antibodies can be produced recombinantly by first isolating the antibodies from the host animals, obtaining the gene sequence, and then using the gene sequence to recombinantly express the antibody in host cells (e.g., CHO cells). Another method that can be

- 78 047726 apply, is the expression of the antibody sequence in plants (e.g., tobacco) or transgenic milk. Suitable methods for recombinant expression of antibodies in plants or milk have been disclosed (see, e.g., Peeters et al. (2001) Production Of Antibodies And Antibody Fragments In Plants, Vaccine 19:2756; Lonberg, N. et al. (1995) Human Antibodies From Transgenic Mice, Int. Rev. Immunol 13:65-93; and Pollock et al. (1999) Transgenic Milk As A Method For The Production Of Recombinant Antibodies, J. Immunol Methods 231:147-157). Suitable methods for producing antibody derivatives, e.g., humanized, single-chain, etc., are known in the art. In another embodiment, antibodies can be produced recombinantly using phage display technology (see, e.g., U.S. Patents 5,565,332, 5,580,717, 5,733,743, 6,265,150 and Winter, G. et al. (1994) Making Antibodies By Phage Display Technology, Annu. Rev. Immunol. 12.433-455).

The antibodies or protein of interest can be sequenced using Edman digestion, which is well known to those skilled in the art. The peptide information obtained from mass spectrometry or Edman digestion can be used to design probes or primers that are used to clone the protein of interest.

An alternative method for cloning a protein of interest is to pan using purified PD-1 or portions thereof to cells expressing an antibody of interest or a protein that contains one or more CDRs from anti-PD-1 MAT 1, anti-PD-1 MAT 2, anti-PD-1 MAT 3, anti-PD-1 MAT 4, anti-PD-1 MAT 5, anti-PD-1 MAT 6, anti-PD-1 MAT 7, anti-PD-1 MAT 8, anti-PD-1 MAT 9, anti-PD-1 MAT 10, anti-PD-1 MAT 11, anti-PD-1 MAT 12, anti-PD-1 MAT 13, anti-PD-1 MAT 14, or anti-PD-1 MAT 15, or an antibody that competes with anti-PD-1 MAT 1, anti-PD-1 MAT 2, anti-PD-1 MAT 3, PD-1, MAT 4 to PD-1, MAT 5 to PD-1, MAT 6 to PD-1, MAT 7 to PD-1, MAT 8 to PD-1, MAT 9 to PD-1, MAT 10 to PD-1, MAT 11 to PD-1, MAT 12 to PD-1, MAT 13 to PD-1, MAT 14 to PD-1, or MAT 15 to PD-1 for binding to human PD1. The panning procedure can be performed by preparing a cDNA library from tissues or cells that express PD-1, overexpressing the cDNA in a second cell type, and screening the transfected second cell type to assess specific binding to PD-1 in the presence or absence of anti-PD-1 MAT 1, anti-PD-1 MAT 2, anti-PD-1 MAT 3, anti-PD-1 MAT 4, anti-PD-1 MAT 5, anti-PD-1 MAT 6, anti-PD-1 MAT 7, anti-PD-1 MAT 8, anti-PD-1 MAT 9, anti-PD-1 MAT 10, anti-PD-1 MAT 11, anti-PD-1 MAT 12, anti-PD-1 MAT 13, anti-PD-1 MAT 14, or anti-PD-1 MAT 15. Detailed descriptions of the methods used in cloning mammalian genes encoding cell surface proteins by panning can be found in the art (see, for example, Aruffo, A. et al. (1987) Molecular Cloning Of A CD28 cDNA By A High-Efficiency COS Cell Expression System, Proc. Natl. Acad. Sci. (U.S.A.) 84:8573-8577 and Stephan, J. et al. (1999) Selective Cloning Of Cell Surface Proteins Involved In Organ Development: Epithelial Glycoprotein Is Involved In Normal Epithelial Differentiation, Endocrinol. 140:5841-5854).

Vectors containing the polynucleotides of interest may be introduced into the host cell by any of a number of suitable methods, including electroporation; transfection using calcium chloride, rubidium chloride, calcium phosphate, DEAE-dextran, or other substances; microparticle bombardment; lipofection; and infection (e.g., if the vector is an infectious agent such as vaccinia virus). The choice of vectors or polynucleotides for introduction will often depend on the characteristics of the host cell. Any host cell capable of overexpressing heterologous DNAs may be used to isolate genes encoding the antibody, polypeptide, or protein of interest. Non-limiting examples of suitable mammalian host cells include, but are not limited to, COS, HeLa, and CHO cells. Preferably, the host cells express the cDNA at a level of about 5 times higher, more preferably 10 times higher, even more preferably 20 times higher than the level of the corresponding endogenous antibody or protein of interest, if present, in the host cell. Screening of the host cells for specific binding to PD-1 is performed by immunoassay or fluorescence-activated cell sorting (FACS). A cell overexpressing the antibody or protein of interest can be identified.

The present invention includes polypeptides comprising the amino acid sequence of the antibodies of the present invention. The polypeptides of the present invention can be prepared by methods known in the art. The polypeptides can be prepared by proteolytic or other type of cleavage of antibodies, by recombinant methods (i.e., single or hybrid polypeptides), as described above, or by chemical synthesis. Antibody polypeptides, especially shorter polypeptides containing no more than about 50 amino acids, are typically prepared by chemical synthesis. Chemical synthesis methods are known in the art and are commercially available. For example, a polypeptide to human PD-1 can be prepared by an automated polypeptide synthesizer using a solid phase method.

Included within the scope of the present invention are antibody variants MAT to PD-1 1, MAT 2 to PD-1, MAT 3 to PD-1, MAT 4 to PD-1, MAT 5 to PD-1, MAT 6 to PD-1, MAT 7 to PD-1, MAT 8 to PD-1, MAT 9 to PD-1, MAT 10 to PD-1, MAT 11 to PD-1, MAT 12 to PD-1, MAT 13 to PD-1, MAT 14 to PD-1 or MAT 15 to PD-1 and

- 79 047 726 polypeptide fragments thereof that bind to PD-1, including functionally equivalent antibodies and hybrid polypeptides that do not significantly affect the properties of these molecules, as well as variants that have increased or decreased activity. Modification of polypeptides is a standard method in the art and does not need to be described in detail herein. Examples of modified polypeptides include polypeptides with conservative substitutions of amino acid residues, one or more deletions or additions of amino acids that do not significantly adversely affect functional activity, or the use of chemical analogs. Amino acid residues that can be conservatively substituted for each other include, but are not limited to: glycine/alanine; serine/threonine; valine/isoleucine/leucine; asparagine/glutamine; aspartic acid/glutamic acid; lysine/arginine; and phenylalanine/tyrosine. Suitable polypeptides also include glycosylated and non-glycosylated polypeptides, as well as polypeptides with other post-translational modifications such as, for example, glycosylation with different sugars, acetylation and phosphorylation. Preferably, the amino acid substitutions will be conservative, i.e., the substituted amino acid will have chemical properties similar to those of the original amino acid. Suitable conservative substitutions are known in the art and examples have been provided above. Amino acid modifications can range from changing or modifying one or more amino acids to completely reorganizing a region such as a variable domain. Changes in the variable domain can alter binding affinity and/or specificity. Other methods of modification include the use of coupling techniques known in the art, including, but not limited to, enzymatic methods, oxidative substitution and chelation. Modifications can be used, for example, to attach labels for immunoassay, such as attaching radioactive fragments for radioimmunoassay. The modified polypeptides are prepared using procedures conventional in the art and can be screened using standard quantitative assays known in the art.

The present invention includes within its scope fusion proteins comprising one or more polypeptides or antibodies of MAT to PD-1 1, MAT 2 to PD-1, MAT 3 to PD-1, MAT 4 to PD-1, MAT 5 to PD-1, MAT 6 to PD-1, MAT 7 to PD-1, MAT 8 to PD-1, MAT 9 to PD-1, MAT 10 to PD-1, MAT 11 to PD-1, MAT 12 to PD-1, MAT 13 to PD-1, MAT 14 to PD-1 or MAT 15 to PD-1 according to the present invention. According to one embodiment, the present invention provides a fusion polypeptide that comprises a light chain, a heavy chain, or both light and heavy chains. According to another embodiment, the present invention provides a fusion polypeptide that comprises a heterologous immunoglobulin constant region. According to another embodiment of the present invention, the hybrid polypeptide comprises a light chain variable domain and a heavy chain variable domain of an antibody produced by hybridomas deposited in publicly available databases. For the purposes of the present invention, the hybrid antibody protein comprises one or more polypeptide domains that specifically bind to PD-1 and another amino acid sequence to which it is not attached in the native molecule, such as a heterologous sequence or a homologous sequence from another region.

XI. Use of PD-1-binding molecules according to the present invention.

Included within the scope of the present invention are compositions, including pharmaceutical compositions, comprising PD-1-binding molecules of the present invention (e.g., anti-PD1 antibodies, anti-PD-1 bispecific diabodies, etc.), polypeptides derived from said molecules, polynucleotides comprising sequences encoding said molecules or polypeptides, and other agents described herein.

A. Therapeutic options.

As discussed above, PD-1 plays an important role in the negative regulation of T cell proliferation, function and homeostasis. Some of the PD-1-binding molecules of the present invention have the ability to inhibit PD-1 function and thereby reverse PD-1-mediated inhibition of the immune system. In this regard, MAT to PD-1 1, MAT 3 to PD-1, MAT 5 to PD-1, MAT 6 to PD-1, MAT 7 to PD-1, MAT 8 to PD-1, MAT 9 to PD-1, MAT 10 to PD-1, MAT 11 to PD-1, MAT 12 to PD-1, MAT 13 to PD-1, MAT 14 to PD-1 and MAT 15 to PD-1, their humanized derivatives and molecules comprising PD-1-binding fragments thereof (e.g., bispecific antibodies, bispecific diabodies (including, but not limited to, DART-A, DART-B, DART-C, DART-D, DART-E, DART-F, DART-G, DART-H, DART-I and DART-J) and etc.), or those that compete for binding to these antibodies, may be used to block PD-1-mediated inhibition of the immune system and thereby promote activation of the immune system.

Said bispecific PD-1 binding molecules according to the present invention, which bind to PD-1 and another molecule involved in the regulation of an immune checkpoint present on the cell surface (e.g., LAG-3), enhance the immune system by blocking the inhibition of the immune system mediated by PD-1 and said checkpoint molecules.

- 80 047726 immune response control point. Accordingly, said PD-1-binding molecules of the present invention can be used to enhance the immune response (e.g., T-cell-mediated immune response) of a subject. In particular, said PD-1-binding molecules of the present invention can be used to treat any diseases or conditions associated with an undesirably suppressed immune system, including cancer and diseases that are associated with the presence of a pathogen (e.g., a bacterial, fungal, viral or protozoan infection).

Cancers that can be treated using said PD-1 binding molecules according to the present invention include cancers characterized by the presence of a cancer cell selected from the group consisting of: adrenal tumor, AIDS-associated cancer, alveolar soft tissue sarcoma, astrocytic tumor, bladder cancer, bone cancer, brain and spinal cord cancer, metastatic brain tumor, breast cancer, carotid artery tumors, cervical cancer, chondrosarcoma, chordoma, chromophobe renal cell carcinoma, clear cell carcinoma, colon cancer, colorectal cancer, cutaneous benign fibrous histiocytoma, desmoplastic small round cell tumor, ependymoma, Ewing's tumor, extraskeletal myxoid chondrosarcoma, fibrogenesis imperfecta of bone, fibrous dysplasia of bone, gallbladder cancer or bile duct, gastric cancer, gestational trophoblastic disease, germ cell tumor, head and neck cancer, hepatocellular carcinoma, islet cell tumor, Kaposi sarcoma, renal cell cancer, leukemia, lipoma/benign lipomatous tumor, liposarcoma/malignant lipomatous tumor, liver cancer, lymphoma, lung cancer, medulloblastoma, melanoma, meningioma, multiple endocrine neoplasia, multiple myeloma, myelodysplastic syndrome, neuroblastoma, neuroendocrine tumors, ovarian cancer, pancreatic cancer, papillary thyroid carcinoma, parathyroid tumor, pediatric cancer, peripheral nerve sheath tumor, pheochromocytoma, pituitary tumor, prostate cancer, posterior uveal melanoma, rare hematological disorder, metastatic renal cell carcinoma, rhabdoid tumor, rhabdomyosarcoma, sarcoma, skin cancer, soft tissue sarcoma, squamous cell carcinoma, gastric cancer, synovial sarcoma, testicular cancer, thymic carcinoma, thymoma, metastatic thyroid cancer, and uterine cancer.

In particular, the PD-1 binding molecules of the present invention can be used to treat colorectal cancer, hepatocellular carcinoma, glioma, renal cancer, breast cancer, multiple myeloma, bladder cancer, neuroblastoma, sarcoma, non-Hodgkin's lymphoma, non-small cell lung cancer, ovarian cancer, pancreatic cancer and rectal cancer.

Pathogen-related diseases that can be treated using said PD-1 binding molecules according to the present invention include chronic viral, bacterial, fungal and parasitic infections. Chronic infections that can be treated using the PD-1 binding molecules according to the present invention include Epstein-Barr virus, hepatitis A virus (HAV); hepatitis B virus (HBV); hepatitis C virus (HCV); herpes viruses (e.g. HSV-1, HSV-2, HHV-6, CMV), human immunodeficiency virus (HIV), vesicular stomatitis virus (VSV), infections caused by

Bacilli, Citrobacter, Cholera, Diphtheria, Enterobacter, Gonococci, Helicobacter pylori, Klebsiella, Legionella, Meningococci, Mycobacterium, Pseudomonas, Pneumonococci, Rickettsia bacteria,

Salmonella, Serratia, Staphylococci, Streptococci, Tetanus,

Aspergillus (fumigatus, niger, etc.), Blastomyces dermatitidis, Candida (albicans, krusei, glabrata, tropicalis, etc.), Cryptococcus neoformans, Mucorales species (mucor, absidia, rhizopus), Sporothrix schenkii, Paracoccidioides brasiliensis, Coccidioides immitis, Histoplasma capsulatum, Leptospirosis, Borrelia burgdorferi, helminth parasites (hookworms, tapeworms, trematodes, flatworms (e.g.

Giardia lambia, trichinella, Dientamoeba Fragilis, Trypanosoma brucei, Trypanosoma cruzi and

Leishmania donovani).

The PD-1 binding molecules of the present invention can be combined with other anti-cancer agents, in particular molecules that specifically bind to a cancer antigen (e.g., antibodies, diabodies). Anti-cancer drugs that can be combined with the PD-1 binding molecules of the present invention include molecules that specifically bind to another cancer antigen, including: 19.9, found in colon cancer, in gastric cancer mucins; 4.2; A33 (colorectal carcinoma antigen, Almqvist, Y. 2006, Nucl Med Biol. Nov; 33(8):991-998); ADAM-9 (US Patent Publication 2006/0172350, PCT Publication WO 06/084075); AH6, found in gastric cancer; ALCAM (PCT Publication WO

- 81 047726

03/093443); APO-1 (human lymphocyte malignant antigen) (Trauth et al. (1989) "Monoclonal

Antibody-Mediated Tumor Regression By Induction of Apoptosis,” Science 245:301-304); Bl (Egloff, A.M. et al. 2006, Cancer Res. 66(l):6-9); B7-H3 (Collins, M. et al. (2005) “The B7 Family Of Immune-Regulatory Ligands”, Genome Biol. 6:223.1-223.7). Chapoval, A. et al. (2001) “B7-NZ: A Costimulatory Molecule For T Cell Activation and IFN-γ Production,” Nature Immunol. 2:269-274; Sun, M. et al. (2002) "Characterization of Mouse and Human B7-H3 Genes", J. Immunol. 168:6294-6297); BAGE (Bodey, B. 2002 Expert Opin Biol Ther. 2(6):57784); beta-catenin (Prange W. et al. 2003 J Pathol. 201(2):250-9);

blood group antigens ALe ^b /Le ^y found in colon adenocarcinoma; Burkitt lymphoma antigen-38.13, C14 found in colon adenocarcinoma; CA125 (ovarian carcinoma antigen) (Bast, RC Jr. et al. 2005 Int J Gynecol Cancer 15 Suppl 3:274-81 ; Yu et al. (1991) "Coexpression Of Different Antigenic Markers On Moieties That Bear CA 125 Determinants", Cancer Res. 51(2):468-475); carboxylpeptide M (US Patent Publication 2006/0166291); CD5 (Calin, GA et al. 2006 Semin Oncol. 33(2): 167-73; CD19 (Ghetie et al. (1994) "Anti-CD19 Inhibits The Growth Of Human B-Cell Tumor Lines In Vitro And Of Daudi Cells In SCID Mice By Inducing Cell Cycle Arrest", Blood 83:1329-1336; Troussard, X et al. 1998 Hematol Cell Ther. 40(4): 139-48); (Kreitman, RJ 2006 AAPS J. 18;8(3):E532-51); CD23 (Rosati, S. et al. 2005 Curr Top Microbiol Immunol. 5;294:91-107); CD25 (Troussard, X. et al. 1998 Hematol Cell Ther. 40(4): 139-48); CD27 (Bataille, R. 2006 Haematologica 91(9): 1234-40); CD28 (Bataille, R. 2006 Haematologica 91(9): 1234-40); CD33 (Sgouros et al. (1993) "Modeling And Dosimetry Of Monoclonal Antibody M195 (Anti-CD33) In Acute Myelogenous Leukemia", J. Nucl. Med. 34:422-430); CD36 (Ge, Y. 2005 Lab Hematol. 11(1):31-7); CD40/CD154 (Messmer, D. etal. 2005 Ann NY Acad Sci. 1062:51-60); CD45 (Jurcic, JG 2005 Curr Oncol Rep. 7(5):339-46); CD56 (Bataille, R. 2006 Haematologica 91(9): 1234-40); CD46 (US Patent No. 7148038; PCT Publication WO 03/032814); CD52 (Eketorp, S.S. et al. (2014) "Alemtuzumab (Anti-CD52 Monoclonal Antibody) As Single-Agent Therapy In Patients With Relapsed/Refractory Chronic Lymphocytic Leukaemia (CLL)-A Single Region Experience On Consecutive Patients", Ann Hematol. 93(10): 1725-1733; Suresh, T. et al. (2014) Antibody Approaches To Lymphoma Therapy," J. Hematol. Oncol. 7:58; Hoelzer, D. (2013) "Targeted Therapy With Monoclonal Antibodies in Acute Lymphoblastic Leukemia," Curr. Opin. 25(6):701-706); CD56 (Bataille, R. 2006 Haematologica 91(9): 1234-40); CD79a/CD79b (Troussard, X. et al. 1998 Hematol Cell Ther. 40(4):139-48; Chu, PG et al. 2001 Appl Immunohistochem Mol Morphol. 9(2):97-106); CD103 (Troussard, X. et al. 1998 Hematol Cell Ther. 40(4): 139-48); CD317 (Kawai, S. et al. (2008) "Interferon-A Enhances CDS 17 Expression And The Antitumor Activity Of Anti-CD317 Monoclonal Antibody In Renal Cell Carcinoma Xenograft Models", Cancer Science 99(12):2461-2466; Wang, W. et al. (2009) HM1.24 (CD317) Is A Novel Target Against L ung Cancer For Immunotherapy Using Anti-HM1.24 Antibody,” Cancer Immunology, Immunotherapy 58(6):967-976; Wang, W. et al. (2009) “Chimeric And Humanized Anti-HM 1.2 4 Antibodies Mediate Antibody-Dependent Cellular Cytotoxicity Against Lung Cancer Cells,” 63(1): 23-31; (2013) "Aberrant Regulation Of The BST2 (Tetherin) Promoter Enhances Cell Proliferation And Apoptosis Evasion In High Grade Breast Cancer Cells", PLoS ONE 8(6)e67191, pp. 1-10); CDK4 (Lee, YM etal. 2006 Cell Cycle 5(18):2110CE CEA (carcinoembryonic antigen;

- 82 047726

Foon et al. (1995) "Immune Response To The Carcinoembryonic Antigen In Patients Treated With An Anti-Idiotype Antibody Vaccine", J. Clin. Invest. 96(l):334-42); Mathelin, C. 2006 Gynecol Obstet Fertil. 34(7-8):638-46; TellezAvila, F.I. et al. 2005 Rev Invest Clin. 57(6):814-9); CEACAM9/CEACAM6 (Zheng, C. et al. (2011) "Α Novel Anti-CEACAM5 Monoclonal Antibody, CC4, Suppresses Colorectal Tumor Growth and Enhances NK Celis-Mediated Tumor Immunity", PLoS One 6(6):e21146, pp. 1-11); CO 17-1A (Ragnhammar et al. (1993) "Effect Of Monoclonal Antibody 17-1A And GM-CSF In Patients With Advanced Colorectal Carcinoma - Long-Lasting, Complete Remissions Can Be Induced", Int. J. Cancer 53:751 -758);

CO-43 (blood group antigen Le ^b ); CO-514 (blood group antigen Le ^a ), found in adenocarcinoma; CTA-1; CTLA-4 (Peggs, KS et al., 2006 Curr Opin Immunol. 18(2):206-13); cytokeratin 8 (PCT publication WO 03/024191); D1.1; D156-22; DR5 (Abdulghani, J. et al. (2010) "TRAIL Receptor Signaling And Therapeutics", Expert Opin. Ther. Targets 14(10): 1091-1108; Andera, L. (2009) "Signaling Activated By The Death Receptors Of The TNFR Family", Biomed. Pap. Med. Fac. Univ. Palacky Olomouc Czech. Repub. 153(3): 173-180; Carlo-Stella, C. et al. (2007) “Targeting TRAIL Agonistic Receptors for Cancer Therapy”, Clin, Cancer 13(8): 2313-2317; Chaudhari, B. R. et al. (2006) “Following the TRAIL to Apoptosis”, Immunologic Res.

the E1 series (blood group B), found in pancreatic cancer; EGFR (epidermal growth factor receptor, Adenis, A. et al., 2003 Bull Cancer. 90 Spec No:S228-32); ephrin (and in particular EphA2 (US Patent 7,569,672, PCT Publication WO 06/084226), Erb (ErbB1, ErbB3, ErbB4, Zhou, H. et al. 2002 Oncogene 21(57):8732-8740; Rimon, E. et al. 2004 Int J Oncol. 24(5): 1325-1338); GAGE (GAGE-1; GAGE-2; Akcakanat, A. et al. 2006 Int J Cancer. 118(1): 123-128); GD2/GD3/GM2 (Livingston, P. O. et al. 2005 Cancer Immunol Immunother. 54(10): 1018-1025); ganglioside GD2 (Gru; Saleh et al. (1993) "Generation Of A Human Anti-Idiotypic Antibody That Mimics The GD2 Antigen", J. Immunol., 151, 3390-3398); ganglioside GD3 (Gd3; Shitara et al. (1993) “A Mouse/Human Chimeric Anti-(Ganglioside GD3) Antibody With Enhanced Antitumor Activities,” Cancer Immunol. Immunother. 36:373-380); GM2 ganglioside (Gw; Livingston et al. (1994) "Improved Survival In Stage III Melanoma Patients With GM2 Antibodies: A Randomized Trial Of Adjuvant Vaccination With GM2 Ganglioside", J. Clin. Oncol. 12:1036-1044); ganglioside GM3 (Gw; Hoon et al. (1993) "Molecular Cloning Of A Human Monoclonal Antibody Reactive To Ganglioside GM3 Antigen On Human Cancers", Cancer Res. 53:5244-5250); GICA 19-9 (Herlyn et al. (1982) "Monoclonal Antibody Detection Of Aulating Circulating Tumor-Associated Antigen. I. Presence Of Antigen In Sera Of Patients With Colorectal, Gastric, And Pancreatic Carcinoma", J. Clin. Immunol. 2:135-140); gplOO (Lotem, M. et al. 2006 J Immunother. 29(6):616-27);

Gp37 (human T-cell leukemia antigen,

Bhattacharya-Chatterjee et al. (1988) “Idiotype Vaccines Against Human T Cell Leukemia. II. Generation And Characterization Of A Monoclonal Idiotype Cascade (Abi, Ab2, and Ab3),” J. Immunol. 141:1398-1403); gp75 (melanoma antigen; Vijayasardahl et al. (1990) "The Melanoma Antigen Gp75 Is The Human Homologue Of The Mouse In (Brown) Locus Gene Product", J. Exp. Med. 171(4): 1375-1380); gpA33 (Heath, J.K. et al. (1997) "The Human A33 Antigen Is A Transmembrane Glycoprotein And A Novel Member Of The Immunoglobulin Superfamily", Proc. Natl. Acad. Sci. (U.S.A.) 94(2):469-474; Ritter, G. et al. (1997) “Characterization of Posttranslational Modifications of Human A33 Antigen, A Novel Palmitoylated Surface Glycoprotein of Human Gastrointestinal Epithelium,” Biochem. Res. Commun. 236(3):682–686; et al. (2006) "EpCAM and gpA33 are Markers of Barrett's Metaplasia", J. Clin. Pathol. 59(3):260-263);

HER2 antigen (HER2/neu, p185 ^HER2 ; Kumar, Pal S et al., 2006 Semin Oncol. 33(4):386-91); HMFG (human milk fat globule antigen, WO1995015171); human papillomavirus-B6/human papillomavirus-E7 (DiMaio, D. et al., 2006 Adv Virus Res. 66:125-59; HMW-MAA (high molecular weight melanoma antigen,

- 83 047726

Natali et al. (1987) “Immunohistochemical Detection Of Antigen In Human Primary And Metastatic Melanomas By The Monoclonal Antibody 140.240 And Its Possible Prognostic Significance”, Cancer 59:55-63;

Mittelman et al. (1990) “Active Specific Immunotherapy In Patients With Melanoma. A Clinical Trial With Mouse Antiidiotypic Monoclonal Antibodies Elicited With Syngeneic Anti-HighMolecular-Weight-Melanoma-Associated Antigen Monoclonal Antibodies,” J. Clin. Invest.

86:2136-2144); antigen I (differentiation antigen, Feizi (1985) “Demonstration By Monoclonal Antibodies That Carbohydrate Structures Of Glycoproteins And Glycolipids Are Onco-Developmental Antigens”, Nature 314:53-57); IL13Ra2 (PCT publication WO 2008/146911; Brown, C.E. et al. (2013) ".Glioma IL13Ra2 Is Associated With Mesenchymal Signature Gene Expression And Poor Patient Prognosis", PLoS One. 18;8(10):e77769;

Barderas, R. et al. (2012) "High Expression Of IL-13 Receptor A2 In Colorectal Cancer Is Associated With Invasion, Liver Metastasis, And Poor Prognosis", Cancer Res. 72(11):27802790; Kasaian, M.T. et al. (2011) "IL-13 Antibodies Influence IL-13 Clearance In Humans By Modulating Scavenger Activity Of IL-13Ra2", J. Immunol. 187(l):561-569; Bozinov, O. et al.

(2010) "Decreasing Expression Of The Interleukin-13 Receptor IL-13Ralpha2 In Treated Recurrent Malignant Gliomas", Neurol. Med. Chir. (Tokyo) 50(8):617-621; Fujisawa, T. et al.

(2009) “A novel role of inter leukin-13 receptor alpha2 in pancreatic cancer invasion and metastasis,” Cancer Res. 69(22):8678-8685);

integrin β6 (PCT Publication No. WO 03/087340); JAM-3 (PCT Publication No. WO 06/084078); KID3 (PCT Publication No. WO 05/028498); KID31 (PCT Publication No. WO 06/076584); carcinoma common antigen KS 1/4 Hellstrom et al. (1986) Monoclonal Mouse Antibodies Raised Against Human Lung Carcinoma, Cancer Res. 46:3917-3923); LEA; LUCA-2 (U.S. Patent Publication Nos. 2006/0172349; PCT Publication No. WO 06/083852); M1:22:25:8; M18; M39; MAGE (MAGE-1; MAGE-3; (Bodey, V. 2002 Expert Opin Biol Ther. 2(6):577-84); MART (Kounalakis, N. et al., 2005 Curr Oncol Rep. 7(5):377-82; mesothelin (Chang K, and Pastan I. 1996 Molecular cloning of mesothelin, a differentiation antigen present on mesothelin elium, mesotheliomas, and ovarian cancers, Proc Natl Acad Sci USA 93:136-40), MUC-1 (Mathelin, C 2006 Gynecol Obstet Fertil. 34(7-8):638-46); Myl; N-acetylglucosaminyltransferase (Dennis, JW 1999 Biochim Biophys Acta. 6;1473(1):21-34); neoglycoprotein; NS-10, found in adenocarcinomas; OFA-1; OFA-2; oncostatin M (oncostatin beta receptor, U.S. Patent No. 7,572,896, PCT Publication WO 06/084092); p15 (Gil, J. et al., 2006 Nat Rev Mol Cell Biol. 7(9):667-77); p97 (melanoma-associated antigen; Estin et al. (1989) Transfected Mouse Melanoma Lines That Express Various Levels Of Human Melanoma-Associated Antigen p97, J. Natl. Cancer Instit. 81(6):445-454); PEM (polymorphic epithelial mucin, Hilkens et al. (1992) Cell MembraneAssociated Mucins And Their Adhesion-Modulating Property, Trends in Biochem. Sci. 17:359-363); PEMA (polymorphic epithelial mucin antigen); PIPA (US Patent No. 7,405,061, PCT Publication WO 04/043239); PSA (prostate specific antigen, Henttu et al. (1989) cDNA Coding For The Entire Human Prostate Specific Antigen Shows High Homologies To The Human Tissue Kallikrein Genes, Biochem. Biophys. Res. Comm. 10(2):903-910; Israeli et al. (1993) Molecular Cloning Of A Complementary DNA Encoding A Prostate-Specific Membrane Antigen, Cancer Res. 53:227-230; Cracco, CM et al., 2005 Minerva Urol Nefrol. 57(4):301-11); PSMA (prostate specific membrane antigen, Ragupathi, G. 2005 Cancer Treat Res. 123:157-180); prostatic acid phosphatase (Tailor et al. (1990) Nucleotide Sequence Of Human Prostatic Acid Phosphatase Determined From A Full-Length cDNA Clone, Nucl. Acids Res. 18(16):4928); R ₂₄ , found in melanoma; ROR1 (U.S. Patent No. 5,843,749); sphingolipids; SSEA1; SSEA-3; SSEA-4; sTn (Holmberg, LA 2001 Expert Opin Biol Ther. 1(5):881-91); T-cell receptor, derived from cutaneous T-cell lymphoma (see Edelson (1998) Cutaneous T-Cell Lymphoma: A Model For Selective Immunotherapy, Cancer J Sci Am. 4:62-71); T5A7, found in myeloid cells; TAG-72 (Yokota et al. (1992) Rapid Tumor Penetration Of A Single-Chain Fv And Comparison With Other Immunoglobulin Forms, Cancer Res. 52:3402–3408); TL5 (blood group A); TNF receptor (TNFα receptor, TNF-β receptor; TNF-γ receptor (van Horssen, R. et al., 2006 Oncologist. 11(4):397-408; Gardnerova, M. et al., 2000 Curr Drug Targets. 1(4):327-64); TRA-1-85 (blood group H), transferrin receptor (US Patent No. 7,572,895, PCT Publication WO 05/121179), 5T4 (TPBG, trophoblast glycoprotein; Boghaert, ER et al. (2008) The Oncofetal Protein, 5T4, Is A Suitable Target For Antibody-Guided Anti-Cancer Chemotherapy With Calicheamicin, Int. J. Oncol. 32(1):221-234; Eisen, T. et al. (2014) Naptumomab Estafenatox: Targeted Immunotherapy with a Novel Immunotoxin, Curr. Oncol. Rep. 16:370, p. 1-6); TSTA (tumor-specific transplantation antigen), such as virally induced tumor antigens, including T antigen of DNA tumor viruses and envelope antigens of RNA tumor

- 84 047726 holoviruses, oncofetal antigen-alpha-fetoprotein, such as colon CEA, oncofetal antigen of bladder tumor (Hellstrom et al. (1985) "Monoclonal Antibodies To Cell Surface

Antigens Shared By Chemically Induced Mouse Bladder Carcinomas", Cancer. Res. 45:22102188); VEGF (Pietrantonio, F. et al. (2015) "Bevacizumab-Based Neoadjuvant Chemotherapy For Colorectal Cancer Liver Metastases: Pitfalls And Helpful Tricks In A Review For Clinicians", Crit. Rev. Oncol. Hematol. 95(3):272- 281; Grabowski, J.P. (2015) ".Current Management of Ovarian Cancer", Minerva Med. 106(3):151-156; Field, K.M. (2015) "Bevacizumab And Glioblastoma: Scientific Review, Newly Reported Updates, And Ongoing Controversies ", Cancer 121(7):997-1007; Suh, D.H. et al. (2015) "Major Clinical Research Advances In Gynecologic Cancer In 2014", J. Gynecol. Oncol. 26(2): 156-167; Liu, K.J. et al.

(2015) "Bevacizumab In Combination With Anticancer Drugs For Previously Treated Advanced Non-Small Cell Lung Cancer", Tumour Biol. 36(3): 1323-1327; Di Bartolomeo, M. et al. (2015) “Bevacizumab treatment in the elderly patient with metastatic colorectal cancer,” Clin. Interv. Aging 10:127-133);

VEGF receptor (O'Dwyer. PJ 2006 Oncologist. 11(9):992-998); VEP8; VEP9; VIM-D5; and hapten Y, Le ^y , found in embryonal carcinoma cells. In some embodiments, the PD-1 binding molecules of the invention are used in combination with one or more molecules that specifically bind to 5T4, B7H3, CD19, CD20, CD51, CD123, DR5, EGFR, EpCam, GD2, gpA33, HER2, ROR-1, TAG-72, an antibody to VEGF-A, and/or VEGFR2.

The PD-1 binding molecules of the present invention may be combined with an immunogenic agent, such as a tumor vaccine. Suitable vaccines may contain purified tumor antigens (including recombinant proteins, peptides and carbohydrate molecules), autologous or allogeneic tumor cells. Several tumor vaccine strategies have been described (see, for example, Palena, S., et al., (2006) “Cancer vaccines: preclinical studies and novel strategies,” Adv. Cancer Res. 95, 115-145; Mellman, I., et al. (2011) “Cancer immunotherapy comes of age,” Nature 480, 480-489; Zhang, X. M. et al. (2008) “The anti -tumor immune response induced by a combination of MAGE-3/MAGE-n-derived peptides,” Oncol. Rep. 20, 245-252; Disis, M. L. et al. (2002) “Generation of T-cell immunity to the HER-2/neu protein after active immunization with HER-2/neu peptide-based vaccines,” J. Clin. 26242632; Vermeij, R. et al. (2012) “Potentiation of a p53-SLP vaccine by cyclophosphamide in ovarian cancer: a single-arm phase II study.” Int. J. Cancer 131, E670-E680).

The PD-1 binding molecules of the present invention can be combined with chemotherapeutic regimens. In these cases, it is possible to reduce the dose of the administered chemotherapeutic reagent (Mokyr et al. (1998) Cancer Research 58: 5301-5304).

The PD-1 binding molecules of the present invention can be combined with other immunostimulatory molecules, such as antibodies, that activate the host immune system's responsiveness to provide increased levels of T cell activation. In particular, anti-PD-1 antibodies, anti-PD-L1 antibodies, and/or anti-CTLA-4 antibodies have been shown to activate the immune system (see, e.g., del Rio, M-L. et al. (2005) "Antibody-Mediated Signaling Through

PD-1 Costimulates T Cells And Enhances CD28-Dependent Proliferation,” Eur. J Immunol 35:3545–3560; Barber, D. L. et al. (2006) “Restoring function in exhausted CD8 T cells during chronic viral infection,” Nature 439, 682-687; Iwai, Y. et al. (2002) "Involvement OfPD-Ll On Tumor Cells In The Escape From Host Immune System And Tumor Immunotherapy By PD-L1 Blockade", Proc. Natl Acad. Sci. USA 99, 12293-12297; Leach, D. R., et al., (1996) "Enhancement Of Antitumor Immunity By CTLA-4 Blockade", Science 271, 1734-1736).

Additional immunostimulatory molecules that can be combined with the PD-1 binding molecules of the present invention include antibodies to molecules on the surface of dendritic cells that activate dendritic cell (DC) function and antigen presentation, antibodies to CD40 that can replace helper T cell activity, and activating antibodies to T cell costimulatory molecules such as PD-L1, CTLA-4, OX-40 4-1BB, and ICOS (see, e.g.,

- 85 047726

Ito et al. (2000) "Effective Priming Of Cytotoxic T Lymphocyte Precursors By Subcutaneous

Administration Of Peptide Antigens In Liposomes Accompanied By Anti-CD40 And Anti-CTLA-4

Antibodies”, Immunobiology 201:527-40; US Patent No. 5811097; Weinberg et al. (2000) “Engagement of the OX-40 Receptor in vivo Enhances Antitumor Immunity,” Immunol 164:2160–2169; Melero et al. (1997) “Monoclonal Antibodies Against The 4-IBB T-Cell Activation Molecule Eradicate Established Tumors,” Nature Medicine 3: 682–685; Hutloff et al.

(1999) "ICOS Is An Inducible T-Cell Co-Stimulator Structurally And Functionally Related To CD28", Nature 397: 263-266; and Moran, A.E. et al. (2013) "The TNFRs 0X40, 4-1BB, and CD40 As Targets for Cancer Immunotherapy", Curr Opin Immunol. 2013 Apr; 25(2): 10.1016/j.coi.2013.01.004), and/or stimulatory chimeric antigen receptors (CARs) comprising an antigen-binding domain directed to a disease antigen fused to one or more intracellular signaling domains from different costimulatory protein receptors ( eg CD28, 4-1BB, ICOS, OX40, etc.), which serve to stimulate T cells upon antigen binding (see, eg,

Tettamanti, S. et al. (2013) "Targeting Of Acute Myeloid Leukaemia By CytokineInduced Killer Cells Redirected With A Novel CD123-Specific Chimeric Antigen Receptor", Br.

J. Haematol. 161:389-401; Gill, S. et al. (2014) “Efficacy Against Human Acute Myeloid Leukemia And Myeloablation Of Normal Hematopoiesis In A Mouse Model Using Chimeric Antigen Receptor-Modified T Cells”, Blood 123(15): 2343-2354; Mardiros, A. et al. (2013) “T Cells Expressing CD123-Specific Chimeric Antigen Receptors Exhibit Specific Cytolytic Effector Functions and Antitumor Effects Against Human Acute Myeloid Leukemia,” Blood 122:31383148; Pizzitola, I. et al. (2014) "Chimeric Antigen Receptors Against CD33/CD123 Antigens Efficiently Target Primary Acute Myeloid Leukemia Cells in Vivo", Leukemia doi:10.1038/leu.2014.62).

The PD-1 binding molecules of the present invention can be combined with an inhibitory chimeric antigen receptor (iCAR) to redirect targeted immunotherapeutic responses. The iCAR comprises an antigen-binding domain directed to a disease antigen fused to one or more intracellular signaling domains from various inhibitory protein receptors (e.g., CTLA-4, PD-1, etc.) that serve to limit T cell responses upon binding to the antigen (see, e.g., Fedorov V.D. (2013) PD-1-and CTLA-4Based Inhibitory Chimeric Antigen Receptors (iCARs) Divert Off-Target Immunotherapy Responses, Sci. Tranl. Med. 5:215ra172 doi:10.1126/scitranslmed.3006597. In particular, the PD-1-binding molecules of the present invention are used in combination with an anti-CD137 antibody, an anti-CTLA-4 antibody, an anti-OX40 antibody, an anti-LAG-3 antibody, an anti- PD-L1, anti-TIGIT antibody, anti-TIM-3 antibody and/or cancer vaccine.

B. Diagnostic and therapeutic-diagnostic utility.

Some of the PD-1-binding molecules of the present invention are unable or exhibit weak ability to block the binding of PD-1 to the ligand PD-1L. In this regard, the antibodies MAB 2 to PD-1 and MAB 4 to PD-1, their humanized derivatives, as well as molecules containing their PD-1-binding fragments (e.g., bispecific diabodies, etc.), or those that compete for binding with said antibodies, can be labeled with a detectable label (e.g., using radioactive, enzymatic, fluorescent, chemiluminescent, paramagnetic, diamagnetic or other labeled moieties) and used to detect PD-1 in samples or visualize PD-1 on cells. Because these molecules do not affect the biological activity of PD-1, they are particularly useful in methods for determining the extent, location, and change in PD-1 expression in subjects (e.g., subjects being treated for cancer associated with PD-1 expression or targeting).

XII. Pharmaceutical compositions.

The compositions of the present invention include bulk drug compositions suitable for the manufacture of pharmaceutical compositions (e.g., crude or non-sterile compositions) and pharmaceutical compositions (i.e., compositions that are suitable for administration to a subject or patient) that can be used in the preparation of unit dosage forms. Suitable compositions comprise a prophylactically or therapeutically effective amount of the PD-1 binding molecules of the present invention or a combination of suitable agents and a pharmaceutically acceptable carrier. Preferably, the compositions of the present invention comprise a prophylactically or therapeutically effective amount of the PD-1-86 047 726 binding molecules of the present invention and a pharmaceutically acceptable carrier. The present invention particularly includes suitable pharmaceutical compositions wherein the PD-1 binding molecule is: anti-PD-1 MAT 1, anti-PD-1 MAT 2, anti-PD-1 MAT 3, anti-PD-1 MAT 4, anti-PD-1 MAT 5, anti-PD-1 MAT 6, anti-PD-1 MAT 7, anti-PD-1 MAT 8, anti-PD-1 MAT 9, anti-PD-1 MAT 10, anti-PD-1 MAT 11, anti-PD-1 MAT 12, anti-PD-1 MAT 13, anti-PD-1 MAT 14, or anti-PD-1 MAT 15; a humanized anti-PD-1 MAT 1; MAT 2 to PD-1, MAT 3 to PD-1, MAT 4 to PD-1, MAT 5 to PD-1, MAT 6 to PD-1, MAT 7 to PD-1, MAT 8 to PD-1, MAT 9 to PD-1, MAT 10 to PD-1, MAT 11 to PD-1, MAT 12 to PD-1, MAT 13 to PD-1, MAT 14 to PD-1 or MAT 15 to PD-1; a PD-1-binding fragment of any of said antibodies; or wherein the PD-1-binding molecule is an anti-PD-1 bispecific diabody (e.g., a PD-1xLAG-3 bispecific diabody). Included within the scope of the present invention are in particular molecules that comprise 3 CDR _L and 3 CDRH of MAT to PD-1 1, MAT 2 to PD-1, MAT 3 to PD-1, MAT 4 to PD-1, MAT 5 to PD-1, MAT 6 to PD-1, MAT 7 to PD-1, MAT 8 to PD-1, MAT 9 to PD-1, MAT 10 to PD-1, MAT 11 to PD-1, MAT 12 to PD-1, MAT 13 to PD-1, MAT 14 to PD-1 or MAT 15 to PD-1; humanized MAT to PD-1 1; MAT 2 to PD-1, MAT 3 to PD-1, MAT 4 to PD-1, MAT 5 to PD-1, MAT 6 to PD-1, MAT 7 to PD-1, MAT 8 to PD-1, MAT 9 to PD-1, MAT 10 to PD-1, MAT 11 to PD-1, MAT 12 to PD-1, MAT 13 to PD-1, MAT 14 to PD-1 or MAT 15 to PD-1. The scope of the present invention also includes such pharmaceutical compositions that further comprise a second therapeutic antibody (e.g., a tumor-specific monoclonal antibody) that is specific for a particular cancer antigen and a pharmaceutically acceptable carrier.

According to a specific embodiment of the present invention, the term pharmaceutically acceptable means approved by a regulatory agency of the federal or state government or listed in the United States Pharmacopeia or other generally recognized pharmacopeia for use in animals and, more particularly, in humans. The term carrier refers to a diluent, an adjuvant, such as Freund's adjuvant (complete and incomplete), an excipient, or a vehicle with which the drug compound is administered. Suitable pharmaceutical carriers can be sterile liquids such as water and oils, including petroleum oils, animal, vegetable, or synthetic oils such as peanut oil, soybean oil, mineral oil, sesame oil, and the like. Water is a preferred carrier when the pharmaceutical composition is administered intravenously. Physiological saline solutions and aqueous dextrose and glycerol solutions can also be used as liquid carriers, particularly for injection solutions. Suitable pharmaceutical excipients include starch, glucose, lactose, sucrose, gelatin, malt, rice, flour, chalk, silica gel, sodium stearate, glycerol monostearate, talc, sodium chloride, dry skim milk, glycerol, propylene, glycol, water, ethanol, etc. The composition, if necessary, can also contain minor amounts of wetting or emulsifying agents or buffering agents to maintain pH. Suitable compositions can be in the form of solutions, suspensions, emulsions, tablets, pills, capsules, powders, sustained-release compositions, etc.

Typically, the ingredients of the compositions of the present invention are supplied separately or mixed to form a unit dosage form, for example, as a dry lyophilized powder or a water-free concentrate, in a hermetically sealed container such as an ampoule or sachet indicating the amount of active agent. In the event that the composition is to be administered by infusion, it can be dispensed with an infusion bottle containing sterile pharmaceutical water or saline. In the event that the composition is to be administered by injection, an ampoule of sterile water for injection or saline can be provided so that the ingredients can be mixed before administration.

The compositions of the present invention may be prepared in neutral or salt forms. Pharmaceutically acceptable salts include, but are not limited to, salts formed with anions such as those derived from hydrochloric, phosphoric, acetic, oxalic, tartaric acids, etc., and those formed with cations such as those derived from sodium, potassium, ammonium, calcium, iron hydroxides, isopropylamine, triethylamine, 2-ethylaminoethanol, histidine, procaine, etc.

The present invention also provides a pharmaceutical package or kit comprising one or more containers filled with a PD-1 binding molecule according to the present invention (and more preferably an anti-PD-1 MAT 1, an anti-PD-1 MAT 2, an anti-PD-1 MAT 3, an anti-PD-1 MAT 4, an anti-PD-1 MAT 5, an anti-PD-1 MAT 6, an anti-PD-1 MAT 7, an anti-PD-1 MAT 8, an anti-PD-1 MAT 9, an anti-PD-1 MAT 10, an anti-PD-1 MAT 11, an anti-PD-1 MAT 12, an anti-PD-1 MAT 13, an anti-PD-1 MAT 14 or an anti-PD-1 MAT 15; humanized anti-PD-1 MAT 1, an anti-PD-1 MAT 2, an anti-PD-1 MAT 3, an anti-PD-1 MAT 4, an anti-PD-1 MAT 5 to PD-1, MAT 6 to PD-1, MAT 7 to PD-1, MAT 8 to PD-1, MAT 9 to PD-1, MAT 10 to PD-1, MAT 11 to PD-1, MAT 12 to PD-1, MAT 13 to PD-1, MAT 14 to PD1 or MAT 15 to PD-1, a PD-1-binding fragment of any of these antibodies, or a PD-1-binding molecule that is a bispecific diabody to PD-1 (e.g., a PD-1xLAG-3 bispecific diabody). Included within the scope of the present invention are, in particular, molecules

- 87 047726 kits that contain 3 CDRL and 3 CDRH1 from MAT to PD-1 1, MAT 2 to PD-1, MAT 3 to PD-1, MAT 4 to PD-1, MAT 5 to PD-1, MAT 6 to PD-1, MAT 7 to PD-1, MAT 8 to PD-1, MAT 9 to PD-1, MAT 10 to PD-1, MAT 11 to PD-1, MAT 12 to PD-1, MAT 13 to PD-1, MAT 14 to PD-1 or MAT 15 to PD-1, alone or with a suitable pharmaceutically acceptable carrier. In addition, one or more other prophylactic or therapeutic agents suitable for treating a disease can also be included in the pharmaceutical package or kit. The present invention also provides a pharmaceutical package or kit comprising one or more containers filled with one or more ingredients of the pharmaceutical compositions of the present invention. A notice in the form prescribed by a governmental agency regulating the manufacture, use or sale of pharmaceuticals or biological products, which reflects approval by said agency of manufacture, use or sale for administration to humans, may optionally be affixed to said container(s).

The present invention also provides kits that can be used in the above methods. The kit may comprise any of the PD-1 binding molecules of the present invention. The kit may further comprise one or more other prophylactic and/or therapeutic agents useful for treating cancer in one or more containers; and/or the kit may further comprise one or more cytotoxic antibodies that bind to one or more cancer antigens associated with cancer. In some embodiments of the present invention, the other prophylactic or therapeutic agent is a chemotherapeutic agent. In other embodiments of the present invention, the prophylactic or therapeutic agent is a biological or hormonal therapeutic agent.

XIII. Methods of administration.

The compositions of the present invention can be provided for the treatment, prevention and amelioration of one or more symptoms associated with a disease, disorder or infection by administering to a subject an effective amount of a fusion protein or conjugated molecule of the present invention, or a pharmaceutical composition comprising a fusion protein or conjugated molecule of the present invention. According to a preferred aspect, said compositions are substantially purified (i.e., substantially free of substances that limit their action or cause undesirable side effects). According to a specific embodiment of the present invention, the subject is an animal, preferably a mammal, such as a non-primate mammal (e.g., bovine, equine, feline, canine, rodent, etc.) or a primate (e.g., a monkey such as a cynomolgus macaque, a human, etc.). According to a preferred embodiment of the present invention, the subject is a human. Various delivery systems are known and can be used to administer the compositions of the present invention, such as encapsulation in liposomes, microparticles, microcapsules, recombinant cells capable of expressing the antibody or fusion protein, receptor-mediated endocystosis (see, for example, Wu et al. (1987) Receptor-Mediated in vitro Gene Transformation By A Soluble DNA Carrier System, J. Biol. Chem. 262:4429-4432), construction of the nucleic acid as part of a retroviral or other vector, etc.

Methods of administration of the molecule of the present invention include, but are not limited to, parenteral administration (e.g., intradermal, intramuscular, intraperitoneal, intravenous and subcutaneous), epidural and mucosal (e.g., intranasal and oral routes). According to a specific embodiment of the present invention, the PD-1 binding molecules of the present invention are administered intramuscularly, intravenously or subcutaneously. The compositions can be administered by any convenient route, such as by infusion or bolus injection, by absorption through epithelial or mucosal membranes (e.g., oral mucosa, rectal and intestinal mucosa, etc.) and can be administered together with other biologically active agents. Administration can be systemic or local. In addition, pulmonary administration can also be used, for example, using an inhaler or nebulizer and an aerosol agent formulation. See, for example, U.S. Patent Nos. 6,019,968; 5,985,320; 5,985,309; 5,934,272; 5,874,064; 5,855,913; 5,290,540; and 4,880,078; and PCT Publications WO 92/19244; WO 97/32572; WO 97/44013; WO 98/31346; and WO 99/66903, all of which are incorporated herein by reference in their entireties.

The present invention also provides PD-1-binding molecules of the present invention packaged in a hermetically sealed container, such as an ampoule or a sachet, indicating the amount of the molecule. According to one embodiment of the present invention, said molecules are supplied as a dry, sterile, lyophilized powder or concentrate that does not contain water in a hermetically sealed container and can be reconstituted, for example, using water or saline to an appropriate concentration for administration to a subject. Preferably, the PD-1-binding molecules of the present invention are supplied as a dry, sterile, lyophilized powder in a hermetically sealed container.

- 88 047726

The lyophilized PD-1 binding molecules of the present invention should be stored at a temperature of 2 to 8°C in their original container, and the molecules should be administered within 12 hours, preferably within 6 hours, within 5 hours, within 3 hours, or within 1 hour after reconstitution. In another embodiment, said molecules are supplied in liquid form in a hermetically sealed container indicating the amount and concentration of the molecule, fusion protein, or conjugated molecule.

Preferably, said PD-1 binding molecules, if provided in liquid form, are supplied in a hermetically sealed container.

The amount of the composition of the present invention that will be effective in treating, preventing or improving one or more symptoms associated with the disorder can be determined using standard clinical techniques. The exact dose to be used in the composition will also depend on the route of administration and the severity of the condition, and should be determined in accordance with the judgment of the practitioner and the circumstances of each patient. Effective doses can be extrapolated from dose response curves obtained from in vitro test systems or animal models.

In the present invention, the term "effective amount" of a pharmaceutical composition in one embodiment is an amount sufficient to obtain beneficial or desired results, including, but not limited to, clinical results such as a reduction in symptoms resulting from a disease, a reduction in a symptom of an infection (e.g., viral load, fever, pain, sepsis, etc.) or a symptom of cancer (e.g., cancer cell proliferation, tumor presence, tumor metastasis, etc.), thereby improving the quality of life of individuals suffering from said disease, a reduction in the dose of other drugs needed to treat the disease, an enhancement of the effect of another drug, such as by targeting and/or internalization, a slowing of the progression of the disease, and/or an extension of the survival period of individuals.

An effective amount can be administered by one or more administrations. For the purposes of the present invention, an effective amount of a drug, compound or pharmaceutical composition is an amount sufficient to reduce the proliferation (or effect) of the virus present and to reduce and/or slow the progression of the viral disease, directly or indirectly. According to some embodiments of the present invention, an effective amount of a drug, compound or pharmaceutical composition can optionally be achieved in combination with another drug, compound or pharmaceutical composition. Accordingly, an effective amount can be considered in relation to the administration of one or more chemotherapeutic agents, and one agent can be considered to be administered in an effective amount if, in combination with one or more other agents, the desired result is or can be achieved.

Although individual needs vary, determination of optimal ranges of effective amounts of each component is within the skill of the art.

For PD-1 binding molecules included within the scope of the present invention, the dosage administered to a patient is preferably determined based on the body weight (kg) of the recipient subject. For PD-1 binding molecules included within the scope of the present invention, the dosage administered to a patient is typically at least about 0.01 μg/kg, at least about 0.05 μg/kg, at least about 0.1 μg/kg, at least about 0.2 μg/kg, at least about 0.5 μg/kg, at least about 1 μg/kg, at least about 2 μg/kg, at least about 5 μg/kg, at least about 10 μg/kg, at least about 20 μg/kg, at least about 50 μg/kg, at least about 0.1 mg/kg, at least about 1 mg/kg, at least about 3 mg/kg, at least about 5 mg/kg, at least about 10 mg/kg, at least about 30 mg/kg, at least about at least about 50 mg/kg, at least about 75 mg/kg, at least about 100 mg/kg, at least about 125 mg/kg, at least about 150 mg/kg or more of the subject's body weight.

The dosage and frequency of administration of the PD-1 binding molecule of the present invention can be reduced or altered by enhancing the uptake and penetration of the molecule into tissues through modifications such as, for example, lipidation.

The dosage of the PD-1 binding molecule according to the present invention administered to a patient can be calculated for use as a monotherapy. In another embodiment, the molecule can be used in combination with other therapeutic compositions, and the dosage administered to the patient is lower than when these molecules are administered as a monotherapy.

The pharmaceutical compositions of the present invention may be administered locally to the area in need of treatment; such administration may be achieved, for example, but not limited to,

- 89 047726 by them, by local infusion, by injection or by means of an implant, wherein said implant is a porous, non-porous or gelatinous material, including membranes such as silastic membranes or fibres. Preferably, when administering the molecule according to the present invention, care should be taken to use materials that do not absorb the molecule.

The compositions of the present invention may be delivered in a vesicle, in particular a liposome (see Langer (1990) New Methods Of Drug Delivery, Science 249:1527-1533); Treat et al., in Liposomes in the Therapy of Infectious Disease and Cancer, Lopez-Berestein and Fidler (eds.), Liss, New York, pp. 353-365 (1989); Lopez-Berestein, ibid., pp. 317-327).

The compositions of the present invention can be delivered in a controlled release system or a sustained release system. Any technique known to one of skill in the art can be used to prepare sustained release formulations containing one or more PD-1 binding molecules of the present invention. See, for example, U.S. Patent No. 4,526,938; PCT Publication No. WO 91/05548; PCT Publication No. WO 96/20698; Ning et al. (1996) Intratumoral Radioimmunotheraphy Of A Human Colon Cancer Xenograft Using A Sustained-Release Gel, Radiotherapy & Oncology 39:179-189, Song et al. (1995) Antibody Mediated Lung Targeting Of Long-Circulating Emulsions, PDA Journal of Pharmaceutical Science & Technology 50:372-397; Cleek et al. (1997) Biodegradable Polymeric Carriers For A bFGF Antibody For Cardiovascular Application, Pro. Int'l. Symp. Control. Rel. Bioact. Mater. 24:853-854; and Lam et al. (1997) Microencapsulation Of Recombinant Humanized Monoclonal Antibody For Local Delivery, Proc. Int'l. Symp. Control Rel. Bioact. Mater. 24:759-760, all of which are incorporated herein by reference in their entireties. According to one embodiment of the present invention, a pump may be used in the controlled release system (see Langer, supra; Sefton, (1987) Implantable Pumps, CRC Crit. Rev. Biomed. Eng. 14:201-240; Buchwald et al. (1980) Long-Term, Continuous Intravenous Heparin Administration By An Implantable Infusion Pump In Ambulatory Patients With Recurrent Venous Thrombosis, Surgery 88:507-516; and Saudek et al. (1989) A Preliminary Trial Of The Programmable Implantable Medication System For Insulin Delivery, N. Engl. J. Med. 321:574-579). According to another embodiment of the present invention, polymeric materials can be used to achieve controlled release of molecules (see, e.g., Medical applications of controlled release, Langer and Wise (eds.), CRC Pres., Boca Raton, Florida (1974); Controlled Drug Bioavailability, Drug Product Design and Performance, Smolen and Ball (eds.), Wiley, New York (1984); Levy et al. (1985) Inhibition Of Calcification Of Bioprosthetic Heart Valves By Local Controlled-Release Diphosphonate, Science 228:190-192; During et al. (1989) Controlled Release Of Dopamine From A Polymeric Brain Implant: in vivo Characterization, Ann. Neurol. 25:351-356; Howard et al. (1989) Intracerebral Drug Delivery In Rats With Lesion-Induced Memory Deficits, J. Neurosurg. 7(1): 105-112); U.S. Patent No. 5,679,377; U.S. Patent No. 5,916,597; U.S. Patent No. 5,912,015; U.S. Patent No. 5,989,463; U.S. Patent No. 5,123,232; PCT Publication WO 99/15154; and PCT Publication WO 99/20253). Examples of polymers used in sustained release preparations include, but are not limited to, poly-2-hydroxyethyl methacrylate, polymethyl methacrylate, polyacrylic acid, ethylene-vinyl acetate copolymer, polymethacrylic acid, polyglycolides (PLG), polyanhydrides, poly-N-vinylpyrrolidone, polyvinyl alcohol, polyacrylamide, polyethylene glycol, polylactides (PLA), poly(lactic-glycolic acid) copolymer (PLGA), and polyorthoesters. The controlled release system can be located near the therapeutic target (e.g., the lungs), resulting in only a fraction of the systemic dose being required (see, e.g., Goodson, in Medical Applications of Controlled Release, supra, vol. 2, pp. 115-138 (1984)). Polymer compositions that can be used as controlled release implants can be used according to Dunn et al. (See U.S. Patent No. 5,945,155). This particular method is based on the therapeutic effect of controlled release of a bioactive material from a polymeric system under in situ conditions. Implantation can typically occur at any location in the body of a patient requiring therapeutic treatment. A non-polymeric sustained delivery system can be used, in which a non-polymeric implant in the body of a subject is used as a drug delivery system. When implanted in the body, the organic solvent of the implant will be distributed, dispersed or leached from the composition into the surrounding tissue fluid, and the non-polymeric material will gradually coagulate or precipitate to form a solid microporous matrix (See U.S. Patent No. 5,888,533).

Controlled release systems are discussed in the review by Langer (1990, New Methods Of Drug Delivery, Science 249:1527-1533). Any technique known to one skilled in the art can be used to prepare sustained release formulations containing one or more therapeutic agents of the present invention. See, for example, U.S. Patent No. 4,526,938; International Publications

- 90 047726

WO 91/05548 and WO 96/20698; Ning et al. (1996) "Intratumoral Radioimmunotheraphy Of A Human Colon Cancer Xenograft Using A Sustained-Release Gel", Radiotherapy & Oncology 39:179-189, Song et al. (1995) "Antibody Mediated Lung Targeting Of Long-Circulating Emulsions", PDA Journal of Pharmaceutical Science & Technology 50:372-397; Cleek et al. (1997) "Biodegradable Polymeric Carriers For A bFGF Antibody For Cardiovascular Application", Pro. Int'l. Symp. Control. Rei. Bioact. Mater. 24:853-854; and Lam et al. (1997) "Microencapsulation Of Recombinant Humanized Monoclonal Antibody For Local Delivery", Proc. Int'l. Symp. Control Rei. Bioact. Mater.

24:759-760, all of which are incorporated herein by reference in their entireties.

If the composition of the present invention is a nucleic acid encoding a PD-1 binding molecule of the present invention, the nucleic acid can be administered in vivo to stimulate expression of the PD-1 binding molecule it encodes by constructing it as part of an appropriate nucleic acid expression vector and administering it so that the vector becomes intracellular, such as by using a retroviral vector (see U.S. Patent No. 4,980,286) or by direct injection or by using microparticle bombardment (e.g., gene gun; Biolistic, Dupont) or by coating with lipids or cell surface receptors or transfection agents, or by administering the vector linked to a homeobox-like peptide known to penetrate the nucleus (see, e.g., Joliot et al. (1991) Antennapedia Homeobox Peptide Regulates Neural Morphogenesis, Proc. Natl. Acad. Sci. (U.S.A.) 88:1864-1868), etc. In another embodiment, the nucleic acid can be administered intracellularly and integrated into the host cell DNA for expression by homologous recombination. Treating a subject with a therapeutically or prophylactically effective amount of a PD-1 binding molecule according to the present invention can comprise a single treatment or, preferably, can comprise a series of treatments. In a preferred example, the subject is treated with said diabody once a week for 1 to 10 weeks, preferably 2 to 8 weeks, more preferably 3 to 7 weeks, and even more preferably for about 4, 5, or 6 weeks. The pharmaceutical compositions according to the present invention can be administered once a day, twice a day, or three times a day. In another embodiment, the pharmaceutical compositions may be administered once a week, twice a week, once every two weeks, once a month, once every six weeks, once every two months, twice a year, or once a year. It will also be understood that the effective dosage of the molecules used for treatment may increase or decrease over the course of a particular treatment.

Examples

The following examples illustrate various methods for the compositions in diagnostic or therapeutic methods according to the present invention. The examples are provided for illustrative purposes and in no way limit the scope of the present invention.

Example 1. Characterization of monoclonal antibodies to human PD-1.

Fifteen murine monoclonal antibodies capable of specifically binding to both human and cynomolgus macaque PD-1 were isolated and designated anti-PD-1 MAb 1, anti-PD-1 MAb 2, anti-PD-1 MAb 3, anti-PD-1 MAb 4, anti-PD-1 MAb 5, anti-PD-1 MAb 6, anti-PD-1 MAb 7, anti-PD-1 MAb 8, anti-PD-1 MAb 9, anti-PD-1 MAb 10, anti-PD-1 MAb 11, anti-PD-1 MAb 12, anti-PD-1 MAb 13, anti-PD-1 MAb 14, and anti-PD-1 MAb 15. The CDRs of these antibodies were found to differ and are shown above. Binding to the extracellular domain of human and cynomolgus monkey PD-1 was assessed as follows: 96-well flat-bottomed Maxi-Sorb™ plates were coated with soluble human or cynomolgus monkey PD-1 (human PD-1 extracellular domain fused to a polyhistidine tag (shPD-1 His) or human Fc region (shPD-1 hFc), or cynomolgus monkey PD-1 extracellular domain fused to human Fc region (scyno-PD1 Fc)), each at a concentration of 0.5 or 1 μg/mL, the plates were washed and incubated with one of the isolated anti-PD-1 antibodies, PD-1 mAbs 1-15. For these studies, anti-PD-1 antibodies were used at concentrations of 3, 1.0, 0.3333, 0.1111, 0.0370, 0.0123, or 0.0041 μg/mL (three-fold serial dilutions). The amount of antibody binding to immobilized PD-1 (human or cynomolgus monkey) was estimated using horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG secondary antibody. All samples were run on a plate reader (Victor 2 Wallac, Perkin Elmers). Representative binding curves for soluble human PD-1 and soluble cynomolgus monkey PD-1 are shown in Figs. 7A–7D and Figs. 8A–8C, respectively. The results of the binding studies (Figs. 7A-7D and Figs. 8A-8C) indicate that all anti-PD-1 antibodies, anti-PD-1 MAbs 1-15, bind to both human soluble PD-1 and cynomolgus macaque soluble PD-1.

To further characterize the murine anti-PD-1 antibodies, their ability to block the binding of soluble human PD-L1 to soluble human PD-1 was assessed in two different assays.

- 91,047,726 quantitative studies. One quantitative study assessed the ability of antibodies to block binding of human PD-1 to surface-immobilized PD-L1. For this quantitative study, each of the anti-PD-1 antibodies, anti-PD-1 1-15 mAbs, or anti-PD-1 control antibody (anti-PD-1 A mAb) was mixed with shPD-1 His fusion protein (at a concentration of 2.5 μg/mL) and separately incubated with biotin-labeled soluble human PD-L1 (the extracellular domain of PD-L1 fused to human Fc (sPD-L1)), at a concentration of 1 μg/mL, immobilized on a streptavidin-coated plate. For these studies, anti-PD-1 antibodies were used at concentrations of 10, 5.0, 2.5, 1.25, 0.625, 0.3125, or 0.1563 μg/mL (two-fold serial dilutions). The amount of shPD-1 His bound to immobilized sPD-L1 was estimated by His-tag using an HRP-conjugated anti-His-tag secondary antibody. All samples were run on a plate reader (Victor 2 Wallac, Perkin Elmers). The results of this experiment are shown in Figs. 9A–9D.

The results of the quantitative inhibition studies (Figs. 9A-9D) indicate that the anti-PD-1 antibodies, anti-PD-1 MAb 1, anti-PD-1 MAb 3, anti-PD-1 MAb 5, anti-PD-1 MAb 6, anti-PD-1 MAb 7, anti-PD-1 MAb 8, anti-PD-1 MAb 9, anti-PD-1 MAb 10, anti-PD-1 MAb 11, anti-PD-1 MAb 12, anti-PD-1 MAb 13, anti-PD-1 MAb 14, and anti-PD-1 MAb 15, were able to block the binding of soluble human PD-L1 to soluble human PD-1 to varying degrees, while anti-PD-1 MAb 2 and anti-PD-1 MAb 4 exhibited little or no blocking activity in this quantitative assay format.

In a second quantitative study, we assessed the ability of murine anti-PD-1 antibodies, PD-1 mAbs 1-15, to block the binding of PD-1 ligand (i.e., human PD-L1 or human PD-L2) to PD-1 expressed on the surface of NSO cells. To perform this quantitative assay, each anti-PD-1 antibody, anti-PD-1 mAb 1-15, or anti-PD-1 control antibody (anti-PD-1 mAb A or anti-PD-1 mAb B) was separately mixed with biotinylated soluble human PD-L1 (shPD-L1 fusion protein) or biotinylated soluble human PD-L2-muIgFc fusion protein (shPD-L2; Ancell, catalog #573-030), each at a concentration of 0.1 μg/assay, and incubated with human PD-1-expressing NSO cells (~250,000 cells/well) in blocking buffer (FACS+10% human serum albumin). For these studies, anti-PD-1 antibodies were used at concentrations of 4.0, 1.0, 2.5x10 ^-1 , 6.25x10 ^-2 , 1.56x10 ^-2 , 3.90x10 ^-3 , 9.76x10 ^-4 , 2.4x10-4, 0.6x10'4 μg/assay (fourfold serial dilutions). The amount of shPDL1 (or shPD-L2) bound to the surface of NSO cells was determined using a phycoerythrin-conjugated streptavidin-coupled secondary antibody by FACS. IC ₅₀ values for inhibition of PD-1/PD-L1 binding were determined, and the mean value (*) of at least two experiments (except where indicated) is presented in Table 6.

Table 6

Antibody to PD1 IC50 (mcg/study) Antibody to PD1 IC50 (mcg/study) MATkPD-Ι A 0,0044 MAT 8 kPD-1 0,6611$ MATkPD-Ι in 0,0064 MAT 9 kPD-1 0,0154 MATkPD-1 1 0,0048 MAT 10kPD-1 0,0057 MAT 2 to PD-1 0,0110 MAT and kPD-1 0,0259$ MAT 3 to PD-1 0,0361$ MAT 12 kPD-1 0,0238$ MAT 4 to PD-1 0,0156$ MAT 13 kPD-1 0,0117 MAT 5 to PD-1 0,0039 MAT 14 kPD-1 0,0149$ MAT 6 to PD-1 0,0051 MAT 15 kPD-1 0,0060 MAT 7 to PD-1 0,0024

Ϊ Results of one experiment.

The results of quantitative shPD-L1 inhibition studies (Table 6) indicate that anti-PD-1 antibodies, anti-PD-1 MAT 1, anti-PD-1 MAT 2, anti-PD-1 MAT 3, anti-PD-1 MAT 4, anti-PD-1 MAT 5, anti-PD-1 MAT 6, anti-PD-1 MAT 7, anti-PD-1 MAT 9, anti-PD-1 MAT 10, anti-PD-1 MAT 11, anti-PD-1 MAT 12, anti-PD-1 MAT 13, anti-PD-1 MAT 14, and anti-PD-1 MAT 15, were able to block the binding of human PD-L1 to human PD-1 expressed on the surface of NSO cells. Specifically, anti-PD-1 mAb 1, anti-PD-1 mAb 5, anti-PD-1 mAb 6, anti-PD-1 mAb 7, anti-PD-1 mAb 10, and anti-PD-1 mAb 15 blocked shPD-L1 binding similarly to or more efficiently than the reference anti-PD-1 antibodies (anti-PD-1 mAb A, anti-PD-1 mAb B), whereas anti-PD-1 mAb 8 had essentially no blocking effect in this assay format. Both anti-PD-1 mAb 2 and anti-PD-1 mAb 4 were able to block PD-1/PD-L1 binding in this assay format.

- 92 047726

Similarly, anti-PD-1 antibodies, anti-PD-1 MAb 1, anti-PD-1 MAb 2, anti-PD-1 MAb 3, anti-PD1 MAb 4, anti-PD-1 MAb 5, anti-PD-1 MAb 6, anti-PD-1 MAb 7, anti-PD-1 MAb 9, anti-PD-1 MAb 10, anti-PD-1 MAb 12, anti-PD-1 MAb 13, and anti-PD-1 MAb 14, were able to block the binding of human PD-L2 to human PD-1 expressed on the surface of NSO cells, whereas anti-PD-1 MAb 8 had essentially no blocking effect in this quantitative assay format. Specifically, anti-PD-1 mAb 1, anti-PD-1 mAb 5, anti-PD-1 mAb 6, anti-PD-1 mAb 7, and anti-PD-1 mAb 10 blocked shPD-L2 binding similarly to or more efficiently than the reference anti-PD-1 antibodies (anti-PD-1 mAb A, anti-PD-1 mAb B). The anti-PD-1 antibodies anti-PD-1 mAb 11 and anti-PD-1 mAb 15 were not tested in this quantitative study. Results for several humanized anti-PD-1 antibodies, including anti-hPD-115, are provided below.

Example 2. Humanization and further characterization.

The variable domains of the anti-PD-1 antibodies, anti-PD-1 MAT 1, anti-PD-1 MAT 2, anti-PD-1 MAT 7, anti-PD-1 MAT 9, and anti-PD-1 MAT 15, were humanized and, where antigenic epitopes were identified, the antibodies were further deimmunized to generate final humanized variable domains. Humanization of anti-PD-1 MAT 1, anti-PD-1 MAT 2, and anti-PD-1 MAT 15 yielded one humanized VH domain and one humanized VL domain for each antibody, designated herein as VH1 anti-hPD-1 MAT 1 and VL1 anti-hPD-1 MAT 1; VH1 anti-hPD-1 MAT 2 and VL1 anti-hPD-1 MAT 2; and VH1 MAT to hPD-1 15 and VL1 MAT to hPD-1 15. Humanization of MAT 7 to PD-1 resulted in two humanized VH domains, designated herein as VH1 MAT to hPD-1 7 and VH2 MAT to hPD-1 7, and three humanized VL domains, designated herein as VL1 MAT to hPD-1 1, VL2 MAT to hPD-1 7, and VL3 MAT to hPD-1 7. Humanization of MAT 9 to PD-1 resulted in two humanized VH domains, designated herein as VH1 MAT to hPD-1 9 and VH2 MAT to hPD-1 9, and two humanized VL domains, designated herein as VL1 MAT to hPD-1 9 and VL2 MAT to hPD-1 1. In cases where there was several humanized variable domains are obtained, the humanized variable domains of the heavy and light chains of a particular anti-PD-1 antibody (e.g., anti-PD-1 Mab 7) can be used in any combination, and particular combinations of humanized chains are referred to with reference to specific VH/VL domains, for example, a humanized antibody comprising VH1 of an anti-hPD-1 Mab 7 and VL2 of an anti-hPD-1 Mab 7 is specifically referred to as an anti-hPD-1 Mab 7 (1.2). Humanized full-length antibodies were obtained using the human IgG1 constant region containing the L234A/L235A substitutions (IgG1 (AA)), or the human IgG4 constant region containing the S228P substitution (IgG4 (P)). Full-length humanized IgG1 heavy chains were constructed as follows: The C-terminus of the humanized VH domain was fused to the N-terminus of a human IgG1 constant region containing a variant CH2-CH3 domain (containing L234A/L235A (AA) substitutions) and lacking the C-terminal lysine residue (SEQ ID NO: 255):

ASTKGPSVFP LAPSSKSTSG GTAALGCLVK DYFPEPVTVS WNSGALTSGV

HTFPAVLQSS GLYSLSSWT VPSSSLGTQT YICNVNHKPS NTKVDKRVEP KSCDKTHTCP PCPAPEAAGG PSVFLFPPKP KDTLMISRTP EVTCVWDVS HEDPEVKFNW YVDGVEVHNA KTKPREEQYN STYRWSVLT VLHQDWLNGK EYKCKVSNKA LPAPIEKTIS KAKG QPREPQ VYTLPPSREE MTKNQVSLTC LVKGFYPSDI AVEWESNGQP ENNYKTTPPV LDSDGSFFLY SKLTVDKSRW QQGNVFSCSV MHEALHNHYT QKSLSLSPG

In SEQ ID NO: 255, amino acid residues 1-98 correspond to the CH1 domain of IgG1 (SEQ ID NO: 10), amino acid residues 99-113 correspond to the hinge region of IgG1 (SEQ ID NO: 32), and amino acid residues 114-329 correspond to the CH2-CH3 domain of IgG1 containing the L234A/L235A substitutions (underlined) (SEQ ID NO: 5) but lacking the C-terminal lysine residue.

The amino acid sequence of the heavy chain of an exemplary humanized antibody ((anti-hPD-1 mAb 7 (1.2)) comprising an IgG1 heavy chain constant region containing the L234A/L235A mutation and lacking the C-terminal lysine residue is as follows (SEQ ID NO: 265):

QVQLVQSGAE VKKPGASVKV SCKASGYSFT SYWMNWVRQA PGQGLEWIGV IHPSDSETWL DQKFKDRVTI TVDKSTSTAY MELSSLRSED TAVYYCAREH YGTSPFAYWG QGTLVTVSSA STKGPSVFPL APSSKSTSGG TAALGCLVKD YFPEPVTVSW NSGALTSGVH TFPAVLQSSG LY SLSSWTV PSSSLGTQTY ICNVNHKPSN TKVDKRVEPK SCDKTHTCPP CPAPEAAGGP SVFLFPPKPK DTLMISRTPE VTCVWDVSH EDPEVKFNWY VDGVEVHNAK TKPREEQYNS TYRWSVLTV LHQDWLNGKE YKCKVSNKAL PAPIEKTISK AKGQPREPQV YTLPPSREEM TKNQVSLTCL VKGFYPSDIA VEWESNGQPE NNYKTTPPVL DSDGSFFLYS KLTVDKSRWQ QGNVFSCSVM HEALHNHYTQ KSLSLSPG

- 93 047726

In SEQ ID NO: 265, amino acid residues 1-119 correspond to the VH domain of the VH1 mAb to hPD-1 7 (SEQ ID NO: 147), amino acid residues 120-217 correspond to the CH1 domain of IgG1 (SEQ ID NO: 10), amino acid residues 218-232 correspond to the hinge region of IgG1 (SEQ ID NO: 32), and amino acid residues 233-448 correspond to the CH2-CH3 domain of IgG1 containing the L234A/L235A substitutions (underlined) (SEQ ID NO: 5) and lacking the C-terminal lysine residue.

Full-length humanized IgG4 antibody heavy chains were constructed as follows: The C-terminus of the humanized VH domain was fused to the N-terminus of a human IgG4 constant region containing a stabilized hinge region (containing the S228P substitution) and lacking the C-terminal lysine residue (SEQ ID NO: 256):

ASTKGPSVFP LAPCSRSTSE STAALGCLVK DYFPEPVTVS WNSGALTSGV HTFPAVLQSS GLYSLSSWT VPSSSLGTKT YTCNVDHKPS NTKVDKRVES KYGPPCPPCP APEFLGGPSV FLFPPKPKDT LMISRTPEVT CVWDVSQED PEVQFNWYVD GVEVHNAKTK PREEQFNSTY LTVLH QDWLNGKEYK CKVSNKGLPS SIEKTISKAK GQPREPQVYT LPPSQEEMTK NQVSLTCLVK GFYPSDIAVE WESNGQPENN YKTTPPVLDS DGSFFLYSRL TVDKSRWQEG NVFSCSVMHE ALHNHYTQKS LSLSLG

In SEQ ID NO: 256, amino acid residues 1-98 correspond to the CH1 domain of IgG4 (SEQ ID NO: 254), amino acid residues 99-110 correspond to the stabilized hinge region of IgG4 containing S228P substitutions (underlined) (SEQ ID NO: 13), and amino acid residues 111-326 correspond to the CH2-CH3 domain of IgG4 (SEQ ID NO: 4) lacking the C-terminal lysine residue.

The amino acid sequence of the heavy chain of an exemplary humanized antibody ((anti-hPD-1 mAb 7 (1.2)) comprising an IgG4 heavy chain constant region comprising a stabilized hinge region with the S228P mutation and lacking the C-terminal lysine residue is as follows (SEQ ID NO: 266):

QVQLVQSGAE VKKPGASVKV SCKASGYSFT SYWMNWVRQA PGQGLEWIGV IHPSDSETWL DQKFKDRVTI TVDKSTSTAY MELSSLRSED TAVYYCAREH YGTSPFAYWG QGTLVTVSSA STKGPSVFPL APCSRSTSES TAALGCLVKD YFPEPVTVSW NSGALTSGVH TFPAVLQSSG LY SLSSWTV PSSSLGTKTY TCNVDHKPSN TKVDKRVESK YGPPCPPCPA PEFLGGPSVF LFPPKPKDTL MISRTPEVTC VWDVSQEDP EVQFNWYVDG VEVHNAKTKP REEQFNSTYR WSVLTVLHQ DWLNGKEYKC KVSNKGLPSS IEKTISKAKG QPREPQVYTL PPSQEEMTKN QVSLTCLVKG FYPSDIAVEW ESNGQPENNY KTTPPVLDSD GSFFLYSRLT VDKSRWQEGN VFSCSVMHEA LHNHYTQKSL SLSLG

In SEQ ID NO: 266, amino acid residues 1-119 correspond to the VH domain of the VH1 mAb to hPD-1 7 (SEQ ID NO: 147), amino acid residues 120-217 correspond to the CH1 domain of IgG4 (SEQ ID NO: 254), amino acid residues 218-229 correspond to the stabilized hinge region of IgG4 containing the S228P substitution (underlined) (SEQ ID NO: 13), and amino acid residues 230-445 correspond to the CH2-CH3 domain of IgG4 (SEQ ID NO: 4) lacking the C-terminal lysine residue.

Full-length humanized light chains were constructed as follows: The C-terminus of the humanized VL domain was fused to the N-terminus of the human kappa light chain (SEQ ID NO: 8). The same light chain was paired with the IgG1 (AA) and IgG4 (P) heavy chains.

The amino acid sequence of the light chain of an exemplary humanized anti-PD-1 antibody (anti-hPD-1 mAb 7(1.2)) containing the kappa chain constant region is as follows (SEQ ID NO:264):

EIVLTQSPAT LSLSPGERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL

LIHAASNQGS GVPSRFSGSG SGTDFTLTIS SLEPEDFAVY FCQQSKEVPY

TFGGGTKVEI KRTVAAPSVF IFPPSDEQLK SGTASWCLL NNFYPREAKV

QWKVDNALQS GNSQESVTEQ DSKDSTYSLS STLTLSKADY EKHKVYACEV

THQGLSSPVTSKSFNRGEC

In SEQ ID NO: 264, amino acid residues 1-111 correspond to the VL domain of the VL2 MAT to hPD-1 7 (SEQ ID NO: 151), and amino acid residues 112-218 correspond to the constant region of the kappa light chain (SEQ ID NO: 8).

Anti-PD-1 antibodies containing alternative constant regions, such as modified Fc regions, can be readily prepared by incorporating different constant regions and/or introducing one or more amino acid substitutions, additions or deletions. For example, if a bispecific antibody is desired, the CH2-CH3 knob and hole domains are used to facilitate heterodimerization. Chimeric anti-PD-1 antibodies containing murine

- 94 047 726 human variable domains and constant regions were prepared as described above. Humanized antibodies (IgG1 (AA) and/or IgG4 (P)) were tested for binding and blocking activity as described above. Binding to human PD-1 (shPD-1 His and shPD-1 hFc) and cynomolgus monkey PD-1 (shPD-L1 hFc) of the humanized antibodies was comparable to that of the corresponding mouse antibody. In addition, the humanized antibodies retained the ability to block human PD-L1 binding to human PD-1 in an ELISA assay. The binding kinetics of mouse anti-PD-1 antibodies mAb 2, anti-PD-1 mAb 7, anti-PD-1 mAb 9, anti-PD-1 mAb 15, humanized anti-hPD-1 mAb 2, anti-hPD-1 mAb 7 (1.2), anti-hPD-1 mAb 9 (1.1), anti-hPD-1 mAb 15, and anti-PD-1 control antibodies mAb PD-1 A and mAb PD-1 B were studied using Biacore analysis. Anti-PD-1 antibodies were captured on immobilized protein A and incubated with His-tagged soluble human PD-1 (shPD-1-His) or soluble human-cynomolgus monkey PD-1 Fc fusion protein (scyno PD-1 hFc) cleaved to remove the Fc region, and the binding kinetics were determined by Biacore analysis. In additional studies, anti-PD-1 antibodies, anti-hPD-1 mAb 7(1.2) IgG1 (AA), anti-hPD-1 mAb 7(1.2) IgG4 (P), anti-hPD-1 mAb 9(1.1) IgG1 (AA), anti-hPD-1 mAb 9(1.1) IgG4 (P), anti-PD-1 mAb A IgG1 (AA), anti-PD-1 mAb A IgG4 (P), anti-PD-1 mAb B IgG1 (AA), and anti-PD-1 mAb B IgG4 (P), were captured onto immobilized goat anti-human F(ab)2 and binding kinetics were determined by Biacore analysis as described above. The calculated _ka , kd , and K _D values from these studies are presented in Table 7.

Table 7

Protein A capture Antibody to PD-1 Man ^a Cynomolgus macaque ¹⁵ to _a (x10 ⁴ ) k _d (xl0' ⁴ ) KD (nM) to _a (x10 ⁴ ) k _d (xl0' ⁴ ) KD (nM) MATkPD-Ι A 60 18 3 14 9.6 6.9 MATkPD-Ι in 140 35 2,5 37 12 3.2 MAT7kPD-1 21 2.8 1,3 17 6 3.5 MAT to chROM- 1 7(1.2) BY 4.3 0.39 37 6.4 1,7 MAT9kPD-1 4.3 4.2 9.8 2,2 16 72.7 MAT to 4PD19(1.1) 1,8 6.5 36.1 1,5 11 73.3 MAT 15 kPD-1 4,5 1,3 2.9 2.7 11 40.7 MATkhPD-I 15 2,4 3.2 13.3 2,3 18 78.3 MAT 2 to PD-1 5.5 5,6 10.2 4.2 6.0 14.3 MAT to hPD-1 2 3.2 1,6 5.0 2,3 3.9 17 Capture of goat F(ab) ₂ fragment to human Fc MAT to PD-1 A IgGl (AA) 13 8.4 6.5 8.1 4,5 5,6 MAT kPD-Ι A IgG4 (P) 13 7.9 6.1 8.4 5.0 6.0

- 95 047726

MATkPD-Ι in IgGl (AA) 25 28 And,2 20 6.4 3.2 MATkPD-Ι B IgG4 (P) 26 25 9.6 20 7.9 4.0 MAT to 4PD17(1.2) IgGl (AA) 25 3.8 1,5 16 7,8 4.9 MAT to 4PD- 1 7(1.2) IgG4 (P) 27 4.1 1,5 17 7,8 4.6 MAT to hPD1 9(1.1) IgGl (AA) 5,6 6.1 10.9 5,6 5.2 9.3 MAT to 4PD- 1 9(1.1) IgG4(P) 6.1 5.8 9.5 4.9 7.4 15.1

^a Polyhistidine tagged soluble human PD-1 (shPD-1 His).

^b Cleaved soluble cynomolgus monkey PD-1 (scyno PD-1 hFc).

The results indicate that anti-PD-1 mAb 7 and humanized anti-hPD-1 mAb 7 (1.2) exhibit superior binding kinetics compared to the anti-PD-1 reference antibodies, anti-PD-1 mAb A and anti-PD-1 mAb B. The binding kinetics of anti-PD-1 mAb 2 and anti-hPD-1 mAb 2 are approximately two-fold higher than those of the anti-PD-1 reference antibodies, while the binding kinetics of anti-PD-1 mAb 9, anti-hPD-1 mAb 9 (1.1), anti-PD-1 mAb 15 and anti-hPD-1 mAb 15 are approximately 2-6-fold higher than those of the anti-PD-1 reference antibodies.

The tissue specificity of the anti-human PD-1 antibody, anti-hPD-1 mAb 7, was studied. Normal tissue was exposed to anti-PD-1 mAb 7 or isotype control antibody (0.313 μg/mL) and the extent of staining was visualized. Bloxall™ was used to block endogenous enzymes to reduce non-specific mucin staining in colon tissue. As shown in Fig. 10A, panels i-xii, anti-PD-1 mAb 7 and isotype control antibody were unable to label cells from normal colon, liver, lung, pancreas, kidney, and heart tissue. Additionally, anti-PD-1 mAb 7 and isotype control antibody were unable to stain normal skin (Fig. 10B, panels i-ii). In contrast, anti-PD-1 mAb 7 was found to intensely stain lymphocytes present in normal tonsil tissue and PDCD1-transfected PD-1-expressing NSO cells (Fig. 10B, panels iii and v), whereas the isotype control antibody was unable to stain either cell type (Fig. 10B, panels iv and vi). Accordingly, the results shown in Figs. 10A–10B indicate that anti-PD-1 mAb 7 is able to specifically bind to lymphocytes and PD-1-expressing cells.

The saturation binding profiles of anti-hPD-1 2 IgG1 mAb (AA), anti-hPD-1 7(1.1) IgG1 mAb (AA), anti-hPD-1 7(1.2) IgG1, (AA), anti-hPD-1 7(1.2) IgG4 mAb (P), anti-hPD-1 9(1.1) IgG1 mAb (AA), anti-hPD-1 9(1.1) IgG4 mAb (P), anti-hPD-1 15 IgG1 mAb (AA), and the anti-PD-1 reference antibodies anti-PD-1 MAB A and anti-PD-1 MAB B were studied. Briefly, each of the anti-PD-1 antibodies, anti-PD-1 1-15 mAbs, or anti-PD-1 reference antibodies (anti-PD-1 MAB A and anti-PD-1 MAB B) was mixed with NSO cells, expressing human PD-1 (~250,000 cells/well) in blocking buffer (FACS + 10% human serum albumin). For these studies, anti- ^PD -1 antibodies were used at a concentration of 50, 12.5, 3.13 ^, 2.0x10-4 ^, 4.9x10-2, 1.2x10-2, 3.0x10-3, 1.9x10-4, 7.6x10-4, 4.75x10-5, or ^1.19x10-5 μg/assay (fourfold serial dilutions). The amount of antibody bound to the NSO cell surface was determined using APC-conjugated goat anti-human Ig secondary antibody by FACS. Representative saturation curves are shown in Fig. 11. The EC ₅₀ and EC ₉₀ values were determined, and the sample mean (SM) and standard deviation (SD, σ) of four individual experiments are given in Table 8.

- 96 047726

Table 8

Saturation of binding EC ₅ o (mcg/study) ECeo (mcg/study) Antibody to PD-1 SM SD σ SM SD σ MAT to PD-1 A IgGl (AA) 0,1991 0,1309 1,4528 0,8040 MAT to PD-1 A IgG4 (P) 0,1581 0,1161 1,5464 1,7690 MAT to PD-1 B IgGl (AA) 0,1347 0,0681 1,3917 0.9573 MAT to PD-1 B IgG4 (P) 0,1398 0,0951 1,1619 1,2681 MAT to hPD-1 2 IgGl (AA) 0,4431 0,1997 2,4374 1,2637 MAT to hPD-1 7(1.1) IgGl (AA) 0,1069 0,0500 0,9102 0,5476 MAT to hPD-1 7(1.2) IgGl (AA) 0,1872 0,1553 0,6810 0.3226 MAT to hPD-1 7(1.2) IgG4 (P) 0,1376 0,0926 0,6609 0.3437 MAT to hPD-1 9(1.1) IgGl (AA) 0.3123 0.2291 1,6486 0,9117 MAT to hPD-1 9(1.1) IgG4 (P) 0,5128 0,2228 3,0563 0,9437 MAT to hPD-1 15 IgGl (AA) 0.2927 0,1333 2,0640 0,6096

Saturation binding studies demonstrate that humanized variants of anti-PD-1 mAb 2, anti-PD-1 mAb 7, anti-PD-1 mAb 9, and anti-PD-1 mAb 15 have a favorable binding profile to surface PD-1. In particular, humanized anti-PD-1 mAb 7 (anti-hPD-1 mAb 7(1.1) and anti-hPD1 mAb 7(1.2) containing the IgG1(AA) or IgG4(P) Fc region have the lowest _EC90 values of all antibodies studied.

To further characterize the humanized anti-PD-1 antibodies, anti-hPD-1 2 IgG1 mAb (AA), anti-hPD-1 7(1.1) IgG1 mAb (AA), anti-hPD-1 7(1.2) IgG1 mAb (AA), anti-hPD-1 7(1.2) IgG4 mAb (P), anti-hPD-1 9(1.1) IgG1 mAb (AA), anti-hPD-1 9(1.1) IgG4 mAb (P), and anti-hPD-1 15 IgG1 mAb (AA) were tested for their ability to block the binding of human PD-L1 (shPD-L1) and human PD-L2 (shPD-L2) to PD-1 expressed on the surface of NSO cells. These assays were performed as described above. Representative inhibition curves of sPD-L1 and sPD-L2 binding to PD-1 expressed in NSO cells are shown in Fig. 12A and 12B, respectively. IC ₅₀ and IC ₉₀ values were determined, and the sample mean (SM) and standard deviation (SD, σ) of three separate experiments are shown in Table 9.

Table 9

sPD-Ll SPD-L2 IC50 (mcg/study) IC90 (mcg/study) IC50 (mcg/study) IC90 (mcg/study) Antibody to PD-1 SM SD σ SM SDo SM SDo SM SDc MAT kPD-Ι A IgGl (AA) 0,0203 0,0089 0.2985 0.3279 0,0414 0,0124 0,1601 0,066 MAT kPD-Ι A IgG4 (P) 0,0156 0,0096 0,0776 0,0208 0,0280 0,0070 0,1594 0,1153 MAT kPD-Ι B IgGl (AA) 0,0148 0,0008 0,1034 0,0100 0,0280 0,0059 0,1190 0,060 MAT kPD-Ι B IgG4 (P) 0,0143 0,0013 0,0798 0,0239 0,0280 0,0055 0,0924 0,0065 MATkhPD-I 2 0,0578 0,0124 0,2480 0,050 0,1294 0,0143 0.3813 0,0656

- 97 047726

IgGl (AA) MAT to 4PD1 7(1.1) IgGl (AA) 0,0166 0,0032 0,0674 0,0041 0,0283 0,0147 0,0886 0,0166 MAT to 4PD17(1.2) IgGl (AA) 0,0118 0,0027 0,0678 0,0031 0,0212 0,0031 0,0672 0,0043 MAT to 4PD17(1.2) IgG4 (P) 0,0103 0,0023 0,0520 0,0033 0,0213 0,0019 0,0616 0,0063 MAT to 4PD1 9(1.1) IgGl (AA) 0,0593 0,0036 0.3238 0,0508 0,4002 0,5000 0,4573 0,1805 MAT to 4PD1 9(1.1) IgG4 (P) 0,0460 0,0118 0,2461 0,0513 0,1105 0,0146 0.2914 0,0526 MAT to 4PD- 1 15 IgGl (AA) 0,0440 0,0092 0,2068 0,035 0,0945 0,0022 0.3093 0,0588

Ligand binding inhibition studies demonstrate that humanized variants of anti-PD-1 mAb 2, anti-PD-1 mAb 7, anti-PD-1 mAb 9, and anti-PD-1 mAb 15 are able to inhibit the binding of sPD-L1 and sPD-L2 to PD-1 on the cell surface. In particular, humanized anti-PD-1 mAb 7 (anti-hPD-1 mAb 7 (1.1) and anti-hPD-1 mAb 7 (1.2)) have the lowest IC3 values of all antibodies studied.

Example 3. Blockade of the PD-1/PD-L1 checkpoint with humanized antibodies to human PD-1.

The ability of hPD-1 2 IgG1 MAb (AA), hPD-1 7(1.1) IgG1 MAb (AA), hPD-1 7(1.2) IgG1 MAb (AA), hPD-1 7(1.2) IgG4 MAb (P), hPD-1 9(1.1) IgG1 MAb (AA), hPD-1 9(1.1) IgG4 MAb (P), hPD-1 15 IgG1 MAb (AA) and the reference anti-PD-1 MAb, anti-PD-1 A and anti-PD-1 B MAb, to act as antagonists of the PD-1/PD-L1 axis (i.e., to block PD-1/PD-L1 interaction and prevent suppression of T cell responses) was investigated in a quantitative luciferase reporter assay Jurkat-luc-NFAT/CHO-PD-L1: Briefly, CHO cells expressing PD-L1 (CHO/PDL1) were seeded at a density of 40,000/well in 100 μl of culture medium (RPMI+10% fetal bovine serum (FBS)+100 μg/ml hygromycin B+100 μg/ml G418) and incubated overnight. The following day, the media were removed and Jurkat MNFAT-luc2/PD-1 cells (Promega) were added to each well at a density of 50,000 cells/well in 40 μl of assay buffer (RPMI+2% FBS) and anti-PD-1 antibodies, anti-PD-1 mAbs 1-15, or anti-PD-1 reference antibodies (anti-PD-1 mAbs A and anti-PD-1 mAbs B) (0-25 μg/ml, eight serial 2.5-fold dilutions in assay buffer) and incubated for 6 h at 37°C with an additional incubation for 5-10 min at ambient temperature. Then 80 μl of BioGlo substrate (Promega) were added to each well and the plate was incubated for an additional 5-10 min at ambient temperature, and the luminescence intensity was measured in a Victor plate reader. Typical saturation curves are shown in Fig. 13. The EC ₅₀ and EC ₉₀ values were determined, and the sample mean (SM) and standard deviation (SD, σ) of four individual experiments are given in Table 10.

- 98 047726

Table 10

Antibody to PD-1 Reporter's signal EC ₅ o (mcg/study) ECeo (mcg/study) SM SD σ SM SD σ MAT to PD-1 A IgGl (AA) 0.2549 0,0480 2,4474 1,2228 MAT to PD-1 A IgG4 (P) 0,2049 0,0719 2,5535 1,2139 MAT to PD-1 B IgGl (AA) 0,2119 0,1781 2,2036 2,0118 MAT to PD-1 B IgG4 (P) 0,1142 0,0323 0,9418 0.2863 MAT to hPD-1 2 IgGl (AA) 0.3539 0,0983 3,8975 2,0054 MAT to hPD-1 7(1.1) IgGl (AA) 0,1080 0,0386 1,1992 0,5103 MAT to hPD-1 7(1.2) IgGl (AA) 0,0944 0,0153 0,6452 0.2615 MAT to hPD-1 7(1.2) IgG4 (P) 0,0965 0,0169 0,6885 0,01858 MAT to hPD-1 9 IgGl (AA) 0.2835 0,0530 2,9968 0,8866 MAT to hPD-1 9 IgG4 (P) 0.3154 0,0872 5,0940 4,0496 MAT to 4PD-1 15 IgGl (AA) 0.2585 0,0592 3,3138 1,0532

Reporter signal intensity studies demonstrate that humanized variants of anti-PD-1 Mab 2, anti-PD-1 Mab 7, anti-PD-1 Mab 9 and anti-PD-1 Mab 15 can block the PD-1/PD-L1 axis and prevent the suppression of T cell responses. In particular, humanized anti-PD-1 Mab 7 (anti-hPD-1 Mab 7(1.1) and anti-hPD-1 Mab 7(1.2) containing the IgG1(AA) or IgG4(P) Fc region have the lowest _EC50 / _EC90 values.

Example 4. Functional activity of antibodies to human PD-1.

Staphylococcus aureus enterotoxin type B (SEB) is a microbial superantigen capable of activating a significant proportion of T cells (5-30%) in SEB-responsive donors. SEB binds to MHC II outside the peptide binding groove and is therefore MHC II dependent but unrestricted and TCR-mediated. SEB stimulation of T cells results in oligoclonal T cell proliferation and cytokine production (although donor variability may occur and some donors will not respond to SEB). Within 48 h of SEB stimulation, PBMCs upregulate PD-1 and LAG-3, with a subsequent upregulation observed at day 5 after secondary culture in a 96-well SEB-stimulated plate. Activation of the immune checkpoint proteins PD-1 and LAG-3 following SEB stimulation of PBMCs limits cytokine release upon restimulation. The ability of anti-PD-1 antibodies alone and in combination with anti-LAG-3 antibodies to enhance cytokine release via checkpoint inhibition was investigated. Briefly, PBMCs were purified from whole blood obtained from healthy donors who signed an informed consent form (Biological Specialty Corporation) using the Ficoll-Paque Plus density gradient centrifugation method (GE Healthcare) according to the manufacturer's instructions, and T cells were then purified using the Dynabeads™ Untouched™ Human T Cell Isolation Kit (Life Technologies) according to the manufacturer's instructions. Purified PBMCs were cultured in RPMI+10% heat-inactivated FBS+1% penicillin/streptomycin in T25 flasks for 2-3 days alone or with SEB (Sigma-Aldrich) at 0.1 ng/ml (primary stimulation). At the end of the first round of SEB stimulation, PBMCs were washed twice with phosphate-buffered saline (PBS) and immediately seeded in 96-well tissue culture plates at 1-5x105 cells/well in media only, media with control antibody or anti-PD-1 antibody, media with SEB at 0.1 ng/ml (secondary stimulation) and no antibody, or media with SEB and control IgG or anti-PD-1 antibody, with or without anti-LAG-3 antibody, and cultured for another 2-3 days. At the end of the second stimulation, supernatants were collected for measurement of cytokine secretion using DuoSet ELISA kits for human IFN-γ, TNF-α, IL-10, and IL-4 (R&D Systems) according to the manufacturer's instructions.

The ability of anti-PD-1 MAT 2, anti-PD-1 MAT 7, anti-PD-1 MAT 9, and anti-PD-1 MAT 15, alone or in combination with the unique anti-LAG-3 MAT 1, to enhance cytokine release via checkpoint inhibition was examined.

These studies also included the use of one or more of the following anti-PD-1 reference antibodies: anti-PD-1A mAb; anti-PD-1B mAb; and anti-LAG-3A mAb, alone or in combination. Figure 14 shows the IFN-γ secretion profiles from SEB-stimulated (0.1 ng/mL) PBMCs from a typical responder donor (D: 38941) treated with: no antibody control; antibody

- 99 047726 for isotype control; MAT 7 to PD-1 and/or MAT to LAG-3 7; MAT 9 to PD-1 and/or MAT to LAG-3 1; MAT 15 to PD-1 and/or MAT to LAG-3 1; MAT 2 to PD-1 and/or MAT to LAG-3 1; or reference antibodies to PD-1, MAT to PD-1 B and/or MAT to LAG-3 A (antibodies were used at a concentration of 10 μg/ml).

Additional studies assessed the ability of humanized anti-PD-1 Mab 2, anti-PD-1 Mab 7, anti-PD-1 Mab 9, and anti-PD-1 Mab 15 (containing human IgG1(AA) or human IgG4(P) Fc), as well as the anti-PD-1 reference antibodies, anti-PD-1 Mab A and anti-PD-1 Mab B, to enhance cytokine release through checkpoint inhibition. For these studies, antibodies were used at concentrations of 0.625, 2.5, and 10 μg/mL. Figures 15A–15B show the secretion profiles of IFN-γ (Figure 15A) and TNF-α; (Fig. 15B) from SEB-stimulated (0.2 ng/ml) PBMCs from a typical responder donor (D: 57709) that were treated with no antibody control or one of the following antibodies: isotype control antibody; anti-hPD-1 MAb 2 IgG1 (AA); anti-hPD-1 MAb 7(1.2) IgG1 (AA); anti-hPD-1 MAb 7(1.2) IgG4 (P); anti-hPD-1 MAb 9(1.1) IgG1 (AA); anti-hPD-1 MAb 9(1.1) IgG4 (P); anti-hPD-1 MAb 15 IgG1 (AA); or the reference antibodies to PD-1, MAb to PD-1 A IgG1 (AA), MAb to PD-1 A IgG4 (P), MAb to PD-1 B IgG1 (AA), MAb to PD-1 B IgG4 (P). The total amount of IFN-γ (pg/mg) in the samples treated with SEB+ antibody was determined for the samples treated with 0.625, 2.5, and 10 μg/mL of PD-1 antibodies, and the mean (SM) and standard deviation (SD, σ) for three different responding donors (except where indicated) are shown in Table 1. 11. The ratio of IFN-γ levels secreted in the sample treated with humanized versions of anti-PD1 MAB 2, anti-PD-1 MAB 7, anti-PD-1 MAB 9, and anti-PD-1 MAB 15 (containing human IgG1(AA) or human IgG4(P) Fc) compared to the levels induced by anti-PD-1 control antibodies, anti-PD-1 MAB A, and anti-PD-1 MAB B (i.e., humanized anti-PD-1 MAB/anti-PD-1A and humanized anti-PD1 MAB/anti-PD-1B) are presented in Table 12 and Table 13, respectively.

Table 11

IFN-γ secretion (pg/ml) Antibody to PD1 (mcg/ml) 0.625 mcg/ml 2.5 mcg/ml 10 mcg/ml Antibody to PD1 SM SD σ SM SD σ SM SD σ MAT to PD-1 A IgGl (AA) 221,18 110.89 341,13 247.93 347,46 144.72 MAT to PD-1 A IgG4 (P) 281.36 132.65 495,15 190.57 399,41 117.56 MAT to PD-1 B IgGl (AA) 366,69 196.64 387,682 215,51 387.32 282.81 MAT to PD-1 B IgG4 (P) 348,40 185.96 433,382 163.23 551,68 125.08 MAT to hPD-1 7(1.2) IgGl (AA) 302,05 185.71 610,70 209.77 414.63 272.65 MAT to hPD-1 7(1.2) IgG4 (P) 384,57$ 323,79J 411,40 398,59 370,06 108,12 MAT to hPD-1 9(1.1) IgGl (AA) 340.81 207.76 442,598 303,70 655.29 567.91 MAT to hPD-1 9(1.1) IgG4 (P) 309.82 130,30 468,62 350,15 424.35 288.95 MAT to 4PD-1 15 IgGl (AA) 360,00 274.28 373.32 160.25 541.83 444.22 MAT to hPD-1 2 IgGl (AA) 275.88 135.23 372,73 53,53 496,70 235.37 Control IgG 137,14 76.61 100.65 48,67 138,10 120.81 Antibody absent 120,05 73.90 120,05 73.90 109.46 85.18

ί Results obtained from two responding donors.

- 100 047726

Table 12

Ratio of IFN-γ secretion levels (new mAb to PD-1/mAb to PD-1 A) Antibody to PD1 (mcg/ml) 0.625 mcg/ml 2.5 mcg/ml 10 mcg/ml Antibody to PD1 SM SD σ SM SD σ SM SD σ MAT to PD-1 A IgGl (AA) 1,00 0,00 1,00 0,00 1,00 0,00 MAT to PD-1 A IgG4 (P) 1,00 0,00 1,00 0,00 1,00 0,00 MAT to PD-1 B IgGl (AA) 1.77 0.92 1.28 0.36 1.07 0.42 MAT to PD-1 B IgG4 (P) MAbchRO-1 7(1.2) IgGl 1.23 0.16 0.92 0.27 1.40 0,12 (AA) 1.36 0.37 2.46 1.85 1,17 0.41 MAT to hPO-1 7(1.2) IgG4 (P) 1,20$ 0,35$ 0.79 0.54 0.95 0.22 MATkchPO-1 9(1.1) IgGl (AA) 1.48 0.19 1.46 0.71 1.70 0.84 MAT to hPO-1 9(1.1) IgG4 (P) 1,13 0.13 0.91 0.42 1,02 0.46 MATkchPO-1 15 IgGl (AA) 1.50 0.39 1.51 1.23 1.48 0.71 MAT to hPO-1 2 IgGl (AA) 1.32 0.53 1.48 0.86 1.42 0,12 Control IgG 0.63 0,2 0.33 0.08 0.39 0.24 Antibody absent 0.54 0,12 0.39 0.14 0.31 0.17

ί Results obtained from two donors who had a response. Table 13

Ratio of IFN-γ secretion levels (new mAb to PD-1/mAb to PD-1 B) Antibody to PD1 (mcg/ml) 0.625 mcg/ml 2.5 mcg/ml 10 mcg/ml Antibody to PD1 SM SD σ SM SD σ SM SD σ MAT to PD-1 A IgGl (AA) 0.37 0.37 0.82 0,20 1.06 0.48 MAT to PD-1 A IgG4 (P) 0.82 0,12 1,16 0.38 0.72 0,07 MAT to PD-1 B IgGl (AA) 1,0 0,00 ι,θ 0,00 ι,θ 0,00 MAT to PD-1 B IgG4 (P) 1,0 0,00 1,0 0,00 ι,θ 0,00 MAT to hPD-1 7(1.2) IgGl (AA) 0.84 0.22 1.77 0.81 ι,ιι 0,07 MAT to hPD-1 7(1.2) IgG4 (P) 0,91$ 0,26J 0.83 0,50 0.68 0.17 MAT to hPD-1 9(1.1) IgGl (AA) 1.04 0.59 1,12 0.29 1.60 0.42 MAT to hPD-1 9(1.1) IgG4 (P) 0.92 0.09 0.99 0.36 0.75 0.39 MAT to hPD-1 15 IgGl (AA) 1,01 0.48 1.07 0.57 1.34 0,15 MAT to hPD-1 2 IgGl (AA) 0.78 0,12 1,10 0.38 1.46 0.53 Control IgG 0.39 0,08 0.27 0,08 0.34 0.13 Antibody absent 0.34 0.11 0.31 0,03 0.28 0.08

ί Results obtained from two responding donors.

The results of these studies demonstrate that anti-PD-1 antibodies, anti-PD-1 MAb 2, anti-PD-1 MAb 7, anti-PD-1 MAb 9, and anti-PD-1 MAb 15, dramatically enhanced the production of IFN-γ (Figs. 14 and 15A and Tables 11–13) and TNF-α (Fig. 15B) from SEB-stimulated PBMCs upon restimulation. In addition, the combination of anti-PD-1 antibodies with anti-LAG-3 antibodies resulted in further enhancement of cytokine release (Fig. 14) from SEB-stimulated PBMCs upon restimulation. In particular, the combination of anti-PD-1 MAT 2, anti-PD-1 MAT 7, anti-PD-1 MAT 9, or anti-PD-1 MAT 15 with a unique anti-LAG-3 antibody, anti-LAG-3 MAT 1, provided the greatest enhancement.

- 101 047726

Example 5. Binding studies of bispecific PD-1xLAG-3 molecules.

A number of PD-1xLAG-3 bispecific molecules were generated, including diabodies containing three-, four-, and five-chain Fc regions and a bispecific antibody. Four four-chain diabodies containing heterodimer-promoting domains with the E/K helix were generated and designated DART A, DART B, DART C, and DART I. Four four-chain diabodies containing CH1/CL domains were generated and designated DART D, DART E, DART J, and DART 1. Two five-chain diabodies containing heterodimer-promoting domains with the E/K helix and CH1/CL domains were generated and designated DART F and DART G. One three-chain antibody containing the E/K helix was generated and designated DART H. One four-chain bispecific antibody was generated and designated BSAB A. The structures and amino acid sequences of these PD-1xLAG-3 bispecific molecules are shown above and summarized in Table 14 below.

Table 14

Name Initial MAT Fc* Chains SEQ ID NO: Other components DART A MAT kmPD-I 7(1.2) MAT to 4LAG- 3 1(1.4) IgG4 (YTE) 4 267 (Xi=A; X ₂ =Y; X ₃ =T; X ₄ =E) and 268 E/K helices; see Figure 3B DART B MAT kmPD-I 7(1.2) MAT to 4LAG- 3 1(1.3) IgG4 (YTE) 4 267 (Xi=G; X ₂ =Y; X ₃ =T; X ₄ =E) and 268 E/K helices; see Figure 3B DART C MAT kmPD-I 7(1.2) MAT to 4LAG- 3 1(1.3) IgG4 4 267 (Xi=G; X ₂ =M; X ₃ =S; X ₄ =T) and 268 E/K helices; see Figure 3B

- 102 047726

DARTD MAT kmPD-I 7(1.2) MAT to mLAG- 3 1(1.4) IgG4 (YTE) 4 269 and 270 CL/CH1; see Figure ZS DARTE MAT to mPD-I 7(1.2) MAT to mLAG- 3 1(1.4) IgG4 (YTE) 4 271 and 272 CL/CH1; see Figure ZS DARTF MAT to mPD-I 7(1.2) MAT to mLAG- 3 1(1.4) IgGl (AA/YTE) 5 273, 274, 275 and 276 CL/CH1 and E/K helices; see Figure 5 DART G MAT to mPD-I 7(1.2) MAT to mLAG- 3 1(1.4) IgGl (AA/YTE) 5 277, 278, 279 and 280 CL/CH1 and E/K helices; see Figure 5 DARTH MAT to mPD-I 7(1.2) MAT to mLAG- 3 1(1.4) IgGl (AA) 3 281, 282 and 283 E/K spirals; see Figure 4A DART I MAT to mPD-I 7(1.2) MAT to mLAG-3 6(1.1) IgG4 (YTE) 4 290 and 291 E/K helices; see Figure 3B DART J MAT kmPD-I 7(1.2) MAT to mLAG- 3 6(1.1) IgG4 (YTE) 4 292 and 293 CL/CH1; see Figure ZS DART 1 MAT to PD-1 A MAT to LAG-3 A IgGl (AA) 4 284 and 285 CL/CH1; see Figure ZS BSAB A MAT kmPD-I 7(1.2) MAT to mLAG- 3 1(1.4) IgGl (AA) 4 286, 287, 288 and 289 MAT containing a charge-modified Fc region

ί Molecules containing the IgG4 Fc regions also contain a stabilized IgG4 hinge region.

Additional PD-1xLAG-3 bispecific molecules containing alternative binding sites for PD-1 and/or LAG-3 epitopes can be readily generated by incorporating different VH and VL domains. Similarly, molecules binding to an antigen other than LAG-3 can be generated by incorporating VH and VL having the desired specificity.

The saturation binding profiles of the PD-1xLAG-3 bispecific diabodies constructs: DART A, DART B, DART D, DART E, DART F, DART G, DART H, DART I, and DART 1; anti-PD-1 antibodies: anti-hPD-1 MAb 7(1.2) IgG4 (P), anti-hPD-1 MAb 7(1.2) IgG1 (AA), anti-PD-1 A MAb IgG1 (AA), and anti-PD-1 A MAb IgG4 (P); and anti-LAG-3 antibodies: anti-hLAG-3 MAb 1(1.4) IgG4 (P), anti-LAG-3 MAb IgG4 (P), anti-hLAG-3 MAb 1(1.4) IgG1 (AA), and anti-LAG-3 A MAb IgG1 (AA), were tested essentially as described above. The PD-1xLAG-3 bispecific antibody constructs were tested for binding to PD-1 and LAG-3, while anti-PD-1 and anti-LAG-3 antibodies were tested only for binding to their respective antigens. NSO cells expressing PD1 or LAG-3 were used for these studies. Diabodies and antibodies were used (170.0–0.013 μM or 85.0–0.0021 μM (fourfold serial dilutions). _EC50 and _EC90 values were determined and presented in Tables 15–16. Sample mean (SM) and standard deviation (SD, σ) are presented when two or more separate experiments were performed.

- 103 047726

Table 15

Molecule PD-1 binding saturation EC ₅ o (μM) EC90 (µM) SM SD σ SM SD σ DART A 1,9297 0,4324 9,6027 0,4801 DART B 1,7640$ 12,2700$ DARTD 2,2267 0,4140 10,9313 2,6351 DARTE 3,2180 0,5742 23,840 3,2385 DARTF 1,4320$ 14,5800$ DARTG 1,1488 0,6227 3,4220 2,4600 DARTH 4,5310$ 22,6600$ DARTI 1,3232 0,4890 7,8135 4,0821 DART 1 2,1329 1,4850 13,8113 9,0256 MAT to hPD-I 7(1.2) IgG4 (P) 1,2083 0,8112 3,9340 1,8746 MAT to PD-1 A IgG4 (P) 2,3470 1,2362 22,7770 15,0690 MAT to hPD-I 7(1.2) IgGl (AA) 1,0879 0.3958 7,4153 3,0794 MAT to PD-1 A IgGl (AA) 1,6733 0,5464 9,9543 6,6569

§ Results obtained in one experiment.

Table 16

Molecule Saturation of LAG-3 binding EC ₅ o (μM) EC90 (μM) SM SD σ SM SD σ DART A 0,8402 0,2231 4,4448 2,4770 DART B $1,0750 9,8580$ DARTD 0.8985 0,5326 5,7967 4,7329 DARTE 0,9250 0,8075 5,6450 5,6809 DARTF 5,0090 0,5770 19,3350 4,7447 DARTG 0.9396 0.3045 8,5507 4,7448 DARTH 2,3840$ 9,7810 4,2412 DARTI 0,5321 0,0547 4,198 3,2188 DART 1 20,0233 2,1454 115.97 15,2425 MAT to mLAG-3 1(1.4) IgG4 (P) 1,0057 0,1969 5,1360 4,7904 MAT to LAG-3 A IgG4 (P) 0.5968 0,1376 2,0833 0.3244 MAT to mLAG-3 1(1.4) IgGl (AA) 0,6069 0,3430 3,6373 2,4762 MAb to LAG-3 A IgGl (AA) 0,4523 0,1660 2,0187 0,7035

§ Results obtained in one experiment.

Saturation binding studies demonstrate that PD-1xLAG-3 bispecific antibody constructs retain binding to PD-1 and have binding profiles that are similar to those of anti-PD-1 antibodies. Similarly, PD-1xLAG-3 bispecific antibody constructs retain binding to LAG-3 and, with the exception of DART 1, have binding profiles that are similar to those of the parental anti-LAG-3 antibodies.

Example 6. Studies of the inhibitory capacity of bispecific PD-1xLAG-3 molecules.

The ability of the PD-1xLAG-3 bispecific molecules: DART A, DART B, DART D, DART E, DART F, DART G, DART H, DART I, DART 1 and B SAB A; and the anti-PD-1 antibodies: anti-hPD-1 MAb 7(1.2) IgG4 (P), anti-hPD-1 MAb 7(1.2) IgG1 (AA), anti-PD-1 A MAb IgG1 (AA) and anti-PD-1 A MAb IgG4 (P) to block the binding of human PD-L1 (shPD-L1) and human PD-L2 (SHPD-L2) to PD-1 expressed on the surface of NSO cells was tested essentially as described above. Diabodies and antibodies were used in concen-

- 104 047726 33.75-0.002 μM or 107.5-0.0001 μM (fourfold serial dilutions).

The IC ₅₀ and IC ₉₀ values were determined and presented in Table 17. The sample mean (SM) and standard deviation (SD, σ) were presented for those cases where 2 or more separate experiments were performed.

Table 17

Molecule Blocking sPD- Ll/PD-1 binding Blocking sPDL2/PD-1 binding IC50 (µM) IC90 (µM) IC50 (µM) IC90 (µM) SM SDn SM SDn SM SDc SM SDn DART A 0.9645 0,1485 5,6312 1,5247 1,6273 0,4285 6,9335 3,9849 DART B 1,1515 0,0007 4,8615 0,2199 2,1150 0.3154 7,9550 0,0933 DARTD 1,5548 0,1692 7,8950 2,5135 3,1255 0.5869 9,2973 5,5426 DARTE 1,6533 0,3307 7,8470 1,1642 2,9460 0,7736 6,6135 0,0177 DARTF 0.5697 0,1729 2,0360 0,1174 0,8389 0,0846 1,7995 0,2171 DARTG 1,6013 0.3581 8,1953 1,5708 2,5540 0,7891 7,4810 0,2333 DARTH 3,3950 0,1018 18,640 9,5742 6,2065 3,6847 29,395 3,8679 DARTI 0,8363 0,1302 5,3115 0.3125 1,286 0.3125 6,2485 1,3951 DART 1 1,7467 0.3097 5,4533 1,0214 2,8355 1,8250 7,2735 3,9831 BSABA 2,1590 0.3097 11,075 0,8132 4,8775 0,5438 15,580 1,3294 MAT to ₄ PD1 7(1.2) IgG4 (P) 0,5186 0,1668 3,8050 1,2227 1,0425 0.2563 3,4880 0.5459 MAT to PD-1 A IgG4 (P) 0,9209 0.3256 4,3023 0,7069 1,3859 0.3882 5,1675 0.2943 MAT to ₄ PD1 7(1.2) IgGl(AA) 0,7320 0,2337 3,2048 1,1479 0.9769 0.2893 2,8437 1,4801 MAT to PD-1 A IgGl (AA) 1,0765 0.2393 5,2775 0.9933 1,9510 0,8814 5,0880 1,3831

Ligand binding inhibition studies demonstrate that the PD-1xLAG-3 bispecific antibody constructs retain the ability to inhibit sPD-L1 and sPD-L2 binding to PD-1 on the cell surface. In addition, the ability of the PD-1xLAG-3 bispecific molecules: DART A, DART B, DART D, DART E, DART F, DART G, DART H, DART I, DART 1 and B SAB A; and anti-LAG-3 antibodies: anti-hLAG-3 MAb 1(1.4) IgG4 (P), anti-LAG-3 MAb A IgG4 (P), anti-hLAG-3 MAb 1(1.4) IgG1 (AA) and anti-LAG-3 MAb A IgG1 (AA) to block human LAG-3 binding to native MHC class II on the surface of Daudi cells was investigated. Briefly, each PD-1xLAG-3 bispecific molecule and anti-LAG-3 control antibody were mixed with biotinylated soluble human LAG-3Fc (shLAG-3) fusion protein (0.5 μg/mL) and separately incubated with MHC II positive Daudi cells (2.5 x 10 ⁶ cells). The amount of LAG-3 bound to the Daudi cell surface was determined using a PE-conjugated streptavidin secondary antibody by FACS. Diabodies and antibodies were used at concentrations of 27.5-0.026 μM (two-fold serial dilutions) or 107.5-0.0001 μM (four-fold serial dilutions) or 35-0.002 μM (four-fold serial dilutions).

IC ₅₀ and IC ₉₀ values were determined and presented in Table 18. Sample mean (SM) and standard deviation (SD, σ) were provided in cases where 2 or more separate experiments were performed.

- 105 047726

Table 18

Molecule Blocking shLAG-3/MHC class II binding IR ₅ o (µM) IC90 (µM) SM SD σ SM SD σ DART A 1,3835 1,6465 8,396102 8,3962 DART B 0,4081 0,1104 3,0645 0.3924 DARTD 1,1843 1,1398 8,0041 7,3317 DARTE 3,2706 2,9177 28,9683 24,1694 DARTF 1,5347 1,2674 10,3920 11,2555 DARTG 2,0618 3,3552 11,4422 12,4964 DARTH 2,8967 4,9817 17,2533 21,1420 DARTI 0,4864 0,1549 2,339 1,1780 DART 1 15,9610 14,0883 87,1486 109,533 BSAB A 0,7101 0,0571 7,2470 1,0706 MAT to mLAG-3 1(1.4) IgG4 (P) 0,4815 0,2176 3,4837 1,7564 MAB to LAG-3 A IgG4 (P) 0,7011 0,1900 2,4232 0.3481 MAT to hLAG-3 1(1.4) IgGl (AA) 0.3637 0,1409 9,4422 7,9319 MAT to LAG-3 A IgGl (AA) 0.5923 0,3407 2,1451 1,1139

Ligand binding inhibition studies demonstrate that PD-1xLAG-3 bispecific antibody constructs retain the ability to inhibit shLAG-3Fc fusion protein binding to cell surface MHC class II. With the exception of DART 1, PD1xLAG-3 bispecific molecules have similar inhibition profiles as the parental LAG-3 antibodies.

Example 7. Blockade of PD-1/PD-L1 checkpoint molecules by bispecific PD-1xLAG-3 molecules.

The ability of bispecific PD-1xLAG-3 molecules: DART A, DART B, DART D, DART E, DART F, DART G, DART H, DART I, DART 1 and BSAB A; and anti-PD-1 antibodies: anti-hPD-1 MAb 7(1.2) IgG4 (P), anti-hPD-1 MAb 7(1.2) IgG1 (AA), anti-PD-1 A MAb IgG1 (AA), and anti-PD-1 A MAb IgG4 (P), to act as antagonists of the PD-1/PD-L1 axis (i.e., block PD-1/PD-L1 interaction and prevent suppression of T cell responses) were assessed in a quantitative Jurkatluc2-NFAT/CHO-PD-L1 luciferase reporter assay (using CHO/PD-L1 cells and Jurkat MNFAT-luc2/PD-1 cells) as described above. Diabodies and antibodies were used at a concentration of 100–0.0065 μM (four-fold serial dilutions) or 100–0.0013 μM (five-fold serial dilutions).

The IC ₅₀ and IC ₉₀ values were determined and presented in Table 19. The sample mean (SM) and standard deviation (SD, σ) were presented in cases where 2 or more separate experiments were performed.

- 106 047726

Table 19

Molecule Reporter's signal IR ₅ o (µM) IC90 (µM) SM 8Ι)σ SM SDo DART A 0,8804 0,1949 7,9115 1,3232 DART B 1,079 0,1535 7,5413 3,1483 DARTD 1,4044 0.2584 12,0786 3,6616 DARTE 1,4060 0,1222 13,7867 1,4981 DARTF 0,3404 0,0103 1,8710 0,481 DARTG 0,6914 0,0206 4,2090 0,7331 DARTH 36,6167 20,8078 968,300 811,8471 DARTI 1,3335 0.3641 12,146 6,8787 DART 1 11,8807 3,4905 1048,2000 1508,9992 BSAB A 9,7825 1,0288 113,3350 22,2951 MAT to hPD-1 7(1.2) IgG4 (P) 0,6460 0.3035 6,0736 2,5513 MAT to PD-1 A IgG4 (P) 1,328 0,7439 16,5138 9,7149 MAT to hPD-1 7(1.2) IgGl(AA) 0,5214 0,1541 4,7592 2,1044 MAT to PD-1 A IgGl (AA) 1,4514 1,0049 35,7382 40,9858

Reporter signal intensity studies demonstrate that most PD-1xLAG-3 bispecific antibody constructs retain the ability to inhibit sPD-L1 binding to cell surface PD-1. The tetravalent PD-1xLAG-3 bispecific diabodies, DART A, DART B, DART D, DART-E, DART F, DART G, and DART I, were the most potent inhibitors in this quantitative assay. Similar results were obtained for several of these bispecific constructs examined in the quantitative PDL2 binding reporter assay.

Example 8. Functional activity of bispecific PD-1xLAG-3 molecules.

The ability of PD-1xLAG-3 bispecific molecules to enhance cytokine release via checkpoint inhibition was examined in SEB-stimulated PBMCs upon restimulation as described above, except where noted.

Initial studies assessed the ability of the PD-1xLAG-3 bispecific molecules: DART A, DART D, DART E, DART F, DART G, DART H; and anti-PD-1 antibodies and anti-LAG antibodies: anti-PD1 MAb A IgG4 (P) and anti-LAG-3 MAb A IgG4 (P), alone or in combination, to enhance cytokine release through checkpoint inhibition. In these quantitative studies, the PD-1xLAG-3 bispecific molecules and antibodies were used at a total concentration of 3.125, 12.5, or 50 nM, and PBMCs were stimulated with 0.2 ng/mL SEB (0.1 ng/mL was used in previous studies). For these studies, when using the antibody combination, the concentration of each antibody was half the total concentration (i.e., 1.563, 6.25, or 25 nM). In Fig. 16A and 16B show the IFN-γ secretion profiles from SEB-stimulated PBMCs from two typical responding donors, D:35644 and D:59697, respectively. As noted, not all donors responded to SEB at 0.1 or 0.2 ng/mL. To enhance SEB stimulation in PBMCs from a larger number of donors, additional studies used SEB at a high concentration of 85 ng/mL or a mid-range concentration of 0.5 ng/mL. At these concentrations, SEB stimulation is more robust across a larger number of donors, although donor response variability may still be observed.

One of the studies assessed the ability of PD-1xLAG3 bispecific molecules: DART A, DART B; anti-PD-1 antibodies: anti-hPD-1 MAb 7 (1.2) IgG4 (P); anti-LAG-3 antibodies: anti-LAG-3 MAb 1 (1.4) IgG4 (P); and combinations of anti-PD-1 MAb A IgG4 (P) and anti-LAG-3 MAb A IgG4 (P) to enhance cytokine release through checkpoint inhibition. In these quantitative studies, PD-1xLAG-3 bispecific molecules and antibodies were used at concentrations of 0.019, 0.078, 0.3125, 1.25, 5, or 20 nM, and PBMCs were stimulated with 85 ng/mL SEB. To perform this quantitative study, when using a combination of antibodies, each antibody was provided at the indicated concentration, and, accordingly, the total antibody concentration was doubled.

- 107 047 726 concentration used for each antibody (i.e., 0.038, 0.156, 0.625, 2.5, 10, or 40 nM). Figures 17A and 17B show the IFN-γ secretion profiles from SEB-stimulated PBMCs from two representative donors D: 55515 and D: 54024, respectively. In another study, the ability of PD-1xLAG-3 bispecific molecules: DART A, DART B, DART C; anti-PD-1 antibodies: anti-hPD-1 MAb 7(1.2) IgG4(P); anti-LAG-3 antibodies: anti-LAG-3 MAb 1(1.4) IgG4(P); and combinations of anti-PD-1 A IgG4 mAb (P) and anti-LAG-3 A IgG4 mAb (P), enhance cytokine release through checkpoint inhibition. In these quantitative studies, the bispecific molecules and PD-1xLAG-3 antibodies were used at a total concentration of 0.048, 0.195, 0.78, 3.125, 12.5, or 50 nM, and PBMCs were stimulated with 0.5 ng/mL SEB. For these studies, when a combination of antibodies was used, the concentration of each antibody was half the total concentration (i.e., 0.024, 0.098, 0.39, 1.563, 6.25, or 25 nM). In Fig. 18A and 18B show the IFN-γ secretion profiles from SEB-stimulated PBMCs from two typical donors D:20990 and D:54947, respectively.

An additional study assessed the release of the cytokine IL-2.

In particular, the ability of bispecific PD-1xLAG-3 molecules: DART D, DART H; antibodies to PD-1: MAT to PD-1 A IgG4 (P), MAT to hPD-1 7(1.2) IgG4(P); antibodies to LAG-3: MAT to LAG-3 A IgG4 (P) and MAT to LAG-3 1(1.4) IgG4(P); and combinations: MAT to PD-1 A IgG4 (P) and MAT to LAG-3 A IgG4 (P), and MAT to hPD-1 7(1.2) IgG4(P) and MAT to LAG-3 1(1.4) IgG4(P), to enhance IL-2 release through checkpoint inhibition was investigated. In these quantitative studies, the PD-1xLAG-3 bispecific molecules and antibodies were used at a total concentration of 3.125, 12.5, or 50 nM, and PBMCs were stimulated with a high concentration of SEB of 85 ng/mL. For these studies that used a combination of antibodies, the concentration of each antibody was half the total concentration (i.e., 1.563, 6.25, or 25 nM). Figure 19 shows the IL-2 secretion profile of SEB-stimulated PBMCs from a typical donor (D: 54024).

Additional studies assessed the ability of PD-1xLAG-3 bispecific molecules: DART B and DART I; anti-PD-1 antibodies: anti-PD-1 MAT A IgG4 (P), and anti-hPD-1 MAT 7(1.2) IgG4(P); anti-LAG-3 antibodies: anti-LAG-3 MAT A IgG4 (P), anti-hLAG-3 MAT 1(1.4) IgG4(P), and anti-hLAG-3 MAT 6(1.1) IgG4 (P); and combinations of anti-PD-1 A IgG4(P) MAb and anti-LAG-3 A IgG4(P) MAb, anti-hPD-1 7(1.2) IgG4(P) MAb and anti-hLAG-3 1(1.4) IgG4(P) MAb, and anti-hPD-1 7(1.2) IgG4(P) MAb and anti-hLAG-3 6(1.1) IgG4(P) MAb, to enhance cytokine release through checkpoint inhibition. In these studies, the PD-1xLAG-3 bispecific molecules and antibodies were used at concentrations of 0.0061, 0.024, 0.09, 0.39, 1.56, 6.25, or 25 nM, and PBMCs were stimulated with 0.5 ng/ml SEB. For these studies, where a combination of antibodies was used, each antibody was provided at the indicated concentration, and accordingly, the total antibody concentration was twice the concentration used for each antibody (i.e., 0.0122, 0.048, 0.18, 0.78, 3.12, 12.5, or 50 nM). Figure 20 shows the IFN-γ secretion profiles from SEB-stimulated PBMCs obtained from a representative donor D: 56041.

The ability of the PD-1xLAG-3 DART I bispecific molecule; a combination of anti-PD-1 antibody, anti-PD-1 A IgG4 mAb and anti-LAG-3 antibody, anti-LAG-3 A IgG4 mAb (P); and a negative control antibody, to enhance antigen-specific T cell responses was investigated using a quantitative tetanus toxoid secondary immune response assay. Specifically, the antigen-specific enhanced cytokine secretion response was measured using tetanus toxoid as the secondary immune response induction antigen in a co-culture-based quantitative assay system. Briefly, memory CD4 T cells (0.5-1.0 x 105 cells/well) were isolated using negative selection-based isolation kits (Miltenyi Biotec, San Diego, CA, and Invitrogen, Carlsbad, CA, USA) from human peripheral blood and cultured for 5-7 days with irradiated monocytes (0.01-0.05 x 105 cells/well, 3500 rad) from the same donor in the presence or absence of 5 μg/ml tetanus toxoid antigen for secondary immune response (TTd) induction and dilutions (starting at 25 nM) of DART I, anti-PD-1 A IgG4 MAb+anti-LAG-3 A IgG4(P) MAb, or isotype control antibody. In parallel plates, proliferation was measured by tertiary thymidine incorporation and IL-2 and IFN-γ were measured by ELISA (R&D systems, Minneapolis, MN) on days 5–7. Figures 21A–D show the secretion profiles of IFN-γ (Figs. 21A, 21C) and IL-2 (Figs. 21B, 21D) on day 7 for two representative donors (D50702 and D54267).

The results of these studies demonstrate that PD1xLAG-3 bispecific molecules significantly enhanced the production of IFN-γ (Figs. 16A-16B, 17A-17B, 18A-18B, 20) and IL-2 (Fig. 19) from SEB-stimulated PBMCs upon restimulation. In addition, PD-1xLAG-3 bispecific molecules significantly enhanced the production of IFN-γ (Figs. 21A and 21C) from tetanus toxoid-stimulated memory CD4 T cells. In particular, the tetravalent PD-1xLAG-3 bispecific molecules provided a more significant enhancement than the combination of anti-PD-1 and anti-LAG-3 antibodies.

- 108 047726

Example 9. Pharmacokinetics of bispecific PD-1xLAG-3 molecules.

The pharmacokinetics of a candidate PD-1xLAG-3 bispecific molecule, DART I, and a candidate PD-1 antibody, anti-PD-1A mAb, were studied in cynomolgus monkeys. Briefly, two cynomolgus monkeys (one male and one female) were infused with a single dose of DART I (5 mg/kg) or anti-PD-1A mAb (10 mg/kg), and serum concentrations of the molecules were monitored over time using a sandwich ELISA. Briefly, 96-well Maxi-Sorb™ quantitative assay plates were coated with soluble human PD-1 (shPD-1), blocked with bovine serum albumin, washed, and incubated with calibration standards, quality control standards, and diluted serum samples. The amount of captured DART I and anti-PD-1A mAb was estimated by sequential addition of a secondary biotinylated goat anti-human IgG Fc antibody and streptavidin-horseradish peroxidase conjugate (SA-HRP). HRP activity was determined using TMB substrate. All samples were analyzed in a microplate reader (SpectraMax M2e, Molecular Devices, Sunnyvale, CA, USA), and the OD signals emitted by the calibration standards were used in a four-parameter logistic model using SoftMax Pro software (version 5.4, Molecular Devices). Concentrations of anti-PD-1A or DART I mAb were determined based on interpolation of the OD signal data using the standard curve equation. The estimated lower limit of quantification (LLOQ) for this assay was 9.775 ng/mL. 22 shows the serum concentration-time curve, lines represent the mean for male (filled symbols) and female (open symbols) monkeys infused with DART I (solid line, triangles) or anti-PD-1A mAb (dashed line, circles). These data suggest that the pharmacokinetics of the bispecific PD-1xLAG-3 molecule are comparable to those of the anti-PD-1 antibody in cynomolgus monkeys.

Example 10. Toxicology study of antibodies to PD-1 and bispecific PD1xLAG-3 molecules.

The safety profile of a representative anti-PD1 antibody, anti-hPD-1 7(1.2) IgG4 (P) mAb, and a representative PD1xLAG-3 bispecific molecule, DART I, was assessed in a non-GLP (Good Laboratory Practice) dose-finding study in cynomolgus monkeys.

This study assessed the potential toxicity and toxicokinetics of an anti-PD-1 antibody (anti-hPD-1 mAb 7(1.2)IgG4(P)) administered via multiple intravenous infusions. In addition, the potential toxicity and pharmacokinetics of the DART molecule PD-1xLAG-3 (DART I) were assessed when administered via a single intravenous infusion. The study design is presented in Table 20.

Table 20

Group No. Research material Dose level (mg/kg) Dosing days Volume of dose Dose (mg/ml) Number of animals Males Females 1 Control 0 1, 8, 15 5 0 1 ^a 1 ^a 2A MAT to hPD1 7(1.2)IgG4 (P) 1 1, 8, 15 5 0,2 1 ^a 1 ^a 2B MAT to hPD-1 7(1.2)IgG4 (P) 1 1, 8, 15 5 0,2 1 ^ь 1 ^ь FOR MAT to hPD-1 7(1.2)IgG4 (P) 100 1, 8, 15 5 20 1 ^a 1 ^a ZV MAT to mPD1 7(1.2)IgG4 (P) 100 1, 8, 15 5 20 1 ^ь 1 ^ь 4 DART I 5 1 5 1 I ^{e 1st}

a In groups 1, 2A and 3A, doses were administered starting on day 1 and necropsy was performed 72 h after the last (third) dose on day 18.

^b In groups 2B and 3B, doses were administered starting on day 1 and necropsy was performed 7 days after the last (third) dose on day 22.

^c In Group 4, the dose was administered on day 1 and animals were then observed for 28 days after the single dose (until day 29); animals were then returned to the colony.

- 109 047726

The following parameters and endpoints were assessed in this study: clinical signs, body weight, food intake, body temperature, clinical pathology parameters (hematology, coagulation, and clinical chemistry), bioanalytical and toxicokinetic parameters, drug antibody levels, flow cytometry data, cytokine levels, macroscopic autopsy results, organ weights, and histopathology results.

All animals survived to scheduled euthanasia on days 18 or 22 or were discontinued from the study on day 29. For the anti-hPD-1 7(1.2)IgG4(P) mAb, there were no study product-related changes in clinical signs, food intake, body weight, body temperature, hematology, coagulation, or clinical chemistry parameters, or gross necropsy findings. On days 18 and 22, animals receiving the anti-hPD-1 7(1.2)IgG4(P) mAb at 1 or 100 mg/kg developed enlarged spleen weights and a dose-dependent mild to moderate lymphohistiocytic infiltration of the red pulp. Compared with surrounding lymphocytes, the lymphohistiocytic cells had clear cytoplasm and irregular nuclei. Rare mitotic figures were evident. At the microscopic level, the infiltrate correlated with an increase in spleen mass.

Serum concentration-time profiles for animals treated with the hPD-1 7(1.2) IgG4 (P) mAb exhibited the profile expected for an antibody in this species, with some exceptions. The slope of the curves after the third dose declined more steeply than after the first dose for two animals in the 1 mg/kg dose group and two animals in the 100 mg/kg dose group, indicating the possible emergence of anti-drug antibodies (ADA) in later cycles. Analysis showed that 2/4 of the animals developed ADA in the 1 mg/kg group and 1/4 of the animals developed ADA in the 100 mg/kg group.

In conclusion, administration of anti-hPD-1 mAb 7(1.2)IgG4(P) by intravenous infusion once weekly for 3 weeks (days 1, 8, and 15) was well tolerated in cynomolgus monkeys at dose levels of 1 and 100 mg/kg. A dose-dependent mild to moderate lymphohistiocytic cell infiltrate of the splenic red pulp was present at doses of 1 and 100 mg/kg anti-hPD-1 mAb 7(1.2)IgG4(P).

For DART I, there were no changes in clinical signs, food intake, body weight, body temperature, hematology, or coagulation parameters. DART I-related changes in clinical chemistry parameters included unfavorable, transient increases in aspartate aminotransferase (AST) and lactate dehydrogenase (LDH) on day 2. The mean change in AST activity was 3.2-fold greater than that in vehicle-treated controls and 7.8-fold greater than pre-study levels, with levels above the control reference range. The mean change in LDH activity was 2.5-fold greater than that in vehicle-treated controls and 6.9-fold greater than pre-study levels. Both parameters returned to baseline levels by day 8. In conclusion, a single administration of DART-I by intravenous infusion was well tolerated in cynomolgus monkeys at the 5 mg/kg dose level.

Example 11. Study of the pharmacokinetics of antibodies to PD-1 after administration of a single dose.

A single dose pharmacokinetic (PK) study with selected toxicology endpoints was conducted in cynomolgus macaques. In this study, anti-hPD-1 7(1.2) IgG4 (P) mAb was compared with two other anti-PD1 IgG4 (P) antibodies: anti-PD-1 A IgG4 (P) mAb and anti-PD-1 B IgG4 (P) mAb. Each antibody was administered to two monkeys (1 male, 1 female) at 10 mg/kg by intravenous infusion over 1 hour and the animals were observed for 65 days. There were no study product-related changes in clinical signs, body weight, food intake, cytokine release, or immunophenotyping associated with administration of anti-hPD-1 7(1.2) IgG4 (P) mAb or anti-PD-1 A IgG4 (P) mAb. The data obtained were similar to those for MAb to PD-1 B IgG4 (P), except that an increase in IL-5 levels was observed with the administration of MAb to PD-1 B IgG4 (P).

Anti-PD-1 antibody binding to PD-1 on T cells was determined by flow cytometry using a competitive binding assay that assessed the mean fluorescence intensity (MFI) of fluorescently labeled anti-hPD-1 7(1.2) IgG4(P) mAb binding to T cells in the absence (control, PBS) or presence of excess competitor (unlabeled anti-hPD-1 7(1.2) IgG4(P) mAb) for blood samples collected at all time points from cynomolgus monkeys treated with anti-hPD-1 7(1.2) IgG4(P) mAb, anti-PD-1 A IgG4(P) mAb, or anti-PD-1 B IgG4(P) mAb. As shown in Fig. 23A-23C, anti-hPD-1 7 (1.2) IgG4 (P) MAb and anti-PD-1 B IgG4 (P) MAb demonstrated durable binding to PD-1 on CD4+ and CD8+ T cells (PD-1 binding was sustained at >80% for 28 days or more) (Figs. 23A and 23C, respectively) compared to anti-PD-1 A IgG4 (P) MAb (PD-1 binding was sustained at >80% for 21 days or less) (Fig. 23B). For each anti-PD-1 antibody, PD-1 binding data on T cells correlated with their serum concentration.

- 110 047726

Example 12. Repeated-dose toxicity study.

To evaluate the safety, toxicokinetic, and pharmacodynamic profile of the therapeutic molecules of the present invention, an exemplary molecule (anti-hPD-1 7(1.2) IgG4(P) mAb) was administered to cynomolgus monkeys and a non-GLP dose-finding study was performed. In this study, four groups of animals (10 per group, 5 males and 5 females) received anti-hPD-1 7(1.2) IgG4(P) mAb or control drug once weekly by infusion at 3 dose levels. Animals were assessed for any potential toxicity during the 4-week drug dosing period and then for an additional 10-week drug-free period. The experimental design of this study is presented in Table 1. 21. Animals were dosed once weekly by intravenous infusion over 1 h using a calibrated infusion pump on study days 1, 8, 15, and 22. One male and one female from each group were sacrificed on day 25, and the remaining animals were sacrificed on study day 95. The effects of hPD-1 7(1.2)IgG4(P) mAb administration on circulating leukocyte subsets, including PD-1 receptor occupancy on T lymphocytes, were assessed. Anti-drug antibody (ADA) profiles were also determined.

Table 21

Group No. The material under study ^{is a} Dose level (mg/kg) Dose volume (ml/kg) Dose (mg/ml) Number of animals ¹⁵ Main research Recovery Study M F M F 1 Control 0 5.88 0 3 3 2 2 2 MAT to 4PD-1 7 (1.2)IgG4(P) 10 5.88 1,7 3 3 2 2 3 MAT to 4PD-1 7 (1.2)IgG4(P) 40 5.88 6.8 3 3 2 2 4 MAT to 4PD-1 7 (1.2)IgG4(P) 150 5.88 25,5 3 3 2 2

^a Control and MAb to hPD-1 7 (1.2) IgG4 (P) were administered weekly by intravenous infusion.

Six monkeys (3 males/3 females) per group were necropsied on day 25, and the remaining monkeys in the recovery group (2 males/2 females) were necropsied on day 95.

Cynomolgus macaques tolerated weekly intravenous (i.v.) infusions of anti-hPD-1 7(1.2)IgG4(P) mAb at doses of 0, 10, 40, and 150 mg/kg well, and all animals survived to scheduled euthanasia on days 25 or 95. No anti-hPD-1 7(1.2)IgG4(P) mAb-related changes in clinical signs, food intake, body weight, physical, ophthalmologic, and neurologic examination parameters, electrocardiologic parameters, body temperature, respiratory rate, blood pressure, and pulse rate, coagulation parameters, clinical chemistry, urinalysis, organ weights, or gross necropsy findings were detected. Anti-hPD-1 7(1.2)IgG4(P) mAb-related changes in hematologic parameters included a transient decrease in lymphocyte titers. Lymphocyte titers were modestly decreased from baseline (1 day predose) levels by day 2 (23 hours post-infusion) in males and females at doses >10 mg/kg, and statistically significant in males at 10 and 40 mg/kg and in females at 40 and 150 mg/kg compared with controls. Lymphocyte titers returned near baseline levels by day 8 predose but were modestly decreased in some males and females at all dose levels (0.47-0.68-fold relative to baseline) by day 9 (23 hours post-infusion). Lymphocyte titers increased predose on days 15 and 22 but were decreased in some males and females (0.36-0.54-fold relative to baseline) on days 16 and 23 (23 hours post-infusion).

Dose-dependent, transient reductions in circulating immune cell populations, including total leukocytes, T cells, B cells, and NK cells, were observed 23 h after the end of infusion in animals receiving the anti-hPD-1 7(1.2)IgG4(P) mAb compared with controls. The greatest magnitude of change was observed after the first dose on day 1; smaller magnitude changes were briefly observed after subsequent doses on days 8, 15, and 22. Immune cell populations generally recovered to or near baseline values by 72 h after the end of infusion and throughout the recovery phase. No changes in circulating monocytes were observed in animals receiving the anti-hPD1 7(1.2)IgG4(P) mAb compared with controls.

Maximal binding of mAb to hPD-1 7 (1.2) IgG4 (P) to PD-1+/CD4+ and PD-1+/CD8+ cells was observed during the treatment phase of the mAb to hPD-1 7 (1.2) IgG4 (P) study at all doses tested (10, 40 or 150 mg/kg). During the recovery period, animals that did not develop

- 111 047726 anti-drug antibodies (ADA), serum hPD-1 mAb 7 (1.2) IgG4 (P) concentrations remained above 29 μg/mL, and peak PD-1+/CD4+ and PD-1+/CD8+ T cell binding was maintained throughout the 10-week recovery period. There was no evidence of PD-1 modulation on T cells in these animals. During the recovery period, MGD012-bound PD-1+ T cell frequencies decreased to baseline in animals that developed ADA responses. The decrease from maximal binding of anti-hPD-1 7(1.2)IgG4(P) mAb to PD-1+/CD4+ and PD1+/CD8+ cells in ADA-positive animals typically occurred when the apparent serum concentration of anti-hPD-1 7(1.2)IgG4(P) mAb decreased below approximately 25 μg/mL. However, it is unknown whether this apparent threshold ratio applied to ADA-negative animals, as the presence of ADA in ADA-positive animals may contribute to blocking binding of anti-PD-1 antibodies to PD-1.

Gender-related differences in the pharmacokinetic responses of mAb to hPD-1 7(1.2)IgG4(P) were minimal and linear across the dose range studied (10-150 mg/kg). For the hPD-1 7(1.2) IgG4 (P) mAb at 10, 40, and 150 mg/kg, the combined mean C _max for both sexes was 240 μg/mL (0.240 mg/mL), 1078 μg/mL (1.08 mg/mL), and 3938 μg/mL (3.94 mg/mL), and the AUC values were 47,310 h-μg/mL (47.3 h-μg/mL), 205,723 h-μg/mL (206 h-μg/mL), and 745,681 h-μg/mL (746 h-μg/mL), respectively. The mean first-cycle noncompartmental clearance (NCA) of anti-hPD-1 7(1.2) IgG4 (P) mAb before ADA was 0.21 mL/h/kg, which is significantly lower than the glomerular filtration rate in cynomolgus monkeys, as expected for a high-molecular-weight protein. The mean steady-state volume of distribution by first-cycle NCA of anti-hPD-1 7(1.2) IgG4 (P) mAb was 68 mL/kg, which is approximately 1.5-fold greater than the serum volume but less than the extracellular aqueous space. This suggests that anti-hPD-1 7(1.2) IgG4 (P) mAb leaks from the vasculature into the tissue extracellular space, but not all of the extracellular space is accessible to this molecule. The mean mean residence time (MRT) from the first cycle NCA of anti-hPD-1 7(1.2) IgG4 (P) mAb was 335 hours or approximately 14 days. The introduction of ADA decreased anti-hPD-1 7(1.2) IgG4 (P) mAb concentrations in cycles 2-4. Evidence of a decrease in serum anti-hPD-1 7(1.2) IgG4 (P) mAb concentrations following repeated dosing of anti-hPD-1 7(1.2) IgG4 (P) mAb was observed in 7/10, 4/10, and 3/10 animals in the 10, 40, and 150 mg/kg dose groups, respectively. The presence of ADA to the anti-hPD-1 mAb 7(1.2)IgG4(P) was confirmed in 4, 2, and 1 of these animals in the 10, 40, and 150 mg/kg dose groups, respectively; all animals in which ADA was not confirmed were in the terminal necropsy group, in which serum anti-hPD-1 mAb 7(1.2)IgG4(P) concentrations likely prevented ADA detection. Accordingly, in subsequent toxicokinetic studies, if the trough concentration was lower than the previous trough concentration, the data from that time point onwards were re-evaluated. Based on the results of two-compartment modeling of data across all cycles for the 3 dosing groups excluding time points affected by ADA, the mean values for the primary TC parameters for the two-compartment model were 0.22 mL/h/kg for clearance, 38.5 mL/kg for initial volume of distribution (V1), and 33.8 mL/kg for _V2 , resulting in a mean steady-state volume of distribution ( _Vss ) of 72.3 mL/kg and an MRT of 329 h. These values are consistent with those derived from the first-dose NCA. The modeling results suggest that, in the absence of ADA, with weekly dosing in cynomolgus monkeys, steady-state would be reached after the 5th dose, with an accumulation index of 2.4.

At day 25, minimal multifocal perivascular mononuclear cell infiltrates associated with anti-hPD-1 mAb 7(1.2)IgG4(P) were present in the superficial dermis at the IV injection site in males at doses >40 mg/kg and in females at doses >10 mg/kg and were the expected response to repeated injection of foreign protein (monoclonal antibody). At day 95, no microscopic changes associated with anti-hPD-1 mAb 7(1.2)IgG4(P) were noted, indicating recovery of the study device-related change observed at day 25.

In conclusion, the results of this study indicate that administration of the anti-hPD-1 7(1.2)IgG4(P) mAb by intravenous infusion once weekly (days 1, 8, 15, and 22) was clinically well tolerated in cynomolgus monkeys at dose levels of 10, 40, or 150 mg/kg. The observed effects were limited to transient decreases in circulating lymphocytes and minimal injection site changes associated with the injection of foreign protein. Based on these results, a no-observed-adverse-effect level (NOAEL) of 150 mg/kg was considered (combined mean C _max for both sexes 3.94 mg/mL and AUC 746 h-mg/mL).

All publications and patents mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication or patent application was specifically and individually indicated to be incorporated by reference in its entirety. Although the present invention has been described with respect to specific embodiments, it is to be understood that it can be further modified, and it is intended that the present application include any variations, uses, or adaptations thereof.

- 112 -

Claims (31)

the present invention, following in general the principles of the present invention and including such departures from the present invention as come within known or customary practice in the art to which the present invention pertains and which can be applied to the essential features set forth above. CLAUSE OF THE INVENTION 1. A multispecific molecule capable of binding a human PD-1 epitope and one or more additional epitopes, wherein said multispecific molecule comprises a human PD-1 binding domain comprising a heavy chain variable domain and a light chain variable domain, wherein: said heavy chain variable domain comprises a CDRH1 domain, a CDRH2 domain and a CDRH3 domain, and said light chain variable domain comprises a CDRL1 domain, a CDRL2 domain and a CDRL3 domain, where: (a) (1) said CDRH1 domain, CDRH2 domain and CDRH3 domain comprise the amino acid sequences of: SEQ ID NO: 139, SEQ ID NO: 140 and SEQ ID NO: 141, respectively; and (2) said CDR _L 1 domain, CDR _L 2 domain and CDR _L 3 domain comprise the amino acid sequences of: SEQ ID NO: 157, SEQ ID NO: 145 and SEQ ID NO: 146, respectively; or (b) (1) said CDRH1 domain, CDRH2 domain and CDRH3 domain comprise the amino acid sequences of: SEQ ID NO: 139, SEQ ID NO: 140 and SEQ ID NO: 141, respectively; and (2) said CDRL1 domain, CDRL2 domain and CDRL3 domain comprise the amino acid sequences of: SEQ ID NO: 157, SEQ ID NO: 158 and SEQ ID NO: 145, respectively; or (c) (1) said CDRH1 domain, CDRH2 domain and CDRH3 domain comprise the amino acid sequences of: SEQ ID NO: 139, SEQ ID NO: 140 and SEQ ID NO: 141, respectively; and (2) said CDR _L 1 domain, CDR _L 2 domain and CDR _L 3 domain comprise the amino acid sequences of: SEQ ID NO: 144, SEQ ID NO: 145 and SEQ ID NO: 146, respectively. 2. A multispecific molecule according to claim 1, characterized in that: (a) said human PD-1 binding domain comprises a heavy chain variable domain comprising the amino acid sequence set forth in SEQ ID NO: 147 or SEQ ID NO: 149; or (b) said human PD-1 binding domain comprises a light chain variable domain comprising the amino acid sequence set forth in SEQ ID NO: 151, SEQ ID NO: 153 or SEQ ID NO: 155. 3. A multispecific molecule according to claim 1, characterized in that: (a) said CDRH1 domain, CDRH2 domain and CDRH3 domain of said PD-1 binding domain comprise the amino acid sequences of: SEQ ID NO: 139, SEQ ID NO: 140 and SEQ ID NO: 141, respectively; and (b) said CDR _L 1 domain, CDR _L 2 domain and CDR _L 3 domain of said PD-1 binding domain comprise the amino acid sequences of: SEQ ID NO: 157, SEQ ID NO: 145 and SEQ ID NO: 146, respectively. 4. A multispecific molecule according to claim 1, characterized in that: (a) said human PD-1 binding domain comprises a heavy chain variable domain comprising the amino acid sequence set forth in SEQ ID NO: 147; or (b) said human PD-1 binding domain comprises a light chain variable domain comprising the amino acid sequence set forth in SEQ ID NO: 153. 5. The multispecific molecule of claim 4, wherein said heavy chain variable domain of said PD-1 binding domain comprises the amino acid sequence presented in SEQ ID NO: 147, and said light chain variable domain of said PD-1 binding domain comprises the amino acid sequence presented in SEQ ID NO: 153. 6. A multispecific molecule according to any one of claims 1 to 5, characterized in that said molecule is a bispecific binding molecule capable of simultaneously binding to an epitope of PD-1 and to a second epitope. 7. A multispecific molecule according to claim 6, characterized in that said second epitope is an epitope of a molecule involved in the regulation of an immune response checkpoint, present on the surface of an immune cell. 8. The multispecific molecule according to claim 6, characterized in that said second epitope is an epitope of B7-H3, B7-H4, BTLA, CD40, CD40L, CD47, CD70, CD80, CD86, CD94, CD137, CD137L, CD226, CTLA-4, Galectin-9, GITR, GITRL, HHLA2, ICOS, ICOSL, KIR, LAG-3, LIGHT, MHC class I or II, NKG2a, NKG2d, OX40, OX40L, PD1H, PD-1, PD-L1, PD-L2, PVR, SIRPa, TCR, TIGIT, TIM-3 or VISTA. 9. The multispecific molecule according to claim 8, characterized in that said second epitope is an epitope of CD137, CTLA-4, LAG-3, OX40, TIGIT or TIM-3. 10. A multispecific molecule according to claim 9, characterized in that said second epitope - 113 047726 is an epitope of LAG-3. 11. A multispecific molecule according to claim 10, characterized in that said multispecific molecule additionally contains a LAG-3 binding domain: (a) (1) a CDRH1 domain, a CDRH2 domain, and a CDRH3 domain that comprise the amino acid sequences of SEQ ID NO: 57, SEQ ID NO: 58, and SEQ ID NO: 59, respectively; and (2) a CDR _I 1 domain, a CDR _L 2 domain, and a CDR _L 3 domain that comprise the amino acid sequences of SEQ ID NO: 298, SEQ ID NO: 62, and SEQ ID NO: 63, respectively; (b) (1) a CDRH1 domain, a CDRH2 domain, and a CDRH3 domain that comprise the amino acid sequences of SEQ ID NO: 42, SEQ ID NO: 43, and SEQ ID NO: 44, respectively; and (2) a CDR _I 1 domain, a CDR _L 2 domain, and a CDR _L 3 domain that comprise the amino acid sequences of SEQ ID NO: 46, SEQ ID NO: 47, and SEQ ID NO: 48, respectively; or (c) (1) a CDRH1 domain, a CDRH2 domain, and a CDRH3 domain that comprise the amino acid sequences of SEQ ID NO: 42, SEQ ID NO: 43, and SEQ ID NO: 44, respectively; and (2) a CDR _I 1 domain, a CDR _L 2 domain, and a CDR _L 3 domain that comprise the amino acid sequences of SEQ ID NO: 55, SEQ ID NO: 47, and SEQ ID NO: 48, respectively; or (d) (1) a CDRH1 domain, a CDRH2 domain, and a CDRH3 domain that comprise the amino acid sequences of SEQ ID NO: 57, SEQ ID NO: 58, and SEQ ID NO: 59, respectively; and (2) a CDR _I 1 domain, a CDR _L 2 domain, and a CDR _L 3 domain that comprise the amino acid sequences of SEQ ID NO: 61, SEQ ID NO: 62, and SEQ ID NO: 63, respectively. 12. The multispecific molecule according to claim 10, characterized in that the LAG3 binding domain comprises: (a) a CDRH1 domain, a CDRH2 domain and a CDRH3 domain that comprise the amino acid sequences of SEQ ID NO: 57, SEQ ID NO: 58 and SEQ ID NO: 59, respectively; and (b) a CDRL1 domain, a CDRL2 domain and a CDRL3 domain that comprise the amino acid sequences of SEQ ID NO: 298, SEQ ID NO: 62 and SEQ ID NO: 63, respectively. 13. The multispecific molecule according to claim 12, characterized in that said LAG-3 binding domain contains: (a) variable heavy chain VH1 MAT 6 to hLAG-3 (SEQ ID NO: 294); and (b) variable light chain VL1 MAT 6 to hLAG-3 (SEQ ID NO: 296). 14. A multispecific molecule according to any one of claims 1 to 13, characterized in that said molecule is: (a) a diabody, which is a covalently linked complex that contains two, three, four, or five polypeptide chains; or (b) a trivalent binding molecule, which is a covalently linked complex that contains three, four, or five polypeptide chains, or (c) a bispecific antibody. 15. A multispecific molecule according to claim 14, characterized in that said molecule is a diabody. 16. A multispecific molecule according to any one of claims 1 to 15, characterized in that said molecule contains an Fc region. 17. A multispecific molecule according to claim 16, characterized in that said Fc region belongs to the IgG1, IgG2, IgG3 or IgG4 isotype. 18. A multispecific molecule according to claim 16 or 17, characterized in that said molecule additionally contains a hinge domain. 19. A multispecific molecule according to claim 18, characterized in that: (a) said Fc region and hinge domain are of the IgG4 isotype, and (b) said hinge domain contains the stabilizing mutation S228P. 20. A multispecific molecule according to any one of claims 16-19, characterized in that: (a) said Fc region is a variant Fc region that comprises one or more amino acid modifications that reduce the affinity of the Fc region for FcyR, and (b) said one or more modifications that reduce the affinity of the Fc region for FcyR comprise the substitution L234A; L235A; or L234A and L235A, wherein the numbering is EU numbering as in Kabat. 21. A multispecific molecule according to any one of claims 16-20, characterized in that: (a) said Fc region is a variant Fc region that comprises one or more amino acid modifications that increase the half-life of said variant Fc region, and (b) said modifications that increase the half-life of said variant Fc region comprise the substitution M252Y; M252Y and S254T; M252Y and T256E; M252Y, S254T and T256E; or K288D and H435K, wherein the numbering is the EU numbering as in Kabat. 22. A multispecific molecule according to any one of claims 14-21, characterized in that said - 114 047726 the molecule is a diabody containing: (a) SEQ ID NOs: 290 and 291; or (b) SEQ ID NO: 267, wherein X ₁ is Ala; X ₂ is Tyr; X3 is Thr; X ₄ is Glu, and SEQ ID NO: 268; or (c) SEQ ID NO: 267, wherein X1 is Gly; X ₂ is Tyr; X3 is Thr; X ₄ is Glu, and SEQ ID NO: 268; or (d) SEQ ID NO: 267, wherein X1 is Gly; X ₂ is Met; X3 is Ser; X ₄ is Thr, and SEQ ID NO: 268; or (e) SEQ ID NOs: 269 and 270; or (f) SEQ ID NOs: 271 and 272; or (g) SEQ ID NO: 273, 274, 275 and 276; or (h) SEQ ID NO: 277, 278, 279 and 280; or (i) SEQ ID NO: 281, 282 and 283; or (j) SEQ ID NOs: 292 and 293. 23. A multispecific molecule according to claim 22, characterized in that said molecule is a diabody containing SEQ ID NO: 290 and 291. 24. A pharmaceutical composition for blocking the interaction of PD-1 and/or PD-L1 in a human subject, comprising: (A) a multispecific molecule according to any one of claims 1 to 23; and (B) a pharmaceutically acceptable carrier. 25. Use of the composition according to claim 24 for stimulating a T-cell mediated immune response in a subject in need thereof. 26. Use of the composition according to claim 24 in the treatment of a disease or condition associated with a suppressed immune system. 27. The use according to claim 26, characterized in that the said disease or condition is cancer or infection. 28. The use according to claim 27, characterized in that the cancer is selected from the group consisting of adrenal tumor, AIDS-associated cancer, alveolar soft tissue sarcoma, astrocytic tumor, bladder cancer, bone cancer, brain and spinal cord cancer, metastatic brain tumor, breast cancer, carotid artery tumor, cervical cancer, chondrosarcoma, chordoma, chromophobe renal cell sarcoma, clear cell sarcoma, colon cancer, colorectal cancer, cutaneous benign fibrous histiocytoma, desmoplastic small round cell tumor, ependymoma, Ewing's tumor, extraskeletal myxoid chondrosarcoma, imperfect bone fibrogenesis, fibrous dysplasia of bone, gallbladder or bile duct cancer, gastric cancer, gestational trophoblastic disease, germ cell tumor, cancer head and neck, hepatocellular carcinoma, islet cell tumor, Kaposi sarcoma, renal cell carcinoma, leukemia, lipoma/benign lipomatous tumor, a liposarcoma/malignant lipomatous tumor, liver cancer, lymphoma, lung cancer, medulloblastoma, melanoma, meningioma, multiple endocrine neoplasia, multiple myeloma, myelodysplastic syndrome, neuroblastoma, neuroendocrine tumor, ovarian cancer, pancreatic cancer, papillary thyroid carcinoma, parathyroid tumor, pediatric cancer, peripheral nerve sheath tumor, pheochromocytoma, pituitary tumor, prostate cancer, posterior uveal melanoma, rare hematological disorder, metastatic renal cell carcinoma, rhabdoid tumor, rhabdomyosarcoma, sarcomas, skin cancer, soft tissue sarcomas, squamous cell sarcomas, gastric cancer, synovial sarcomas, testicular cancer, thymic carcinoma, thymoma, metastatic thyroid cancer and uterine cancer. 29. The use according to claim 27, characterized in that the cancer is colorectal cancer, hepatocellular carcinoma, glioma, kidney cancer, breast cancer, multiple myeloma, bladder cancer, neuroblastoma; sarcoma, non-Hodgkin's lymphoma, non-small cell lung cancer, ovarian cancer, pancreatic cancer, rectal cancer, acute myeloid leukemia (AML), chronic myelogenous leukemia (CML), B-lineage acute lymphoblastic leukemia (B-ALL), chronic lymphocytic leukemia (CLL), hairy cell leukemia (HCL), blastic plasmacytoid dendritic cell neoplasm (BPDCN), non-Hodgkin's lymphomas (NHL) including mantle cell leukemia (MCL) and small lymphocytic lymphoma (SLL), Hodgkin's lymphoma, systemic mastocytosis, or Burkitt's lymphoma. 30. A multispecific molecule according to any one of claims 1 to 23, characterized in that said molecule contains a detectable label. 31. A nucleic acid expression vector comprising a nucleic acid molecule encoding a multispecific molecule according to any one of claims 1 to 23. - 115 -