NZ782254A - Bispecific molecules having immunoreactivity with pd-1 and ctla-4, and methods of use thereof - Google Patents

Abstract

The present invention is directed to bispecific molecules (e.g., diabodies, bispecific antibodies, trivalent binding molecules, etc.) that possess at least one epitope-binding site that is immunospecific for an epitope of PD-1 and at least one epitope-binding site that is immunospecific for an epitope of CTLA-4 (i.e., a “PD-1 x CTLA-4 bispecific molecule”). The PD-1 x CTLA-4 bispecific molecules of the present invention are capable of simultaneously binding to PD-1 and to CTLA-4, particularly as such molecules are arrayed on the surfaces of human cells. The invention is directed to pharmaceutical compositions that contain such PD-1 x CTLA-4 bispecific molecules, and to methods involving the use of such bispecific molecules in the treatment of cancer and other diseases and conditions. The present invention also pertains to methods of using such PD-1 x CTLA-4 bispecific molecules to stimulate an immune response.

Description

TITLE OF THE INVENTION: Bispecific Molecules Having Immunoreactivity with PD-1 and CTLA-4, and Methods of Use Thereof CROSS-REFERENCE TO RELATED APPLICATIONS This application is a divisional of New Zealand Application No. , filed on 1 June 2018, and is related to International Patent Application No. , filed on 12 December 2016 and claims priority to, U.S. Patent Appln. Serial No. 62/266,944 (filed: December 14, 2015; pending), which application is incorporated herein in its entirety.

REFERENCE TO SEQUENCE G This application includes one or more Sequence Listings pursuant to 37 C.F.R. 1.821 et seq., which are disclosed in computer-readable media (file name: 1301_0134PCT_ST25.txt, created on December 4, 2016, and having a size of 186,040 bytes), which file is herein incorporated by reference in its entirety.

FIELD OF THE INVENTION The present invention is directed to ific molecules (e.g., diabodies, bispecific dies, ent binding les, etc.) that possess at least one epitope-binding site that is immunospecific for an epitope of PD-1 and at least one epitope-binding site that is immunospecific for an epitope of CTLA-4 (i.e., a “PD-1 x CTLA-4 bispecific molecule”). The present invention concerns such PD-1 x CTLA-4 bispecific molecules that possess two epitope-binding sites that are immunospecific for one (or two) epitope(s) of PD-1 and two epitope-binding sites that are specific for one (or two) epitope(s) of CTLA-4. The present ion also is directed to such PD-1 x CTLA-4 ific molecules that additionally comprise an globulin Fc Region. The PD-1 x CTLA-4 bispecific molecules of the present invention are capable of simultaneously binding to PD-1 and to CTLA-4, particularly as such molecules are arrayed on the surfaces of human cells. The invention is directed to pharmaceutical compositions that contain such PD-1 x CTLA-4 bispecific molecules, and to methods involving the use of such bispecific molecules in the ent of cancer and other diseases and conditions. The present invention also ns to methods of using such PD-1 x CTLA-4 bispecific molecules to stimulate an immune response.

BACKGROUND OF THE ION 1. The Immune System Response to Cancer The mammalian immune system is lly poised to recognize and eliminate cancerous cells (Topalian, S.L. et al. (2015) “Immune Checkpoint Blockade: A Common Denominator ch to Cancer Therapy,” Cancer Cell 271450-461). In y duals, the immune system is in a quiescent state, inhibited by a repertoire of diverse inhibitory receptors and ligands. Such immune “checkpoint” pathways are important in maintaining self- tolerance (i.e., in preventing a subject from mounting an immune system attack against his/her own cells (an “autoimmune” reaction) and in ng collateral tissue damage during anti- microbial or anti-allergic immune responses. Upon recognition of a cancer antigen, detection of a microbial pathogen, or the presence of an allergen, an array of activating receptors and ligands induce the activation of the immune system. Such activation leads to the activation of macrophages, Natural Killer (NK) cells and antigen-specific, cytotoxic, T—cells, and promotes the release of various cytokines, all of which act to counter the perceived threat to the health ofthe subject (Dong, C. et al. (2003) “Immune Regulation by Novel Costimulatory Molecules,” Immunolog. Res. 28(1):39-48; tta, V. et al. (2007) “Modulating Co-Stimulation,” herapeutics 41666-675; Korman, AJ. et al. (2007) “Checkpoint Blockade in Cancer Immunotherapy,” Adv. Immunol. 90:297-339). The immune system is capable of returning to its normal quiescent state when the countervailing inhibitory immune signals outweigh the activating immune s.

Thus, the disease state of cancer (and indeed the disease states of infectious diseases) may be considered to reflect a failure to adequately activate a subject’s immune system. Such failure may reflect an inadequate presentation of activating immune signals, or it may reflect an inadequate ability to ate inhibitory immune s in the subject. In some instances, researchers have determined that cancer cells can co-opt the immune system to evade being detected by the immune system (Topalian, S.L. et al. (2015) “Immune Checkpoint Blockade: A Common Denominator Approach to Cancer y,” Cancer Cell -461).

Of particular importance is binding between the B7.1 (CD80) and B72 (CD86) ligands of the Antigen-Presenting Cell and the 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 ulatory Molecules,” Immunolog. Res. 28(1):39—48, Lindley, P.S. et al. (2009) “Ihe Clinical Utility biting CD28-Mediated ulation,” Immunol. Rev. 229:307-321), Binding of B7,] or of B72 to CD28 ates T-cell activation; binding ofB71 or B72 to CTLA-4 inhibits such activation (Dong, C. et al. (2003) “Immune Regulation by Novel Costimulatory Molecules,” Immunolog.

Res. 28(1):39-48, Lindley, P.S. et al. (2009) “The Clinical Utility OfInhibiting CD28-Mediated Costimulation,” Immunol. Rev. 229:307-321; Greenwald, R]. et al. (2005) “Ihe B7 Family Revisited,” Ann. Rev. Immunol. 23:515-548). CD28 is constitutively expressed on the surface of T-cells (Gross, J., et al. (1992) “Identification And Distribution Of Ihe 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 Trafiicking OfCTLA4 AndFocal Localization Towards Sites OfTCR Engagement,” Immunity 4:535—543). Since CTLA-4 is the higher affinity receptor (Sharpe, AH. et al. (2002) “The B 7- CD28 amily,” Nature Rev. Immunol. 2: 1 , an, S.L. et al. (2015) “Immune oint Blockade: A Common Denominator Approach to Cancer Therapy,” Cancer Cell 27:450-461), binding first initiates T-cell proliferation (via CD28) and then inhibits it (via nascent expression of CTLA-4), thereby dampening the effect when proliferation is no longer needed.

II. CTLA-4 Cytotoxic T-lymphocyte associated n—4 (CTLA-4; CD152) is a single pass type I ne protein that forms a de linked homo-dimer rtz J.C,, et al. (2001) “Structural Basis For Co—Stimulation By The Human CTLA-4/B7-2 Complex,” Nature 410:604-608). Alternate splice variants, encoding ent isoforms, have been characterized including a soluble isoform which functions as a monomer (Magistrelli G., et al. (1999) “A Soluble Form OfCTLA-4 Generated By Alternative Splicing Is sed By Nonstimulated Human T Cells,” Eur. J. l. 29:3596-3602, Oaks MK. et al. (2000) “A Native Soluble Form OfCTLA-4,” Cell Immunol 201:144-153).

CTLA-4 is primarily an intracellular antigen whose surface expression is tightly regulated by restricted trafficking to the cell e and rapid internalization (Alegre M-L, et al. AndIntracellular Expression OfCTLA4 OnMouse TCells,” , (1996) “Regulation OfSurface J. Immunol. 157:4762—4770; y, P.S. et al. (1996) “Intracellular Trayficking OfCTLA-4 AndFocal Localization Towards Sites Of TCR Engagement,” Immunity 4:53 5—543). CTLA- 4 is expressed at low levels on the surface of naive effector T-cells (Alegre, M.L., et al. (1996) “Regulation Of Surface And Intracellular Expression Of CTLA4 On Mouse T Cells,” J Immunol 157:4762-70), and constitutively expressed on T regulatory cells (Wang, X.B., et al. (2002) “Expression OfCTLA-4 By Human Monocytes,” Scand. J. Immunol. 55:53-60).

The extracellular region of CTLA-4 comprises a single extracellular Ig(V) domain, followed by a embrane (TM) region and a small intracellular cytoplasmic tail (37 amino acids). The intracellular tail ns two tyrosine-based motifs, which interact with several ellular proteins, including the lipid kinase phosphatidylinositol 3-kinase (PI3K), the phosphatases SHP-2 and PP2A and clathrin adaptor proteins AP-l and AP-2 (Rudd, C.E. et al. (2003) “Unifi/ing Concepts In CD28, ICOS And CTLA4 Co-Receptor Signalling,” Nat Rev l. 3:544-56). CTLA-4 is related to CD28, with the two proteins having approximately 29% identity at the amino acid level (Harper, K. (1991) “CTLA-4 And CD28 Activated Lymphocyte Molecules Are Closely d In Mouse AndHuman As T0 ce, Message Expression, Gene Structure, And Chromosomal Location,” J. Immunol. 147:1037- 1044) When a naive T effector cell is activated through its T-cell receptors (TCRs), CTLA-4 is recruited to the cell surface (Linsley, P.S., et al. (1996) “Intracellular Trajficking OfCTLA-4 AndFocal Localization Towards Sites OfTCR Engagement,” Immunity 4:53 5-43).

Once CTLA—4 is expressed on the T-cell surface, it competes with CD28 (constitutively expressed on T-cells) for CD80/CD86, thereby shutting off further signaling through the TCR and thus down—regulating any further T—cell response by TCR signaling (Carreno, B.M., et al. (2000) “CTLA-4 ) Can Inhibit T Cell Activation By Two Dierrent Mechanisms Depending On Its Level OfCell Surface Expression,” J l 165: 13 52-6;Chuang, E., et al. (1999) “Regulation Of Cytotoxic T cyte-Associated Molecule-4 By Src Kinases,” J Immunol 162: 1270-7). Thus, CTLA-4 acts as a negative regulator of T effector cell activation that diminishes effector function and dictates the efficacy and duration of a T-cell response (Linsley, P.S., et al. (1996) cellular Traficking Of CTLA-4 And Focal Localization Towards Sites OfTCR Engagement,” ty 43). [001 1] In addition, CTLA-4 may play a role in enhancing the negative effect of regulatory T—cells on the immune response to cancer (Tai, Y.T., et al., (2012) “Potent in vitro And in vivo Activity OfAn Fc-Engineered HumanizedAnti-HM1.24 Antibody AgainstMultiple a viaAugmentedEjfector Function,” Blood 1 4-82). CTLA-4 has a much higher ty for members of the B7 family than for CD28, and therefore its expression on a T-cell dictates a dominant negative regulation of the T-cell (Allison, J.P., et al. (1995) “Manipulation Of Costimulatory Signals To Enhance Antitumor T-Cell Responses,” Curr Opin Immunol 7 :682-6). The ism by which CTLA-4 contributes to the suppressor on of T regulatory cells is incompletely understood, but the expression of CTLA-4 on T regulatory cells enhances the suppressive function of these cells (Tai, Y.T., et al, (2012) “Potent in vitro And in vivo ty OfAn Fc-Engineered HumanizedAnti-HMI. 24 Antibody tMultiple Myeloma via Augmented Effector Function,” Blood 1 19:2074-82).

Blockage of CTLA-4 is ed to enhance T-cell responses in vitro (Walunas, T.L., et al. (1994) “CTLA-4 Can Function As A Negative Regulator Of T Cell Activation,” Immunity 1:405-413) and in vivo (Kearney, E.R., et al. (1995) “Antigen-Dependent Clonal Expansion OfA Trace Population OfAntigen-Specific CD4+ T Cells in vivo Is Dependent On CD28 Costimulation And Inhibited By CTLA-4,” J. Immunol. 155:1032-1036) and also to increase mor immunity , D.R. et al. (1996) “Enhancement OfAntitumor Immunity By CTLA-4 Blockade,” Science 271:1734-1736). Thus, blockage of CTLA—4 using anti- CTLA-4 antibodies has been proposed to provide new treatments for disease, especially human diseases where immune stimulation might be beneficial such as for treatment of cancers and infectious diseases (see, Leach, D.R., et al. (1996) “Enhancement OfAntitumor Immunity By CTLA-4 Blockade,” Science. 271:1734-1736, and PCT Publications No. WO 01/14424; WO 04). Development of blockers of CTLA-4 function has focused on the use of monoclonal antibodies such as umab (see, e. g., Hodi, F.S., et al., (2003) “Biologic Activity Of xic T Lymphocyte-Associated Antigen 4 Antibody Blockade In Previously Vaccinated Metastatic Melanoma And Ovarian Carcinoma ts,” Proc. Natl. Acad. Sci.

(USA) 100:4717-4717) and tremelimumab (Ribas, A. et al. (2005) “Antitumor Activity In MelanomaAndAnti-SelfResponses InA Phase I Trial With The Anti-Cytotoxic TLymphocyte- AssociatedAntigen 4 Monoclonal Antibody CP-675,206,” Oncologist 12: 3). 111. Programmed Death-1 (“PD-1”) Programmed Death-1 (“PD-1,” also known as “CD279”) is type I membrane protein member of the extended CD28/CTLA-4 family of T-cell regulators that broadly negatively tes immune ses (Ishida, Y. et al. (1992) “Induced sion OfPD- 1, A Novel Member Of The Immunoglobulin Gene Superfamily, Upon Programmed Cell Death,” EMBO J. 11:3887-3895; United States Patent Application Publications No. 2007/0202100, 311117, 2009/00110667; United States Patents No. 6,808,710, 550, 7,488,802, 7,635,757; 7,722,868; PCT Publication No. wo 01/14557).

The receptor-ligand ctions of the PD-l system appear to be even more complex than those of the CD28/CTLA-4 system. PD-1 is expressed on the cell surface of activated T-cells, B-cells, and monocytes (Agata, Y. et al. (1996) “Expression Of The PD-I Antigen On The Surface OfStimulatedMouse TAndB Lymphocytes,” Int. Immunol. 8(5):765- 772; Yamazaki, T. et al. (2002) “Expression OfProgrammed Death 1 Ligands By Murine T- Cells AndAPC,” J. Immunol. 169:5538-5545) and at low levels in l killer (NK) T-cells (Nishimura, H. et al. (2000) “Facilitation Of Beta Selection And Modification Of Positive Selection In The Thymus -DeficientMice,” 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 ellular region of PD-l consists of a single immunoglobulin (Ig)V domain with 23% identity to the equivalent domain in CTLA-4 (Martin-Orozco, N. et al. (2007) “Inhibitory Costimulation AndAnti-Tumor Immunity,” Semin. Cancer Biol. 17(4):288- 298). The extracellular IgV domain is followed by a transmembrane region and an intracellular tail. The intracellular tail ns two phosphorylation sites d in an immunoreceptor tyrosine-based inhibitory motif and an immunoreceptor tyrosine—based switch motif, which suggests that PD-l negatively regulates TCR signals (Ishida, Y. et al. (1992) “Induced Expression Of PD-I, A Novel Member Of The Immunoglobulin Gene Superfamily, Upon Programmed Cell Death,” E1Vfl30 J. 11:3887-3895, Blank, C. et al. (2006) “Contribution Of The PD-LI/PD-I Pathway To T-Cell Exhaustion: An Update On Implications For c Infections And Tumor n Cancer,” Immunol. Immunother. 56(5):739—745).

PD-l mediates its inhibition of the immune system by binding to B7—H1 and B7- DC (also known as PD-Ll and PD-L2, Flies, D.B. et al. (2007) “The New B 7s: Playing a Pivotal Role in Tumor ty,” J. Immunother. 30(3):251-260; United States Patents Nos. 6,803,192, 7,794,710, United States Patent Application Publication Nos. 059051, 2009/0055944, 2009/0274666, 2009/0313687, PCT Publication Nos. WO 01/39722; WO 02/086083).

B7-H1 and B7-DC are broadly expressed on the surfaces of many types of human and murine tissues, such as heart, placenta, muscle, fetal liver, spleen, lymph nodes, and thymus as well as murine liver, lung, kidney, islets cells of the pancreas and small intestine (Martin- Orozco, N. et al. (2007) “Inhibitory Costimulation AndAnti-Tumor Immunity,” Semin. Cancer Biol. 17(4):288-298). In humans, B7-H1 protein expression has been found in human endothelial cells (Chen, Y. et al. (2005) “Expression ofB7-HI in Inflammatory Renal Tubular Epithelial ” n. Exp. Nephrol. 102ze81-e92; de Haij, S. et al. (2005) “Renal Tubular Epithelial Cells te T-Cell Responses Via ICOS—L And B7—HI” Kidney Int. 68:2091-2102; Mazanet, M.M. et al. (2002) “B 7-H] Is ExpressedBy Human Endothelial Cells And Suppresses T-Cell Cytokine Synthesis,” J. Immunol. 169:3581-3588), dium (Brown, J.A. et al. (2003) “Blockade OfProgrammed Death-1 Ligands On Dendritic Cells Enhances T-Cell Activation And Cytokine Production,” J. Immunol. 170: 1257-1266), syncyciotrophoblasts (Petroff, M.G. et al. (2002) “B7 Family Molecules: Novel Immunomodulators At The Maternal-Fetal Interface,” Placenta 23:895-8101). The molecules are also expressed by nt macrophages of some tissues, by macrophages that have been activated with eron (IFN)-y or tumor necrosis factor (TNF)—u (Latchman, Y. et al. (2001) “PD-L2 Is A Second Ligand For PD-I And Inhibits T-Cell Activation,” Nat. Immunol 2:261- 268), and in tumors (Dong, H. (2003) “B7—HI Pathway And Its Role In The Evasion OfTumor Immunity,” J. Mol. Med. 81:281-287).

The interaction between B7-H1 and PD-l has been found to provide a crucial negative costimulatory signal to T and B-cells (Martin-Orozco, N. et al. (2007) “Inhibitory Costimulation AndAnti-Tumor Immunity,” Semin. Cancer Biol. 17(4):288-298) and functions as a cell death inducer a, Y. et al. (1992) “InducedExpression OfPD-I, A NovelMember Of The globulin Gene Superfamily, Upon Programmed Cell Death,” EMBO J. 11:3887-3 895, Subudhi, S.K. et al. (2005) “The Balance OfImmune Responses: Costimulation Verse Coinhibition,” J . Molec. Med. 83:193—202). More specifically, interaction between low concentrations of the PD-l receptor and the B7-H1 ligand has been found to result in the transmission of an inhibitory signal that strongly inhibits the proliferation of antigen-specific CD8+ T-cells, at higher concentrations the interactions with PD-l do not inhibit T—cell proliferation but ly reduce the tion of multiple cytokines e, AH. et al. (2002) “The B7-CD28 Superfamily,” Nature Rev. Immunol. 2:116—126). T-cell eration and cytokine production by both resting and previously ted CD4 and CD8 s, and even naive T—cells from umbilical-cord blood, have been found to be ted by soluble B7- Hl-Fc fusion proteins (Freeman, G.J. et al. (2000) “Engagement Of The PD-I Immunoinhibitory Receptor By A Novel B7 Family Member Leads To Negative Regulation Of Lymphocyte tion,” J. Exp. Med. 192: 1-9, Latchman, Y. et al. (2001) “PD-L2 Is A Second Ligand For PD-I And Inhibits T-Cell Activation,” Nature Immunol. 2:261—268; Carter, L. et al. (2002) “PD-I.'PD-L Inhibitory Pathway Aflects Both CD4(+) and CD8(+) T-cells And Is Overcome By IL-2,” Eur. J. l. 32(3):634-643; , A.H. et al. (2002) “The B 7-CD28 Superfamily,” Nature Rev. Immunol. 2: 1 16—126).

The role of B7-H1 and PD-l in inhibiting T-cell activation and proliferation has suggested that these biomolecules might serve as therapeutic targets for treatments of inflammation and cancer. Thus, the use of anti-PD—l antibodies to treat infections and tumors and to up-modulate an adaptive immune response has been proposed (see, United States Patent Application Publication Nos. 2010/0040614, 2010/0028330, 2004/0241745, 2008/0311117, 2009/0217401, United States Patent Nos. 7,521,051, 7,563,869, 7,595,048, PCT Publication 1 have been reported by Agata, T. et al. (1996) “Expression Of The PD-I Antigen On The Surface OfStimulatedMouse TAnd8 Lymphocytes,” Int. Immunol. 8(5):?65-772; and Berger, R. et al. (2008) “Phase I Safety AndPharmacokinetic Study OfCT-01 I A HumanizedAntibody Interacting With PD-I, In Patients With Advanced logic Malignancies,” Clin. Cancer Res. 14(10):3044-3051 (see, also, United States Patents No. 8,008,449 and 8,552,154; US Patent Publications No. 166281, 2012/0114648, 2012/0114649, 2013/0017199, 2013/0230514 and 2014/0044738; and PCT Patent Publication Nos. 2004/004771, W0 83174, W0 2009/014708; W0 2009/073533, 2012/145549, and However, despite all such prior advances, a need remains for ed compositions capable of more vigorously directing the body’ s immune system to attack cancer cells or pathogen-infected cells, especially at lower therapeutic concentrations and/or with reduced side effects. Although the ve immune system can be a potent defense mechanism against cancer and disease, it is often hampered by immune suppressive/evasion mechanisms in the tumor microenvironment, such as the expression of PD-l and .

Furthermore, co-inhibitory molecules expressed by tumor cells, immune cells, and stromal cells in the tumor milieu can dominantly attenuate T-cell responses against cancer cells. In on, the use of TLA-4 antibodies induces well-identified side effects ed to as “immune-related adverse events” (irAEs). IrAEs include colitis/diarrhea, dermatitis, hepatitis, endocrinopathies, and atory myopathy. These unique side effects are reported to arise due to breaking immune tolerance upon CTLA-4 blockade (Di Giacomo, A.M., et al. (2010) “The Emerging Toxicity Profiles Of Anti-CTLA-4 Antibodies Across Clinical Indications,” Semin Oncol. 37:499-507). ingly, therapies which overcome these limitations would be of great benefit.

As described in detail below, the present invention addresses this need by ing PD-l X CTLA-4 bispecifrc molecules. Such bispecifrc les are capable of binding to PD-l and CTLA-4 molecules that are present on the surfaces of exhausted and tolerant infiltrating lymphocytes and other cell types, and of thereby impairing the y of such cell-surface molecules to respond to their tive ligands. As such, the PD- 1 X CTLA-4 bispecifrc molecules of the present ion act to block PD-l- and CTLA mediated immune system inhibition, so as to promote the activation or continued tion of the immune system. These attributes permit such bispecifrc molecules to have utility in stimulating the immune system and particularly in the treatment of cancer and pathogen- associated diseases and conditions. The present invention is directed to these and other goals.

SUMNIARY OF THE INVENTION The present invention is directed to bispecifrc molecules (e.g., diabodies, bispecifrc antibodies, trivalent binding molecules, etc.) that possess at least one e-binding site that is immunospecifrc for an epitope of PD-l and at least one epitope-binding site that is immunospecific for an epitope of CTLA-4 (i.e., a “PD-l X CTLA—4 bispecific molecule”). The present invention concerns such PD-l X CTLA—4 bispecifrc molecules that possess two epitope-binding sites that are immunospecifrc for one (or two) epitope(s) of PD-l and two epitope-binding sites that are specifrc for one (or two) epitope(s) of . The present invention also is directed to such PD-l X CTLA-4 ific les that additionally comprise an immunoglobulin Fc Region. The PD-l X CTLA-4 bispecific molecules of the present invention are capable of simultaneously binding to PD-l and to CTLA-4, particularly as such molecules are arrayed on the surfaces of human cells. The ion is directed to pharmaceutical compositions that contain such PD—l X CTLA-4 bispecific molecules, and to methods involving the use of such bispecifrc molecules in the treatment of cancer and other diseases and conditions. The present invention also pertains to methods of using such PD—l X CTLA-4 bispecifrc molecules to stimulate an immune response.

In detail, the invention provides a bispecific molecule possessing both one or more epitope—binding sites capable of immunospecific binding to (an) epitope(s) of PD-l and one or more epitope-binding sites capable of immunospecific binding to (an) epitope(s) of CTLA-4, wherein the molecule comprises: (A) a Heavy Chain Variable Domain and a Light Chain Variable Domain of an antibody that binds PD-l, and (B) a Heavy Chain Variable Domain and a Light Chain Variable Domain of an antibody that binds CTLA-4; wherein the bispecific binding molecule is: (i) a diabody, the diabody being a covalently bonded complex that comprises two, three, four or five polypeptide chains; or (ii) a trivalent binding molecule, the ent binding molecule being a covalently bonded complex that comprises three, four, five, or more polypeptide chains.

The invention concerns the embodiment of such bispecific molecules, wherein the ific binding molecule ts an activity that is enhanced relative to such activity exhibited by two monospecific molecules one of which possesses the Heavy Chain Variable Domain and the Light Chain Variable Domain of the antibody that binds PD-l and the other of which possesses the Heavy Chain Variable Domain and the Light Chain Variable Domain of the antibody that binds CTLA-4.

The invention ns the embodiment of all such bispecific molecules, wherein the molecule elicits fewer immune-related e events ) when stered to a t in need thereof ve to such iREs elicited by the administration of a monospecific antibody that binds CTLA-4 such as ipilimumab.

The invention additionally concerns the embodiment of such bispecific molecules in which the molecule comprises an Fc . The invention additionally concerns the embodiment of such bispecific molecules wherein the Fc Region is a variant Fc Region that comprises: (A) one or more amino acid modifications that reduces the affinity of the variant Fc Region for an FcyR; and/or (B) one or more amino acid ations that enhances the serum half-life of the variant Fc Region.

The invention additionally concerns the embodiment of such ific molecules wherein the modifications that reduces the affinity of the variant Fc Region for an FcyR comprise the substitution of L234A; L23 5A; or L234A and L23 5A, wherein the numbering is that of the EU index as in Kabat.

The invention additionally concerns the embodiment of such bispecific molecules wherein the modifications that that enhances the serum ife of the variant Fc Region comprise the tution ofM252Y; M252Y and 8254T; M252Y and T256E; M252Y, 8254T and T256E; or K288D and H43 5K, wherein the numbering is that of the EU indeX as in Kabat.

The invention onally concerns the embodiment of all such bispecific molecules wherein the le is the diabody and comprises two epitope-binding sites capable of immunospecific binding to an epitope ofPD-l and two epitope-binding sites e of immunospecific binding to an epitope of CTLA-4.

The invention additionally concerns the embodiment of all such bispecific les wherein the molecule is the trivalent binding molecule and ses two epitope- binding sites capable ofimmunospecific binding to an epitope ofPD-l and one epitope-binding site capable of immunospecific binding to an e of CTLA-4.

The invention additionally concerns the embodiment of all such bispecific molecules wherein the molecule is capable of binding to PD-l and CTLA-4 molecules present on the cell surface.

The invention additionally concerns the embodiment of all such bispecific molecules wherein the molecule is capable of simultaneously binding to PD-l and CTLA-4.

The invention additionally concerns the embodiment of all such bispecific molecules wherein the molecule promotes the stimulation of immune cells, and particularly wherein the stimulation of immune cells results in: (A) immune cell eration; and/or (B) immune cell production and/or release of at least one cytokine; and/or (C) immune cell production and/or release of at least one lytic molecule; and/or (D) immune cell expression of at least one activation marker.

The invention additionally concerns the embodiment of all such bispecific molecules wherein the immune cell is a T-lymphocyte or an NK-cell.

The invention additionally concerns the embodiment of all such bispecific molecules wherein the epitope-binding sites capable of immunospecific binding to an epitope of PD-l comprise: (A) the VH Domain of PD-l mAb l (SEQ ID NO:47) and the VL Domain of PD- 1 mAb l (SEQ ID NO:48); or (B) the VH Domain of PD-l mAb 2 (SEQ ID NO:49) and the VL Domain of PD- 1 mAb 2 (SEQ ID NO:50); or (C) the VH Domain of PD-l mAb 3 (SEQ ID NO:51) and the VL Domain of PD- 1 mAb 3 (SEQ ID NO:52); or (D) the VH Domain of PD-l mAb 4 (SEQ ID NO:53) and the VL Domain of PD- 1 mAb 4 (SEQ ID NO:54); or (E) the VH Domain of PD-l mAb 5 (SEQ ID NO:55) and the VL Domain of PD- 1 mAb 5 (SEQ ID NO:56); or (F) the VH Domain of PD-l mAb 6 (SEQ ID NO:57) and the VL Domain of PD- 1 mAb 6 (SEQ ID NO:58); or (G) the VH Domain of PD-l mAb 6-I VH (SEQ ID NO:86) and the VL Domain of PD-l mAb 6-SQ VL (SEQ ID NO:87); or (H) the VH Domain of PD-l mAb 7 (SEQ ID NO:59) and the VL Domain of PD- 1 mAb 7 (SEQ ID NO:60); or (I) the VH Domain of PD-l mAb 8 (SEQ ID NO:61) and the VL Domain of PD- 1 mAb 8 (SEQ ID NO:62).

The invention onally concerns the embodiment of all such bispecific molecules n the epitope-binding site(s) capable of immunospecific binding to an epitope of CTLA-4 comprise: (A) the VH Domain of CTLA-4 mAb 1 (SEQ ID NO:76) and the VL Domain of CTLA-4 mAb l (SEQ ID NO:77); or (B) the VH Domain of CTLA—4 mAb 2 (SEQ ID NO:78) and the VL Domain of CTLA-4 mAb 2 (SEQ ID ; or (C) the VH Domain of CTLA-4 mAb 3 (SEQ ID NO:90) and the VL Domain of CTLA-4 mAb 3 (SEQ ID NO:91).

The invention additionally concerns the embodiment of such bispecific molecules wherein: (A) the epitope-binding sites capable of immunospecific binding to an epitope of PD-l comprise the VH Domain ofPD-l mAb 6-I VH (SEQ ID NO:86) and the VL Domain of PD-l mAb 6—SQ (SEQ ID ; and (B) the epitope-binding site(s) capable of immunospecific binding to an epitope of CTLA-4 comprise(s) the VH Domain of CTLA-4 mAb 3 (SEQ ID NO:90) and the VL Domain of CTLA-4 mAb 3 (SEQ ID NO:91).

The invention additionally concerns the embodiment of all such bispecific molecules wherein the molecule comprises: (A) two polypeptide chains having SEQ ID NO:95, and two polypeptide chain having SEQ ID NO:96; or (B) two polypeptide chains having SEQ ID NO:97, and two polypeptide chain having SEQ ID NO:98; or (C) two polypeptide chains having SEQ ID NO:99, and two polypeptide chain having SEQ ID NO:100; or (D) two polypeptide chains having SEQ ID NO:102, and two polypeptide chain having SEQ ID NO:103; or (E) two polypeptide chains having SEQ ID , and two polypeptide chain having SEQ ID NO:100; or (F) one polypeptide chains having SEQ ID NO:104, one polypeptide chain having SEQ ID NO:105, one polypeptide chain having SEQ ID NO:106, and one polypeptide chain having SEQ ID NO:107; or (CD one polypeptide chains having SEQ ID NO:108, one polypeptide chain having SEQ ID NO:105, one polypeptide chain having SEQ ID NO:109, and one polypeptide chain having SEQ ID NO:107.

The invention additionally concerns the ment of such bispecific molecules in which the molecule comprises an Albumin-Binding Domain, and especially a nized Albumin-Binding Domain.

The invention additionally concerns a pharmaceutical ition that comprises an ive amount of any of such ific molecules and a pharmaceutically acceptable carrier.

The invention additionally concerns the use of such ceutical composition or the use of any of the above-described bispecific molecules to promote stimulation of an immune-mediated response of a subject in need thereof, and in particular, wherein such molecule promotes the stimulation of immune cells, and in ular, stimulation of NK-cells and/or T-lymphocytes. The invention particularly concerns the embodiments wherein such stimulation results in immune cell proliferation, immune cell production and/or release of cytokines (e.g., IFNy, IL-2, TNFOL, etc), immune cell production and/or release of lytic molecules (e.g., granzyme, perforin, etc), and/or immune cell sion of activation markers (e.g., CD69, CD25, CD107a, etc). The invention further concerns methods of treating cancer or other diseases that involve the use or stration of any of the described PD-l X CTLA-4 bispecific molecules to stimulate an immune mediated response. The invention particularly concerns the embodiments in which the immune stimulatory activity of any of the above-described PD-l X CTLA-4 bispeciflc molecules is more potent than the joint or combined administration of a separate anti-PD—l antibody and a separate anti-CTLA-4 antibody ially, wherein such antibodies are monospeciflc for such molecules). The invention also concerns embodiments in which immune cells, particularly NK-cells and/or T- lymphocytes, stimulated by the above-described PD-l X CTLA-4 bispeciflc molecules exhibit enhanced proliferation, altered tion and/or release of cytokines (e.g., IFNy, IL-2, TNFOL, etc), altered production and/or release of lytic les, and/or d expression of activation markers relative to that exhibited by such cells stimulated by the joint or combined administration of a separate anti-PD-l antibody and a separate anti-CTLA-4 antibody. The invention also concerns embodiments in which the above-described PD-l X CTLA-4 ific molecules have a reduced incidence of irAEs. The invention additionally concerns the embodiments in which any of the described PD-l X CTLA-4 bispecific molecules are used in the treatment of a disease or condition ated with a ssed immune system, especially cancer or an infection.

The invention additionally concerns such a use to treat a disease or ion associated with a suppressed immune system, or in the treatment of such a disease or condition.

The invention particularly concerns such a use in in the treatment of a disease or condition associated with a suppressed immune system, or wherein the disease or condition is cancer or an infection cularly, an ion characterized by the ce of a bacterial, fungal, viral or protozoan pathogen).

The invention particularly concerns such a use wherein: (A) the use is in the treatment of cancer, and the cancer is characterized by the ce of a cancer cell selected from the group ting of a cell of: an adrenal gland tumor, an AIDS-associated cancer, an alveolar soft part sarcoma, an astrocytic tumor, bladder cancer, bone cancer, a brain and spinal cord cancer, a metastatic brain tumor, a breast cancer, a carotid body tumors, a al cancer, a chondrosarcoma, a chordoma, a chromophobe renal cell carcinoma, a clear cell carcinoma, a colon cancer, a ctal cancer, a cutaneous benign fibrous cytoma, a desmoplastic small round cell tumor, an ependymoma, a Ewing’s tumor, an extraskeletal myxoid chondrosarcoma, a fibrogenesis imperfecta ossium, a fibrous dysplasia of the bone, a gallbladder or bile duct cancer, gastric cancer, a gestational trophoblastic disease, a germ cell tumor, a head and neck cancer, hepatocellular carcinoma, an islet cell tumor, a Kaposi’s Sarcoma, a kidney cancer, a leukemia, a lipoma/benign lipomatous tumor, a liposarcoma/malignant lipomatous tumor, a liver cancer, a lymphoma, a lung cancer, a medulloblastoma, a melanoma, a ioma, a multiple endocrine neoplasia, a le myeloma, a ysplastic syndrome, a neuroblastoma, a neuroendocrine tumors, an ovarian cancer, a pancreatic , a papillary thyroid carcinoma, a parathyroid tumor, a pediatric cancer, a peripheral nerve sheath tumor, a hromocytoma, a pituitary tumor, a prostate cancer, a posterior uveal melanoma, a rare hematologic disorder, a renal metastatic cancer, a id tumor, a rhabdomyosarcoma, a sarcoma, a skin cancer, a issue sarcoma, a squamous cell cancer, a stomach cancer, a synovial sarcoma, a testicular cancer, a thymic carcinoma, a thymoma, a thyroid metastatic cancer, and a uterine cancer; (B) the use is in the ent of infection, and the infection is a chronic viral, ial, fungal and parasitic infection, terized the presence of 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), Bacilli, Citrobacter, Cholera, Diphtheria, Enterobacter, Gonococci, Helicobacter pylori, Klebsiella, Legionella, Meningococci, mycobacteria, Pseudomonas, Pneumonococci, rickettsia bacteria, Salmonella, Serratia, Staphylococci, ococci, Tetanus, Aspergillus (A. fumigatus, A. niger, etc.), Blastomyces dermatitidis, Candida (C. albicans, C. krusei, C. glabrata, C. tropicalis, etc.), Cryptococcus neoformans, Genus Mucorales (mucor, absidia, rhizopus), hrix schenkii, Paracoccidioides brasiliensis, Coccidioides immitis, Histoplasma capsulalum, Leptospirosis, Borrelia burgdorferi, helminth parasite (hookworm, tapeworms, fiukes, flatworms (e.g. Schislosomia), Giardia Zambia, lrichinella, Dientamoeba Fragilis, Trypcmosoma brucei, Trypcmosoma cruzz', or Leishmania donovani.

The invention particularly concerns such use in the treatment of , wherein the cancer is ctal cancer, hepatocellular carcinoma, glioma, kidney cancer, breast cancer, multiple myeloma, bladder cancer, neuroblastoma; sarcoma, non-Hodgkin’s ma, non- small cell lung cancer, ovarian cancer, pancreatic cancer, a rectal cancer, acute myeloid leukemia (AML), chronic myelogenous leukemia (CML), acute B lymphoblastic leukemia (B- ALL), chronic lymphocytic leukemia (CLL), hairy cell leukemia (HCL), blastic plasmacytoid dendritic cell neoplasm (BPDCN), non-Hodgkin’s lymphomas (NHL), including mantel cell leukemia (MCL), and small lymphocytic lymphoma (SLL), Hodgkin’s ma, systemic mastocytosis, or Burkitt’ s lymphoma.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 provides a schematic of a entative covalently bonded diabody having two epitope—binding sites composed of two polypeptide , each having an E—coil or K-coil Heterodimer-Promoting Domain (alternative Heterodimer-Promoting Domains are provided below). A cysteine residue may be present in a linker and/or in the dimer- Promoting Domain as shown in Figure 3B. VL and VH s that recognize the same e are shown using the same shading or fill pattern.

Figure 2 provides a schematic of a representative covalently bonded diabody molecule having two epitope-binding sites composed of two polypeptide chains, each having a CH2 and CH3 Domain, such that the associated chains form all or part of an Fc Region. VL and VH Domains that recognize the same epitope are shown using the same shading or fill pattern.

Figures 3A-3C provide schematics g representative covalently bonded alent diabodies having four epitope-binding sites composed of two pairs of polypeptide chains (i.e., four polypeptide chains in all). One polypeptide of each pair possesses a CH2 and CH3 , such that the associated chains form all or part of an Fc Region. VL and VH Domains that recognize the same epitope are shown using the same shading or fill pattern. The two pairs of ptide chains may be same. In such ments wherein the two pairs of polypeptide chains are the same and the VL and VH s recognize different epitopes (as shown in Figures 3A-3B), the resulting molecule possesses four e—binding sites and is bispeciflc and bivalent with t to each bound epitope. In such embodiments wherein the VL and VH Domains recognize the same epitope (e.g., the same VL Domain CDRs and the same VH Domain CDRs are used on both chains) the resulting molecule possesses four epitope-binding sites and is monospecific and alent with respect to a single epitope.

Alternatively, the two pairs of polypeptides may be different. In such embodiments wherein the two pairs of polypeptide chains are different and the VL and VH Domains of each pair of polypeptides recognize ent epitopes (as shown by the different shading and ns in Figure 3C), the resulting molecule possesses four e-binding sites and is tetraspeciflc and monovalent with respect to each bound epitope. Figure 3A shows an Fc Region-containing diabody which contains a peptide Heterodimer-Promoting Domain comprising a cysteine residue. Figure 3B shows an Fc Region-containing diabody, which contains E-coil and K-coil Heterodimer—Promoting Domains comprising a cysteine residue and a linker (with an al cysteine residue). Figure 3C, shows an Fc-Region-Containing diabody, which contains antibody CH1 and CL domains.

Figures 4A and 4B provide schematics of a entative ntly bonded diabody molecule having two epitope-binding sites composed of three polypeptide chains.

Two ofthe polypeptide chains possess a CH2 and CH3 Domain, such that the associated chains form all or part of an Fc Region. The polypeptide chains comprising the VL and VH Domain further comprise a Heterodimer-Promoting Domain. VL and VH Domains that recognize the same epitope are shown using the same shading or fill pattern.

Figure 5 provides the schematics of a representative covalently bonded diabody molecule having four e-binding sites composed of five polypeptide chains. Two of the polypeptide chains possess a CH2 and CH3 Domain, such that the associated chains form an Fc Region that comprises all or part of an Fc Region. The polypeptide chains sing the linked VL and VH Domains further se a Heterodimer-Promoting Domain. VL and VH Domains that recognize the same epitope are shown using the same shading or fill n.

Figures 6A-6F provide schematics of representative Fc Region-containing trivalent binding les having three epitope-binding sites. Figures 6A and 6B, respectively, illustrate schematically the domains of trivalent binding molecules comprising two diabody-type binding domains and a Fab-type binding domain having different domain orientations in which the diabody-type binding domains are inal or inal to an Fc Region. The molecules in Figures 6A and 6B comprise four chains. s 6C and 6D, tively, illustrate schematically the domains of trivalent binding molecules comprising two diabody-type g domains inal to an Fc Region, and a linked Fab-type binding domain, or an scFv—type binding domain. The ent binding molecules in Figures 6E and 6F, respectively illustrate schematically the domains of trivalent binding molecules comprising two diabody—type binding domains C-terminal to an Fc Region, and a Fab-type binding domain in which the light chain and heavy chain are linked via a polypeptide spacer, or an scFv—type g domain. The trivalent binding molecules in Figures 6C-6F comprise three chains. VL and VH Domains that recognize the same epitope are shown using the same shading or fill pattern.

Figure 7 illustrates the principles of the present ion by showing that an exemplary bispecific molecule (a PD-1 X LAG-3 bispecific molecule, designated as DART A) is able to stimulate cytokine production to levels higher than those observed upon the joint or combined administration of the parental anti-PD-l and anti-LAG—3 antibodies. Shown are IFNy secretion profiles of PBMCs from a representative donor, stimulated with SEB (0.5 ng/mL) and treated with the exemplary bispecific molecule (PD-1 X LAG-3 bispecific molecule DART A) or with the anti-PD-1 and anti-LAG—3 antibodies alone or in combination.

Figures 8A-8D show the results of ELISA studies measuring the binding of serially diluted binding molecules to human CTLA-4 and human PD-1. Figures 8A-8B show the g curves of CTLA-4 mAb 3 G4P, DART D, TRIDENT A or DART B to soluble hCTLAAvi-His (1 ug/mL) (Figure 8A) or hPD—l-His (1 ug/mL) (Figure 8B) that had been coated onto t plates. Goat anti-human-Fc-HRP (1 : 10,000) was ed as the ary ion molecule to detect binding. Figures 8C-8D show the results of a study on the effect of altering orientations and binding domains on binding. PD-1 X CTLA-4 bispecific molecules comprising the CTLA-4 binding domains of CTLA-4 mAb 1 (e.g., DART B) and CTLA-4 mAb 3 (e.g., DART C and DART D) were incubated in the presence of soluble human PD-l (Figure 8C) or soluble human CTLA—4-Avi-His (Figure 8D), that had been coated onto support plates. Goat anti-human-Fcy-HRP was employed as the secondary detection le to detect binding using PICO chemiluminescent substrate.

Figures 9A-9E show the results of an evaluation of the ability of DART D, TRIDENT A, PD-l mAb 6 G4P, and CTLA-4 mAb 3 G4P and a control trident (having two binding sites for RSV and one binding site for CTLA-4) to block binding ligand binding to PD- 1 and CTLA-l, alone and in combination. Blockade of PD-Ll binding to PD-l was evaluated in the ce of equal amounts of an irrelevant antigen (Figure 9A) and in the presence of equal amounts of CTLA—4 (Figure 9B), and blockade of B7-l binding to CTLA-4 was evaluated evaluated in the presence of equal amounts of an irrelevant antigen (Figure 9C) and in the ce of equal amounts of PD-l (Figure 9D) and in the presence of four fold more PD-l (Figure 9E) using an ELISA assay.

Figures 10A-10B show the s of an evaluation of the ability of DART B, DART D, TRIDENT A, the anti-CTLA-4 antibodies CTLA-4 mAb l, CTLA-4 mAb 3 G4P, and an hIgG control antibody to bind to CHO cells expressing cynomolgus monkey CTLA-4 (Figure 10A) or human CTLA-4 (Figure 10B). Binding was detected using an anti-human Fc ary dy. s 11A-11B show the results of an tion of the ability of DART C, DART D, DART E, TRIDENT A, the anti-CTLA-4 antibodies CTLA-4 mAb l, CTLA-4 mAb 3 GlAA, and the anti-PD-l antibody PD-l mAb 6 G4P to bind to Jurkat cells (which express huCTLA-4 but not PD-l on their surface). Binding of the DART and TRIDENT molecules to human CTLA-4 was detected using anti-human FC secondary Ab (FACS). Figure 11A shows the results for DART C, DART D, DART E, CTLA-4 mAb l, CTLA-4 mAb 3 GlAA, and PD-l mAb 6 G4P. Figure 11B shows the results for CTLA-4 mAb l, CTLA-4 mAb 3 GlAA, PD—l mAb 6 G4P and TRIDENT A.

Figures 12A-12B show the results of an evaluation of the ability of DART D, TRIDENT A and the anti-CTLA-4 antibodies CTLA-4 mAb l, CTLA-4 mAb 3 GlAA to block the CTLA-4 s B7-l and B7-2 in a ased assay. His-tagged derivatives of B7-l and B7-2 were incubated in the presence of the Jurkat cells and artificial antigen presenting cells (Promega). Binding of His-B7-l and -2 was ed using an anti-His antibody. The results of this tion are shown in Figure 12A (His-B7-l) and Figure 12B (His-B7-2).

Figure 13 shows the results of an evaluation of the ability of DART C, DART D, TRIDENT A, CTLA-4 mAb 3 GlAA and PD-l mAb 6 G4P to reverse CTLA-4 immune checkpoint inhibitory signal as demonstrated in a IL-2/Luc-Jurkat-CTLA-4 reporter assay by increased luciferase expression. uc-Jurkat-CTLA-4 cells were incubated in the presence of the listed binding molecules (R:S= l : 0.3) for 30 min at 37 °C, after which time artificial antigen presenting Raji cells were added and the incubation ued for 6 hours. Reversal of CTLA-4 immune oint inhibitory signal was determined by the luciferase assay.

Figure 14 shows the results of an evaluation of the y ofDART D, TRIDENT A, PD-l mAb 6 G4P, and CTLA-4 mAb 3 GlAA to bind NSO cells that express PD-l but not CTLA-4. Binding molecules were incubated in the presence of the cells and the mean fluorescence index of the cells was measured. s ISA-15B show the results of an evaluation of the ability of DART D, TRIDENT A, PD-l mAb 6 G4P, and CTLA-4 mAb 3 GlAA to block binding between PD-l and its ligands PD-Ll and PD-L2 in a cell based assay. PD—Ll-PE or PD-L2-PE was incubated in the presence of such binding molecules and their ability to bind to NSO-PD-l cells was evaluated using FACS. Figure 15A (PD-Ll); Figure 15B (PD-L2).

Figure 16 shows the results of an evaluation ofthe ability ofDART D, TRIDENT A, CTLA-4 mAb 3 GlAA, and PD-l mAb 6 G4P to block immune inhibition resulting from a PD-l / PD-Ll interaction. Binding les were ted in the presence of PD-L1+ CHO and Jurkat effector cells, and the ability of the binding molecules to block immune inhibition (by blocking the PD-l / PD-Ll interaction) was assessed by following the extent of CD3- mediated activation (as demonstrated by sed luciferase expression in the NFAT-luc/PD- l Jurkat assay; Promega).

Figure 17 shows the results of an evaluation ofthe ability ofDART D, TRIDENT A, and a negative control antibody to co—ligate PD-l and CTLA-4 in an enzyme-fragment complementation assay by DiscoverX. Aliquots of the UZOS CTLA-4(l-l95)—PK PD-l(l- A cell line #9 were plated in quadruplicate at 5,000 cells / well in DiscoverX CPS plating media on 384-well plates. Cells were allowed to attach for 4 hours at 37 oC/ 5% C02. 11 point, 1:3 dilution series of each of the binding les were then added to the PD-l — CTLA-4 cells and the DART D and TRIDENT A samples were added to the PD-l — LAG-3 cells. The plates were incubated overnight (16 hrs) at 37 °C / 5% C02. PathHunter detection reagent was added to the wells, which were then incubated for 1 hour at room temperature in the dark, and the plate was then read on an Envision luminometer.

Figure 18 shows the results of an evaluation ofthe ability ofDART D; TRIDENT A; CTLA-4 mAb 3 GlAA; PD-l mAb 6 G4P and the combinations of CTLA—4 mAb 3 G1AA/PD-1 mAb 6 G4P (Ab Combo 1) to enhance the response of a Mixed Lymphocyte Reaction. Monocyte-derived dendritic cells were ted by treating CD14+ monocytes with GM—CSF (provided at day 1 of the incubation period) and IL-4 (provided at day 7 of the tion period). At day 8 of the tion period, a MLR was set up by incubating the CD4+ T cells with the monocyte-derived dendritic cells ded at day 8 of the incubation period) and the anti-CTLA-4 and anti-PD-l binding molecules (provided at day 8 of the incubation perod). The release of IFN—y is plotted in Figure 18. Both the bispecific DART D and TRIDENT A molecules were found to enhance the MLR response to the same extent or slightly better than the combination of individual parental antibodies. The presented data comprises seven series (each relating to a different binding molecule: hIgG4 control; PD-l mAb 6 G4P; CTLA-4 mAb 3 GlAA; a combination of CTLA-4 mAb 3 GlAA/PD-l mAb 6 G4P (Ab Combo 1); DART D; TRIDENT A; and an hIgGl control; respectively from left to right); each series is ed of six s (each relating to a different concentration of the provided molecule: 0.016; 0.08; 0.4; 2; 10 or 50 nM; respectively from left to right).

Figures 19A-19D show the effect of administration of DART D; T A; CTLA-4 mAb 3 GlAA; PD-l mAb 6 G4P and the combination of CTLA-4 mAb l/PD-l mAb 1 (Ab Combo 1) on T-cell responses using a Staphylococcus aureus toxin type B (SEB) re-stimulation assay. Figures 19A-19B show cence-activated cell sorting (FACS) dot plots of the sion of PD-l vs. CTLA—l by such PBMCs in the absence (Figure 19A) or presence (Figure 19B) of SEB stimulation. Figure 19C shows the effect of the SEB stimulation on IFN—y secretion. PBMCs were stimulated with Staphylococcus aureus toxin type B (SEB) at 0.5 ng/ml for 48 hours. Cells were then harvested; washed and re- plated in 96 well plates with antibodies at various concentrations with fresh SEB for an additional 48 hours. The supernatant was then harvested and analyzed by flow cytometry ELISA for IFN—y tion. Both the bispecific DART and the TRIDENT protein showed an increase in IFN—y response that recapitulated the response observed with the combination of the individual parental mAbs. Similar results were seen in a SEB Stimulation assay in which the PBMCs were cultured with a high concentration (500 ng/mL) of SEB for 72 hours.

Presented are six series; each relating to a different binding molecule. Each series is composed of seven columns; which relate to the result obtained with 25 nM; 6.25 nM; 1.56 nM; 0.39 nM; 0.09 nM; 0.02 nM or 0.006 nM binding molecule (respectively; from left to right). Figure 19D shows the release of IL-2 for a representative donor. PBMCs were stimulated with 0.5 ng/ml SEB for 48 hours, harvested, washed and re—plated in 96-well plates with fresh SEB and either DART D, TRIDENT A, CTLA-4 mAb 3 GlAA, PD-l mAb 6 G4P or the combination of CTLA-4 mAb 3 GlAA / PD-l mAb 6 G4P (Ab Combo 1) for an additional 48 hours, and the released 1L-2 was measured. Presented are seven series, each relating to a different binding le or condition. Each series is composed of three columns, which relate to the result obtained with 0.5 nM, 5 nM or 50 nM binding molecule ctively, from left to right). When antibodies were used in combination, each antibody was added at the indicated concentration so that the total concentration of antibody added is doubled.

Figures 20A-20B show the activity of a PD-l x CTLA-4 bispeciflc molecule in a PBMC implanted NOG murine model of Graft Versus Host e (GVHD). CD3+ T cell counts were performed via FACS on study day (Figure 20A) on mice that had ed DART D at a dose of 50 mg/kg or 500 mg/kg (Figure 20A). Survival was monitored over the course of the study and is plotted as percent survival in Figure 20B.

Figures 21A-21C show serum concentration-time profiles for cynomolgus monkeys (coded using a acter alphanumeric code) that had received DART D at 50 mg/kg on days 1, 8 and 15 of the study (Figure 21A), DART D at 75 mg/kg on days 1, 8 and of the study (Figure 21B) or Trident A at 5 mg/kg on day 1 e 21C).

Figures 22A-22B show the effect of administration of DART D on absolute lymphocyte count (ALC) in treated cynomolgus monkeys. Figure 22A shows the ALC in thousands of cells/ul (th/ul). Figure 22B shows the percent change in the ALC normalized to Day 1 (D1).

Figures 23A-23B show CD4+ T cell proliferation and PD-l occupancy on T cells in lgus monkeys that had received DART D administered at 50 mg/kg (Figure 23A) or DART D stered at 75 mg/kg (Figure 23B). s 24A-24B show the effect of DART D administration on CD4+ T cell proliferation in cynomolgus monkeys that had received DART D administered at 50 mg/kg (Figure 24A) or DART D administered at 75 mg/kg (Figure 24B).

DETAILED DESCRIPTION OF THE INVENTION The t invention is directed to bispecific molecules (e.g, diabodies, bispeciflc antibodies, trivalent binding molecules, etc.) that possess at least one epitope-binding site that is immunospecific for an epitope of PD-l and at least one epitope-binding site that is specific for an epitope of CTLA-4 (i.e., a “PD-l X CTLA-4 bispecific molecule”). The t invention concerns such PD-l X CTLA-4 bispeciflc molecules that possess two epitope-binding sites that are immunospeciflc for one (or two) epitope(s) of PD—l and two epitope-binding sites that are immunospecific for one (or two) epitope(s) of CTLA-4. The present ion also is directed to such PD-l X CTLA-4 bispecific molecules that additionally comprise an immunoglobulin Fc Region. The PD-l X CTLA-4 bispeciflc molecules of the present invention are e of simultaneously binding to PD-l and to CTLA-4, particularly as such molecules are arrayed on the surfaces of human cells. The invention is directed to pharmaceutical compositions that contain such PD-l X CTLA-4 bispecific molecules, and to methods involving the use of such bispecific molecules in the treatment of cancer and other diseases and conditions. The present invention also pertains to methods of using such PD-l X CTLA-4 bispeciflc molecules to stimulate an immune response.

T-cell tion requires two distinct signals (Viglietta, V. et a]. (2007) “Modulating Co-Stimulation,” Neurotherapeutics 4:666-675, Korman, A]. et a]. (2007) “Checkpoint Blockade in Cancer Immunothempy,” Adv. Immunol. 90:297-339), The first signal is provided by a T-Cell Receptor (TCR) molecule, expressed on the surface of a T—cell, that has recognized a peptide n that has become associated with a human leukocyte antigen (HLA) sed on the surface of an n-Presenting Cell (APC). The second signal is provided by the interaction of cognate pairs of co-stimulatory ligands: B7-l and B7-2 expressed on APCs and their corresponding ors: CD28 and CTLA-4 expressed on T- cells.

The binding of B7-l and B7-2 molecules to CD28 stimulates T-cell eration and additionally induces increased expression of CTLA-4. CTLA-4 is a negative-regulator that competes with B7-l and B7-2 for g to CD28. Thus, the process responds to disease in two phases: the initial phase involves stimulating T-cell proliferation; the subsequent phase “winds down” the immune response and returns the subject to a ent immune state.

Antibodies that bind CD28 can mimic the binding of B7-l or B7-2 and thus induce or enhance T-cell effector function and the generation of tumor eradicating immunity; such antibodies are co-stimulatory. Conversely, antibodies that block CTLA-4 from binding to B7-1 and B7-2 can prevent s from returning to a quiescent state; such T-cells thus maintain a sustained proliferation that can lead to autoimmunity and the development of immune-related adverse events" (irAEs) (Wang, L. et al. (March 7, 2011) , A Novel Mouse Ig amily Ligand That vely 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 g between the B71 (CD80) and B72 (CD86) ligands of the Antigen-Presenting Cell and the CD28 and CTLA—4 ors of the CD4+ T lymphocyte (Sharpe, AH. et al. (2002) “The B7—CD28 Superfamily,” Nature Rev. Immunol. 126, Dong, C. et al. (2003) “Immune Regulation by Novel Costimulatory Molecules,” Immunolog.

Res. 28(1):39-48, Lindley, P.S. et al. (2009) “The Clinical Utility OfInhibiting CD28-Mediated ulation,” Immunol. Rev. 229:307-321). Binding ofB71 or ofB72 to CD28 stimulates T-cell activation, g of B71 or B7.2 to CTLA—4 ts such activation (Dong, C. et al. (2003) “Immune Regulation by Novel Costimulatory Molecules,” log. Res. 28(1):39- 48; Lindley, PS. et al. (2009) “The Clinical Utility Of Inhibiting CD28-Mediated Costimulation,” Immunol. Rev. 229:307-321, Greenwald, R]. et al. (2005) “Ihe B7 Family Revisited,” Ann. Rev, Immunol. 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 Traficking OfCTLA4 AndFocal Localization Towards Sites OfTCR Engagement,” Immunity 4:535—543). Since CTLA-4 is the higher affinity receptor (Sharpe, AH. et al. (2002) “The B7- CD28 amily,” Nature Rev. Immunol. 2:116-126) binding f1rst initiates T-cell proliferation (via CD28) and then inhibits it (via nascent expression of ), thereby dampening the effect when proliferation is no longer needed.

In el with the above-described interactions, a second set of receptors and binding ligands function to inhibit the immune system, thereby g as a brake to slow the CD28/B7-1/B7mediated enhancement of the immune response. This auxiliary response involves the binding of the programmed cell death-1 protein (PD-1) receptor, expressed on the surface of T—cells, to corresponding ligands: PD-Ll, expressed on Antigen—Presenting Cells (APCs) and PD-L2, expressed on epithelial cells (Chen L. et al. (2013) “Molecular Mechanisms Of T-Cell Co-Stimulation And Co-Inhibition,” Nature Reviews Immunology l3(4):227-242). In contrast to agonist antibodies that bind to CD28 to directly stimulate T-cell responses, antibodies that bind to either PD-l or PD-Ll antagonize or block PD-l/PD—Ll engagement and thus maintain T-cell activation by preventing the ry of a ve signal to the T-cell. As such, dies that bind to either PD-l or PD-Ll augment or maintain T- cell proliferation, cytotoxicity, and/or cytokine secretion. Taken together agonist antibodies, such as anti-CD28, target positive signal pathways and are therefore co-stimulators, while antagonistic antibodies, such as anti—CTLA—4 and anti-PD—l, target negative signal pathways and are called checkpoint inhibitors.

As provided above, CTLA-4 and PD-l ent the canonical checkpoint inhibitors which exert distinct inhibitory effects on T-cell tion. The PD-l X CTLA-4 bispeciflc molecules of the present invention are capable of binding to PD-l and CTLA-4 cell- surface molecules that are present on the surfaces of lymphocytes, and of thereby impairing the ability of such cell-surface molecules to respond to their respective receptors. Without being bound by by any theory or mechanism, the inventors believe that PD-l binding can e T-cell inhibition (e.g, at tumor sites and/or as a result of ion) and that CTLA-l binding can stimulate polyclonal activation and stimulation. As such, the PD-l X CTLA-4 bispeciflc molecules of the present invention are able to attenuate PD-1 and CTLA—4-mediated immune system inhibition, and promote continued immune system tion. It has been demonstrated herein that bispecific molecules which target two immunomodulatory pathways are more potent than the combination of separate antibodies. The instant invention also provides PD-l X CTLA-4 bispecific molecules having PD-l :CTLA-4 binding ratios of l : 1, 1:2, 2:2 and 2:1 which allow for full blockade of both PD-l and CTLA-4 as well as blockade that is biased toward CTLA-4 when co—eXpressed with PD-l. Accordingly, the PD-l X CTLA-4 iflc molecules of the present invention provide unexpected superiority as compared to the combination of separate anti-PD-l and anti-CTLA-4 antibodies. Additionally, the PD-l X CTLA-4 bispeciflc molecules of the t invention may provide immune stimulation with reduced risk of irAEs. 1. Antibodies and Their Binding Domains The antibodies of the t invention are immunoglobulin molecules capable of specific binding to a target, such as a carbohydrate, polynucleotide, lipid, polypeptide, eta, h at least one antigen recognition site, located in the Variable Domain of the globulin molecule. As used , the terms “antibody” and “antibodies” refer to monoclonal dies, multispecific antibodies, human antibodies, humanized antibodies, synthetic antibodies, chimeric antibodies, polyclonal antibodies, camelized antibodies, single- chain Fvs (scFv), single-chain antibodies, Fab fragments, F(ab’) fragments, disulfrde-linked bispecific Fvs (dev), intrabodies, and epitope-binding nts of any of the above. In particular, the term “antibody” includes immunoglobulin les and immunologically active fragments of immunoglobulin molecules, i.e., les that contain an epitope-binding site. Immunoglobulin molecules can be of any type (e.g., IgG, IgE, IgM, IgD, IgA and IgY), class (e.g., IgGi, Ing, IgG3, IgG4, IgAi and IgAz) or subclass. As used herein, an Fc Region is said to be of a particular IgG isotype, class or subclass if its amino acid sequence is most homologous to that isotype relative to other IgG isotypes. In addition to their known uses in diagnostics, antibodies have been shown to be useful as therapeutic agents. Antibodies are capable of specifrcally g to a ptide or n or a non-protein molecule due to the presence on such molecule of a particular domain or moiety or conformation (an “epitope”). An epitope-containing molecule may have immunogenic activity, such that it elicits an antibody production response in an animal, such molecules are termed “antigens”.

The last few decades have seen a revival of interest in the therapeutic potential of antibodies, and antibodies have become one of the leading classes of biotechnology-derived drugs (Chan, C.E. etal. (2009) “The Use OfAntiboa’ies In The Treatment OfInfectious Diseases,” Singapore Med. J. 50(7):663-666). Over 200 dy-based drugs have been approved for use or are under development.

The term “monoclonal antibody” refers to a homogeneous antibody population wherein the monoclonal antibody is comprised of amino acids (naturally occurring and non- naturally occurring) that are involved in the selective binding of an n. Monoclonal dies are highly specific, being directed against a single epitope (or antigenic site). The term “monoclonal antibody” encompasses not only intact monoclonal antibodies and full- length monoclonal antibodies, but also fragments thereof (such as Fab, Fab', F(ab’)2 Fv), single- chain (scFv), s thereof, fusion proteins comprising an antibody n, humanized monoclonal antibodies, chimeric monoclonal dies, and any other modified configuration of the immunoglobulin molecule that comprises an antigen recognition site of the required specificity and the ability to bind to an antigen. It is not intended to be limited as regards to the source of the antibody or the manner in which it is made (e.g., by hybridoma, phage selection, recombinant expression, enic animals, etc.) The term includes whole immunoglobulins as well as the fragments etc. described above under the definition of “antibody. 77 s of making onal antibodies are known in the art. One method which may be employed is the method of 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 developed in mice, rats or rabbits.

The antibodies are produced by immunizing an animal with an immunogenic amount of cells, cell extracts, or protein preparations that contain the desired epitope. The immunogen can be, but is not d to, primary cells, cultured cell lines, cancerous cells, proteins, peptides, nucleic acids, or tissue. Cells used for zation may be cultured for a period oftime (e.g., at least 24 hours) prior to their use as an immunogen Cells may be used as immunogens by themselves or in combination with a non-denaturing adjuvant, such as Ribi (see, e.g., Jennings, V.M. (1995) “Review ofSelectedAdjuvants Used in Antibody Production,” ILAR J. 37(3): 1 19- 125). In general, cells should be kept intact and preferably viable when used as immunogens.

Intact cells may allow ns to be better detected than ruptured cells by the immunized animal. Use of ring or harsh adjuvants, e.g., Freud's adjuvant, may rupture cells and therefore is discouraged. The immunogen may be administered multiple times at periodic intervals such as, bi weekly, or weekly, or may be administered in such a way as to maintain viability in 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 ed recombinantly by any means known in the art. In one embodiment, such an antibody is sequenced and the polynucleotide sequence is then cloned into a vector for expression or propagation. The sequence encoding the antibody of st may be maintained in a vector in a host cell and the host cell can then be expanded and frozen for future use. The polynucleotide ce of such antibodies may be used for genetic manipulation to te the monospecific or multispecific (e.g., bispecific, trispecific and tetraspecific) molecules of the invention as well as an affinity optimized, a chimeric antibody, a humanized antibody, and/or a caninized antibody, to e the y, or other characteristics of the dy. The l principle in humanizing an antibody involves retaining the basic sequence of the antigen-binding portion of the antibody, while swapping the non-human remainder of the antibody with human antibody ces. l antibodies (such as IgG antibodies) are composed of two Light Chains complexed with 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 Region located between the CH1 and CH2 Domains. The basic structural unit of naturally occurring immunoglobulins (e.g., IgG) is thus a tetramer having two light chains and two heavy chains, usually expressed as a glycoprotein of about 150,000 Da. The amino-terminal (“N-terminal”) portion of each chain includes a Variable Domain of about 100 to 110 or more amino acids primarily responsible for antigen recognition. The carboxy-terminal (“C-terminal”) portion of each chain defines a constant region, with light chains having a single Constant Domain and heavy chains usually having three Constant Domains and a Hinge . Thus, the structure of the light chains of an IgG molecule is n-VL-CL-c and the structure of the IgG heavy chains is n-VH-CHl-H-CHZ-CH3- c (where H is the Hinge Region, and n and c represent, tively, the N—terminus and the C- terminus of the polypeptide). The Variable Domains of an IgG molecule consist of the mentarity determining regions (CDR), which contain the residues in contact with e, and non-CDR segments, referred to as framework segments (FR), which in general maintain the structure and determine the positioning of the CDR loops so as to permit such contacting (although certain framework residues may also contact antigen). Thus, the VL and VH s have the structure CDRl-FR2-CDR2-FR3-CDR3-FR4-c. Polypeptides that are (or may serve as) the first, second and third CDR of an antibody Light Chain are herein respectively designated CDRLl Domain, CDRLZ Domain, and CDRL3 Domain. Similarly, polypeptides that are (or may serve as) the first, second and third CDR of an antibody heavy chain are herein respectively designated CDRHI Domain, CDRHZ Domain, and CDRH3 Domain. Thus, the terms CDRLl Domain, CDRL2 , CDRL3 Domain, CDRHl , CDRHZ Domain, and CDRH3 Domain are directed to polypeptides that when incorporated into a n cause that protein to be able to bind to a specific epitope regardless of whether such protein is an antibody having light and heavy chains or a diabody or a single-chain binding molecule (e.g., an scFv, a BiTe, etc), or is another type of n. Accordingly, as used herein, 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 a portion of a molecule comprising an epitope-binding fragment. An epitope-binding nt may contain 1, 2, 3, 4, 5 or all 6 of the CDR Domains of such antibody and, although e of immunospecifically binding to such epitope, may exhibit an immunospecificity, affinity or selectivity toward such epitope that differs from that of such antibody. Preferably, however, an epitope-binding fragment will contain all 6 of the CDR Domains of such antibody, An epitope-binding fragment of an antibody may be a single polypeptide chain (e.g., an scFv), or may comprise two or more polypeptide , each having an amino terminus and a carboxy terminus (e.g, a diabody, a Fab fragment, an Fabz nt, etc). Unless specifically noted, the order of domains of the protein molecules described herein is in the N—terminal to C- Terminal ion.

The invention particularly encompasses PD—l X CTLA-4 bispeciflc binding molecules comprising one, two, or more than two single-chain Variable Domain fragments ”) of an anti-PD-l antibody and one, two, or more than two single-chain Variable Domain fragments of an anti-CTLA-4 antibody. Single-chain Variable Domain fragments are made by linking Light and Heavy chain Variable Domains using a short linking peptide.

Linkers can be modified to provide additional functions, such as to permit the attachment of drugs or attachment to solid supports. The single-chain ts can be produced either recombinantly or synthetically. For synthetic production of scFv, an automated synthesizer can be used. For recombinant production of scFV, a suitable plasmid containing polynucleotide that encodes the scFV can be introduced into a suitable host cell, either eukaryotic, such as yeast, plant, insect or mammalian cells, or prokaryotic, such as E. coli. Polynucleotides ng the scFV of interest can be made by routine manipulations such as ligation of polynucleotides. The resultant scFV can be isolated using standard protein ation techniques known in the art.

The invention also particularly encompasses PD-l X CTLA-4 iflc molecules comprising humanized anti-PD-l and anti—CTLA-4 dies. The term “humanized” antibody refers to a chimeric molecule, generally ed using recombinant techniques, having an antigen—binding site of an immunoglobulin from a man species and a remaining immunoglobulin structure of the molecule that is based upon the structure and /or sequence of a human immunoglobulin. The polynucleotide sequence of the variable domains of such dies may be used for genetic manipulation to generate such derivatives and to improve the ty, or other characteristics of such antibodies. The general principle in humanizing an antibody involves retaining the basic ce of the antigen-binding portion of the antibody, while swapping the non-human remainder of the dy with human antibody sequences. There are four general steps to humanize a onal antibody. These are: (l) determining the nucleotide and predicted amino acid sequence of the starting antibody light and heavy variable domains; (2) designing the humanized antibody or caninized antibody, i.e., deciding which antibody framework region to use during the humanizing or canonizing s, (3) the actual humanizing or caninizing methodologies/techniques; and (4) the transfection and expression of the zed antibody. See, for example, US. Patents Nos. 4,816,567, 5,807,715; 5,866,692; and 6,331,415.

The antigen-binding site may se either a complete Variable Domain fused onto Constant Domains or only the complementarity determining regions (CDRs) of such Variable Domain grafted to riate framework s. Antigen-binding sites may be wild-type or modified by one or more amino acid substitutions. This eliminates the constant region as an immunogen in human individuals, but the possibility of an immune response to the foreign variable domain remains (LoBuglio, A.F. et al. (1989) “Mouse/Human Chimeric Monoclonal Antibody In Man: Kinetics And Immune Response,” Proc. Natl. Acad. Sci.

(USA) 86:4220-4224). Another approach focuses not only on providing human-derived constant regions, but modifying the variable domains as well so as to reshape them as closely as le to human form. It is known that the valiable domains ofboth heavy and light chains contain three complementarity determining regions (CDRs) which vary in response to the antigens in question and determine binding capability, flanked by four framework regions (FRs) which are relatively conserved in a given species and which putatively provide a scaffolding for the CDRs. When man antibodies are prepared with t to a particular n, the variable s can be “reshaped” or “humanized” by grafting CDRs derived from non-human antibody on the FRs present in the human antibody to be modified.

Application of this approach to various antibodies has been ed by 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: GraftingAn Antilysozyme Activity,” Science 239: 1534-1536, Kettleborough, C. A. et al. (1991) “Humanization Of A Mouse Monoclonal Antibody By CDR-Grafting: The Importance Of Framework Residues On Loop Conformation,” Protein Engineering 4:773—3783; Maeda, H. et al. (1991) “Construction Of Reshaped Human Antibodies With HIV-Neutralizing ty,” Human Antibodies oma 2:124—134; Gorman, S. D. et al. (1991) “Reshaping A Therapeutic CD4 Antibody,” Proc, Natl. Acad. Sci. (USA) 88:4181-4185, Tempest, PR. et al. (1991) “ReshapingA Human MonoclonalAntibody To InhibitHuman 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. (USA) 9-2873; Carter, P. et al. (1992) “Humanization OfAn Anti-p185her2 Antibody For Human Cancer Therapy,” Proc.

Natl. Acad. Sci. (USA) 5-4289, and Co, MS. etal. (1992) “Chimeric AndHumanized Antibodies With Specificity For The CD33 Antigen,” J. Immunol. 148:1149-1154. In some embodiments, humanized antibodies preserve all CDR sequences (for example, a humanized mouse antibody which contains all six CDRs from the mouse antibodies). In other embodiments, humanized antibodies have one or more CDRs (one, two, three, four, five, or six) which differ in sequence ve to the original antibody.

A number of“humanized” dy molecules comprising an antigen-binding site derived from a non-human immunoglobulin have been described, including ic antibodies having rodent or modified rodent Variable Domain and their associated complementarity determining regions (CDRs) fused to human constant domains (see, for example, 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. (USA) 86:4220-4224 , Shaw et al. (1987) “Characterization OfA Mouse/Human Chimeric Monoclonal Antibody (17-1A) T0 A Colon Cancer Associated Antigen,” J. Immunol, 138:4534-4538, and Brown et al. (1987) “Tumor-Specific cally Engineered /Human Chimeric Monoclonal Antibody,” Cancer Res. 47:3577-3583). Other references describe rodent CDRs grafted into a human supporting framework region (FR) prior to fusion with an appropriate human antibody Constant Domain (see, for example, Riechmann, L. et al. (1988) “Reshaping Human Antibodiesfor Therapy,” Nature 332:323-327, Verhoeyen, M. et al. (1988) pingHuman 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 inantly ed rodent framework regions. See, for example, European Patent Publication No. 519,596. These “humanized” molecules are designed to minimize unwanted immunological response towards rodent anti-human antibody molecules, which limits the duration and effectiveness of therapeutic applications of those moieties in human recipients. Other methods of humanizing antibodies that may also be utilized are disclosed by Daugherty et al. (1991) “Polymerase Chain Reaction Facilitates The Cloning, CDR-Grafting, And Rapid Expression Of A Murine onal Antibody Directed Against The CD18 ent OfLeukocyte Integrins,” Nucl. Acids Res. 19:2471-2476 and in US. Patents Nos. 377, 6,054,297, 5,997,867, and 5,866,692.

II. Fcy Receptors (FcyRs) The CH2 and CH3 s of the two heavy chains interact to form the Fc Region, which is a domain that is recognized by cellular Fc Receptors, including but not limited to Fc gamma Receptors (FcyRs). As used herein, the term “Fc Region” is used to define a C-terminal region of an IgG heavy chain. The amino acid sequence of the CH2-CH3 Domain of an exemplary human IgG1 is (SEQ ID NO:1): 231 240 250 260 270 280 APELLGGPSV FLFPPKPKDT RM SRTBfiVT CVVVDVSHED PEVKFNWYVD 290 300 310 320 330 GVEVHNAKTK NSTY RVVSVLTVLH KEYK ALPA 340 350 360 370 380 PIEKTISKAK GQPREPQVYT LPPSREEMTK NQVSLTCLVK GFYPSDIAVE 390 400 410 420 430 WESNGQPENN YKTTPPVLDS DGSFFLYSKL TVDKSRWQQG NVFSCSVMHE 440 447 ALHNHYTQKS LSLSPGE as numbered by the EU index as set forth in Kabat, wherein X is a lysine (K) or is absent.

The amino acid sequence of the CH2-CH3 Domain of an ary human IgG2 is (SEQ ID NO:2): 231 240 250 260 270 280 GPSV FLFPPKPKDT RM SRTBfiVT CVVVDVSHED PEVQFNWYVD 290 300 310 320 330 GVEVHNAKTK PRfifiQhNSTF RVVSVLTVVH QDWLNGKEYK CKVSNKGLPA 340 350 360 370 380 PIEKTISKTK GQPREPQVYT LPPSREEMTK NQVSLTCLVK GFYPSDISVE 390 400 410 420 430 PENN YKTTPPMLDS DGSFFLYSKL TVDKSRWQQG NVFSCSVMHE 440 447 ALHNHYTQKS LSLSPGE as numbered by the EU index as set forth in Kabat, wherein X is a lysine (K) or is absent.

The amino acid sequence of the CH2—CH3 Domain of an exemplary human IgG3 is (SEQ ID NO:3): 231 240 250 260 270 280 APELLGGPSV FLFPPKPKDT LMISRTPEVT SHED PEVQFKWYVD 290 300 310 320 330 GVEVHNAKTK PRfifiQYNSTF RVVSVLTVLH QDWLNGKEYK CKVSNKALPA 340 350 360 370 380 PIEKTISKTK GQPREPQVYT LPPSREEMTK NQVSLTCLVK GEYPSD AVfi 390 400 410 420 430 WESSGQPENN YNTTPPMLDS DGSFFLYSKL TVDKSRWQQG NIFSCSVMHE 440 447 ALHNRFTQKS LSLSPG§ as numbered by the EU index as set forth in Kabat, wherein X is a lysine (K) or is absent.

The amino acid sequence of the CH2—CH3 Domain of an exemplary human IgG4 is (SEQ ID NO:4): 231 240 250 260 270 280 APEFLGGPSV FLFPPKPKDT LMISRTPEVT SQED PEVQFNWYVD 290 300 310 320 330 GVEVHNAKTK PRfiflQFNSTY RVVSVLTVLH QDWLNGKEYK CKVSNKGLPS 340 350 360 370 380 S fiKTISKAK GQPREPQVYT LBPSQfifiMTK NQVSLTCLVK GEYPSD AVfi 390 400 410 420 430 WESNGQPENN YKTTPPVLDS DGSFFLYSRL WQEG NVFSCSVMHE 440 447 ALHNHYTQKS LSLSLGE as numbered by the EU index as set forth in Kabat, wherein X is a lysine (K) or is absent.

Throughout the t specification, the numbering of the residues in the constant region of an IgG heavy chain is that of the EU index as in Kabat et al., SEQUENCES OF NS OF IMMUNOLOGICAL INTEREST, 5th Ed. Public Health Service, NHl, MD (1991) t”), expressly incorporated herein by references. The term “EU index as in Kabat” refers to the numbering of the human IgG1 EU antibody. Amino acids from the Variable Domains of the mature heavy and light chains of immunoglobulins are designated by the position of an amino acid in the chain. Kabat described numerous amino acid sequences for antibodies, identified an amino acid consensus sequence for each subgroup, and assigned a e number to each amino acid, and the CDRs are fied as defined by Kabat (it will be understood that CDRHl as defined by Chothia, C. & Lesk, A. M. ((1987) “Canonical structures for the hypervariable regions of immunoglobulins,” J. Mol. Biol. 196:901-917) begins five residues earlier). Kabat’s numbering scheme is ible to antibodies not included in his dium by aligning the antibody in question with one ofthe consensus sequences in Kabat by reference to conserved amino acids. This method for ing residue numbers has become standard in the field and readily identifies amino acids at equivalent positions in different antibodies, including chimeric or humanized variants. For example, an amino acid at position 50 of a human antibody light chain occupies the equivalent position to an amino acid at position 50 of a mouse antibody light chain.

Polymorphisms have been observed at a number of ent positions within antibody constant regions (e. g., Fc positions, including but not limited to positions 270, 272, 312, 315, 356, and 358 as numbered by the EU index as set forth in Kabat), and thus slight ences between the presented sequence and sequences in the prior art can exist.

Polymorphic forms of human immunoglobulins have been haracterized. At present, 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, b3, b0, b3, b4, 5, t, g], c5, u, v, g5) (Lefranc, et al. , “The Human IgG Subclasses: Molecular Analysis OfStructure, Function And Regulation.” Pergamon, Oxford, pp. 43 -78 (1990); Lefranc, G. et al., 1979, Hum. Genet: 50, 199-211). It is specifically contemplated that the antibodies of the present invention may incorporate any allotype, isoallotype, or haplotype of any globulin gene, and are not limited to the allotype, isoallotype or ype ofthe sequences provided herein. Furthermore, in some expression systems the C-terminal amino acid residue (bolded above) of the CH3 Domain may be post-translationally removed. Accordingly, the C-terminal residue of the CH3 Domain is an al amino acid residue in the PD-l X CTLA-4 bispecific molecules of the ion. Specifically encompassed by the instant invention are PD-1 X CTLA-4 bispecific molecules lacking the C-terminal residue of the CH3 Domain. Also specifically encompassed by the t invention are such constructs comprising the C-terminal lysine residue of the CH3 Domain.

As stated above, the Fc Region of natural IgG antibodies is capable of binding to cellular Fc gamma Receptors (FcyRs). Such binding results in the transduction of activating or inhibitory signals to the immune system. The ability of such binding to result in diametrically opposing functions reflects structural differences among the different FcyRs, and in particular reflects whether the bound FcyR possesses an immunoreceptor tyrosine-based activation motif (ITAM) or an immunoreceptor tyrosine-based inhibitory motif (ITIM). The tment of different cytoplasmic enzymes to these structures dictates the outcome of the FcyR—mediated cellular responses. ontaining FcyRs include FcyRI, FcyRIIA, FcyRIIIA, and activate the immune system when bound to an Fc Region. FcyRIIB is the only tly known l ITIM-containing FcyR, it acts to dampen or inhibit the immune system when bound to an Fc . Human neutrophils express the A gene. FcyRIIA clustering via immune complexes or specific antibody linking serves to aggregate ITAMs with receptor-associated kinases which facilitate ITAM phosphorylation. ITAM phosphorylation serves as a docking site for Syk kinase, the activation of which results in the activation of downstream substrates (e. g., PI3K). Cellular activation leads to release of pro- inflammatory mediators. The B gene is expressed on B lymphocytes; its extracellular domain is 96% identical to FcyRIIA and binds IgG complexes in an indistinguishable manner.

The presence of an ITIM in the cytoplasmic domain ofFcyRIIB defines this inhibitory subclass of FcyR. Recently the molecular basis of this inhibition was established. When co-ligated along with an activating FcyR, the ITIM in FcyRIIB becomes phosphorylated and attracts the SH2 domain of the ol polyphosphate 5’—phosphatase (SHIP), which hydrolyzes oinositol messengers released as a consequence of ITAM—containing FcyR- mediated ne kinase activation, consequently preventing the influx of intracellular Ca”. Thus cross- g of FcyRIIB dampens the activating response to FcyR ligation and inhibits cellular responsiveness. B-cell activation, B-cell proliferation and dy secretion is thus aborted.

III. Bispecific Antibodies, Multispecific Diabodies and DART® Diabodies The ability of an antibody to bind an epitope of an n depends upon the presence and amino acid ce of the antibody’s VL and VH Domains. Interaction of an antibody’s Light Chain and Heavy Chain and, in particular, interaction of its VL and VH Domains forms one of the two e-binding sites of a natural dy, such as an IgG.

Natural antibodies are capable of binding to only one epitope species (116., they are monospeciflc), although they can bind multiple copies of that species (i.e., exhibiting bivalency or multivalency).

The binding domains of an antibody, and of the PD-l X CTLA-4 bispecific molecules of the t invention, bind to epitopes in an “immunospecific” . As used herein, an antibody, diabody or other epitope-binding molecule is said to “immunospecifically” bind a region of r molecule (i.e., an epitope) if it reacts or ates more ntly, more rapidly, with greater duration and/or with greater affinity with that epitope relative to alternative epitopes. For example, an antibody that immunospecifically binds to a viral epitope is an antibody that binds this viral e with r affinity, avidity, more readily, and/or with greater duration than it immunospecifically binds to other viral epitopes or non-viral epitopes. It is also understood by reading this definition that, for example, an antibody (or moiety or epitope) that immunospecifically binds to a first target may or may not specifically or preferentially bind to a second target. As such, “immunospecific g” does not necessarily e (although it can include) exclusive binding. Generally, but not necessarily, reference to binding means “immunospecific” binding. Two molecules are said to be capable of g to one another in a “physiospecific” manner, if such binding exhibits the specificity with which receptors bind to their respective ligands.

One aspect of the present invention reflects the recognition that the functionality of antibodies can be ed by generating multispecific antibody—based molecules that can simultaneously bind to one or more epitope(s) of PD-l and also one or more epitope(s) of CTLA-4. For molecules having more than one epitope-binding site specific for an epitope of PD-l, such epitopes may be identical to one another, overlapping, or distinct from one another; binding to one such epitope may compete with or not compete with g to another of such epitopes. Likewise, for molecules having more than one epitope—binding site immunospecific for an epitope of CTLA-4, such epitopes may be cal to one another, overlapping, or distinct from one another; binding to one such epitope may compete with or not compete with binding to the second of such epitopes. It is expressly contemplated that such characteristics may be independently varied to yield PD—l X CTLA-4 bispecific molecules that, for example, possess: (1) the ability to bind to two identical epitopes of PD-l and to: (a) two cal epitopes of CTLA-4, or (b) two overlapping epitopes of CTLA-4, or (c) two distinct epitopes of CTLA—4; (2) the ability to bind to two overlapping epitopes of PD-l and to: (a) two cal epitopes of CTLA-4, or (b) two overlapping epitopes of CTLA-4; or (c) two distinct es of CTLA-4, (3) the ability to bind to two distinct epitopes of PD-l and to: (a) two identical epitopes of CTLA-4; or (b) two overlapping epitopes of CTLA-4, or (c) two distinct epitopes of CTLA-4.

In order to provide les having greater capability than natural antibodies, a wide variety of recombinant bispecific antibody s have been developed (see, e. g., PCT Publication Nos. either to fuse a further epitope-binding fragment (e.g., an scFv, VL, VH, etc.) to, or within the dy core (IgA, IgD, IgE, IgG or IgM), or to fuse multiple epitope-binding fragments (e.g., two Fab nts or . Alternative formats use linker peptides to fuse an e-binding fragment (e.g, an scFv, VL, VH, etc.) to a dimerization domain such as the CH2-CH3 Domain or alternative polypeptides ( 2007/046893). PCT Publications Nos. 2010/136172 disclose a trispecific antibody in which the CL and CH1 Domains are switched from their respective natural positions and the VL and VH Domains have been diversified (WO 2008/027236, Publications Nos. WO 63427 and Domain to contain a fusion protein adduct comprising a binding domain. PCT ations Nos. WO 28797, W02010028796 and antibodies whose Fc Regions have been replaced with additional VL and VH Domains, so as to form ent binding molecules. PCT Publications Nos. WO 2003/025018 and W02003012069 se recombinant diabodies whose individual chains contain scFv Domains. PCT Publications No. are synthesized as a single polypeptide chain and then subjected to proteolysis to yield heterodimeric structures. PCT Publications Nos. 2012/162583, additional binding domains or functional groups to an dy or an antibody portion (e.g., adding a diabody to the antibody’s light chain, or adding additional VL and VH Domains to the antibody’s light and heavy chains, or adding a heterologous fusion protein or chaining multiple Fab Domains to one another).

The art has additionally noted the lity to produce diabodies that differ from such natural antibodies in being capable of binding two or more ent epitope species (i. e., ting bispeciflcity or peciflcity in addition to bivalency or multivalency) (see, e.g., Holliger et al. (1993) odies Small Bivalent AndBispecific Antibody Fragments,” Proc.

Natl. Acad. Sci. (USA) 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(1-2):90-94, Lu, D. et al. (2005) “A Fully Human Recombinant IgG-Like Bispecific Antibody T0 Both The Epidermal Growth Factor Receptor And The Insulin-Like Growth Factor Receptor For EnhancedAntitumor ty,” J. Biol. Chem. 280(20):]9665-19672; WO 02/02781 (Mertens et al.), Olafsen, T. et al (2004) “Covalent Disulfide-Linked Anti-CEA Diabody Allows Site- Specific Conjugation And Radiolabeling For Tumor Targeting Applications,” Protein Eng.

Des. Sel. 17(1):21-27, Wu, A. et al. (2001) merization OfA Chimeric Anti-CD20 Single Chain Fv-Fv Fusion Protein Is Mediated Through Variable Domain Exchange,” Protein Engineering 14(2):1025-1033, 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 y (Small Recombinant Bispecific Antibody) Using A Refolding System,” Protein Eng. 13(8):583-588, Baeuerle, PA. et al. (2009) cific T-Cell Engaging Antibodies For Cancer Therapy,” Cancer Res. 69(12):4941-4944).

The design of a diabody is based on the antibody derivative known as a singlechain Variable Domain fragment (scFv). Such molecules are made by linking Light and/ or Heavy Chain Variable s using a short linking peptide. Bird et al. (1988) (“Single- Chain Antigen-BindingProteins,” Science 242:423-426) describes example of linking peptides which bridge imately 3.5 nm between the carboxy terminus of one Variable Domain and the amino terminus of the other Variable Domain. Linkers of other sequences have been designed and used (Bird et al. (1988) e-Chain Antigen-Binding Proteins,” Science 242:423-426). Linkers can in turn be modified for additional functions, such as attachment of drugs or attachment to solid supports. The -chain variants can be produced either recombinantly or tically. For synthetic production of scFv, an automated synthesizer can be used. For inant production of scFv, a suitable plasmid containing polynucleotide that s the scFv can be introduced into a suitable host cell, either eukaryotic, such as yeast, plant, insect or mammalian cells, or prokaryotic, such as E. coli. Polynucleotides encoding the scFv of interest can be made by routine manipulations such as ligation of polynucleotides. The resultant scFv can be isolated using standard protein purification techniques known in the art.

The provision of bispecific binding molecules (e.g., non-monospecific diabodies) es a significant advantage over antibodies, including but not limited to, a “trans” binding capability sufficient to co-ligate and/or co-localize different cells that s different epitopes and/or a “cis” binding capability sufficient to co—ligate and/or co-localize different molecules expressed by the same cell. Bispecific binding molecules (e.g., non-monospecific diabodies) thus have wide-ranging applications including therapy and immunodiagnosis. Bispecificity allows for great flexibility in the design and engineering of the diabody in various applications, providing enhanced avidity to multimeric antigens, the cross-linking of differing antigens, and directed targeting to specific cell types g on the presence of both target antigens. Due to their increased valency, low dissociation rates and rapid clearance from the circulation (for diabodies of small size, at or below ~50 kDa), diabody molecules known in the art have also shown particular use in the field of tumor imaging (Fitzgerald et al. (1997) “Improved Tumour Targeting By Disulpliide Stabilized Diabodies Expressed In Pichia pastoris, ” Protein Eng. : 1221).

The ability to produce ific diabodies has led to their use (in “trans”) to co- ligate two cells together, for example, by co-ligating receptors that are present on the surface of different cells (e.g., linking cytotoxic T-cells to tumor cells) (Staerz et al. (1985) d dies Can Target Sites For Attack By T Cells,” Nature 314:628-631, and Holliger et al. (1996) “Specific Killing OfLymphoma Cells By xic T-Cells MediatedBy A ific Diabody, ” Protein Eng. 9:299-305, Marvin et al. (2005) “Recombinant Approaches To IgG-Like Bispecifzc Antibodies,” Acta Pharmacol. Sin. 26:649-65 8).

Alternatively, or additionally, bispecific diabodies can be used (in “cis”) to co-ligate les, such as ors, etc, that are present on the surface of the same cell. Co—ligation of different cells and/or receptors is useful to modulation or functions and/or immune cell signaling.

However, the above advantages come at a salient cost. The formation of such non- monospecific diabodies requires the successful assembly of two or more ct and different polypeptides (i. e., such formation requires that the diabodies be formed through the heterodimerization of different polypeptide chain species). This fact is in contrast to monospecific ies, which are formed through the homodimerization of identical polypeptide . Because at least two dissimilar polypeptides (i.e., two polypeptide species) must be provided in order to form a non-monospecific diabody, and because homodimerization of such polypeptides leads to inactive les (Takemura, S. et al. (2000) “Construction Of A Diabody (Small Recombinant Bispecific Antibody) UsingA Refolding ,” Protein Eng. l3(8):583-588), the production of such polypeptides must be accomplished in such a way as to prevent covalent g between polypeptides of the same species (i.e., so as to prevent merization) (Takemura, S. et al. (2000) “Construction Of A Diabody (Small Recombinant Bispecific Antibody) Using A Refolding System,” Protein Eng. l3(8):583-588).

The art has therefore taught the non-covalent association of such polypeptides (see, e.g., Olafsen et al. (2004) “Covalent ide-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 onal ement 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 ,” Protein Eng. l3(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):]9665-19672). r, the art has ized that bispecific diabodies composed of non- covalently associated polypeptides are unstable and readily dissociate into nctional monomers (see, e.g., 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 EnhancedAntitumor ty,” J. Biol. Chem. ): 19665-19672).

In the face of this challenge, the art has succeeded in developing stable, covalently bonded heterodimeric non-monospecific diabodies, termed DART® (Qual Affinity Be- largeting Reagents) diabodies; see, e.g., United States Patent Publications No. 2013- 0295121, 2010-0174053 and 2009—0060910, European Patent Publication No. EP 2714079, EP 2601216, EP 2376109; EP 2158221 and PCT Publications No. 2012/018687; Infected CD4 T Cells by Dual-Affinity Re-targeting Molecules (DARTs) that Bind HIV Envelope and Recruit Cytotoxic T Cells,” PLoS . 11(11):e1005233. doi: . 1371/joumal.ppat. 1005233; Al Hussaini, M. et al. (2015) “Targeting CD123 In AML Using A T-Cell DirectedDual-Afinizy Re-Targeting (DART®) Platform,” Blood pii: 2014 575704; Chichili, G.R. et al. (2015) “A CD3xCD123 Bispecific DARTFor Redirecting Host T Cells To enous ia: Preclinical Activity And Safety In Nonhuman Primates,” Sci.

Transl. Med. 7(289):289ra82; Moore, RA. et al. (2011) “Application Of Dual Ayfmity Retargeting Molecules To Achieve Optimal Redirected T-Cell Killing OfB-Cell Lymphoma,” Blood 117(17):4542-4551; Veri, MC. et al. (2010) “Therapeutic Control OfB Cell tion Via Recruitment Of chamma Receptor IIb (CD323) Inhibitory Function With A Novel ific Antibody Scaflol ,” Arthritis Rheum. 62(7):1933-1943; Johnson, S. et al. (2010) “Eflector Cell Recruitment With NovelFv-BasedDual-Aflinity Re-Targeting Protein Leads To Potent Tumor sis And in vivo B—Cell Depletion,” J. Mol. Biol. 399(3):436-449). Such diabodies comprise two or more covalently complexed polypeptides and involve engineering one or more cysteine residues into each of the employed polypeptide species that permit disulfide bonds to form and thereby covalently bond one or more pairs of such polypeptide chains to one another. For e, the addition of a cysteine residue to the C—terminus of such constructs has been shown to allow disulfide bonding between the involved polypeptide chains, stabilizing the ing y t interfering with the diabody’s binding characteristics.

Many variations of such molecules have been described (see, e. g., United States Patent Publications No. 2015/0175697; 2014/0255407; 2014/0099318; 2013/0295121; 2010/0174053; 2009/0060910; 2007-0004909; European Patent ation No. EP 2714079; EP 2601216; EP 2376109; EP 2158221; EP 1868650; and PCT Publications No. WO 2012/162068; herein.

Alternative constructs are known in the art for applications where a tetravalent molecule is desirable but an PC is not required including, but not limited to, tetravalent tandem antibodies, also referred to as “TandAbs” (see, e.g. United States Patent Publications Nos. 2005-0079170, 2007-0031436, 2010-0099853, 20667 2013—0189263; European Patent ation Nos. EP 1078004, EP 2371866, EP 2361936 and EP 1293514; PCT Publications homo-dimerization of two identical chains each sing a VH1, VL2, VH2, and VL2 Domain.

IV. Preferred PD-l x CTLA-4 Bispecific Molecules One embodiment of the present invention relates to PD-1 X CTLA-4 bispecific molecules that are capable of binding to a “first epitope” and a “second epitope,” such epitopes not being identical to one another. Such bispecific molecules comprise “VLl” / “VH1” domains that are capable of binding to the first epitope and “VL2” / “VH2” domains that are capable of binding to the second epitope. The notations “VLl” and “VH1” denote, respectively, the Light Chain Variable Domain and Heavy Chain Variable Domain that bind the “first” epitope of such bispecific molecules. Similarly, the notations “VL2” and “VH2” denote, respectively, the Light Chain Variable Domain and Heavy Chain Variable Domain that bind the “second” epitope of such ific molecules. It is irrelevant whether a particular epitope is designated as the first vs. the second epitope, such notations having relevance only with respect to the ce and orientation of domains of the ptide chains of the binding molecules of the present ion. In one embodiment, one of such epitopes is an epitope of human PD-l and the other of such epitopes is an e of CTLA-4. In certain ments, a bispecific molecule comprises more than two epitope-binding sites. Such bispecific molecules will bind at least one e of PD-1 and at least one epitope of CTLA-4 and may further bind additional epitopes of PD-1 and/or additional epitopes of CTLA-4.

The t invention particularly relates to PD-1 X CTLA-4 bispecific molecules (e.g, bispecific antibodies, bispecific diabodies, ent binding molecules, etc.) that s epitope-binding fragments of antibodies that enable them to be able to coordinately bind to at least one epitope of PD-l and at least on epitope of CTLA-4. Selection of the VL and VH Domains of the polypeptide domains of such molecules is coordinated such that the VL Domain and VH Domain of the same polypeptide chain are not capable of forming an epitope- binding site capable of binding either PD-1 or CTLA-4. Such selection is additionally coordinated so that polypeptides chains that make up such PD-l X CTLA-4 bispecific molecules assemble to form at least one functional antigen binding site that is c for at least one epitope of PD-1 and at least one functional n binding site that is specific for at least one epitope of CTLA-4.

The present invention particularly relates to such PD-1 X CTLA—4 bispecific molecules that exhibit an activity that is enhanced relative to such activity oftwo monospecific molecules one of which possesses the Heavy Chain Variable Domain and the Light Chain Variable Domain of the antibody that binds PD-1 and the other of which possesses the Heavy Chain le Domain and the Light Chain Variable Domain of the antibody that binds CTLA-4. Examples of such activity es attenuating the activity of PD-1, attenuating the activity of CTLA-4, enhancing immune system activation, ing effector function, enhancing anti-tumor activity. As used herein, such attenuation of activity refers to a se of 10% or more, a decrease of 20% or more, a decrease of 50% or more, a decrease of 80% or more, or a decrease of 90% or more in a PD-l and/or CTLA-4 inhibitory activity, or the complete elimination of such PD-l and/or CTLA-4 inhibitory activity. As used herein, such enhancement of activity refers to an enhancement of 10% or more, an enhancement of 20% or more, an enhancement of 50% or more, an enhancement of 80% or more, or an enhancement of 90% or more in an immune system-activating activity mediated by or affected by the expression or ce of PD-l and/or CTLA-4, relative to the activity ted by two monospecific molecules one of which possesses the Heavy Chain Variable Domain and the Light Chain Variable Domain ofthe antibody that binds PD-1 and the other of which possesses the Heavy Chain Variable Domain and the Light Chain Variable Domain of the antibody that binds CTLA-4. Examples of immune system-activating activity include, but are not limited to immune cell (e. g., hocyte, NK-cell) eration, immune cell production and/or release of cytokines, immune cell production and/or release of lytic molecules (e.g., me, perforin, etc), and/or immune cell expression of activation markers. Cytokines which are released upon activation of the immune system are known in the art and include, but are not limited to: IFNy, IL—2, and TNFOL, (see, e.g., Janeway, CA. ei al. 2011) IMMUNOBIOLOGY” 8th ed. Garland e Publishing, NY; Banyer, J.L. (2000) “Cytokines in innate and adaptive immunity,” Rev Immunogenet. 2359-3 73). Activation s expressed by immune cells are known in the art and include, but are not limited to, CD69, CD25, and CD107a (see, e.g., Janeway, C.A. el al. (2011) IMMUNOBIOLOGY” 8th ed. Garland Science Publishing, NY, Shipkova, M. and Wieland, E. (2012) “Surface markers oflymphocyie activation and markers ofcellproliferation,” Clin Chim Acta 413: 1338-1349).

A. PD-l x CTLA-4 Bispecific Antibodies The instant invention asses bispeciflc antibodies capable of simultaneously binding to PD-1 and CTLA-4. In some embodiments, the bispeciflc antibody capable of simultaneously binding to PD-l and CTLA-4 is produced using any of the s described in PCT Publications No. 2007/146968, 2012/009544, WO 03652, reference in its entirety.

B. PD-l x CTLA-4 Bispecific Diabodies Lacking Fc Regions One embodiment of the present invention relates to bispecific ies that comprise, and most preferably are composed of, a first polypeptide chain and a second polypeptide chain, whose sequences permit the polypeptide chains to covalently bind to each other to form a covalently associated diabody that is capable of simultaneously binding to PD- 1 and to CTLA-4.

The first polypeptide chain of such an embodiment of bispecific diabodies comprises, in the N—terminal to C-terminal ion, an N—terminus, the VL Domain of a monoclonal antibody capable of binding to either PD-l or CTLA-4 (i.e., either VLPD-l or VLan-4), a first intervening spacer e (Linker l), a VH Domain of a monoclonal antibody capable of binding to either CTLA—4 (if such first polypeptide chain ns VLPD- 1) or PD-l (if such first polypeptide chain contains VLCTLA-4), a second intervening spacer peptide (Linker 2) optionally containing a cysteine residue, a Heterodimer—Promoting Domain and a C-terminus (Figure 1).

The second polypeptide chain of this embodiment of ific diabodies comprises, in the N—terminal to C-terminal direction, an N—terminus, a VL Domain of a monoclonal antibody capable of binding to either PD-l or CTLA-4 (i.e., either VLPD-l or VLCTLA-4, and being the VL Domain not selected for inclusion in the first polypeptide chain of the diabody), an intervening spacer peptide (Linker l), a VH Domain of a monoclonal antibody capable of binding to either CTLA-4 (if such second polypeptide chain contains VLPD-l) or to PD-l (if such second polypeptide chain contains -4), a second intervening spacer peptide (Linker 2) ally containing a cysteine residue, a Heterodimer-Promoting Domain, and a C-terminus (Figure 1), 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., either PD-l or CTLA-4). se, the VL Domain of the second polypeptide chain interacts with the VH Domain of the first polypeptide chain in order to form a second functional n—binding site that is specific for a second antigen (i.e., either CTLA- 4 or PD-l). Thus, the selection of the VL and VH Domains of the first and second polypeptide chains is coordinated, such that the two polypeptide chains ofthe diabody collectively comprise VL and VH Domains capable ofbinding to both an epitope ofPD-l and to an epitope of CTLA- 4 (i.e., they collectively comprise VLPD.1/VHPD-1 and -4/VHCTLA-4).

Most preferably, the length of the intervening linker peptide (Linker 1, which separates such VL and VH Domains) is selected to ntially or completely prevent the VL and VH Domains ofthe polypeptide chain from binding to one r (for example consisting of from O, l, 2, 3, 4, 5, 6, 7, 8 or 9 intervening linker amino acid residues). Thus the VL and VH s ofthe first polypeptide chain are substantially or completely incapable of binding to one another. Likewise, the VL and VH s of the second polypeptide chain are substantially or completely incapable of binding to one r. A preferred intervening spacer peptide (Linker 1) has the sequence (SEQ ID NO:5): GGGSGGGG.

The length and composition of the second intervening spacer peptide (Linker 2) is selected based on the choice of one or more ptide domains that promote such dimerization (i.e., a “Heterodimer-Promoting Domain”). Typically, the second intervening spacer peptide (Linker 2) will comprise 3—20 amino acid es. In particular, where the employed Heterodimer-Promoting Domain(s) do/does not comprise a cysteine e a cysteine-containing second intervening spacer peptide (Linker 2) is ed. A cysteine- containing second intervening spacer peptide (Linker 2) will contain 1, 2, 3 or more cysteines.

A preferred cysteine—containing spacer peptide (Linker 2) has the sequence is SEQ ID NO:6: GGCGGG. Alternatively, Linker 2 does not comprise a cysteine (e.g., GGG, GGGS (SEQ ID NO:7), LGGGSG (SEQ ID NO:8), GGGSGGGSGGG (SEQ ID NO:9), ASTKG (SEQ ID NO:10), LEPKSS (SEQ ID NO:11), APSSS (SEQ ID NO:12), etc.) and a Cysteine- Containing Heterodimer-Promoting , as described below is used. Optionally, both a cysteine-containing Linker 2 and a cysteine-containing Heterodimer-Promoting Domain are used.

The Heterodimer-Promoting Domains may be C (SEQ ID NO:13) or VEPKSC (SEQ ID NO:14) or AEPKSC (SEQ ID NO:15) on one polypeptide chain and GFNRGEC (SEQ ID NO:16) or FNRGEC (SEQ ID NO:17) on the other polypeptide chain (USZOO7/OOO4909).

] In a preferred embodiment, the Heterodimer—Promoting Domains will comprise tandemly repeated coil domains of opposing charge for example, “E-coil” l domains (SEQ ID NO:18: EVAALEK—EVAALEK—EVAALEK—EVAALEK), whose glutamate residues will form a negative charge at pH 7, and “K-coil” s (SEQ ID NO:19: EVAALEE— EVAAT .541 —§VAAT .E'i—EVAALEE), whose lysine residues will form a positive charge at pH 7.

The presence of such charged domains promotes association between the first and second polypeptides, and thus fosters heterodimer formation. Heterodimer-Promoting s that se modifications of the above-described E-coil and K-coil sequences so as to include one or more cysteine residues may be utilized. The presence of such cysteine residues permits the coil present on one polypeptide chain to become covalently bonded to a complementary coil present on another polypeptide chain, thereby covalently bonding the polypeptide chains to one another and increasing the stability of the diabody. es of such particularly preferred are Heterodimer—Promoting Domains include a Modified E-Coil having the amino acid ce EVAAQEK—EVAALEK—§VAAL§K—§VAAL§K (SEQ ID NO:20), and a modified K-coil having the amino acid sequence EVAAggE—EVAALEE—§VAAL§E— EVAALEE (SEQ ID NO:21).

As disclosed in WO 2012/018687, in order to e the in vivo pharmacokinetic properties of diabodies, a diabody may be modified to contain a polypeptide portion of a serum—binding protein at one or more of the termini of the y. Most preferably, such polypeptide portion of a serum-binding protein will be installed at the C- us of the diabody. Albumin is the most nt protein in plasma and has a half-life of 19 days in humans. Albumin possesses several small molecule binding sites that permit it to non-covalently bind to other proteins and thereby extend their serum half-lives. The Albumin-Binding Domain 3 (ABD3) of protein G of Streptococcus strain G148 consists of 46 amino acid residues forming a stable three-helix bundle and has broad albumin-binding specificity sson, M.U. er al. (2002) “Structure, Specificity, AndMode OfInteraction For Bacterial Albumin-BindingModules,” J. Biol. Chem. 277(10):8114-8120. Thus, a particularly preferred polypeptide portion of a serum-binding protein for improving the in vivo pharmacokinetic properties of a diabody is the Albumin-Binding Domain (ABD) from streptococcal protein G, and more preferably, the Albumin-Binding Domain 3 (ABD3) of protein G of Streptococcus strain G148 (SEQ ID NO:22): LAEAKVLANR ELDKYGVSDY NAKS LIDE ILAALP.

As disclosed in WO 2012/162068 (herein incorporated by reference), “deimmunized” variants of SEQ ID NO:22 have the ability to attenuate or eliminate MHC class II binding. Based on ational mutation results, the following combinations of substitutions are considered to be preferred substitutions for forming such a deimmunized ABD: 66D/7OS +71A; 66S/7OS +71A; 66S/7OS +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. Variant ABDs having the modifications L64A, I65A and D79A or the modifications N668, T708 and D79A. Variant deimmunized ABD having the amino acid sequence: LAEAKVLANR ELDKYGVSDY YKNLIQ66NAK§70 éuEGVKALIDE ILAALP (SEQ ID NO:23), or the amino acid sequence: LAEAKVLANR ELDKYGVSDY §65NNAKT VEGVKALI§79E ILAALP (SEQ or the amino acid sequence: LAEAKVLANR ELDKYGVSDY YKNLI§66NAK§70 VEGVKAL é79l12 TIAALB (SEQ ID NO:25), are particularly preferred as such deimmunized ABD exhibit substantially wild-type binding while providing attenuated MHC class II binding. Thus, the first polypeptide chain of such a diabody having an ABD contains a third linker r 3) ably positioned C—terminally to the E-coil (or K-coil) Domain of such polypeptide chain so as to intervene between the E- coil (or K-coil) Domain and the ABD (which is preferably a deimmunized ABD). A preferred sequence for such Linker 3 is SEQ ID NO:7: GGGS.

C. PD-l x CTLA-4 ific Diabodies ning Fc Regions One embodiment of the t invention relates to bispeciflc diabodies capable of simultaneously g to PD-l and CTLA-4 that comprise an Fc Region. The addition of an IgG CH2-CH3 Domain to one or both of the diabody polypeptide chains, such that the complexing 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. Incorporating an IgG CH2-CH3 Domains onto both of the diabody polypeptides will permit a ain bispecific ion- containing diabody to form (Figure 2).

Alternatively, incorporating an IgG 3 Domains onto only one of the diabody polypeptides will permit a more x four-chain bispecific Fc Region-containing diabody to form (Figures 3A-3C). Figure 3C shows a representative four-chain diabody sing the Constant Light (CL) Domain and the Constant Heavy CH1 Domain, however fragments of such domains as well as other polypeptides may alternatively be employed (see, e.g., Figures 3A and 3B, United States Patent Publications No. 2013-0295121; 2010-0174053 and 2009-0060910; European Patent Publication No. EP 2714079; EP 2601216; EP 2376109; EP 2158221 and PCT Publications No. 2010/080538). Thus, for example, in lieu of the CH1 Domain, one may employ a peptide having the amino acid ce GVEPKSC (SEQ ID NO:13) VEPKSC (SEQ ID NO:14), or AEPKSC (SEQ ID , derived from the Hinge Region of a human IgG, and in lieu of the CL Domain, one may employ the C-terminal 6 amino acids ofthe human kappa light chain, GFNRGEC (SEQ ID NO:16) or FNRGEC (SEQ ID NO:17). A representative peptide containing four-chain diabody is shown in Figure 3A. Alternatively, or in addition, one may employ a peptide sing tandem coil domains of opposing charge such as the “E-coil” helical s (SEQ ID NO:18: EVAALEK—§VAAL§K—§VAAL§K—§VAAL§K or SEQ ID NO:19: EK—EVAALEK—EVAALEK—EVAALEK); and the “K-coil” domains (SEQ ID NO:20: EVAATET.—§VAAT.§E—§VAAL§?—§VAAT.§E or SEQ ID NO:21: EVAAggE— EVAALEE—EVAALEE—EVAALEE). A representative coil domain-containing four-chain diabody is shown in Figure 3B.

The ific Fc Region-containing molecules of the present invention may include additional intervening spacer peptides rs), generally such Linkers will be incorporated between a peptide Heterodimer-Promoting Domain (e.g., an E-coil or K-coil) and CH2-CH3 Domains and/or between CH2-CH3 Domains and a Variable Domain (i.e., VH or VL). Typically, the additional Linkers will comprise 3-20 amino acid residues. Linkers that may be ed in the bispecific Fc Region-containing diabody molecules of the present invention include: GGGS (SEQ ID NO:7), LGGGSG (SEQ ID NO:8), GGGSGGGSGGG (SEQ ID NO:9), ASTKG (SEQ ID NO:10), DKTHTCPPCP (SEQ ID NO:26), EPKSCDKTHTCPPCP (SEQ ID NO:27), LEPKSS (SEQ ID NO:11), APSSS (SEQ ID NO:28), and APSSSPME (SEQ ID NO:29), LEPKSADKTHTCPPC SEQ ID NO:30), GGC, and GGG. SEQ ID NO:11 may be used in lieu of GGG or GGC for ease of cloning. Additionally, the amino acids GGG, or SEQ ID NO:11 may be immediately followed by SEQ ID NO:26 to form the alternate linkers: GGGDKTHTCPPCP (SEQ ID NO:31); and LEPKSSDKTHTCPPCP (SEQ ID NO:32). Bispecific Fc Region-containing molecules of the present invention may incorporate an IgG Hinge Region in addition to or in place of a linker. Exemplary Hinge Regions e: EPKSCDKTHTCPPCP (SEQ ID NO:33) from IgGl, ERKCCVECPPCP (SEQ ID NO:34) from IgG2, ESKYGPPCPSCP (SEQ ID NO:35) from IgG4, and ESKYGPPCPECP (SEQ ID NO:36) an IgG4 hinge variant comprising a stabilizing 8228P tution (as numbered by the EU index as set forth in Kabat) to reduce strand exchange.

] As provided in Figure 3A-3C, bispecifrc Fc Region-containing diabodies of the ion may comprise four different chains. The first and third polypeptide chains of such a diabody contain three domains: (i) a VLl-containing Domain, (ii) a VH2-containing Domain, (iii) Heterodimer-Promoting Domain and (iv) a Domain containing a CH2—CH3 sequence. The second and fourth polypeptide chains contain: (i) a VL2—containing Domain, (ii) a VH1- containing Domain and (iii) a dimer-Promoting Domain, where the dimer- Promoting Domains promote the dimerization of the first/third ptide chains with the second/fourth polypeptide chains. The VL and/or VH Domains of the third and fourth polypeptide chains, and VL and/or VH Domains ofthe first and second polypeptide chains may be the same or different so as to permit tetravalent binding that is either monospeciflc, bispecifrc or tetraspecifrc. The notations “VL3” and “VH3” denote, tively, the Light Chain le Domain and Variable Heavy Chain Domain that bind a “third” epitope of such diabody. Similarly, the notations “VL4” and “VH4” denote, respectively, the Light Chain Variable Domain and Variable Heavy Chain Domain that bind a “fourth” epitope of such diabody. The general ure of the ptide chains of a representative four-chain bispecifrc Fc Region-containing diabodies of invention is provided in Table 1: \Hz-VLZ-VH l-HPD-COOH \Hz-VL l -VH2-HPD—CH2-CH3-COOH 1fic. .

\Hz-VL l -VH2-HPD-CH2-CH3-COOH NHz-VLZ-VHI-HPD-COOH \Hz-VLZ-VHI-HPD-COOH KHz-VL l -VH2—HPD—CH2-CH3-COOH Tetraspecific. 7 \Hz-VL3-VH4-HPD-CH2-CH3-COOH \Hz—VL4-VH3-HPD-COOH HPD = Heterodimer—Promoting Domain In a specific embodiment, ies of the present invention are bispecific, tetravalent (i.e., possess four epitope—binding sites), Fc-containing diabodies that are composed of four total polypeptide chains (Figures 3A-3C). The bispecific, tetravalent, Fc-containing diabodies of the invention comprise two epitope-binding sites immunospecific for PD-l (which may be capable of binding to the same epitope of PD-l or to different epitopes of PD-l), and two epitope-binding sites immunospecific for CTLA-4 (which may be capable of binding to the same epitope of CTLA-4 or to different es of CTLA-4).

In a further embodiment, the ific Fc Region—containing ies may comprise three polypeptide chains. The first polypeptide of such a diabody ns three domains: (i) a VLl—containing Domain, (ii) a VH2-containing Domain and (iii) a Domain containing a CH2-CH3 sequence. The second polypeptide of such a y ns: (i) a VL2-containing Domain, (ii) a VHl-containing Domain and (iii) a Domain that promotes heterodimerization and covalent bonding with the diabody’s first polypeptide chain. The third polypeptide of such a diabody comprises a CH2—CH3 sequence. Thus, the first and second polypeptide chains of such a diabody associate together to form a VLl/VHl binding site that is capable of binding to the first epitope (i.e., either PD-l or CTLA-4), as well as a 2 binding site that is e of binding to the second epitope (i.e., either CTLA-4 or PD-l). The first and second ptides are bonded to one another through a disulfide bond involving cysteine residues in their respective Third Domains. Notably, the first and third polypeptide chains complex with one another to form an Fc Region that is stabilized via a disulfide bond.

Such bispecific diabodies have enhanced potency. Figures 4A and 4B illustrate the structures of such diabodies, Such Fc-Region-containing bispecific diabodies may have either of two orientations (Table 2): Table 2 NH2-CH2-CH3—COOH First NHz-VL l -VH2-HPD-CH2—CH3-COOH Orientation NHz-VLZ—VHl-HPD-COOH NHz-CHZ-CH3-COOH Second NHz-CHZ-CH3-VLl-VH2—HPD-COOH Orientation NH2-VL2-VH1-HPD-COOH HPD = Heterodimer—Promoting Domain In a specific embodiment, diabodies of the present invention are bispecific, bivalent (i.e., possess two epitope-binding sites), Fc-containing diabodies that are composed of three total polypeptide chains (Figures 4A-4B). The bispecific, bivalent Fc-containing diabodies of the invention comprise one epitope-binding site immunospecific for PD-l, and one epitope—binding site specific for CTLA-4.

In a further ment, the bispecific Fc Region-containing diabodies may comprise a total of five polypeptide chains. In a particular embodiment, two of the five ptide chains have the same amino acid sequence. The first polypeptide chain of such a diabody contains: (i) a VH1-containing domain, (ii) a CHl-containing domain, and (iii) a Domain containing a CH2—CH3 sequence. The first polypeptide chain may be the heavy chain of an antibody that contains a VH1 and a heavy chain constant region. The second and fifth polypeptide chains of such a y contain: (i) a VLl-containing domain, and (ii) a CL- containing domain. The second and/or fifth polypeptide chains of such a y may be light chains of an antibody that contains a VLl complementary to the VH1 of the first/third polypeptide chain. The first, second and/or fifth polypeptide chains may be isolated from a lly occurring antibody. atively, they may be constructed recombinantly. The third polypeptide chain of such a y contains: (i) a VH1-containing domain, (ii) a CH1- containing domain, (iii) a Domain containing a CH2-CH3 sequence, (iv) a ntaining Domain, (v) a VH3-containing Domain and (vi) a Heterodimer-Promoting , where the Heterodimer-Promoting Domains promote the dimerization of the third chain with the fourth chain. The fourth ptide of such diabodies contains: (i) a VL3-containing Domain, (ii) a VH2-containing Domain and (iii) a Domain that es heterodimerization and covalent bonding with the diabody’s third polypeptide chain.

Thus, the first and second, and the third and fifth, polypeptide chains of such diabodies associate together to form two VLl/VHl binding sites e of binding a first epitope. The third and fourth polypeptide chains of such diabodies associate together to form a VL2/VH2 binding site that is capable of binding to a second epitope, as well as a VL3/VH3 binding site that is e of binding to a third e. The first and third polypeptides are bonded to one another through a disulfide bond involving cysteine residues in their tive constant regions. Notably, the first and third polypeptide chains complex with one another to form an Fc Region. Such bispecific diabodies have enhanced potency. Figure 5 illustrates the structure of such diabodies. It will be understood that the VLl/VHl, VLZ/VHZ, and VL3/VH3 Domains may be the same or different so as to permit binding that is monospecific, bispecific or trispecific. However, as provided , these domains are preferably selected so as to bind PD-l and CTLA-4.

The VL and VH Domains of the ptide chains are selected so as to form VL/VH binding sites specific for a desired epitope. The VL/VH binding sites formed by the association of the polypeptide chains may be the same or different so as to permit tetravalent binding that is monospecific, bispecific, trispecific or tetraspecific. In particular, the VL and VH Domains may be ed such that a bispecific diabody may comprise two binding sites for a first epitope and two binding sites for a second epitope, or three binding sites for a first epitope and one binding site for a second epitope, or two g sites for a first epitope, one binding site for a second epitope and one g site for a third epitope (as depicted in Figure ). The general structure of the polypeptide chains of representative five-chain Fc Region- containing diabodies of invention is provided in Table 3: Table 3 l\H2-VL1-CL COOH 1\’H2 VHl-CHI CH2 CH3 COOH BiSPecific (2x2) kHz-VH1-CH1-CH2-CH3-VL2-VH2-HPD-COOH NHz-VLl—CL-COOH NH2-VL2-VH2-HPD-COOH I\;H2 VL1 CL COOH NHz-VLZ-VHl-HPD—COOH NH2-VL3-VH2-HPD—COOH HPD = Heterodimer—Promoting Domain In a c embodiment, diabodies of the present invention are bispecific, tetravalent (i.e., possess four epitope-binding sites), Fc-containing diabodies that are ed of five total polypeptide chains having two e-binding sites immunospecific for PD-l (which may be capable of binding to the same epitope of PD-l or to different epitopes of PD- 1), and two epitope-binding sites specific for CTLA—4 (which may be capable of binding to the same epitope of CTLA-4 or to ent epitopes of CTLA-4). In another embodiment, the ific, tetravalent, Fc-containing diabodies of the invention comprise three e-binding sites immunospecific for PD-l (which may be capable of binding to the same epitope of PD-l or to two or three different epitopes of PD—l), and one epitope-binding site specific for CTLA- 4. In another embodiment, the bispecific, tetravalent, Fc-containing diabodies of the invention comprise one epitope-binding sites immunospecific for PD-l, and three epitope-binding sites specific for CTLA-4 (which may be capable of binding to the same epitope of CTLA-4 or to two or three different es of CTLA-4).

D. PD-l x CTLA-4 Bispecific Trivalent g Molecules Containing Fc ] A further embodiment of the present invention relates to bispecific trivalent binding molecules comprising an Fc Region capable of simultaneously binding to an epitope of PD-l and an epitope present on CTLA-4. Such bispecific trivalent binding molecules comprise three epitope-binding sites, two ofwhich are Diabody-Type Binding Domains, which provide binding Site A and binding Site B, and one of which is a Fab-Type Binding Domain (or an scFv-Type Binding Domain), which es binding Site C (see, e.g., Figures 6A-6F, and PCT Application No: PCT/USIS/33081, and PCT/US 1 5/33076). Such bispeciflc trivalent molecules thus comprise “VLl” / “VH1” domains that are capable of binding to the first epitope and “VL2” / “VH2” domains that are capable of binding to the second epitope and “VL3” and “VH3” domains that are capable of binding to the “third” epitope of such trivalent molecule. A “Diabody-Type Binding Domain” is the type of epitope-binding site present in a diabody, and especially, a DART® y, as described above. Each of a ype Binding Domain” and an “scFv-Type Binding Domain” are e—binding sites that are formed by the interaction of the VL Domain of an immunoglobulin light chain and a menting VH Domain of an globulin heavy chain. Fab-Type Binding Domains differ from Diabody- Type g Domains in that the two polypeptide chains that form a Fab-Type Binding Domain comprise only a single epitope-binding site, whereas the two polypeptide chains that form a y-Type Binding Domain comprise at least two epitope-binding sites. rly, scFv-Type Binding s also differ from Diabody-Type Binding Domains in that they comprise only a single epitope-binding site. Thus, as used herein Fab-Type, and scFv-Type Binding Domains are distinct from Diabody-Type Binding Domains.

Typically, the trivalent binding molecules of the present invention will se four different polypeptide chains (see s 6A-6B), however, the molecules may comprise fewer or greater numbers ofpolypeptide chains, for example by fusing such polypeptide chains to one another (e.g, via a peptide bond) or by dividing such ptide chains to form additional polypeptide chains, or by associating fewer or additional polypeptide chains via disulfide bonds. Figures 6C-6F illustrate this aspect of the present invention by schematically depicting such molecules having three polypeptide chains. As ed in Figures 6A-6F, the ent binding molecules of the present invention may have alternative orientations in which the Diabody—Type Binding Domains are N—terminal (Figures 6A, 6C and 6D) or C-terminal (Figures 6B, 6E and 6F) to an Fc Region.

In certain embodiments, the first polypeptide chain of such trivalent binding molecules of the present invention contains: (i) a VLl-containing Domain, (ii) a VH2- containing Domain, (iii) a Heterodimer-Promoting Domain, and (iv) a Domain containing a CH2-CH3 sequence. The VLl and VL2 Domains are located N—terminal or C-terminal to the CH2—CH3-containing domain as presented in Table 4 (also see, Figures 6A and 6B). The second polypeptide chain of such embodiments contains: (i) a VL2-containing Domain, (ii) a VHl—containing Domain, and (iii) a Heterodimer-Promoting Domain. The third polypeptide chain of such embodiments contains: (i) a VH3-containing Domain, (ii) a CH1-containing Domain and (iii) a Domain containing a CH2-CH3 sequence. The third polypeptide chain may be the heavy chain of an antibody that contains a VH3 and a heavy chain constant region, or a polypeptide that contains such s. The fourth polypeptide of such embodiments contains: (i) a VL3-containing Domain and (ii) a taining Domain. The fourth polypeptide chains may be a light chain of an antibody that contains a VL3 complementary to the VH3 of the third polypeptide chain, or a polypeptide that contains such domains. The third or fourth polypeptide chains may be isolated from naturally occurring antibodies.

Alternatively, they may be ucted recombinantly, synthetically or by other means.

The Light Chain Variable Domain of the first and second polypeptide chains are separated from the Heavy Chain Variable Domains of such polypeptide chains by an intervening spacer peptide having a length that is too short to permit their VL1/VH2 (or their VL2/VH1) domains to associate together to form epitope-binding site capable of binding to either the first or second e. A preferred intervening spacer peptide (Linker 1) for this purpose has the sequence (SEQ ID NO:5): GGGSGGGG. Other Domains of the trivalent g molecules may be separated by one or more intervening spacer es (Linkers), optionally comprising a cysteine e. In particular, as provided above, such Linkers will typically be incorporated between Variable Domains (i.e., VH or VL) and peptide Heterodimer-Promoting Domains (e.g., an E-coil or K-coil) and between such e Heterodimer-Promoting Domains (e.g., an E-coil or K-coil) and CH2-CH3 Domains. ary linkers useful for the tion of ent binding molecules are provided above and are also provided in PCT Application Nos: PCT/US15/33081; and PCT/US15/33076.

Thus, the first and second polypeptide chains of such trivalent binding molecules associate together to form a VL1/VH1 binding site capable of binding a first epitope, as well as a 2 binding site that is capable of g to a second epitope. The third and fourth polypeptide chains of such trivalent binding molecules associate together to form a VL3/VH3 binding site that is capable of binding to a third epitope.

As described above, the trivalent binding molecules of the present invention may comprise three polypeptides. ent g molecules comprising three polypeptide chains may be obtained by linking the domains of the fourth polypeptide N-terminal to the ntaining Domain of the third polypeptide (e.g., using an intervening spacer peptide (Linker 4)). Alternatively, a third polypeptide chain of a trivalent binding molecule of the invention containing the following s is utilized: (i) a ntaining Domain, (ii) a VH3- containing Domain, and (iii) a Domain containing a 3 sequence, wherein the VL3 and VH3 are spaced apart from one another by an intervening spacer peptide that is sufficiently long (at least 9 or more amino acid residues) so as to allow the association of these domains to form an epitope-binding site. One preferred intervening spacer peptide for this purpose has the ce: GGGGSGGGGSGGGGS (SEQ ID NO:37).

It will be understood that the VLl/VHl, VL2/VH2, and VL3/VH3 Domains of such trivalent binding molecules may be different so as to permit binding that is bispeciflc or trispecific. However, as ed herein, these domains are selected so as to provide a trivalent binding molecule capable of binding PD-l and .

In particular, the VL and VH Domains may be selected such that a trivalent binding molecule comprises two binding sites for PD-l (which may be capable of binding to the same epitope of PD-l or to different epitopes of PD-l) and one binding sites for a CTLA- 4, or one g site for PD-l and two binding sites for CTLA-4 (which may be capable of binding to the same epitope of CTLA-4 or to different epitopes of ), or one binding site for PD-l, one binding site for CTLA-4 and one binding site for a third antigen that is not PD-l or CTLA-4. The general structure of the polypeptide chains of representative trivalent binding molecules of invention is provided in Figures 6A-6F and in Table 4: 21101 Chain NHz-VLZ-VHl-HPD-COOH Four Chain 1st Chain l\‘H2-VL1-VH2-HPD-CH2-CH3-COOH Orientation 3r01 Chain NH2-VH3-CH1-CH2-CH3-COOH 2ncl Chain NHz-VL3-CL-COOH 211d Chain 2-VHl-HPD-COOH Four Chain 1st Chain NHz-CHZ-CH3-VL1-VH2-HPD-COOH Orientation 3rcl Chain NHz—VH3-CH1-CH2-CH3-COOH 2““1 Chain 3-CL—COOH th-VLZ-VHl-HPD-COOH Three Chain lst NHz-VLI-VH2-HPD-CH2-CH3-COOH onematlon NHz-VL3-VH3-HPD-CH2-CH3-COOH 2nd h ’ NHZ VL— 2-VH l-I‘IPD-COOH Three Chain 2nd NH2-CH2-CH3-VL1-VH2-HPD-COOH O'“en a 1°“tt. 3-VH3-HPD-CH2-CH3-COOH HPD = Heterodimer—Promoting Domain One embodiment of the present invention relates to bispecific trivalent g molecules that se two e-binding sites for PD-l and one epitope-binding site for CTLA-4.

The two epitope-binding sites for PD-l may bind the same epitope or different epitopes. Another embodiment of the present invention relates to bispecific trivalent g molecules that comprise, one epitope-binding site for PD-l and two epitope-binding sites for CTLA-4. The two epitope-binding sites for CTLA-4 may bind the same epitope or different epitopes of CTLA-4. As provided above, such bispecific trivalent binding molecules may comprise three, four, five, or more polypeptide chains.

V. nt Domains and Fe Regions Provided herein are antibody Constant Domains useful in the generation of the PD-l X CTLA-4 bispecific molecules (e.g, antibodies, diabodies, trivalent binding les, etc.) of the invention.

A preferred CL Domain is a human IgG CL Kappa Domain. The amino acid sequence of an exemplary human CL Kappa Domain is (SEQ ID NO:38): RTVAAPSVFI FPPSDEQLKS GTASVVCLLN NFYPREAKVQ WKVDNALQSG NSQfiSVTfiQD SKDSTYSLSS TLTLSKADYE CEVT HQGLSSPVTK SFNQGEC Alternatively, an ary CL Domain is a human IgG CL Lambda Domain.

The amino acid sequence of an exemplary human CL Lambda Domain is (SEQ ID NO:39): SVTL FPPSSEELQA NKATLVCLIS DFYPGAVTVA WKADSSPVKA GVETTPSKQS NNKYAASSYL SLTPEQWKSH QSYSCQVTHE GSTVEKTVAP As provided herein, the PD-l X CTLA-4 bispecific molecules of the invention may comprise an Fc Region. The PC Region of such molecules of the invention may be of any isotype (e.g., IgG], IgG2, IgG3, or IgG4). The PD-l X CTLA-4 bispecific molecules of the invention may further comprise a CH1 Domain and/or a Hinge Region. When present, the CH1 Domain and/or Hinge Region may be of any isotype (e.g., IgGl, IgG2, IgG3, or IgG4), and is preferably of the same isotype as the desired Fc Region.

] An exemplary CH1 Domain is a human IgG1 CH Domain. The amino acid sequence of an ary human IgG1 CH1 Domain is (SEQ ID NO:40): ASTKGPSVFP LAPSSKSTSG GTAALGCLVK DYFPEPVTVS WNSGALTSGV HTFPAVLQSS GLYSLSSVVT GTQT YICNVNHKPS NTKVDKRV An exemplary CH1 Domain is a human IgG2 CH Domain. The amino acid sequence of an exemplary human IgG2 CH1 Domain is (SEQ ID : ASTKGPSVFP LAPCSRSTSE STAALGCLVK DYFPEPVTVS TSGV LQSS GLYSLSSVVT VPSSNFGTQT YTCNVDHKPS NTKVDKTV An exemplary CH1 Domain is a human IgG4 CH1 Domain. The amino acid sequence of an exemplary human IgG4 CH1 Domain is (SEQ ID NO:42): ASTKGPSVFP LAPCSRSTSE STAALGCLVK DYFPE?VTVS TSGV HTFPAVLQSS GLYSLSSVVT 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 (SEQ ID NO:33): EPKSCDKTHTCPPCP.

Another exemplary Hinge Region is a human IgG2 Hinge Region. The amino acid sequence of an exemplary human IgG2 Hinge Region is (SEQ ID NO:34): ERKCCVECPPCP .

Another exemplary Hinge Region is a human IgG4 Hinge Region. The amino acid sequence of an exemplary human IgG4 Hinge Region is (SEQ ID NO:35): ESKYGPPCPSCP. As described herein, an IgG4 Hinge Region may comprise a stabilizing on such as the $228P substitution. The amino acid sequence of an exemplary stabilized IgG4 Hinge Region is (SEQ ID NO:36): ESKYGPPCPPCP .

The Fc Region ofthe Fc Region-containing molecules (e.g., antibodies, diabodies, ent molecules, etc.) of the present invention may be either a complete Fc Region (e.g., a te IgG Fc Region) or only a fragment of an Fc Region. Optionally, the Fc Region of the Fc Region-containing molecules of the present ion lacks the C-terminal lysine amino acid residue.

In traditional immune function, the interaction of antibody-antigen complexes with cells of the immune system results in a wide array of responses, ranging from effector functions such as antibody dependent cytotoxicity, mast cell degranulation, and phagocytosis to immunomodulatory signals such as ting lymphocyte proliferation and dy secretion. All of these interactions are ted through the binding of the Fc Region of antibodies or immune complexes to specialized cell surface receptors on hematopoietic cells.

The diversity of cellular responses triggered by antibodies and immune complexes results from the structural geneity of the three Fc receptors: FcyRI (CD64), FcyRII (CD32), and FcyRIII (CD16). FcyRI (CD64), A (CD32A) and FcyRIII (CD16) are activating (i.e., immune system enhancing) receptors, FcyRIIB (CD32B) is an inhibiting (i.e., immune system dampening) or. In addition, interaction with the neonatal Fc Receptor (FcRn) mediates the recycling ofIgG molecules from the endosome to the cell surface and release into the blood.

The amino acid sequence of exemplary wild-type IgGl (SEQ ID NO:1), IgG2 (SEQ ID NO:2), IgG3 (SEQ ID NO:3), and IgG4 (SEQ ID NO:4) are presented above.

Modification of the Fc Region may lead to an d phenotype, for example altered serum half—life, d ity, altered tibility to cellular enzymes or altered effector function. It may therefore be desirable to modify an Fc Region—containing PD—l X CTLA-4 bispecific molecule of the t invention with respect to or function, for e, so as to enhance the effectiveness of such molecule in ng cancer. Reduction or elimination of effector function is desirable in certain cases, for example in the case of antibodies whose mechanism of action involves blocking or antagonism, but not killing of the cells bearing a target antigen. Increased effector function is generally desirable when directed to undesirable cells, such as tumor and foreign cells, where the 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). Molecules of the invention possessing such conferred or altered effector on activity are useful for the treatment and/or prevention of a disease, disorder or infection in which an enhanced efficacy of effector function activity is desired.

Accordingly, in certain embodiments, the Fc Region of the Fc Region-containing les of the present invention may be an engineered variant Fc Region. Although the Fc Region of the bispecific Fc Region-containing molecules of the t invention may possess the ability to bind to one or more Fc receptors (e.g, FcyR(s)), more preferably such variant PC Region have altered binding to chRIA (CD64), chRIIA (CD32A), FcvRIIB (CD32B), FcyRIIIA (CD16a) or IB (CD16b) ive to the g exhibited by a wild-type Fc Region), e.g., will have enhanced binding to an activating receptor and/or will have substantially reduced or no ability to bind to inhibitory receptor(s). Thus, the Fc Region of the Fc Region-containing molecules of the present invention may include some or all of the CH2 Domain and/or some or all of the CH3 Domain of a complete Fc Region, or may comprise a variant CH2 and/or a variant CH3 ce (that may include, for example, one or more insertions and/or one or more deletions with respect to the CH2 or CH3 domains of a complete Fc Region). Such Fc Regions may comprise non—Fc polypeptide portions, or may se portions of non-naturally te Fc Regions, or may comprise non-naturally occurring orientations of 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 linked to a CH2 Domain, etc.).

Fc Region modifications identified as ng effector function are known in the art, including modifications that increase binding to activating ors (e.g., FcyRIIA (CD16A) and reduce binding to inhibitory receptors (e.g., FcyRIIB (CD32B) (see, e.g., Stavenhagen, J.B. et al. (2007) “Fc Optimization 0f Therapeutic Antibodies Enhances Their Ability To Kill Tumor Cells In Vitro Ana’ Controls Tumor Expansion In Vivo Via Low-Afinity Activating chamma Receptors,” Cancer Res. 57(18):8882-8890). Table 5 lists exemplary single, , triple, quadruple and quintuple substitutions (relative to the amino acid sequence of SEQ ID NO:1) of exemplary modification that increase binding to activating ors and/or reduce binding to tory receptors.

Table 5 F243L and R292P mm and P396L R292P and P396L mm and P396L ——— Table 5 Variations of Preferred Activatin- Fc Re'ions mm,mm andmm F243L, R292P and WM F243L, R292P and V3051 F243L, R292P and P396L F243L, mm and P396L V284M R292L and K370N L234F, F243L, R292P and Y300L L234F, F243L, R292P and Y300L L23 51, F243L, R292P and Y300L L23 5Q, F243L, R292P and WM P247L, D270E, Y300L and N421K R255L, D270E, R2926 and P396L R255L, D270E, Y300L and P396L D270E, G316D, P396L and R416G — L235V, F243L, R292P, Y300L and P396L F243L, R292P, V3051, Y300L and P396L L235P, F243L, R292P, WM and P396L— Exemplary variants of human IgG1 Fc Regions with reduced binding to CD32B and/or sed binding to CD16A contain F243L, R292P, Y300L, V3051 or P296L tutions. These amino acid substitutions may be present in a human IgG1 Fc Region in any combination. In one embodiment, the human IgG1 Fc Region variant ns a F243L, R292P and Y300L substitution. In another embodiment, the human IgG1 Fc Region variant contains a F243L, R292P, Y300L, V3051 and P296L tution.

] In certain embodiments, it is preferred for the Fc Regions of PD-l X CTLA-4 bispeciflc molecules of the present invention to exhibit decreased (or substantially no) binding to FcyRIA (CD64), FcyRIIA (CD32A), FcyRIIB (CD32B), IA (CD16a) or FcyRIIIB (CD16b) (relative to the binding exhibited by the wild-type IgGl Fc Region (SEQ ID NO:1).

In a specific embodiment, the PD—l X CTLA-4 iflc molecules of the present invention comprise an IgG Fc Region that exhibits reduced ADCC effector function. In a preferred embodiment the CH2-CH3 Domains of such PD-l X CTLA-4 bispeciflc molecules include any 1, 2, 3, or 4 of the substitutions: L234A, L235A, D265A, N297Q, and N297G. In another embodiment, the 3 Domains contain an N297Q substitution, an N297G substitution, L234A and L235A substitutions or a D265A substitution, as these mutations abolish FcR binding. Alternatively, a CH2-CH3 Domain of a naturally occurring Fc region that inherently ts decreased (or substantially no) binding to FcyRIIIA (CD16a) and/or reduced effector function ive to the binding and or function exhibited by the wild-type IgGl Fc Region (SEQ ID NO:1)) is utilized. In a specific embodiment, the PD-l X CTLA—4 bispecific molecules of the present invention se an IgG2 Fc Region (SEQ ID NO:2) or an IgG4 Fc Region (SEQ ID:NO:4). When an IgG4 Fc Region is utilized, the instant invention also encompasses the introduction of a stabilizing mutation, such as the Hinge Region $228P substitution described above (see, e.g., SEQ ID NO:36). Since the N297G, N297Q, L234A, L235A and D265A substitutions abolish effector function, in circumstances in which effector function is desired, these substitutions would preferably not be employed.

A preferred IgGl sequence for the CH2 and CH3 Domains of the Fc Region- containing molecules of the present invention having reduced or abolished effector function will comprise the substitutions L235A (SEQ ID NO:43): APEA_AGGPSV FLFPPKPKUT TM SRTBHV'I' CVVV)VSHH) NWYV.) GVEVHNA<TK PRfifiQYNSTY VLH QDWLWGKEYK CKVSNKALPA PIEKTISKAK GQBRiPQVYT LPPSREEMTK CLVK IAVE WESNGQPENN YKTT?PVLDS DGSFFLYSKL TVDKSRWQQG NVFSCSVMHE AL-INHYTQKS LSLSPG§ wherein, X is a lysine (K) or is absent.

The serum half—life of proteins comprising Fc Regions may be increased by increasing the binding affinity of the Fc Region for FcRn. The term “half-life” as used herein means a pharmacokinetic property of a le that is a measure of the mean survival time ofthe molecules following their administration. Half-life can be expressed as the time required to eliminate fifty percent (50%) of a known quantity of the molecule from the subject’s body (e.g., human patient or other ) or a specific compartment f, for example, as ed in serum, i.e., circulating half-life, or in other tissues. In general, an increase in half- life results in an increase in mean residence time (MRT) in circulation for the molecule stered.

In some embodiments, the PD-l X CTLA-4 bispecific molecules of the present invention comprise a variant Fc Region, wherein the variant Fc Region comprises at least one amino acid modification relative to a wild-type Fc Region, such that the le has an increased half-life (relative to a molecule comprising a wild-type Fc Region). In some embodiments, the PD-l X CTLA-4 bispecific molecules of the present ion comprise a variant IgG Fc Region, wherein the variant Fc Region comprises a half-live extending amino acid substitution at one or more positions selected from the group consisting of 238, 250, 252, 254,256,257, 256,265, 272, 8, 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 mutations capable of increasing the half-life of an Fc Region-containing molecule are known in the art and include, for e M252Y, S254T, T256E, and combinations thereof. For example, see the mutations bed in US. Patents No. 6,277,375, 7,083,784, 7,217,797, 8,088,376, US.

Publication Nos. 2002/0147311; 2007/0148164; and International Publication Nos. WO 98/23289, reference in their ties. PD-l X CTLA-4 bispeciflc molecules with enhanced half-life also include those possessing t Fc Regions comprising tutions at two or more of PC Region residues 250, 252, 254, 256, 257, 288, 307, 308, 309, 311, 378, 428, 433, 434, 435 and 436. In particular, two or more substitutions selected from: T250Q, M252Y, S254T, T256E, K288D, T307Q, V308P, A378V, M428L, N434A, H435K, and Y436I.

In a specific embodiment, a PD-l X CTLA—4 bispeciflc molecule possesses a t IgG Fc Region comprising substitutions of: (A) M252Y, $254T 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 Y43 61, (H) V308P and N434A, or (I) K288D and H43 5K.

In a preferred embodiment PD-l X CTLA-4 bispecific molecules possess a variant IgG Fc Region comprising any 1, 2, or 3 of the substitutions: M252Y, S254T and T256E. The invention further encompasses PD-l X CTLA-4 bispecific molecules possessing variant Fc Regions comprising: (A) one or more mutations which alter effector function and/or chR, and (B) one or more mutations which extend serum half-life.

A preferred IgGl sequence for the CH2 and CH3 Domains of the Fc Region- containing les of the present ion having increased serum ife will comprise the substitutions M252Y, $254T and T256E (SEQ ID NO:80): ABEA_AGGJ:’SV b'Lh'PBKk’KJT LY__TREPJL'V'1' cvvv )VSHH ) BHVKENWYV.) GVEVHNA<TK PRfifiQYNSTY QVVSVLTVLH QDWLWGKEYK CKVSNKALPA PIEKTISKAK GQBRiPQVYT LPPSREEMTK NQVSLTCLVK GFYPSDIAVE WESNGQPENN YKTT?PVLDS DGSFFLYSKL TVDKSRWQQG NVFSCSVMHE ALHNHYTQKS LSLSPG§ wherein, X is a lysine (K) or is absent.

As will be noted, the CH2-CH3 Domains of SEQ ID NO:80 includes substitutions at positions 234 and 235 with alanine, and thus form an Fc Region exhibit decreased (or substantially no) binding to FcyRIA (CD64), A (CD32A), FcyRIIB (CD32B), FcyRIIIA (CDl6a) or FcyRIIIB (CDl6b) ive to the binding exhibited by the wild-type Fc Region (SEQ ID NO:1). The invention also encompasses such IgGl CH2—CH3 Domains, which comprise the ype alanine residues, alternative and/or additional substitutions which modify or function and/or FyR binding activity of the Fc region.

A preferred IgG4 sequence for the CH2 and CH3 Domains of the Fc Region- ning molecules of the present invention having increased serum half—life will comprise the tutions M252Y, $254T and T256E (SEQ ID NO:81): APEFLGGPSV FLF?PKPKDT LYITREPEVT CVVVDVSQED PZVQFNWYVDJ.

GVEViNA<TK PQfifiQFNSTY QVVSVLTVLH QDWLNGKEYK C(VSNKGLPS SIEKTISKAK GQPREPQVYT LPPSQEEMTK NQVSLTCLVK GFYPSDIAVE WESNGQPENN Y<TTPPVLDS DGSFFLYSRL TVDKSRWQEG VMHE ALHNHYTQKS LSLSLG§ wherein, X is a lysine (K) or is absent.

For certain antibodies, diabodies and trivalent binding molecules whose Fc Region-containing first and third polypeptide chains are not identical, it is desirable to reduce or t homodimerization from occurring between the CH2-CH3 Domains of two first polypeptide chains or between the CH2-CH3 Domains of two third polypeptide chains. The CH2 and/or CH3 Domains of such polypeptide chains need not be identical in sequence, and advantageously are modified to foster complexing between the two polypeptide chains. For example, an amino acid substitution (preferably a substitution with an amino acid comprising a bulky side group g a “knob”, e.g., tryptophan) can be introduced into the CH2 or CH3 Domain such that steric interference will t interaction with a similarly mutated domain and will obligate the mutated domain to pair with a domain into which a complementary, or accommodating mutation has been engineered, i. e., “the hole” (e. g., a substitution with e). Such sets of ons can be engineered into any pair of polypeptides comprising 3 Domains that forms an Fc Region to foster heterodimerization. Methods of protein engineering to favor heterodimerization over homodimerization are well known in the art, in particular with respect to the ering of immunoglobulin-like molecules, and are encompassed herein (see e.g., y et al. (1996) “‘Knobs—Into-Holes’ Engineering Of Antibody CH3 Domains For Heavy Chain Heterodimerization, ” Protein Engr. 9:617-621, Atwell et al. (1997) ”Stable Heterodimers From Remodeling Yhe Domain ace Of A Homodimer Using/1 Phage Display Library, ” J. Mol. Biol. 270: 26—3 5, and Xie et al. (2005) “A New Format OfBispecijic Antibody: Highly Eflicient Heterodimerization, Expression And Tumor Cell Lysis, ” J. Immunol. Methods 296:95-101, each of which is hereby incorporated herein by nce in its ty), A preferred knob is created by modifying an IgG Fc Region to contain the modification T366W. A preferred hole is d by modifying an IgG Fc Region to contain the modification T366S, L368A and Y407V. To aid in purifying the hole-bearing third polypeptide chain homodimer from the final bispecific heterodimeric Fc Region-containing molecule, the protein A binding site of the hole-bearing CH2 and CH3 Domains of the third polypeptide chain is preferably mutated by amino acid substitution at position 435 (H435R).

Thus, the hole-bearing third ptide chain homodimer will not bind to protein A, whereas the bispecific heterodimer will retain its ability to bind protein A via the protein A binding site on the first polypeptide chain. In an alternative embodiment, the hole-bearing third polypeptide chain may incorporate amino acid substitutions at positions 434 and 435 (N434A/N43 5K).

A preferred IgGl amino acid sequence for the CH2 and CH3 Domains of the first ptide chain of an Fc Region-containing molecule of the present invention will have the “knob-bearing” sequence (SEQ ID NO:44): APEééGGPSV FLFP?KPKDT EM SRTPfiVT CVVVDVSiED PEVKFWWYVD GVEVHNAXTK PREEQYNSTY RVVSVLTVLH KEYK CKVSN<ALPA PIEKTISKAK GQPREPQVYT LPPSREEWTK NQVSTKCWVK GTYPSDIAVE WLSNGQBLNN VLDS DGSFFLYSKL TVDKSRWQQG NVFSCSVMHE ALHNHYTQKS LSLS?G§ wherein X is a lysine (K) or is .

A preferred IgGl amino acid sequence for the CH2 and CH3 Domains of the second polypeptide chain of an Fc Region-containing molecule of the present invention having two polypeptide chains (or the third polypeptide chain of an Fc Region-containing le having three, four, or five polypeptide chains) will have the “hole-bearing” sequence (SEQ AEEééGGPSV ELHPPKEKJT LMHSRTPEVT CVVV)VSHH) BHVKENWYVJ GVEVHNA<TK PRfifiQYNSTY QVVSVLTVLH QDWLWGKEYK ALPA PIEKTISKAK GQBRiPQVYT LPPSREEMTK NQVSL§C§VK IAVE WESNGQPENN YKTT?PVLDS DGSFFLYSKL TVDKSRWQQG NVFSCSVMHE ALHNBYTQKS LSLSPG§ wherein X is a lysine (K) or is absent.

As will be noted, the CH2-CH3 Domains of SEQ ID NO:44, and SEQ ID NO:45 include substitutions at positions 234 and 235 with alanine, and thus form an Fc Region exhibit decreased (or substantially no) binding to FcyRIA (CD64), A (CD32A), FcyRIIB (CD32B), FcyRIIIA (CDl6a) or FcyRIIIB (CDl6b) (relative to the binding ted by the wild-type Fc Region (SEQ ID NO:1). The invention also encompasses such IgGl CH2—CH3 s, which se the wild-type alanine residues, alternative and/or additional substitutions which modify effector function and/or FyR binding activity of the Fc region. The invention also encompasses such CH2-CH3 Domains, which further comprise one or more half—live extending amino acid substitutions. In particular, as provided above, the invention encompasses such hole-bearing and such knob-bearing CH2-CH3 Domains which further comprise the M252Y/S254T/T256E.

A red IgGl amino acid ce, for the CH2 and CH3 Domains further comprising M252Y/S254T/T256E, of the first polypeptide chain of an Fc -containing le of the present invention will have the “knob-bearing” sequence (SEQ ID NO:82): APEééGGPSV FLFP?KPKDT LI IREPfiVT CVVVDVSiED PEVKFWWYVD GVEVHNAKTK PREEQYNSTY RVVSVLTVLH QDWLNGKEYK CKVSNXALPA PIEKTISKAK GQPREPQVYT TPPSREEWTK NQVSVECTVK GTYPSDIAVE WESNGQPENN YKTTPPVLDS DGSFFLYSKL TVDKSRWQQG NVFSCSVMHE ALHNHYTQKS § wherein X is a lysine (K) or is absent.

A preferred IgGl amino acid sequence, for the CH2 and CH3 Domains further comprising M252Y/S254T/T256E, of the second polypeptide chain of an Fc Region- containing le of the present invention having two polypeptide chains (or the third polypeptide chain of an Fc Region-containing molecule having three, four, or five polypeptide chains) will have the “hole-bearing” sequence (SEQ ID NO:83): APEééGGPSV FLFP?KPKDT PX TREPfiVT CVVVDVSIED PEVKFWWYVD GVEVHNAKTK PREEQYNSTY RVVSVLTVLH KEYK ALPA PIEKTISKAK GQPQEPQVYT WPPSRFHWTK NQVSL§C§VK GFYPSD"AV? WESNGQPENN YKTTPPVLDS DGSFFLZSKL TVDKSRWQQG NVFSCSVMHE ALHNEYTQKS LSLS?G§ wherein X is a lysine (K) or is absent.

] A preferred IgG4 amino acid sequence for the CH2 and CH3 Domains, comprising M252Y/SZS4T/T256E, of the first polypeptide chain of an F0 Region-containing molecule of the present invention will have the “knob-bearing” sequence (SEQ ID NO:84): APEFLGGPSV FL??PKPKDT LYITREPEVT CVVVDVSQED PEVQFVWYVD GVEViNA<TK BQfifiQFNSTY QVVSVLTVLH QDWLNGKEYK CKVSN<GLPS SIEKTISKAK GQPREPQVYT LPPSQEEMTK NQVSLECLVK GHYPS) AVE WESNGQPENN Y<TTPPVLDS YSRL TVDKSRWQEG NVFSCSVMHE ALHNHYTQKS LSLSLG§ wherein X is a lysine (K) or is .

A preferred IgG4 amino acid sequence, for the CH2 and CH3 Domains comprising M252Y/SZS4T/T256E, of the second polypeptide chain of an Fc Regioncontaining molecule of the t invention having two polypeptide chains (or the third polypeptide chain of an Fc Region-containing molecule having three, four, or five polypeptide chains) will have the “hole-bearing” sequence (SEQ ID NO:85): APEFLGGPSV FLFPPKPKDT PEVT CVVV)VSQH) PHVQENWYVJ GVEVHNA<TK PRfifiQFNSTY QVVSVLTVLH KEYK CKVSNKGLPS SIEKTISKAK GQPREPQVYT LPPSQEEMTK NQVSL§CA¥K GFYPSDIAVE WESNGQPENN YKTTPPVLDS YSRL TVDKSRWQEG NVFSCSVMHE ALINBYTQKS LSLSLG§ wherein X is a lysine (K) or is absent.

As will be noted, the CH2-CH3 Domains of SEQ ID NO:84, and SEQ ID NO:85 e the M252Y/8254T/T256E tutions, and thus form an IgG4 Fc Region exhibiting sed serum half-life. The invention also encompasses IgG4 CH2-CH3 Domains, which comprise the wild-type ZS4/T256 residues.

It is red that the first polypeptide chain will have a “knob-bearing” CH2- CH3 sequence, such as that of SEQ ID NO:44. However, as will be recognized, a “hole- bearing” CH2-CH3 Domain (e.g., SEQ ID NO:45) could be employed in the first polypeptide chain, in which case, a “knob-bearing” CH2-CH3 Domain (e.g., SEQ ID NO:44) would be employed in the second polypeptide chain of an Fc Region-containing molecule of the present invention having two polypeptide chains (or in the third ptide chain of an Fc Region- ning molecule having three, four, or five polypeptide chains).

In other embodiments, the invention encompasses PD-l X CTLA—4 bispecific molecules sing CH2 and/or CH3 Domains that have been engineered to favor heterodimerization over homodimerization using mutations known in the art, such as those disclosed in PCT Publication No.

VI. Anti-PD-l Binding lities Antibodies that are immunospecific for PD-l are known (see, e.g., United States Patent Applications No. 62/198,867; 62/239,559; 62/255,140 United States s No. 8,008,449, 8,552,154, PCT Patent ations 2013/014668). Preferred PD-l binding capabilities useful in the generation of the PD-l X CTLA-4 bispecific molecules of the present invention are e of binding to a continuous or discontinuous (e.g., conformational) portion (epitope) of human PD-l (CD279) and will preferably also t the ability to bind to PD-l molecules of one or more non-human species, in particular, primate species (and especially a primate species, such as cynomolgus monkey).

Additional desired antibodies may be made by isolating antibody-secreting hybridomas elicited using PD-l or a peptide fragment thereof. A entative human PD-l polypeptide (NCBI Sequence 009.2; ing a 20 amino acid residue signal sequence, shown underlined) and the 268 amino acid residue mature protein) has the amino acid sequence (SEQ ID NO:46): MQIPQAPWPV VWAVLQLGWR PGWFLDSPDR PWNPPTFSPA_LWVVT?GDNA TFTCSFSNTS ESFVLNWYRM SPSNQTD<LA AFPEDRSQPG QDCRFRVTQL PNGRDFHMSV VRARRNDSGT YWCGATSWAP KAQIKFSTRA FTRVTFRRAF VPTAiPSPSP RPAGQFQTLV LLGS LVTLVWVLAV "CSRAAQGT" GARRTGQPLK EDPSAVPVES VJYGELDEQW PBVP CVPEQTLYAT MGTS SPARRGSADG PQSAQPLRPE DGHCSWPL Preferred anti-PD-l binding molecules (e.g., antibodies) useful in the generation of the PD-l X CTLA—4 bispecific molecules of the instant invention possess the VL and/or VH Domains of the anti-human PD-l monoclonal antibody “PD-1 mAb 1” (nivolumab, CAS Reg.

No.:9464l4—94-4, also known as 5C4, BMS—936558, ONO—4538, MDX-llO6, and ed as OPDIVO® by Bristol-Myers Squibb), “PD-1 mAb 2” (pembrolizumab, (formerly known as lizumab), CAS Reg. No.:13748534, also known as MK-3475, SCH-900475, and marketed as KEYTRUDA® by Merck); “PD-1 mAb 3” (EH12.2H7; Dana ), “PD-1 mAb 4” (pidilizumab, CAS Reg. No: 1036730-42—3 also known as CT-Oll, CureTech,), or any of the anti-PD-l antibodies in Table 6; and more preferably possess 1, 2 or all 3 of the CDRLS of the VL Region and/or 1, 2 or all 3 of the CDRHS of the VH Domain of such anti-PD- 1 monoclonal antibodies. onal anti-PD-l antibodies possessing unique binding characteristics useful in the methods and compositions of the instant inventions have recently been identified (see, United States Patent Application Nos. 62/198,867; 62/239,559; 62/255,140). Particularly, preferred are PDbinding molecules which possess a humanized VH and/or VL Domain of the anti-PD-l antibody “PD-1 mAb 5” (hPD—l mAb 2, MacroGenics); “PD-1 mAb 6” (hPD-1 mAb 7, MacroGenics); “PD-1 mAb 7” (hPD-1 mAb 9, MacroGenics); or “PD-1 mAb 8” (hPD-1 mAb 15, MacroGenics); and more ably possess 1, 2 or all 3 of the CDRLS of the VL Region and/or 1, 2 or all 3 of the CDRHs of the VH Domain of such humanized D-l monoclonal antibodies.

A. PD-l mAb 1 ] The amino acid sequence of the VH Domain of PD-l mAb 1 (SEQ ID NO:47) is shown below (CDRH residues are shown underlined).

QVQLVESGGG SLRL 0C ASG"TFS NSGMHWVRQA PGKGLEWVAZ IWYDGSKRYY ADSVKGRFTI SRDNSKNTLF LQMNSLRAED TAVYYCATEE EWGQGTLVT vss The amino acid sequence of the VL Domain of PD-1 mAb 1 (SEQ ID NO:48) is shown below (CDRL residues are shown underlined).

EIVLTQSPAT LSLSPGERAT LSCRASQSVS SYLAWYQQKP GQAPRLLIYD ASNRATGIPA RFSGSGSGTD FTLTISSLEP EDFAVYYCQQ SSNWPRTFGQ GTKVEIK B. PD-l mAb 2 The amino acid ce of the VH Domain of PD-1 mAb 2 (SEQ ID NO:49) is shown below (CDRH es are shown underlined).

QVQLVQSGVE VKKPGASVKV SC<ASGYTFT NYYMYWVRQA PGQGLEWMGQ INPSNGGTNF NEKFKNRVTL TTDSSTTTAY MELKSLQFDD TAVYYCARED YRFDMGFDYW GQGTTVTVSS The amino acid sequence of the VL Domain of PD-1 mAb 2 (SEQ ID NO:50) is shown below (CDRL residues are shown underlined).

EIVLTQSPAT LSLSPGERAT KGVS TSGYSYLHWY QQKPGQAPRL LIYLASYLES GVPARFSGSG SGTDFTLTIS SLEPEDFAVY YCQHSRDLPL TFGGGTKVEIK C. PD-1 mAb 3 The amino acid sequence of the VH Domain of PD-1 mAb 3 (SEQ ID NO:51) is shown below (CDRH residues are shown ined).

QVQLQQSGAE LAKPGASVQM SCKASGYSFT SSWIHWVKQR PGQGLEWIGY IYPSTGFTEY NQKFKDKATL TADKSSSTAY MQLSSLTSED SAVYYCARWR AMDY WGQGTSVTVSS The amino acid sequence of the VL Domain of PD-1 mAb 3 (SEQ ID NO:52) is shown below (CDRL residues are shown underlined).

DIVLTQSPAS LTVSLGQRAT ISCRASQSVS MHWY QQKPGQPPKL NLES GIPARFSGSG SGTDFTLNIH PVEEEDTATY YCQHSWEIPY TFGGGTKLEI K D. PD-1 mAb 4 The amino acid sequence of the VH Domain of PD-1 mAb 4 (SEQ ID NO:53) is shown below (CDRH residues are shown underlined).

SGSE LKKPGASVKI SCKASGYTFT NYGMNWVRQA PGQGLQWMGW INTDSGESTY AEEFKGRFVF NTAY LQITSLTAED TGMYFCVRVG YDALDYWGQG TLVTVSS The amino acid sequence of the VL Domain of PD-1 mAb 4 (SEQ ID NO:54) is shown below (CDRL residues are shown underlined).

EIVLTQSPSS LSASVGDRVT ITCSARSSVS YMHWFQQKPG KAPKLWIYRT SNLASGVPSR FSGSGSGTSY CLTINSLQPE DFATYYCQQR SSFPLTFGGG TKLEIK E. PD-1 mAb 5 The amino acid sequence of the VH Domain of PD-1 mAb 5 (SEQ ID NO:55) is shown below (CDRH es are shown underlined).

EVQLVESGGG LVQPGGSLRL SCAASGFVFS SFGMHWVRQA PGKGLEWVAY SISY ADTVKGRFTI SRDNAKNTLY LQMNSLRTED TALYYCASLS DYFDYWGQGT TVTVSS The amino acid sequence of the VL Domain of PD-l mAb 5 (SEQ ID NO:56) is shown below (CDRL residues are shown underlined).

DWMTQSPLS LPVTLGQPAS IISCRSSQSLV HSTGNTYLHW YLQKPGQSPQ SNRF §GVPDRFSGS GSGTDFTLKI SRVEAEDVGV YYCSQTTHVP VEFGQGTKLE IK F. PD-l mAb 6 The amino acid sequence of the VH Domain of PD-l mAb 6 (SEQ ID NO:57) is shown below (CDRH residues are shown underlined).

QVQLVQSGAE VKKPGASVKV SCXASGYSFT SYWMNWVRQA PGQGLEWXlGZ IHPSDSETWL DQKFKDRVTI TVDKSTSTAY MELSSLRSED TAVYYCAREE YGTSPFAYWG QGTLVTVSS n X1 is I or A The amino acid sequence of the VL Domain of PD-l mAb 6 (SEQ ID NO:58) is shown below (CDRL residues are shown ined).

EIVLTQSPAT LSLSPGERAT LSCRAXLESVD MNWF QQKPGQPPKL L:HAASNX2GS GVPSRFSGSG SGTDFTLTIS SLEPEDFAVY EVPY EFGGGTKVEI K wherein: X1 is N or S and X2 is Q or R; or X1 is N and X2 is Q; or X1 is S and X2 is Q; or X1 is S and X2 is R In particular embodiments the amino acid sequence of PD-l mAb 6 comprises: (a) SEQ ID NO:57; wherein X1 is I; and SEQ ID NO:58; wherein X1 is N and X2 is Q; or (b) SEQ ID NO:57; wherein X1 is I; and SEQ ID NO:58; wherein X1 is S and X2 is Q.

An exemplary anti-PD—l VH Domain designated “PD-1 mAb 6-I VH” ses SEQ ID NO:57 wherein X1 is I and has the amino acid sequence (SEQ ID NO:86): QVQLVQSGAE VKKPGASVKV SC<ASGYSFT SYWMNWVRQA PGQGLEWIGV IHPSDSETWL DQKFKDRVTI TVDKSTSTAY MELSSLRSED TAVYYCAREH YGTSPFAYWG QGTLVTVSS An exemplary anti-PD-l VL Domain designated “PD-1 mAb 6-SQ VL” comprises SEQ ID NO:58 wherein X1 is S and X2 is Q and has the amino acid sequence (SEQ ID NO:87): EIVLTQSPAT LSLSPGERAT LSCRASESVD NYGMSFMNWF PPKL LIHAASNQGS GVPSRFSGSG SGTDFTLT S SLJPEDEAVY ECQQSKEVPY TFGGGTKVEI K An exemplary anti-PD-l antibody that ses a PD—l mAb 6-I VH domain and a PD-l mAb 6-SQ VL domain is designated as “PD-1 mAb 6-ISQ.” G. PD-l mAb 7 The amino acid sequence of the VH Domain of PD-l mAb 7 (SEQ ID NO:59) is shown below (CDRH residues are shown ined).

EVQLVESGGG LXlRPGGSLKL SCAASGFTFS SYLVX¢WVRQA PGKGLEWXgAE ISGGGGNTYY SDSVKGRFTI SRDNAKNSLY LQMNSX4RAED TATYYCARXE FDGAWFAYWG QGTLVTVSS wherein: X1 is V or A; X2 is S or G; X3 is V or T; X4 is L or A; or X1 is V, X2 is S, X3 is V, and X4 is L; or X1 is A, X2 is G, X3 is T, and X4 is A The amino acid sequence of the VL Domain of PD-l mAb 7 (SEQ ID NO:60) is shown below (CDRL residues are shown underlined).

SPSS LSASVGDRVT ITCRASENIY XIYLAWYQQKP GKAPKLLIY§£ GVPS RFSGSGSGTD FTLTISSLQP EDFATYYCQH HYAVPWTFGQ GTKLEIK wherein: X1 is S or N and X2 is N or D; or X1 is S and X2 is N; or X1 isN and X2 isD In particular embodiments PD-l mAb 7 comprises: (a) SEQ ID NO:59, wherein X1 is V, X2 is S, X3 is V, and X4 is L; and SEQ ID NO:60, wherein X1 is S and X2 is N; or (b) SEQ ID NO:59, wherein X1 is A, X2 is G, X3 is T, and X4 is A; and SEQ ID NO:60, wherein X1 is N and X2 is D.

H. PD-l mAb 8 The amino acid sequence of the VH Domain of PD-l mAb 8 (SEQ ID NO:61) is shown below (CDRH residues are shown underlined).

EVQLVESGGG LVRPGGSLRL SCAASGFTFS SYLISWVRQA PGKGLEWVAé DTYY ADSVKGRFTI SRDNAKNSLY LQMNSLRAED TATYYCARBE TYAMDYWGQG TLVTVSS The amino acid sequence of the VL Domain of PD-l mAb 8 (SEQ ID NO:62) is shown below (CDRL residues are shown ined).

DIQMTQSPSS LSASVG)RVT TCRASENIY NYLAWYQQKP GKAPKLLIYE AKTLAAGVPS SGTD SLQP EDFATYYCQE HYAVPWTFGQ GTKLEIK 1. Additional Anti-PD-l Antibodies Additional anti-PD-l antibodies which may be utilized to generate the PD-l X CTLA—4 bispecific molecules of the instant invention are ed in Table 6.

Table 6: Additional Anti-PD-l Antibodies PD-l Antibodies Reference / Source PDl-l7; PDl-28; ; PD1-35; and PDl-F2 US Patents No. 7,488,802; 7,521,051; and 8,088,905; PCT Patent Publication WO 2004/056875 l7D8; 2D3; 4H1; 5C4; 4A1 l; 7D3; and 5F4 US Patents No. 8,008,449; 8,779,105; and 9,084,776; PCT Patent Publication WO 2006/121168 hPD-l.08A; hPD-1.09A; 109A, KO9A, 409A; US Patents No. 8,354,509; h409Al l; 6; h409Al7; Codon optimized 8,900,587; and 5,952,136; PCT 109A; and Codon optimized 409A Patent Publication WO 2008/156712 1E3; 1E8, and 1H3 US Patent ation 2014/004473 8; PCT Patent Publication 9A2; 10B11; 6E9; APE1922; APE1923; APE1924; APE1950; APE1963; and APE2058 2014/179664 GAl; GA2; GBl, GB6; GHl; A2; C7; H7; SH-A4; US Patent ation SH-A9; RG1H10; RG1H11; RG2H7; RGZHIO; 2014/0356363; PCT Patent RG3E12, RG4A6; RG5D9; RG1H10-H2AIS; Publication wo 2014/194302 RG1H10-H2AZS; RG1H10-3C; RG1H10-16C; RG1H10-17C; RG1H10-19C; RG1H10-21C; and RGlH10-23C2 Table 6: Additional Anti-PD-l Antibodies PD-l Antibodies nce / Source H1M7789N; H1M7799N; H1M78OON' US Patent Publication H2M7780N; H2M7788N; H2M7790N; 2015/0203579; PCT Patent 1N; H2M7794N; H2M7795N; Publication WO 12800 H2M7796N; H2M7798N; H4H9019P; H4xH9034P2;H4xH9035P2;H4xH9037P2 H4XH9045P2; H4XH9O48P2; H4H9057P2; H4H9068P2; H4XH91 19P2; 20P2; H4Xh9128p2; H4Xh9135p2; H4Xh9l45p2; H4Xh8992 o; H4Xh8999 o; and H4Xh9008 .; PD-l mAb 1; PD-l mAb 2; hPD-l mAb 2; PD-l US Patent Applications No. mAb 3; PD-l mAb 4; PD-l mAb 5; PD-1 mAb 6; 62/198,867 and 62/239,559 PD—l mAb 7; hPD—l mAb 7; PD-l mAb 8; PD-l mAb 9; hPD-l mAb 9; PD-l mAb 10; PD-l mAb 11; PD-l mAb 12; PD-l mAb 13; PD-l mAb 14; PD—l mAb 15; and hPD-l mAb 15 J. Exemplary anti-PD-l Antibody An exemplary anti-PD-l antibody designated “PD-1 mAb 6 G4P” comprises: a heavy chain having the VH Domain of PD-l mAb 61 (SEQ ID NO:86), an IgG4 CH1 Domain (SEQ ID NO:42); a stabilized IgG 4 Hinge (SEQ ID NO:36); and IgG4 CH2-CH3 Domains lacking the C-terminal lysine (SEQ ID NO:4); and a light chain having the VL Domain ofPD- 1 mAb 6SQ (SEQ ID NO:87) and a kappa CL (SEQ ID NO:38), The amino acid sequence of the complete heavy chain of PD-l mAb 6 G4P (SEQ ID NO:88) is shown below.

QVQLVQSGAE VKKPGASVKV SCKASGYSFT SYWMNWVRQA WIGV IHPSDSETWL DQKFKDRVTI TVDKSTSTAY MELSSLRSED TAVYYCAREH AYWG QGTLVTVSSA STKGPSVFPL APCSRSTSES TAALGCLVKD YFPEPVTVSW NSGALTSGVH TFPAVLQSSG LYSLSSVVTV PSSSLGTKTY TCNVDHKPSN TKVDKRVESK BCPA RflELGGPSVF LEPPKPKDTL MISRTPEVTC VVVDVSQEDP EVQFNWYVDG VEVHNAKTKP REEQFNSTYR VVSVLTVLiQ KVSNKGLBSS fiKTTS {AKG Q?REPQVYTL PPSQEEMT<N LVKG EYPSDIAVLW LSNGQRflNNY KTTPPVLDSD GSFFTYSRTT QEGN MHEA WHNHYTQKSL LSLG The amino acid sequence of the complete light chain of PD-l mAb 6 G4P (SEQ ID NO:89) is shown below.

EIVLTQSPAT LSLSPGERAT ESVD NYGMSFMNWF QQKPGQ?PKL LIHAASNQGS GVPSRFSGSG SGTDFTLT S SLfiP fiDhAVY ECQQSKEVPY TFGGGTKVEI KRTVAABSVE iFPBSDEQLK SGTASVVCLL NNFYPREAKV QWKVDNALQS GNSQ?SVT?Q YSLS STLTLSKADY EKHKVYACEV THQGLSSPVT KSFNRGEC VII. Anti-CTLA-4 Binding Capabilities Antibodies that are immunospecific for CTLA-4 are known (see, e.g., United States Patents No. 6,984,720, 6,682,736; 7,034,121, 7,109,003, 7,132,281; 7,411,057, 7,605,238; 7,807,797; 7,824,679; 8,017,114; 8,143,379; 8,318,916; 895; 8,784,815; and 8,883,984; US Patent Publications 2009/0123477; 2009/0252741; and 105914; PCT Patent Publications No. WO 00/37504; WO 01/14424; WO 01/54732; 2006/066568; and useful in the generation of the PD-l X CTLA-4 bispecific molecules of the present invention are capable ofbinding to a continuous or discontinuous (e.g., conformational) n (epitope) of human CTLA-4 and will preferably also t the ability to bind to CTLA-4 les of one or more non-human species, in particular, primate species (and especially a primate species, such as cynomolgus monkey), Additional desired antibodies may be made by isolating antibody-secreting hybridomas elicited using CTLA-4 or a peptide fragment thereof. A representative human CTLA-4 polypeptide (NCBI Sequence NP_005205.2; including a 35 amino acid residue signal sequence (shown underlined) and the 188 amino acid residues of the mature protein) has the amino acid sequence (SEQ ID NO:75): MACLGFQRHK AQLNLATRTW PCTLLFFLLF IPVFCKAMHV AQPAVVLASS RGHASFVCLY ASPGKATLVR VTVLRQADSQ VTEVCAATYM MGNELTFLDD SICTGTSSGN QVNLTIQGLR AMDTGLY CK PPYY LGIGNGTQIY VIDPEPCPDS DFLLWILAAV SSGLFFYSFL LTAVSLSKML KKRSPLTTGV YVKMPPTfiPfi CfiKQEQPYEi PHW Preferred anti-CTLA-4 binding les (e.g., antibodies) useful in the generation of the PD-l X CTLA-4 bispecific molecules of the t invention possess the VL and/or VH Domains of the anti-human CTLA-4 monoclonal dy “CTLA-4 mAb 1” mumab, CAS Reg. No: 9, also known as MDXOIO, and ed as YERVOY® by Bristol-Myers Squibb); “CTLA-4 mAb 2” (tremelimumab, CAS Reg. No.: 745013-59—6, also known as CP-675206); “CTLA-4 mAb 3” (4B6 as provided in Table 7) or any of the other anti—CTLA—4 antibodies in Table 7; and more preferably possess 1, 2 or all 3 of the CDRLS of the VL Region and/or 1, 2 or all 3 of the CDRHs of the VH Domain of such anti-CTLA-4 monoclonal antibodies.

A. CTLA-4 mAb 1 The amino acid sequence of the VH Domain of CTLA-4 mAb 1 (SEQ ID NO:76) is shown below (CDRH residues are shown underlined).

QVQLVESGGG VVQPGRSLRL SCAASGFTFS SYTMHWVRQA PGKGLEWVTF ISYDGNNKYY ADSVKGRFTI SRDNSKNTLY LQMNSLRAED TAIYYCARTG WLGPFDYWGQ GTLVTVSS The amino acid sequence of the VL Domain of CTLA-4 mAb 1 (SEQ ID NO:77) is shown below (CDRL residues are shown underlined).

EIVLTQSPGT LSLSPGERAT LSCRASQSVG SSYLAWYQQK PGQAPRLLIY GAFSRATGIP DRFSGSGSGT DFTLTISRLE PEDFAVYYCQ WTFG B. CTLA-4 mAb 2 The amino acid sequence of the VH Domain of CTLA-4 mAb 2 (SEQ ID NO:78) is shown below (CDRH residues are shown underlined).

QVQLVESGGG VVQPGRSLRL SCAASGFTFS SYGMHWVRQA PGKGLEWVAV NKYY ADSVKGRFTI SRDNSKNTLY RAED ARDP RGATLYYYYY GMDVWGQGTT VTVSS The amino acid sequence of the VL Domain of CTLA-4 mAb 2 (SEQ ID NO:79) is shown below (CDRL residues are shown ined).

DIQMTQSPSS LSASVGDRVT ITCRASQSIN SYLDWYQQKP LIYA ASSLQSGVPS RFSGSGSGTD FTLTISSLQP EDFATYYCQQ YYSTPFTFGP GTKVEIK C. CTLA-4 mAb 3 The amino acid sequence of the VH Domain of CTLA-4 mAb 3 (SEQ ID NO:90) is shown below (CDRH residues are shown underlined).

SGGG VVQPGRSLRL SCAASGFTFS VRQA PGKGLEWVTF ISYDGSNKHY ADSVKGRFTV SRDNSKNTLY LQMNSLRAED TAIYYCARTG WLGPFDYWGQ GTLVTVSS The amino acid sequence of the VL Domain of CTLA-4 mAb 3 (SEQ ID NO:91) is shown below (CDRL residues are shown underlined).

EIVLTQSPGT LSLSPGERAT LSCRASQSVS SSFLAWYQQK PGQAPRLLIY GASSRATGIP DRFSGSGSGT DFTLTISRLE PEDFAVYYCQ QYGSSPWTFG QGTKVEIK D. Additional Anti-CTLA-4 Antibodies Additional anti-CTLA-4 antibodies which may be utilized to generate the PD-1 X CTLA-4 bispecific molecules of the instant invention are provided in Table 7.

Table 7: Additional Anti-CTLA-4 dies CTLA-4 Antibodies Reference / Source mAb 26 US Patent No. 7,034,121; PCT Patent Publication WO 01/54732 10D1, 1E2; and 4B6 US Patents No. 6,984,720; 238, 8,017,114; 8,318,916, and 8,784,815, PCT Patent Publication W0 01/14424 2.13, 3.1.1, 4.1.1; 4.8.1, 4.91, 4102, US Patents No. 6,682,736, 7,109,003, 4.13.1,4.14.3,6.1.1,11.2.1,11.6.1,11.7.1, 7,132,281; 7,411,057, 7,807,797; 7,824,679, 12.241, 1231, 1231.1, 129.1, and 12.911 8,143,379; 8,491,895, and 8,883,984, PCT Patent Publication WO 00/3 7504 3B10; 8H5; 8H5-1B1, 3B10-4F7; 3; US Patent Publication 2014/0105914, PCT 2C7-1G10; 3B10-6E3, and 1 Patent Publication 3.7F10A2; 4.3F6B5, 4.4A7F4, 4.6C1E3, US Patent Publication 2009/0123477, PCT 4.7A8H8; 4.7E11F1; 4.8H10H5; TGN2122; Patent Publication and TGN2422 L3D10, L1B11, K4G4, KMlO, and YL2 US Patent Publication 2009/0252741, PCT Patent Publication E. Exemplary anti-CTLA-4 dies An ary anti-CTLA-4 antibody designated “CTLA-4 mAb 3 GlAA” comprises a heavy chain having the VH Domain ofCTLA-4 mAb 3 (SEQ ID NO:90), an IgGl CH1 Domain (SEQ ID NO:40), an IgGl Hinge (SEQ ID , and IgG1 CH2-CH3 Domains the substitutions L234A/L235A (SEQ ID NO:43).

The amino acid ce of the complete heavy chain of CTLA-4 mAb 3 GlAA (SEQ ID NO:92) is shown below.

SGGG VVQPGRSLRL SCAASGFTFS SYTMHWVRQA PGKGLEWVTF ISYDGSNKHY ADSVKGRFTV SRDNSKNTLY LQMNSLRAED TAIYYCARTG WLGPFDYWGQ GTLVTVSSAS FPLA PSSKSTSGGT AALGCLVKDY FPEPVTVSWN GV-T FPAVLQSSGL YSLSSVVTVP SSSLGTQTYI CNVNHKPSWT KVDKRVEPKS CDKTHTCPPC PAPEAAGGPS VFLFPPKPKD TLW SRTPfiV TCVVVDVS E WWYV DGVEVHNAKT KPREEQYNST YRVVSVLTVL HQDWLNGKEY KC<VSNKALP AP EKT SKA KGQPREPQVY TLPPSREEWT KNQVSLTCLV D AV fiWflSNGQPEN NYKTTPPVLD SDGSFFLYSK LTVDKSRWQQ GNVFSCSVMH EALHNHYTQK SLSLSPGK An alternative exemplary anti-CTLA—4 antibody designated “CTLA-4 mAb 3 G4P” comprises a heavy chain having the VH Domain of CTLA-4 mAb 3 (SEQ ID NO:90), an IgG4 CH1 Domain (SEQ ID NO:42), a stabilized IgG4 Hinge (SEQ ID NO:36), and IgG4 CH2—CH3 Domains g the C-terminal lysine (SEQ ID NO:4). The amino acid sequence of the complete heavy chain of CTLA-4 mAb 3 G4P is shown below (SEQ ID NO:93).

QVQLVESGGG VVQPGRSLRL SCAASGFTFS SYTMHWVRQA PG<GLEWVTF ISYDGSNKHY ADSVKGRFTV NTLY LQMNSLRAED TAIYYCARTG WLGBEJYWGQ GTLVTVSSAS TKGPSVFPLA SEST AALGCLVKDY FPL‘LJ N SGALTSGVHT FPAVLQSSGL YSLSSVVTVP SSSLGTKTYT CNVDflXPSNT KVDKRVESKY GPBCPPCBAP SVEL EBBKPKDTLM IS?T?EVTCV EDPE VQFNWYVDGV EVHNAKTKPR TYRV VSVLTVLHQD WLWGKEYKCK VSNKGLPSSI EKTISKAKGQ PEEPQVYULP PSQEEWTKNQ VSLTCLVKGF YPSDIAVEWE SNGQPENNYK TT?PVLDSDG SFFLYSRLTV DKSRWQEGNV FSCSVMHEAL {NHYTQKSLS LSLG The amino acid sequence of the te light chain of CTLA-4 mAb 3 GlAA and CTLA-4 mAb 3 G4P (SEQ ID NO:94) is shown below.

EIVLTQSPGT WSLSPGERAT LSCRASQSVS SSFLAWYQQK PGQAPQLLIY GASSRATGIP DRFSGSGSGT DFTLTISRLE PEDFAVYYCQ QYGSSPWTFG QGTKVEIKRT VAAPSVTIFP ?SDEQLKSGT ASVVCLLNNF YPREA<VQWK VDNALQSGNS QESVTEQDSK DSTYSLSSTL TLSKADYEKH KVYACQVUHQ.L GLSSPVTKSF NRGEC The exemplary anti-CTLA-4 antibodies, CTLA-4 mAb 3 GlAA and CTLA-4 mAb 3 G4P, both comprise a light chain having the VL Domain of CTLA-4 mAb 3 (SEQ ID NO:91) and a kappa CL (SEQ ID NO:38).

VIII. Exemplary PD-l x CTLA-4 Bispecific Molecules A. Exemplary Four Chain Fc Region-Containing Diabodies Having E/K- Coils Three exemplary PD-l X CTLA-4 bispecific, four-chain, Fc Region—containing ies, comprising E/K—coil Heterodimer-Promoting Domains were generated (designated “DART B,” “DART C,” and “DART D”). The structure of these Fc Region-containing ies is detailed below. These ary PD-l X CTLA-4 diabodies are intended to illustrate, but in no way limit, the scope of the invention. 1. DART B DART B is a bispecific, four—chain, Fc Region-containing diabody having two binding sites specific for PD-l, two binding sites specific for CTLA-4, a variant IgG4 Fc Region engineered for extended ife, and E/K-coil Heterodimer—Promoting Domains. The first and third polypeptide chains of DART B comprise, in the N—terminal to C-terminal direction: an N—terminus, a VL Domain of a monoclonal antibody capable ofbinding to CTLA- 4 (VLCTLA-4 CTLA-4 mAb l VL) (SEQ ID NO:77); an intervening linker peptide (Linker 1: GGGSGGGG (SEQ ID NO:5)); a VH Domain of a monoclonal antibody capable of binding to PD-l (VHPD-l PD-l mAb 6—1 VH) (SEQ ID NO:86); a cysteine-containing intervening linker peptide (Linker 2: GGCGGG (SEQ ID ; a cysteine—containing Heterodimer-Promoting (E4mfl) Dommn (HVAACEK—HVAALHK-HVAALHK—HVAALHK (SEQ ID ; a stabilized IgG4 hinge region (SEQ ID NO:36); a variant of an IgG4 CH2-CH3 Domain comprising substitutions M252Y/8254T/T256E and lacking the C-terminal residue (SEQ ID NO:81); and a C-terminus.

] The amino acid ce of the first and third polypeptide chains of DART B is (SEQ ID NO:95): EIVLTQSPGT LSLSPGERAT LSCRASQSVG YQQK PGQAPRLLIY GAFSRATGIP DRFSGSGSGT UFTTTTSRTF PEDFAVYYCQ QYGSSPWTFG IKGG GSGGGGQVQL VQSGAEVK<P SGYSFTSYWM NWVRQAPGQG LEWiGViHPS JSETWLDQKF KDRVTITVDK SWRSEDTAVY YCAREHYGTS PFAYWGQGTL VTVSSGGCGG AALEKEVAAL HKfiVAALfiKfi SKYGPBCPRC BABLELGGBS (Ill—ZIKJl—J DY TREPflV TCVVVUVSQ? NWYV DGVEVHNAKT RVVSVLTVL HQDWLNG<E KC<VSNKGLP ISKA KGQPREPQVY LPPSQEEWT KNQVSVTCVV KGTYPSDIAV QPEN NYKTT??VLD YSR LTVDKSRWQE GNVFSCSVMH EALHNHYTQK G The second and fourth ptide chains ofDART B comprise, in the N—terminal to C—terminal direction: an N—terminus, a VL Domain of a monoclonal antibody capable of g to PD-l (VLPD-l PD-l mAb 6-SQ VL) (SEQ ID NO:87); an intervening linker peptide (Linker 1: GGGSGGGG (SEQ ID NO:5)); a VH Domain of a monoclonal antibody capable of binding CTLA-4 (VHCTLA-4 CTLA-4 mAb l VH) (SEQ ID NO:76); a cysteine-containing intervening linker peptide (Linker 2: GGCGGG (SEQ ID NO:6)); a cysteine-containing Heterodimer—Promoting (K—coil) Domain (KVAACKE—KVAALKE—KVAALKE—KVAALKE (SEQ ID NO:21); and a C-terminus.

The amino acid sequence of the second and fourth polypeptide chains of DART B is (SEQ ID NO:96): EIVLTQSPAT ERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL LIHAASNQGS GVPSRFSGSG SGTDETLTiS SLEPEDEAVY ECQQSKEVPY TFGGGTKVEI KGGGSGGGGQ VQLVESGGGV VQPGRSLRLS CAASGFTFSS YTMHWVRQAP VTFI KYYA FTIS TLYL QMNSLRAEDT RTGW LG?FDYWGQG TLVTVSSGGC GGGKVAACKE KVAALKEKVA ALKEKVAALK E 2. DART C DART C is a bispecific, four-chain, Fc Region-containing diabody having two binding sites specific for PD-l, two binding sites specific for CTLA-4, a variant IgG4 Fc Region engineered for extended half-life, and E/K-coil Heterodimer—Promoting Domains. The first and third polypeptide chains of DART C comprise, in the N—terminal to C-terminal direction: an N—terminus, a VL Domain of a monoclonal antibody capable ofbinding to CTLA- 4 (VLCTLA-4 CTLA-4 mAb 3 VL) (SEQ ID NO:91); an intervening linker peptide (Linker 1: GGGSGGGG (SEQ ID NO:5)); a VH Domain of a monoclonal antibody capable of binding to PD—l (VHPD-l PD-l mAb 6-I VH) (SEQ ID N0:86); a cysteine-containing intervening linker peptide (Linker 2: GGCGGG (SEQ ID NO:6)); a cysteine—containing Heterodimer-Promoting l) Domain (EVAACEK—41VAAT.fitK—EVAALEK—RVAATEK (SEQ ID NO:20)); a stabilized IgG4 hinge region (SEQ ID NO:36); a variant of an IgG4 3 Domain comprising substitutions M252Y/SZS4T/T256E and lacking the C-terminal residue (SEQ ID NO:81); and a inus.

The amino acid sequence of the first and third polypeptide chains of DART C is (SEQ ID NO:97): EIVLTQSPGT TSLSPGERAT WSCRASQSVS SSTLAWYQQK PGQAPRLLIY GASSRATGIP DQFSGSGSGT DFTLT"SRLF YYCQ QYGSSPWTFG QGTKVEIKGG GSGGGGQVQL VQSGAEVKKP GASVKVSCKA SGYSFTSYWM NWVQQAPGQG TfiW GV {PS DSLTWLDQKF TVDK STSTAYMELS SLRSEDTAVY YCAREHYGTS PFAYWGQGTL VTVSSGGCGG GEVAACEKEV AALEKEVAAW EKfiVAALfiK? CPPC GGPS VFLFPPKPKD TLY TRHBfiV VSQS DPEVQEWWYV DGVEVHNAKT KPREEQFNST YRVVSVTTVT HQDWLNGKEY KCKVSNKGLP SSIEKTISKA KGQPREPQVY TLPBSQLLWT KNQVSLTCLV KGEYPSD AV fiWfiSNGQPEN NYKTTPPVLD SDGSFFLYSR LTVDKSRWQE GNVFSCSVMH EALHNHYTQK SLSLSLG The second and fourth polypeptide chains ofDART C se, in the N—terminal to inal direction: an N—terminus, a VL Domain of a monoclonal antibody capable of binding to PD-l (VLPD-i PD-l mAb 6-SQ VL) (SEQ ID NO:87); an intervening linker peptide (Linker 1: GGGSGGGG (SEQ ID NO:5)); a VH Domain of a monoclonal antibody capable of binding CTLA-4 (VHCTLA-4 CTLA-4 mAb 3 VH) (SEQ ID NO:90); a cysteine-containing intervening linker peptide (Linker 2: GGCGGG (SEQ ID N0:6)); a cysteine-containing Heterodimer-Promoting (K-coil) Domain (KVAACKE—KVAALKE—KVAALKE—KVAALKE (SEQ ID NO:21); and a inus.

The amino acid sequence of the second and fourth polypeptide chains of DART C is (SEQ ID NO:98): EIVLTQSPAT ERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL NQGS GVPSRFSGSG SGTJETLT S SLHBKJEAVY ECQQSKEVPY KVfi KGGGSGGGGQ VQLVESGGGV VQPGRSLRLS CAASGFTFSS YTMHWVRQAP GKGLEWVTFI SYDGSNKHYA DSVKGRFTVS RDNSKNTLYL QMNSLRAfiDT A YYCARTGW LGPFDYWGQG TLVTVSSGGC GGGKVAACKE KVAALKEKVA ALKEKVAALK E 3. DART D ] DART D is a bispecific, four-chain, Fc Region-containing diabody having two binding sites specific for PD-l, two binding sites specific for CTLA-4, a variant IgG4 Fc Region engineered for extended ife, and E/K-coil Heterodimer—Promoting Domains. The first and third polypeptide chains of DART D comprise, in the N—terminal to C-terminal direction: an inus, a VL Domain of a monoclonal antibody capable of g to PD-l (VLPD.1 PD-l mAb 6-SQ VL) (SEQ ID NO:87); an intervening linker peptide (Linker 1: GGGSGGGG (SEQ ID NO:5)); a VH Domain of a monoclonal antibody capable of binding CTLA-4 (VHCTLA-4 CTLA—4 mAb 3 VH) (SEQ ID N0:90); a cysteine-containing intervening linker peptide (Linker 2: GGCGGG (SEQ ID NO:6)); a cysteine-containing Heterodimer- Promoting (E-coil) Domain (EVAACEK— fiVAAL fiK-FVAALFK- fiVAAL fiK (SEQ ID NO:20)); a ized IgG4 hinge region (SEQ ID NO:36); a variant of an IgG4 CH2—CH3 Domain comprising substitutions 8254T/T256E and lacking the C-terminal residue (SEQ ID NO:81); and a C-terminus.

The amino acid sequence of the first and third polypeptide chains of DART D is (SEQ ID NO:99): EIVLTQSPAT LSLS?GERAT LSCRASESVD NYGMSTMNWF QQKPGQ?PKL LIHAASNQGS GVPSRFSGSG SGTDFTLTLS SLEPEDEAVY ECQQSKEVPY TFGGGTKV? _ KGGGSGGGGQ VQLVESGGGV VQ?GRSLRLS CAASGFTFSS YTMHWVRQAP GKGLEWVTFI SYDGSNKHYA DSVKGRFTVS RDNSKNTLYL QMNSLRAEDT AIYYCARTGW LGPFDYWGQG SGGC GGGEVAACEK FVAATFIKEVA ATE3KEVAALR K fiSKYGPBCP PCBAPLELGG PPKP KDTLYITREP EVTCVVVDVS QEDPEVQFNW YVDGVEVHNA KTKPREEQFN STYRVVSVLT VL GK EYKCKVSNKG LPSSIEKTIS KAKGQPREPQ VYTLPPSQEE MTKNQVSLTC LVKGFYPSDI AVEWESNGQP TPPV LDSDGSFFLY SRLTVDKSRW QEGNVFSCSV NHYT QKSLSLSLG ] The second and fourth polypeptide chains of DART D se, in the N- terminal to C-terminal direction: an N-terminus, a VL Domain of a onal dy e of binding to CTLA-4 (VLCTLA-4 CTLA-4 mAb 3 VL) ( SEQ ID NO:91); an intervening linker peptide r 1: GGGSGGGG (SEQ ID NO:5)); a VH Domain of a monoclonal antibody capable of binding to PD-1 (VHPD-1 PD-1 mAb 6-I VH) (SEQ ID NO:86); a cysteine-containing intervening linker peptide (Linker 2: GGCGGG (SEQ ID NO:6)); a cysteine -containing Heterodimer-Promoting (K-coil) Domain (KVAACKE- KVAALKE-KVAALKE-KVAALKE (SEQ ID NO:21); and a C-terminus.

] The amino acid sequence of the second and fourth polypeptide chains of DART D is (SEQ ID NO:100): EIVLTQSPGT LSLSPGERAT LSCRASQSVS SSFLAWYQQK PGQAPRLLIY GASSRATGIP DRFSGSGSGT DFTLTISRLE PEDFAVYYCQ QYGSSPWTFG QGTKVEIKGG GSGGGGQVQL VQSGAEVKKP GASVKVSCKA SGYSFTSYWM NWVRQAPGQG LEWIGVIHPS DSETWLDQKF KDRVTITVDK STSTAYMELS SLRSEDTAVY YCAREHYGTS QGTL VTVSSGGCGG GKVAACKEKV AALKEKVAAL KEKVAALKE 4. DART F DART F is a bispecific, four-chain, Fc Region-containing diabody having two binding sites specific for PD-1, two binding sites ic for CTLA-4, a variant IgG1 Fc Region engineered to reduce/eliminate effector function and to extend half-life, and il Heterodimer-Promoting Domains. The first and third polypeptide chains of DART F comprise, in the inal to C-terminal direction: an N-terminus, a VL Domain of a monoclonal antibody capable of binding to PD-1 (VLPD-1 PD-1 mAb 6-SQ VL) (SEQ ID ; an intervening linker peptide (Linker 1: GGGSGGGG (SEQ ID NO:5)); a VH Domain of a monoclonal antibody capable of binding CTLA-4 (VHCTLA-4 CTLA-4 mAb 3 VH) (SEQ ID NO:90); a cysteine-containing intervening linker peptide (Linker 2: GGCGGG (SEQ ID NO:6)); a cysteine-containing Heterodimer-Promoting (E-coil) Domain (EVAACEK- EVAALEK-EVAALEK-EVAALEK (SEQ ID NO:20)); a linker (SEQ ID NO:30); a variant of an IgG1 CH2-CH3 Domain comprising substitutions L235A/L235A/M252Y/S254T/T256E and lacking the C-terminal residue (SEQ ID NO:80); and a C-terminus.

The amino acid sequence of the first and third polypeptide chains of DART F (SEQ ID NO:101) is: EIVLTQSPAT LSLS?GERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL LIHAASNQGS GVPSRFSGSG SGTDFTLTTS SLEPEJEAVY ECQQSKEVPY TFGGGTKV?’ KGGGSGGGGQ VQLVESGGGV LRLS CAASGFTFSS YTMHWVRQAP GKGLEWVTFI SYDGSN<€YA DSVKGRFTVS RDNSKNTLYL QMNSLRAEDT AIYYCARTGW LG?FDYWGQG TLVTVSSGGC GGGEVAACEK EVAALEKT.L A ATMZKEVAATM AD<T {TCPPCPAPE VFLF PPKPKDT1Y_ TREPEVTCVV VDVSHED?E KFNWYVDGVE VHNAKTKPRE EQYNSTYRVV SVLTVLHQDW KC<V SNKALPAPIE KTISKAKGQP REPQVYTLPP SREEMTKNQV SLTCLVKGFY PSDIAVEWES NGQPENNYKT TPPVLDSDGS FFLYSKLTVD KSRWQQGNVF SCSVMHEALH NHYTQKSLSL The second and fourth polypeptide chains ofDART F comprise, in the N—terminal to C-terminal direction: an N—terrninus, a VL Domain of a monoclonal antibody capable of binding to CTLA-4 A-4 CTLA-4 mAb 3 VL) (SEQ ID NO:91), an intervening linker peptide (Linker 1: GG (SEQ ID NO:5)); a VH Domain of a monoclonal antibody capable of binding to PD—l (VHPD-l PD-l mAb 6-I VH) (SEQ ID NO:86); a cysteine- containing intervening linker peptide (Linker 2: GGCGGG (SEQ ID NO:6)); a cysteine- containing Heterodimer-Promoting (K-coil) Domain (KVAACKE—KVAALKE—KVAALKE— KVAALKE (SEQ ID NO:21), and a C-terminus.

The amino acid sequence of the second and fourth polypeptide chains of DART F is the same as that ofthe econd and fourth polypeptide chains ofDART D (SEQ ID NO: 100).

B. Exemplary Four-Chain Fc Region-Containing Diabodies Having CL/CHl s: DART E An exemplary PD-l X CTLA-4 bispecific, four-chain, Fc Region-containing y sing CL/CHl s designated “DART E” was generated. The structure of this Fc Region-containing diabodies is detailed below. This exemplary PD-l X CTLA-4 diabody is intended to illustrate, but in no way limit, the scope of the invention.

DART E is a bispecific, four—chain, Fc -containing diabody having two binding sites specific for PD-l, two binding sites specific for CTLA-4, CL/CHl Domains, and a variant IgG4 Fc Region engineered for extended half-life. The first and third polypeptide chains of DART E se, in the inal to inal direction: an N—terminus; a VL Domain of a monoclonal antibody capable of binding to CTLA—4 (VLCTLA-4 CTLA-4 mAb 3 VL) (SEQ ID NO:91); an ening linker peptide (Linker 1: GGGSGGGG (SEQ ID N0:5)); a VH Domain of a monoclonal antibody capable of binding to PD-l (VHPD-l PD-l mAb 6-I VH) (SEQ ID NO:86); an intervening linker peptide (Linker 2: LGGGSG (SEQ ID NO:8)); an IgG4 CH1 Domain (SEQ ID NO:42); a ized IgG4 hinge region (SEQ ID NO: 36); a variant of an IgG4 CH2-CH3 Domain sing substitutions M252Y/SZS4T/T256E and lacking the C-terminal e (SEQ ID NO:81); and a C-terminus.

The amino acid sequence of the first and third polypeptide chains of DART E is (SEQ ID NO:102): EIVLTQSPGT isLSPGERAT WSCRASQSVS SSTLAWYQQK PGQAPRLLIY GASSRATG__P JRESGSGSGT DFTLTISRLE PEDFAVYYCQ QYGSSPWTFG ’KGG GSGGGGQVQL VQSGAEVKKP GASVKVSCKA SGYSFTSYWM NWVRQAPGQG LfiW GV HES DQKF KDQVTITVDK STSTAYMELS SLRSEDTAVY YCAREHYGTS PFAYWGQGTL VTVSSLGGGS GASTKGPSVF PLAPCSRSTS ESTAALGCLV {DYFPEPVTV LTSG VHTFPAVLQS SGLYSLSSVV TVPSSSLGTK TYTCNVDHKP SNTKVDKRVE SKYGPPCPPC GGPS VELEPPKPKD TLYITQEPEV TCVVVDVSQE DPEVQFNWYV DGVEVHNA<T KPREEQFNST YQVVSVLTVL HQDWLNGKEY {CKVSNKGLP SSIEKTIS<A TWPPSQEEMT KNQVSWTCTV {GFYPSDIAV EWLSNGQPLN NYKTTPPVLD SDGSFFLYSR RWQE GNVFSCSVMH The second and fourth polypeptide chains ofDART E se, in the N—terminal to C-terminal direction: an N—terminus; a VL Domain of a monoclonal antibody capable of binding to PD-l (VLPD-l PD-l mAb 6-SQ VL) (SEQ ID NO:87); an intervening linker peptide (Linker 1: GGGSGGGG (SEQ ID NO:5)); a VH Domain of a monoclonal antibody capable of g CTLA-4 (VHCTLA-4 CTLA-4 mAb 3 VH) (SEQ ID NO:90); an intervening linker peptide (Linker 2: LGGGSG (SEQ ID NO:8)); a Kappa CL Domain (SEQ ID NO:38); and a C-terminus.

The amino acid sequence of the second and fourth polypeptide chains of DART E is (SEQ ID NO:103): EIVLTQSPAT LSLS?GERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL NQGS GVPSRFSGSG SGTDFTLTLS SLEPEDEAVY ECQQSKEVPY TFGGGTKVEI KGGGSGGGGQ VQLVESGGGV VQPGRSLRLS CAASGFTFSS YTWHWVRQAP GKGLEWVTFI SYDGSN<€YA DSVKGQFTVS TLYL QMVSLRAEDT AIYYCARTGW LG?FDYWGQG TLVTVSSLGG GSGRTVAAPS VEHEPPSDEQ LKSGTASVVC LLWNFYPQEA {VQWKVDNAL QSGNSQESVT EQ)SK)STYS LSSTLTLSKA DYEKHKVYAC EVTHQGLSSP VTKSFNRGEC C. Exemplary Trivalent Binding Molecules Containing Fc s Two exemplary PD-l X CTLA-4 bispecific, four—chain, Fc Region-containing trivalent binding molecules were generated (designated NT A” and “TRIDENT B”).

The structure of these Fc Region-containing trivalent binding molecules is detailed below.

Also presented below is a three chain variant designated “TRIDENT C,” which may be generated. These ary PD-l X CTLA-4 trivalent binding molecules are intended to rate, but in no way limit, the scope of the invention. 1. TRIDENT A TRIDENT A is a bispecific, four chain, Fc Region-containing trivalent binding molecule having two binding sites specific for PD-l, one binding sites specific for CTLA-4, a variant knob/hole-bearing IgG4 Fc Region ered for extended half-life, E/K—coil Heterodimer—Promoting Domains and CL/CHl Domains. The first polypeptide chain of T A ses, in the N—terminal to C—terminal direction: a VL Domain of a monoclonal antibody e of binding to PD-l I PD-l mAb 6-SQ VL) (SEQ ID NO:87); an intervening linker peptide (Linker 1: GGGSGGGG (SEQ ID NO:5)); a VH Domain of a monoclonal antibody capable of binding to PD-l (VHPD-I PD-l mAb 6-I VH) (SEQ ID NO:86); a cysteine-containing intervening linker peptide (Linker 2: GGCGGG (SEQ ID ; a cysteine-containing Heterodimer-Promoting (E-coil) Domain (EVAACEK— +1VAAT. +1K—EVAATEK— 41VAAT. 41K (SEQ ID NO:20)); a ized IgG4 hinge region (SEQ ID NO: 36); a knob-bearing IgG4 CH2-CH3 Domain comprising substitutions V1252Y/8254T/T256E and lacking the C-terminal residue (SEQ ID NO:84); and a C-terminus.

The amino acid sequence of the first polypeptide chain of TRIDENT A is (SEQ ID NO:104): EIVLTQSPAT ERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL LIHAASNQGS SGSG SGTDETLT S SLHBK)EAVY ECQQSKEVPY TFGGGTKVZ.L KGGGSGGGGQ VQLVQSGAEV KKPGASVKVS C(ASGYSFTS YWMNWVRQAP GQGLEWLGVL HBSDSETWLD QKEKDRVTLT VDKSTSTAYM FLSSWRSEDT AVYYCAREiY GTS?FAYWGQ GTLVTVSSGG AACE KFVAALF<EV AALEKFVAAL H<fiSKYGP3C PPCPABEHLG GBSVELEBPK PKDTLYITQE PEVTCVVVDV S QEDPEVQTN WYVDGVEVHN AKTKPREEQF NSTYQVVSVL TVLiQDWLWG KEY<CKVSWK GLPSS fiKTi S<AKGQPREP QVYTLPPSQE EMT<NQVSLW CLV<GFYPSD :AVEWESNGQ PENNYKTTPP VLDSDGSFFL DKSR WQEGNVFSCS VMHEALiNHY TQKSLSLSLG The second polypeptide chain of TRIDENT A comprises, in the N—terminal to C- terminal direction: an N—terminus, a VL Domain of a monoclonal antibody capable of binding to PD-l (VLPD-l PD-l mAb 6-SQ VL) (SEQ ID NO:87), an intervening linker e (Linker 1: GG (SEQ ID NO:5)); a VH Domain of a monoclonal antibody capable of binding to PD-l l PD-l mAb 6-1 VH) (SEQ ID NO:86); a cysteine-containing intervening linker peptide (Linker 2: GGCGGG (SEQ ID ; a ne-containing Heterodimer- Promoting (K-coil) Domain (KVAACKE—KVAALKE—KVAALKE—KVAALKE (SEQ ID NO:21)), and a C-terminus.

] The amino acid sequence of the second polypeptide chain of TRIDENT A is (SEQ ID NO:105): EIVLTQSPAT LSLSPGERAT LSCRASESVD NYGMSFMNWF PPKL LIHAASNQGS SGSG SGTDETLTiS SLLPEJEAVY ECQQSKEVPY TFGGGTKVEI KGGGSGGGGQ VQLVQSGAEV KK?GASVKVS CKASGYSFTS YWMNWVRQAP GQGLEWIGVI HPSDSETWLD QKFKDRVTIT TAYM ELSSLRSZDT AVYYCAREHY GTS?FAYWGQ GTLVTVSSGG CGGGKVAACK.L EKVAALKE<V AALKEKVAAL KB The third polypeptide chains of TRIDENT A comprises, in the N—terrninal to C- terminal direction: an N—terminus; a VH Domain of a monoclonal antibody capable of binding CTLA-4 (VHCTLA-4 CTLA-4 mAb 3 VH) (SEQ ID NO:90), an IgG4 CH1 Domain (SEQ ID NO:42), a stabilized IgG4 hinge region (SEQ ID NO: 36), a hole-bearing IgG4 CH2-CH3 Domain comprising substitutions M252Y/SZS4T/T256E and lacking the C-terminal residue (SEQ ID NO:81); and a inus.

The amino acid sequence of the third polypeptide chain of TRIDENT A (SEQ ID NO:106): SGGG VVQPGRSLRL SCAASGFTFS SYTMHWVQQA PGKGLEWVTF NKHY ADSVKGRFTV SRDNSKNTLY LQMNSLRAED TAIYYCARTG WLGPFDYWGQ GTLVTVSSAS TKGPSVF?LA SEST AALGCLVKUY FPEPVTVSWN SGALTSGVHT FPAVLQSSGL YSLSSVVTVP SSSLGTKTYT CNVDHKPSNT KVDKRVESKY GPPCPPC?AP EFLGGPSVFL DTLY ITQEPEVTCV VVDVSQfiDPfi VQENWYVDGV EVTNA<TKPR fifiQFNSTYRV WLNGKEYKCK VSNKGLPSSI EKTISKAKGQ PREPQVYTLP VSLSCAV<GT YPSDIAVEWE SNGQPENNYK TTPPVLDSDG DKSRWQEGNV FSCSVMHEAL HNRYTQKSLS LSLG The fourth polypeptide chain of TRIDENT A comprises, in the N—terrninal to C- terminal direction: an N—terminus; a VL Domain of a monoclonal antibody capable of binding to CTLA-4 (VLCTLA-4 CTLA-4 mAb 3 VL) (SEQ ID NO:91); a Kappa CL Domain (SEQ ID NO:38); and a C-terminus.

The amino acid sequence ofthe fourth polypeptide chain of TRIDENT A is (SEQ ID NO:107): EIVLTQSPGT LSLSPGERAT LSCRASQSVS SSFLAWYQQK LLIY GASSRATGIP DQFSGSGSGT DFTWT"SRLE PEDFAVYYCQ QYGSSPWTFG QGTKVEIKRT VAABSVEHEP BSDEQLKSGT ASVVCLLNNF YPREAKVQWK VDNALQSGNS QESVTEQDSK DSTYSLSSTL TLSKADYEKH KVYACEVTHQ GLSSBVTKSE NQGLC 2. TRIDENT B TRIDENT B is a bispecific, four-chain, Fc Region-containing trivalent binding molecule having two binding sites specific for PD-l, one binding sites specific for CTLA—4, a variant knob/hole-bearing IgGl Fc Region engineered to reduce/eliminate effector function and to extend half-life, E/K—coil Heterodimer—Promoting Domains and CL/CHl Domains. The first polypeptide chain of TRIDENT B comprises, in the N—terminal to C-terminal direction: a VL Domain of a monoclonal antibody capable of binding to PD-l (VLPD-l PD-l mAb 6-SQ VL) (SEQ ID NO:87); an intervening linker peptide (Linker 1: GGGSGGGG (SEQ ID NO:5)); a VH Domain of a monoclonal antibody capable of binding to PD-l l PD-l mAb 6-I VH) (SEQ ID NO:86); a cysteine-containing intervening linker e (Linker 2: GGCGGG (SEQ ID NO:6)); a cysteine-containing Heterodimer-Promoting (E—coil) Domain (EVAACEK— RVAATEK—FVAATEK—FVAATRK (SEQ ID NO:20)); a linker (SEQ ID NO: 31); a knob- g IgGl 3 Domain comprising substitutions L234A/L235A/M252Y/8254T/ T256E and lacking the C-terminal residue (SEQ ID N0:82); and a C-terminus.

The amino acid sequence of the first polypeptide chain of T B is (SEQ ID NO:108): SPAT LSLSPGERAT LSCRASESVD NYGMSTMNWF QQKPGQ?PKL LIHAASNQGS GVPSRFSGSG SGTDFTLTLS SLEPEDEAVY ECQQSKEVPY KVE: GGGQ VQLVQSGAEV KK?GASVKVS CKASGYSFTS YWMNWVRQAP GQGLEWIGVI TWLD QKFKDQVTIT TAYM ELSSLRSEDT AVYYCAREHY GTSPFAYWGQ GTLVTVSSGG CGGGEVAACE EKEV AALEKEVAAW EKGGGDKTHT CPPCPAPEAA GGPSVFLFPP KPKDTLYLTR EBEVTCVVVD VSflEDPEVKF NWYVDGVEVH NAKTKPREEQ YNSTYRVVSV LTVTHQDWLW KVSN XALPAP EKT SKAKGQPRE PQVYTLPPS? EEWTKNQVSL WCLVKGEYBS D AVEWfiSNG QPENNYKTTP PVLDSDGSFF LYSKLTVDKS RWQQGNVFSC SVMHEALHNH YTQKSLSLSP The second polypeptide chain of T B comprises, in the N—terminal to C- terminal direction: an N—terminus, a VL Domain of a monoclonal antibody capable of binding to PD-l (VLPD-l PD—l mAb 6-SQ VL) (SEQ ID NO:87), an intervening linker peptide (Linker 1: GGGSGGGG (SEQ ID NO:5)); a VH Domain of a onal antibody capable of binding to PD-l (VHPD-l PD—l mAb 6-I VH) (SEQ ID NO:86); a cysteine-containing intervening linker peptide (Linker 2: GGCGGG (SEQ ID NO:6)), a cysteine-containing Heterodimer- Promoting (K-coil) Domain (KVAACKE—KVAALKE—KVAALKE—KVAALKE (SEQ ID NO:21)), and a C-terminus.

The amino acid sequence of the second polypeptide chain of TRIDENT B is the same as that of the second polypeptide chain of TRIDENT A (SEQ ID NO:105): ] The third polypeptide chains of T B ses, in the N—terminal to C- terminal direction: an N—terminus, a VH Domain of a monoclonal antibody capable of binding CTLA-4 (VHCTLA-4 CTLA-4 mAb 3 VH) (SEQ ID N0:90), an IgGl CH1 Domain (SEQ ID NO:40), an IgGl hinge region (SEQ ID NO:33); a hole—bearing IgGl CH2-CH3 Domain comprising substitutions L234A/L235A/M252Y/S254T/T256E and lacking the C-terminal residue (SEQ ID NO:83); and a C—terminus.

The amino acid sequence of the third ptide chain of TRIDENT B is (SEQ ID NO:109): QVQLVESGGG VVQPGRSLRL SCAASGFTFS SYTMHWVRQA PGKGLEWVTF ISYDGSNKHY ADSVKGRFTV SQDNSKNTLY LQMNSLRAED TAIYYCARTG WLGBEJYWGQ GTLVTVSSAS TKGPSVFPLA PSSXSTSGGT AALGCLVKDY FPEPVTVSWN SGALTSGVHT FPAVLQSSGL YSLSSVVTVP SSSLGTQTYI PSNT EP<S CDKTHTCPBC PABjAAGGPS VELEBPKPKD mar-<HLYITREPEV TCVVVDVSEE D?EVKFVWYV DGVEVHNAKT KPREEQYNST RVVSVLTVL HQDWLNGKEY KCKVSN<ALP AP fiKT SKA PQVY EMT KNQVSLSCAV KGEYBS) AV HWHSNGQPEN PVLD DGSFFLVSK LTVDKSRWQQ GNVFSCSVMH EALHNRYTQK GK The fourth polypeptide chain of TRIDE\IT B comprises, in the N—terminal to C- terminal direction: an N—terminus; a VL Domain of a monoclonal antibody capable of binding to CTLA-4 (VLCTLA.4 CTLA-4 mAb 3 VL) (SEQ ID NO:91), a Kappa CL Domain (SEQ ID NO:38); and a C-terminus.

The amino acid ce of the fourth polypeptide chain of TRIDENT B is the same as that of the second polypeptide chain of TRIDENT A (SEQ ID NO:107). 3. TRIDENT C As provided herein, trivalent binding les sing three polypeptide chain may be generated by combining (e.g., fusing encoding polynucleotides, etc.) the binding domains of two te polypeptide chains into one chain. One bispecific, chain, Fc Region-containing trivalent binding molecule that may be generated has two binding sites specific for PD-l, one binding sites specific for CTLA-4, a variant knob/hole-bearing IgG4 Fc Region engineered for extended half-life, and E/K-coil Heterodimer-Promoting Domains (“TRIDENT C”). The first and second polypeptide chains of TRIDENT C may be identical to those of TRIDENT A provided above.

Where the first and second chains are identical to those of TRIDENT A, the third polypeptide chain of TRIDENT C may comprise, in the N-terminal to C-terminal direction: an N—terminus; a VL Domain of a monoclonal antibody capable of g to CTLA—4 (VLCTLA-4 CTLA-4 mAb 3 VL) (SEQ ID , an intervening spacer peptide (GGGGSGGGGSGGGGS (SEQ ID NO:37)); a VH Domain of a onal antibody capable of binding CTLA-4 (VHCTLA-4 CTLA-4 mAb 3 VH) (SEQ ID N0:90), a stabilized IgG4 hinge region (SEQ ID NO: 36); a hole-bearing IgG4 CH2-CH3 Domain comprising substitutions M252Y/8254T/T256E and lacking the C-terminal e (SEQ ID NO:85), and a C-terminus.

Thus, the amino acid sequence of the third polypeptide chain of TRIDENT C is (SEQ ID NO:110): EIVLTQSPGT LSLSPGERAT LSCRASQSVS SSFLAWYQQK LLIY GASSRATGIP GSGT DFTLTISRLE PEDFAVYYCQ QYGSSPWTFG QGTKVEIKGG GGSGGGGSGG GGSQVQLV:.L GGGVVQPGRS LQLSCAASGF TFSSYTMHWV RQAPGKGLEW VTFISYDGSN KAYADSVKGR FTVSRDNSKN TLYLQWNSLR AEDTAIYYCA RTGWLGPFDY WGQGTLVTVS SESKYGPPCP PCPAPLELGG PSVELEPBKP KDTLYLTREP LVTCVVVDVS QfiDBfiVQENW YVDGVEVHNA KTKPREEQFN STYRVVSVLT VLHQDWLWGK EYKCKVSNKG LESS EKT S KA<GQPREPQ VYTLPPSQEE SLSC AVKGFYPSD: AVEWESNGQP ENWYKTTPPV LDSDGSFFLV KSRW QEGNVFSCSV MHEALHNRYT QKSLSLSLG IX. Methods of Production The PD-1 X CTLA-4 bispecifrc molecules of the present invention are most preferably produced through the recombinant expression of nucleic acid molecules that encode such polypeptides, as is well-known in the art.

Polypeptides of the invention may be conveniently prepared using solid phase peptide synthesis (Merrifield, B. (1986) “SolidPhase Synthesis,” Science 232(4748):341-347, en, RA. (1985) “General Method For Ihe Rapid Solid-Phase sis 0f Large s OfPeptides: Specificity OfAntigen-Antibody Interaction At The Level OfIndividual Amino Acids,” Proc. Natl. Acad. Sci. (USA) :5131-5135; Ganesan, A. (2006) “Solid- Phase Synthesis In The Twenty-First Century,” Mini Rev. Med. Chem. 6(1):3-10).

In an ative, dies may be made recombinantly and expressed using any method known in the art. Antibodies may be made recombinantly by first isolating the antibodies made from host animals, obtaining the gene ce, and using the gene sequence to express the antibody inantly in host cells (e.g., CHO cells). Another method that may be employed is to express the antibody sequence in plants (e. g., tobacco) or transgenic milk. Suitable methods for expressing antibodies recombinantly in plants or milk have been disclosed (see, for example, Peeters et al. (2001) “Production OfAntibodies And Antibody nts 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 MilkAs A Method For The tion OfRecombinant Antibodies,” J. Immunol Methods 231:147- 157). Suitable methods for making tives of antibodies, e.g., humanized, single-chain, etc. are known in the art, and have been described above. In r alternative, antibodies may be made recombinantly by phage y technology (see, for example, US. Patents No. ,565,332, 5,580,717, 5,733,743, 6,265,150; and Winter, G. et al. (1994) “MakingAntibodies By Phage Display Technology,” Annu. Rev. Immunol. 12433-455).

Vectors containing polynucleotides of interest (e.g., polynucleotides encoding the polypeptide chains of the PD-1 X CTLA-4 bispeciflc molecules of the present invention) can be introduced into the host cell by any of a number of appropriate means, including electroporation, transfection employing calcium chloride, rubidium chloride, calcium phosphate, DEAE-dextran, or other sub stances, microproj ectile bombardment, lipofection; and infection (e.g., where the vector is an infectious agent such as ia virus). The choice of introducing s or cleotides will often depend on features of the host cell.

Any host cell capable of overexpressing heterologous DNAs can be used for the purpose of expressing a polypeptide or n of st. Non-limiting examples of suitable mammalian host cells include but are not limited to COS, HeLa, and CHO cells.

The invention includes polypeptides comprising an amino acid ce of the PD-l X CTLA-4 bispecific molecule of this invention. The polypeptides of this invention can be made by procedures known in the art. The polypeptides can be produced by proteolytic or other degradation of the antibodies, by inant methods (i.e., single or fusion polypeptides) as described above or by chemical synthesis. Polypeptides of the antibodies, especially shorter ptides up to about 50 amino acids, are conveniently made by chemical synthesis. s of chemical synthesis are known in the art and are commercially available.

] The invention includes variants of PD-l X CTLA-4 bispecific molecules, including functionally equivalent polypeptides that do not significantly affect the properties of such molecules as well as variants that have enhanced or sed activity. Modification of polypeptides is e practice in the art and need not be described in detail . Examples of modified polypeptides include polypeptides with conservative tutions of amino acid residues, one or more deletions or additions of amino acids which do not significantly deleteriously change the functional activity, or use of chemical analogs. Amino acid residues that can be vatively substituted for one another include but are not limited to: glycine/alanine, serine/threonine; /isoleucine/leucine, asparagine/glutamine, aspartic acid/glutamic acid; lysine/arginine; and phenylalanine/tyrosine. These polypeptides also include glycosylated and non-glycosylated ptides, 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 would be conservative, i.e., the substituted amino acid would possess similar chemical properties as that of the original amino acid. Such conservative substitutions are known in the art, and es have been provided above. Amino acid modifications can range from changing or modifying one or more amino acids to complete redesign of a region, such as the Variable Domain.

Changes in the Variable Domain can alter binding affinity and/or specificity. Other methods of modification include using coupling techniques known in the art, including, but not limited to, enzymatic means, oxidative substitution and chelation. Modifications can be used, for example, for attachment of labels for immunoassay, such as the attachment of radioactive moieties for radioimmunoassay. Modified polypeptides are made using established procedures in the art and can be screened using standard assays known in the art.

The invention encompasses fusion proteins comprising one or more of the polypeptides or antibodies of this invention. In one embodiment, a fusion polypeptide is provided that ses a light chain, a heavy chain or both a light and heavy chain. In r embodiment, the fusion polypeptide contains a heterologous immunoglobulin constant region.

In another embodiment, the fusion polypeptide contains a Light Chain Variable Domain and a Heavy Chain le Domain of an antibody produced from a publicly-deposited hybridoma.

For purposes of this invention, an antibody fusion protein contains one or more polypeptide domains that specifically bind to PD-l and/or CTLA-4 and r amino acid sequence to which it is not attached in the native molecule, for example, a heterologous sequence or a gous sequence from another region.

X. Uses of the PD-l x CTLA-4 Bispecific Molecules of the Present Invention The present invention encompasses compositions, including ceutical compositions, comprising the PD—l X CTLA-4 bispecific molecules of the present invention (e.g., bispeciflc antibodies, bispeciflc diabodies, trivalent binding molecules, etc), polypeptides derived from such molecules, polynucleotides comprising ces encoding such les or polypeptides, and other agents as described herein.

As discussed above, both PD—l and CTLA-4 play important roles in negatively regulating immune responses (e.g., immune cell proliferation, function and homeostasis). The PD-l X CTLA-4 bispecific molecules of the present invention have the ability to inhibit PD-l function, and thus reverse the PD-l-mediated immune system inhibition. In addition, the PD- 1 X CTLA-4 bispeciflc les of the present invention have the ability to inhibit CTLA-4 function and thus augment the immune system by blocking immune system tion mediated by PD-l and CTLA-4. The PD-l X CTLA-4 bispeciflc molecules of the present invention also allow for full de of both PD—l and CTLA-4, as well as blockade that is biased toward CTLA-4 when co-expressed with PD-l. Thus, the PD-l X CTLA-4 bispeciflc les ofthe invention are useful for relieving T-cell exhaustion and/or augmenting an immune response (e.g., a T-cell and/or NK-cell mediated immune se) of a subject. In particular, the PD-l X CTLA-4 bispecific molecules of the invention and may be used to treat any e or condition 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).

The cancers that may be treated by the PD-1 x CTLA-4 bispecific molecules of the present invention include s characterized by the presence of a cancer cell selected from the group consisting of a cell of: an adrenal gland tumor, an AIDS-associated cancer, an alveolar soft part sarcoma, an astrocytic tumor, bladder cancer, bone cancer, a brain and spinal cord cancer, a atic brain tumor, a breast cancer, a carotid body tumors, a cervical cancer, a chondrosarcoma, a chordoma, a chromophobe renal cell carcinoma, a clear cell carcinoma, a colon cancer, a colorectal cancer, a cutaneous benign fibrous histiocytoma, a desmoplastic small round cell tumor, an ependymoma, a Ewing’s tumor, an extraskeletal myxoid chondrosarcoma, a fibrogenesis ecta ossium, a fibrous dysplasia of the bone, a gallbladder or bile duct cancer, gastric cancer, a ional trophoblastic e, a germ cell tumor, a head and neck cancer, hepatocellular carcinoma, an islet cell tumor, a Kaposi’s Sarcoma, a kidney cancer, a leukemia, a lipoma/benign lipomatous tumor, a liposarcoma/malignant lipomatous tumor, a liver cancer, a lymphoma, a lung cancer, a medulloblastoma, a melanoma, a meningioma, a multiple endocrine neoplasia, a multiple myeloma, a myelodysplastic syndrome, a neuroblastoma, a ndocrine tumors, an ovarian cancer, a atic cancer, a ary thyroid carcinoma, a parathyroid tumor, a pediatric , a peripheral nerve sheath tumor, a phaeochromocytoma, a pituitary tumor, a prostate , a posterior uveal melanoma, a rare hematologic disorder, a renal metastatic cancer, a rhabdoid tumor, a rhabdomyosarcoma, a sarcoma, a skin cancer, a softtissue sarcoma, a squamous cell cancer, a h , a synovial sarcoma, a testicular cancer, a thymic carcinoma, a thymoma, a thyroid metastatic cancer, and a uterine cancer.

In particular, PD-1 x CTLA-4 bispecific molecules of the present invention may be used in the treatment of colorectal cancer, hepatocellular carcinoma, glioma, kidney cancer, breast , multiple myeloma, bladder cancer, neuroblastoma; sarcoma, dgkin’s ma, all cell lung cancer, ovarian cancer, pancreatic cancer and rectal cancer.

Infections that may be treated by the PD-1 X CTLA-4 bispecific molecules of the present invention include chronic viral, bacterial, fungal and tic infections. Chronic infections that may be treated by the PD-1 X CTLA-4 bispecific molecules of 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 deficiency Virus (HIV), Vesicular Stomatitis Virus (VSV), Bacilli, acler, Cholera, Diphtheria, Enterobacter, Gonococci, Helicobacler pylori, Klebsiella, ella, Meningococci, mycobacteria, Pseudomonas, Pneumonococci, rickettsia bacteria, Salmonella, Serratia, lococci, Streptococci, Tetanus, Aspergillus (A. fumigatus, A. niger, etc), Blasz‘omyces dermatilia’is, Candida (C. ns, C. , C. glabraia, C. lropicalis, eta), Cryptococcus neoformans, Genus Mucorales (mucor, absia’ia, rhizopus), Sporothrix ii, Paracoccidioides brasiliensis, ioia’es immitis, Histoplasma capsulalum, pirosis, ia burgdorferi, helminth parasite (hookworm, tapeworms, flukes, flatworms (e. g.

Schistosomia), Giardia Zambia, trichinella, Dientamoeba Fragilis, Trypanosoma brucei, Trypanosoma cruzi, and Leishmania donovani.

XI. Pharmaceutical Compositions The compositions of the invention include bulk drug compositions useful in the manufacture of pharmaceutical compositions (e.g., impure or non-sterile compositions) and pharmaceutical compositions (i.e., compositions that are suitable for stration to a subject or patient) that can be used in the preparation of unit dosage forms. Such compositions comprise a prophylactically or therapeutically effective amount of the PD-l X CTLA-4 ific molecules of the present invention, or a combination of such agents and a pharmaceutically acceptable carrier. Preferably, compositions of the invention comprise a prophylactically or therapeutically effective amount of the PD-l X CTLA-4 bispeciflc molecules of the present invention and a pharmaceutically acceptable carrier. The invention also encompasses such pharmaceutical itions that additionally include a second eutic antibody (e.g., tumor-specific monoclonal antibody) that is specific for a particular cancer antigen, and a pharmaceutically acceptable carrier.

In a specific embodiment, the term “pharmaceutically acceptable” means approved by a regulatory agency of the Federal or a state government or listed in the US.

Pharmacopeia or other generally recognized pharmacopeia for use in s, and more particularly in humans. The term “carrier” refers to a diluent, adjuvant (e.g, Freund’ s adjuvant (complete and incomplete), excipient, or vehicle with which the therapeutic is administered.

Generally, the ingredients of compositions of the invention are supplied either separately or mixed together in unit dosage form, for example, as a dry lyophilized powder or water free concentrate in a hermetically sealed container such as an ampoule or sachette indicating the ty of active agent. Where the composition is to be administered by infusion, it can be dispensed with an on bottle containing sterile pharmaceutical grade water or saline.

Where the composition is administered by injection, an ampoule of sterile water for injection or saline can be provided so that the ingredients may be mixed prior to administration.

The invention also provides a pharmaceutical pack or kit comprising one or more containers filled with a PD-l X CTLA-4 bispecific molecule of the present invention, alone or with such pharmaceutically acceptable carrier. Additionally, one or more other prophylactic or therapeutic agents useful for the treatment of a disease can also be included in the pharmaceutical pack or kit. The invention also provides a pharmaceutical pack or kit comprising one or more containers filled with one or more of the ingredients of the ceutical compositions of the invention. Optionally associated with such container(s) can be a notice in the form prescribed by a governmental agency regulating the manufacture, use or sale of ceuticals or biological ts, which notice reflects approval by the agency of manufacture, use or sale for human administration.

The t invention es kits that can be used in the above methods. A kit can comprise any of the PD—l X CTLA-4 bispecific molecules of the present invention. The kit can further comprise one or more other prophylactic and/or therapeutic agents useful for the ent of cancer, in one or more containers.

XII. Methods of Administration The compositions of the present invention may be provided for the treatment, prophylaxis, 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 a conjugated molecule of the invention, or a pharmaceutical composition comprising a fusion protein or a conjugated molecule of the invention. In a preferred aspect, such compositions are substantially purified (i.e., ntially free from substances that limit its effect or produce undesired side effects). In a specific ment, the subject is an animal, preferably a mammal such as non-primate (e.g., bovine, equine, feline, canine, rodent, etc.) or a primate (e.g., monkey such as, a cynomolgus monkey, human, etc). In a preferred embodiment, the t is a human. s ry systems are known and can be used to administer the compositions ofthe invention, e.g., encapsulation in liposomes, microparticles, microcapsules, recombinant cells capable of expressing the antibody or fusion n, receptor-mediated endocytosis (See, e.g., Wu et a]. (1987) “Receptor-MediatedIn Vitro Gene Transformation By A Soluble DNA Carrier System, ”J. Biol. Chem. 262:4429-4432), construction of a nucleic acid as part of a retroviral or other vector, etc.

Methods of administering a molecule of the 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). In a specific embodiment, the PD-l X CTLA-4 ific molecules of the present ion are administered intramuscularly, intravenously, or subcutaneously. The compositions may be administered by any ient route, for example, by infusion or bolus injection, by absorption through epithelial or mucocutaneous s (e.g., oral , rectal and intestinal mucosa, etc.) and may be administered together with other biologically active .

Administration can be systemic or local. In on, pulmonary administration can also be employed, e.g., by use of an inhaler or nebulizer, and formulation with an aerosolizing agent.

See, e.g., US. Patents No. 6,019,968; 5,985, 320, 5,985,309, 5,934,272, 5,874,064; 5,855,913, ,290,540, and 4,880,078; and PCT Publication Nos. WO 92/19244, WO 97/32572; WO 97/44013, WO 98/31346, and WO 99/66903, each h is incorporated herein by reference in its entirety.

The invention also provides that preparations of the PD-l X CTLA—4 bispeciflc molecules of the present invention are ed in a hermetically sealed ner such as an ampoule or te indicating the quantity of the le. In one embodiment, such molecules are supplied as a dry sterilized lyophilized powder or water free concentrate in a hermetically sealed container and can be reconstituted, e.g., with water or saline to the appropriate concentration for administration to a subject. Preferably, the PD-l X CTLA-4 bispeciflc molecules of the present ion are supplied as a dry sterile lyophilized powder in a hermetically sealed container.

] The lyophilized preparations of the PD-l X CTLA-4 bispeciflc molecules of the present invention should be stored at between 2°C and 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 being reconstituted. In an alternative embodiment, such molecules are supplied in liquid form in a hermetically sealed container indicating the quantity and concentration of the molecule, fusion protein, or ated molecule. Preferably, such PD-l X CTLA-4 bispecific molecules when ed in liquid form are supplied in a hermetically sealed container.

The amount of such preparations of the invention that will be effective in the treatment, prevention or amelioration of one or more symptoms associated with a disorder can be determined by standard clinical techniques. The precise dose to be ed in the formulation will also depend on the route of administration, and the seriousness of the condition, and should be decided according to the judgment of the practitioner and each patient’s circumstances. ive doses may be extrapolated from dose-response curves derived from in vitro or animal model test systems.

As used herein, an “effective ” of a pharmaceutical ition, in one embodiment, is an amount sufficient to effect beneficial or desired results including, without limitation, al results such as decreasing ms resulting from the disease, attenuating a symptom of infection (e.g., viral load, fever, pain, sepsis, etc.) or a symptom of cancer (e.g., the proliferation, of cancer cells, tumor presence, tumor metastases, etc), thereby increasing the quality of life of those suffering from the disease, decreasing the dose of other medications required to treat the disease, enhancing the effect of another medication such as via targeting and/or internalization, delaying the progression of the disease, and/ or prolonging survival of individuals.

An effective amount can be administered in one or more administrations. For purposes of this invention, an effective amount of drug, compound, or pharmaceutical composition is an amount sufficient: to kill and/or reduce the proliferation of cancer cells, and/or to eliminate, reduce and/or delay the development of metastasis from a primary site of ; or to reduce the proliferation of (or the effect of) an ious pathogen and to reduce and/or delay the development of the pathogen-mediated disease, either directly or indirectly.

In some embodiments, an effective amount of a drug, compound, or pharmaceutical ition may or may not be achieved in conjunction with another drug, compound, or ceutical composition. Thus, an “effective amount” may be considered in the context of administering one or more chemotherapeutic , and a single agent may be considered to be given in an effective amount if, in conjunction with one or more other agents, a desirable result may be or is achieved. While individual needs vary, determination of l ranges of effective amounts of each component is within the skill of the art.

For the PD-1 x CTLA-4 bispecific molecules encompassed by the invention, the dosage administered to a t is preferably determined based upon the body weight (kg) of the recipient subject. For the PD-1 x CTLA-4 ific molecules encompassed by the ion, the dosage administered to a patient is typically from about 0.01 μg/kg to about 150 mg/kg or more of the subject’s body weight.

The dosage and frequency of administration of a PD-1 x CTLA-4 bispecific molecule of the present invention may be reduced or altered by enhancing uptake and tissue penetration of the molecule by modifications such as, for example, lipidation.

The dosage of a PD-1 x CTLA-4 ific molecule of the invention administered to a patient may be calculated for use as a single agent therapy. Alternatively, the molecule may be used in combination with other therapeutic compositions and the dosage administered to a patient are lower than when the molecules are used as a single agent therapy.

The pharmaceutical compositions of the invention may be administered locally to the area in need of treatment; this may be achieved by, for example, and not by way of limitation, local infusion, by injection, or by means of an t, the implant being of a porous, non-porous, or gelatinous material, including membranes, such as sialastic membranes, or fibers. Preferably, when administering a molecule of the invention, care must be taken to use materials to which the le does not absorb.

] The compositions of the invention can be delivered in a vesicle, in particular a liposome (See Langer (1990) “New s Of Drug Delivery,” Science 249:1527-1533); Treat et al., in LIPOSOMES IN THE THERAPY OF INFECTIOUSDISEASE AND , Lopez-Berestein and Fidler (eds.), Liss, New York, pp. 353- 365 (1989); Berestein, ibid., pp. 3 17-327).

] Where the composition of the invention is a nucleic acid encoding a PD-1 x CTLA- 4 bispecific molecule of the t invention, the nucleic acid can be administeredin vivo to e expression of its encoded PD-1 x CTLA-4 bispecific molecule by constructing it as part of an appropriate nucleic acid expression vector and administering it so that it becomes intracellular,e.g., by use of a retroviral vector (See U.S. Patent No. 4,980,286), or by direct injection, or by use of microparticle bombardment (e.g., a gene gun; Biolistic, Dupont), or coating with lipids or cell surface receptors or transfecting agents, or by administering it in linkage to a homeobox-like peptide which is known to enter the nucleus (See e.g., Joliot et al. (1991) ‘Mntennapedia Homeobox Peptide Regulates Neural Morphogenesis,” Proc. Natl.

Acad. Sci. (USA) 88:1864—1868), etc. Alternatively, a nucleic acid can be introduced intracellularly and orated within host cell DNA for expression by homologous recombination.

] Treatment of a subject with a eutically or prophylactically effective amount of a PD-l X CTLA-4 bispecific molecule ofthe t invention can include a single treatment or, preferably, can include a series of treatments. In a preferred example, a subject is treated with such a y one time per week for between about 1 to 10 weeks, preferably between 2 to 8 weeks, more preferably between about 3 to 7 weeks, and even more preferably for about 4, 5, or 6 weeks. The pharmaceutical compositions of the invention can be administered once a day, twice a day, or three times a day. Alternatively, the pharmaceutical itions can 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 per year. It will also be appreciated that the effective dosage of the molecules used for treatment may increase or decrease over the course of a ular treatment.

XIII. Exemplary Embodiments The invention is particularly directed to the embodiments E1-E26: E1. A bispecific molecule possessing both one or more epitope-binding sites capable of immunospeciflc binding to (an) epitope(s) of PD—l and one or more epitope-binding sites capable of immunospeciflc binding to (an) epitope(s) of CTLA-4, wherein such molecule comprises: (A) a Heavy Chain Variable Domain and a Light Chain Variable Domain of an antibody that binds PD—l; and (B) a Heavy Chain Variable Domain and a Light Chain Variable Domain of an dy that binds CTLA-4, wherein such molecule is: (i) a diabody, such diabody being a covalently bonded complex that comprises two, three, four or five polypeptide chains; or (ii) a trivalent binding molecule, such trivalent binding le being a covalently bonded x that comprises three, four, five, or more polypeptide chains.

E2. The bispecific molecule of Embodiment El, wherein such molecule exhibits an activity that is enhanced relative to such activity exhibited by two monospecific molecules one ofwhich possesses such Heavy Chain Variable Domain and such Light Chain Variable Domain of such antibody that binds PD-l and the other of which possesses such Heavy Chain Variable Domain and such Light Chain Variable Domain of such antibody that binds CTLA-4.

E3. The bispecific molecule ofEmbodiment E1 or E2, wherein such le elicits fewer immune-related e events (irAEs) when administered to a subject in need thereof relative to such iREs elicited by the administration of a monospecific antibody that binds CTLA-4.

E4. The bispecific molecule of Embodiment E3, n said monospecific antibody that binds CTLA-4 is umab.

E5. The ific molecule of any one of Embodiments E1-E4, wherein such molecule comprises an Fc Region.

E6. The ific molecule ofEmbodiment E5, wherein such Fc Region is a variant Fc Region that comprises: (A) one or more amino acid modifications that reduces the affinity of the variant Fc Region for an FcyR, and/or (B) one or more amino acid modifications that enhances the serum ife of the variant Fc Region.

E7. The bispecific molecule of Embodiment E6, wherein such modifications that reduces the affinity of the variant Fc Region for an FcyR comprise the tution of L234A, L23 5A, or L234A and L23 5A, wherein such numbering is that of the EU index as in Kabat.

E8. The bispecific molecule of Embodiment E6 or E7, wherein such modifications that that enhances the serum ife of the variant Fc Region comprise the substitution of M252Y, M252Y and SZS4T, M252Y and T256E, M252Y, $254T and T256E; or K288D and H43 5K, wherein such numbering is that of the EU index as in Kabat. -lOO- E9. The ific molecule of any one of Embodiments E1-E8, wherein such molecule is such diabody and comprises two epitope-binding sites capable of immunospecific binding to an epitope of PD-l and two epitope-binding sites capable of immunospecific binding to an epitope of CTLA-4.

E10. The bispecific molecule of any one of Embodiments E1-E8, wherein such molecule is such ent binding molecule and comprises two epitope-binding sites capable of immunospecific binding to an epitope of PD-l and one epitope- binding site e of immunospecific binding to an epitope of CTLA-4.

E11. The ific molecule of any one of Embodiments El—ElO, wherein such molecule is capable of binding to PD-l and CTLA-4 les present on the cell e.

E12. The ific molecule of any one of Embodiments El—Ell, wherein such molecule is capable of simultaneously binding to PD-l and CTLA-4.

E13. The bispecific le of any one of Embodiments E1-E12, wherein such molecule promotes the stimulation of immune cells.

E14. The bispecific molecule of Embodiment E13, wherein such stimulation of immune cells results in: (A) immune cell proliferation; and/or (B) immune cell production and/or release of at least one cytokine; and/or (C) immune cell production and/or release of at least one lytic molecule; and/or (D) immune cell expression of at least one activation marker.

E15. The bispecific molecule of Embodiment E13 or E14, wherein such immune cell is a T-lymphocyte or an NK-cell.

E16. The bispecific molecule of any one of Embodiments El-ElS, wherein such epitope-binding sites capable of immunospecific binding to an epitope of PD-l (A) the VH Domain of PD-l mAb l (SEQ ID NO:47) and the VL Domain ofPD—l mAb l (SEQ ID NO:48); or (B) the VH Domain of PD-l mAb 2 (SEQ ID NO:49) and the VL Domain of PD—l mAb 2 (SEQ ID NO:50); or (C) the VH Domain of PD-l mAb 3 (SEQ ID NO:51) and the VL Domain ofPD-l mAb 3 (SEQ ID ; or (D) the VH Domain of PD-l mAb 4 (SEQ ID NO:53) and the VL Domain of PD-l mAb 4 (SEQ ID NO:54); or (E) the VH Domain of PD-l mAb 5 (SEQ ID NO:55) and the VL Domain ofPD-l mAb 5 (SEQ ID NO:56); or (F) the VH Domain of PD—l mAb 6 (SEQ ID NO:57) and the VL Domain of PD-l mAb 6 (SEQ ID NO:58); or (G) the VH Domain of PD-l mAb 6—1 VH (SEQ ID NO:86) and the VL Domain of PD-l mAb 6-SQ VL (SEQ ID NO:87); or (H) the VH Domain of PD-l mAb 7 (SEQ ID NO:59) and the VL Domain of PD—l mAb 7 (SEQ ID NO:60); or (I) the VH Domain of PD-l mAb 8 (SEQ ID NO:61) and the VL Domain of PD-l mAb 8 (SEQ ID NO:62).

E17. The bispecific molecule of any one of Embodiments El—El6, wherein such epitope-binding site(s) capable of immunospecific binding to an e of CTLA-4 comprise: (A) the VH Domain of CTLA-4 mAb 1 (SEQ ID NO:76) and the VL Domain of CTLA-4 mAb l (SEQ ID NO:77); or (B) the VH Domain of CTLA—4 mAb 2 (SEQ ID NO:78) and the VL Domain of CTLA-4 mAb 2 (SEQ ID NO:79); or (C) the VH Domain of CTLA—4 mAb 3 (SEQ ID NO:90) and the VL Domain of CTLA-4 mAb 3 (SEQ ID NO:91).

E18. The bispecific molecule of Embodiment 17, n: (A) such e-binding sites capable of immunospecific binding to an epitope of PD-l comprise the VH Domain of PD-l mAb 6—1 VH (SEQ ID NO:86) and the VL Domain of PD-l mAb 6-SQ (SEQ ID NO:87); (B) such epitope-binding site(s) capable of immunospecific binding to an epitope of CTLA-4 comprise(s) the VH Domain of CTLA-4 mAb 3 -lO2- (SEQ ID NO:90) and the VL Domain of CTLA-4 mAb 3 (SEQ ID NO:91).

E19. The bispecific molecule of any one of Embodiments El—E18, n such molecule comprises: (A) two polypeptide chains having SEQ ID NO:95, and two polypeptide chain having SEQ ID NO:96; or (B) two polypeptide chains having SEQ ID NO:97, and two polypeptide chain having SEQ ID NO:98; or (C) two polypeptide chains having SEQ ID NO:99, and two polypeptide chain having SEQ ID NO:100; or (D) two polypeptide chains having SEQ ID NO:102, and two polypeptide chain having SEQ ID NO:103; or (E) two polypeptide chains having SEQ ID , and two polypeptide chain having SEQ ID NO:100; or (F) one polypeptide chains having SEQ ID NO:104, one polypeptide chain having SEQ ID NO:105, one polypeptide chain having SEQ ID NO:106, and one polypeptide chain having SEQ ID NO:107; or (G) one polypeptide chains having SEQ ID NO:108, one ptide chain having SEQ ID NO:105, one polypeptide chain having SEQ ID NO:109, and one polypeptide chain having SEQ ID NO:107.

E20. A pharmaceutical composition that comprises an effective amount of the bispeciflc molecule of any of Embodiments E1-E19 and a pharmaceutically acceptable carrier.

E21. The bispecific molecule of any one of Embodiments El—El9, wherein such molecule is used to promote ation of an immune-mediated response of a subject in need thereof.

E22. The bispeciflc le of any one of ments , wherein such molecule is used in the treatment of a disease or condition associated with a suppressed immune system.

E23. The bispecific le of Embodiment E22, wherein the disease or condition is cancer or an infection. -lO3- E24. The bispecific molecule of Embodiment E23, wherein such cancer is characterized by the presence of a cancer cell selected from the group consisting of a cell of: an l gland tumor, an ssociated , an alveolar soft part sarcoma, an astrocytic tumor, bladder cancer, bone cancer, a brain and spinal cord cancer, a atic brain tumor, a breast cancer, a carotid body tumors, a cervical cancer, a chondrosarcoma, a chordoma, a chromophobe renal cell carcinoma, a clear cell carcinoma, a colon cancer, a colorectal cancer, a cutaneous benign fibrous histiocytoma, a desmoplastic small round cell tumor, an ependymoma, a Ewing’s tumor, an keletal myxoid chondrosarcoma, a fibrogenesis imperfecta ossium, a fibrous sia of the bone, a gallbladder or bile duct cancer, gastric cancer, a gestational trophoblastic disease, a germ cell tumor, a head and neck cancer, hepatocellular carcinoma, an islet cell tumor, a Kaposi’s Sarcoma, a kidney , a leukemia, a lipoma/benign lipomatous tumor, a liposarcoma/malignant lipomatous tumor, a liver cancer, a lymphoma, a lung cancer, a medulloblastoma, a melanoma, a meningioma, a multiple endocrine neoplasia, a multiple myeloma, a myelodysplastic syndrome, a neuroblastoma, a neuroendocrine tumors, an ovarian , a pancreatic cancer, a papillary d carcinoma, a parathyroid tumor, a pediatric cancer, a peripheral nerve sheath tumor, a hromocytoma, a pituitary tumor, a prostate , a posterior uveal melanoma, a rare hematologic disorder, a renal metastatic , a rhabdoid tumor, a rhabdomyosarcoma, a sarcoma, a skin cancer, a soft-tissue a, a squamous cell cancer, a stomach cancer, a synovial sarcoma, a testicular cancer, a thymic carcinoma, a thymoma, a thyroid metastatic cancer, and a uterine cancer.

E25. The bispecific le of Embodiment E24, wherein such infection is characterized by the presence of a bacterial, fungal, viral or protozoan pathogen.

E26. The bispecific molecule of Embodiment E25, wherein such infection is characterized by the presence of 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), Bacilli, acter, Cholera, Diphtheria, Enterobacter, Gonococci, Helicobacter pylori, Klebsiella, Legionella, Meningococci, cteria, Pseudomonas, Pneumonococci, rickettsia bacteria, Salmonella, Serrall'a, Staphylococci, Streptococci, Tetanus, Aspergillus (A. fumigalus, A. niger, etc), myces dermatitia’is, Candida (C. albicans, C. krusez', C. ta, C. tropicalz's, etc. ), Cryptococcus neoformans, Genus Mucorales (mucor, absidl'a, rhizopus), Sporothrz'x schenkn', Paracoccz'dioia’es brasilz'ensis, Coccidioz'a’es immitis, Histoplasma capsulatum, Leptospirosis, Borrell'a burga’orferi, helminth parasite (hookworm, tapeworms, flukes, flatworms (e.g. Schistosomia), Giara’l'a lambia, z'nella, moeba Fragilz's, Trypanosoma brucei, Trypanosoma cruzz', and Leishmania donovani.

Having now generally described the invention, the same will be more readily understood h reference to the following Examples. The following examples illustrate various methods for compositions in the diagnostic or treatment s of the invention. The examples are intended to illustrate, but in no way limit, the scope of the invention.

Example 1 Bispecific Molecules Provide Enhanced Stimulation of Immune Responses A bispeciflc molecule having specificity for distinct cell surface ns that modulate two immunomodulatory pathways, PD-l and LAG-3, was ted and designated “DART A.” DART A is a bispecific, four chain, Fc Region-containing y having two binding sites specific for PD-l, two binding sites specific for LAG-3, a variant IgG4 Fc Region ered for ed half-life, and cysteine-containing E/K-coil Heterodimer—Promoting Domains. As provided in more detail below, DART A comprises the binding specificities (i.e., the VH and VL Domains) of a humanized anti-PD-l antibody (hPD-l mAb 6) and a humanized anti-LAG—3 dy 3 mAb 1). The amino acid sequence of the first and third polypeptide chains of DART A is (SEQ ID NO:63): DIQWTQSPSS LSASVGJRVT TCRASQDVS SVVAWYQQKP GKAPKLLIYS ASYRYTGVPS RFSGSGSGTD FTLTISSLQP EDFATYYCQQ HYSTPWTTGG GTKLEIKGGG SGGGGQVQLV QSGAEVKKPG ASVKVSCXAS GYSFTSYWMN WVRQAPGQGT EW GVTHPSD SjTWLDQKFK DRVTITVDKS TSTAYWEUSS AVYY CAREHYGTS? FAYWGQGTLV TVSSGGCGGG fiVAACfiKfiVA ALEKEVAAL? KEVAALEKES KYGPPCPPCP A?EFLGG?SV FLFPPK?KDT LY TREPfiVT SQflD PfiVQFNWYVD GVEVHNAKTK PREEQFNSTY RVVSVLTVLH QDWLNGKEYK C(VSNKGLPS S fiKT SKAK GQPREPQVYT LPPSQEEWTK NQVSUTCWVK GTYPSDIAVE WESNGQPENN VLDS DGSFFLYSRL TVDKSRWQEG NVFSCSVMHE ALINHYTQKS LSLSLG In SEQ ID NO:63, amino acid residues 1-107 pond to the amino acid sequence of a VL Domain of a humanized monoclonal antibody capable of binding to LAG-3 3 mAb 1); residues 108-115 correspond to the intervening spacer peptide (Linker 1: GGGSGGGG (SEQ ID NO:5)); residues 116-234 correspond to the VH Domain of a monoclonal antibody capable of binding to PD-1 (hPD-1 mAb 6, SEQ ID NO:57, wherein X1 is I); residues 235-240 correspond to an ening spacer peptide (Linker 2: GGCGGG (SEQ ID ; residues 8 correspond to a ne-containing Heterodimer-Promoting (E- coil) Domain (41VAAC +1K—RVAALFK— 41VAAT. +1K— 41VAAL *ZK (SEQ ID ); es 0 correspond to a stabilized IgG4 Hinge Region (SEQ ID NO:36); residues to 281-496 correspond to a variant of IgG4 CH2-CH3 Domain comprising substitutions M252Y/SZS4T/T256E and lacking the C-terminal residue.

The amino acid sequence of the second and fourth polypeptide chains of DART A is (SEQ ID NO:64): EIVLTQSPAT LSLS?GERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL LIHAASNQGS GVPSRFSGSG SGTDFTLTIS SLEPEDEAVY ECQQSKEVPY TFGGGTKVEI KGGGSGGGGQ VQLVQSGAEV KKPGASVKVS CKASGYTFTD YNMDWVRQAP GQGLEWMGDI WPDNGVTIYN QKFEGQVTMT TDTSTSTAYM ELRSLRSJJT AVYYCAREAD YFYFDYWGQG TTLTVSSGGC GGGKVAACKE KVAALKE<VA.ALKEKVAAL< E In SEQ ID NO:64, amino acid residues 1-111 correspond to the amino acid sequence of a VL Domain of a monoclonal antibody capable of binding to PD-1 (hPD-1 mAb 6, SEQ ID NO:58 wherein X1 is S and X2 is Q); residues 112-119 correspond to an intervening spacer peptide (Linker 1: GGGSGGGG (SEQ ID NO:5)); residues 120-237 correspond to a VH Domain of a humanized onal antibody e of binding LAG-3 (hLAG—3 mAb 1); residues 23 8-243 correspond to a cysteine-containing spacer linker peptide (Linker 2: GGCGGG (SEQ ID NO:6)); residues 244-271 correspond to a cysteine-containing Heterodimer—Promoting (K-coil) Domain (KVAACKR—KVAALKT.—KVAAT.KF.—KVAAT.KT.

(SEQ ID NO:21)).

The ability of DART A to stimulate T-cells was examined in a Staphylococcus aureus enterotoxin type B (“SEB”) assay. SEB is a microbial superantigen capable of activating a large proportion of s (5-30%) in SEB-responsive donors. SEB binds to MHC II outside the peptide binding grove and thus is 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 be observed). Within 48 hours of SEB-stimulation PMBCs upregulate PD-1 and LAG—3 with a further enhancement at day 5, post-secondary culture in 96-well plate with SEB-stimulation. Upregulation of the immune check point proteins PD-l and LAG-3 following SEB-stimulation of PBMCs limits cytokine release upon SEB ulation. The ability of DART A to enhance cytokine release through checkpoint inhibition was examined and compared to the activity of the parental D-l and anti—LAG- 3 antibodies alone and in combination.

] Briefly, PBMCs were purified using the Ficoll-Paque Plus (GE Healthcare) density gradient centrifugation method according to manufacturer’s instructions from whole blood ed under informed consent from healthy donors (Biological Specialty Corporation) and T-cells were then purified using the Dynabeads® Untouched Human T-Cells Kit (Life Technologies) ing to manufacturer’s instructions. Purifled PBMCs were cultured in RPMI—media + 10% heat inactivated FBS + 1% Penicillin/Streptomycin in T-25 bulk flasks for 2-3 days alone or with SEB -Aldrich) at 0.5 ng/mL (primary stimulation). At the end of the first round of SEB-stimulation, PBMCs were washed twice with PBS and immediately plated in 96-well tissue culture plates at a tration of 1-5 X 105 cells/well in media alone, media with a control antibody, media with SEB at 0.5 ng/mL (secondary stimulation) and no antibody, or media with SEB and DART A, a control IgG or an anti-PD-l antibody +/- an anti- LAG—3 mAb, and ed for an onal 2-3 days. At the end of the second stimulation, tants were harvested to measure cytokine secretion using human DuoSet ELISA Kits for IFNy, TNFOL, IL-10, and IL-4 (R&D Systems) according to the manufacturer’s ctions.

In these assays DART A (a PD-l X LAG-3 bispeciflc molecule) and the anti-PD- 1 and anti-LAG—3 antibodies were used at a concentration of 0.0061, 0.024, 0.09, 0.39, 1.56, 6.25 or 25 nM. For these studies, where a combination of dies is used each antibody is provided at the indicated concentration and thus the total antibody concentration is twice the concentration used for each dy (i.e., 0.0122, 0.048, 0.18, 0.78, 3.12, 12.5 or 50 nM).

Figure 7 shows the IFNy secretion profiles from SEB-stimulated PBMCs from a representative donor (D: 56041). Similar results were seen for PD—l X LAG-3 bispecific molecules comprising VH/VL domains from alternative PD—l and LAG-3 antibodies, and for PD—l X LAG-3 ific molecules have alternative ures (see, e.g., Figure 3C, and for numerous donors.

The results of these studies demonstrate that PD-l X LAG-3 bispecific molecules dramatically enhanced IFNy production from SEB-stimulated PBMCs upon restimulation.

These results show that ific molecules that target two modulatory pathways were more potent than the combination of separate antibodies targeting the same pathways.

Example 2 PD-l x CTLA-4 Bispecific Molecules Bispecific molecules having specificity for PD-l and CTLA—4 may be generated using methods provided herein and known in the art. The general structure of the polypeptide chains of l PD-l X CTLA-4 bispecific molecules is provided in Table 8. In ular, bispecific bivalent diabody molecules, comprising two polypeptide chains, having one binding site for PD-l and one g site for CTLA-4 may be generated n the polypeptide chains have the general structure of Variation I (also see, e.g., Figure 1). Bispecific bivalent diabody molecules, comprising three polypeptide chains, having one binding site for PD-l, one g site for CTLA-4 and an Fc Region may be generated wherein the polypeptide chains have the general structure of Variation II (also see, e. g., Figure 4A). Bispecific tetravalent diabody molecules, comprising four polypeptide chains, having two identical binding sites for PD-l, two identical binding sites for CTLA-4 and an Fc Region may be ted wherein the polypeptide chains have the l structure of Variations III or IV (also see, e.g., Figures 3A-3C). In addition, bispecific trivalent molecules, comprising four polypeptide chains, having two binding sites for PD-l and one binding site for CTLA-4 (or two binding sites for CTLA-4 and one binding site for PD-l), and an Fc Region may be generated wherein the polypeptide chains have the general structure of Variation V (also see, e.g., Figure 6A). In addition, bispecific bivalent antibody les comprising four polypeptide chains having one binding site for PD-l, one binding site for CTLA-4 and an Fc Region may be ted wherein the polypeptide chains have the general structure of Variation VI (also see, e.g., United States Patent No. 7,695,936 and PCT Patent Publication . . Polypeptide First VL1 — Linkerl — VH2 — Linker2 — I'D S—econdVL2—Linkerl — VH1 — Linker2 — HPD Fir (VL1)— (Linker 1)— (VH2)— (Linker 2)— (HPD)— (Linker 3 — modified CH2-CH3 Domain 11 Second VL2 — Linkerl — VH1 — Linker2 — HPD Third (Linker3) — (modified CH2-CH3 Domain) a—ndFirst (VL1)— (Linker 1)— (VH2)— (Linker 2)— (HPD)— r Third 3 — CH2-CH3 Domain Second and (VL2)— (Linker 1)— (VH1)— (Linker 2)— (HPD) Fourth First and (VL1) — (Linker 1) — (VH2) — (Linker 2) — (CH1) — (Hinge) Third — CH2-CH3 Domain Second and (VL2) — (Linker 1) — (VH1) — (Linker 2) — (CL) Fourth (—VL1)—(Linker1)—(VH2)—(Linker2)—(HPD)— (Linker F'1rst V Second 3VL— modified CH2-CH3 Domain Third Fourth (VL1) — (Linker 1) — (VH2) — r 2) — (HPD) — (Linker First 3 — modified CH2-CH3 Domain VI Second VL2 — l — VH1 — 2 — HPD (VL3)— (Linker 4)— (VH3)— (CH1) — (Hinge) — (modified Th' d1r C-—H2CH3Domain First second Third Fourth HPD = Heterodimer—Promoting Domain For each Variation of the bispecific molecules provided in Table 8: (a) VL1 and VH1 are the variable domains of an anti-PD-l antibody and VL2 and VH2 are the variable s of an anti-CTLA-4 antibody; or (b) VL1 and VH1 are the variable domains of an anti-CTLA-4 antibody and VL2 and VH2 are the variable domains of an anti-PD-l antibody.

For Variations V and VI: VL3 and VH3 are the variable domains of an anti-PD-l antibody or are the variable domains of an anti-CTLA-4 dy.

] Linkers, Heterodimer—Promoting s and constant regions (e.g., CH1, Hinge, CH2-CH3 Domains) useful in the generation of such bispecific molecules are provided above. In particular, as detailed herein, for molecules whose first and third polypeptide chains are not identical the CH2—CH3 s are modified to promote heterodimerization and reduce or prevent homodimerization, for example by modifying the CH2-CH3 Domain one chain to comprise a “hole” and modifying the 3 Domains on the other chain to comprise a ” As detailed above, the Hinge and/or CH2-CH3 Domains may comprise amino acid substitutions, which stabilize the bispecific molecules and/or alter effector on and/or enhance serum half-life.

Example 3 Universal Bispecific Adaptor (“UBA”) Molecules Alternatively, a bispecific molecule (e. g., a bispecific antibody, a ific diabody, trivalent binding molecule, etc.) may be constructed that ses one epitope- binding site that specifically binds to PD—l (or CTLA-4) and a second epitope-binding site that specifically binds a hapten, e.g. fiuorescein isothiocyanate (also known as fluoroisothiocyanate or FITC). Such a bispecific molecule serves as a universal bispecific adaptor (“UBA”) molecule able to co—ligate a binding domain specific for PD—l (or CTLA-4) with a fiuorescein- conjugated binding molecule (e.g., an antibody, scFv, etc.) specific for CTLA-4 (or PD-l). For example, the eactive arm of such a universal bispecific adaptor molecule may be used to bind to a FITC labeled dy that binds CTLA-4 (or PD-l) thereby generating a universal ific r molecule that is adapted to bind PD-l and CTLA-4. Such universal bispecific adaptor molecules are useful for the rapid assessment of bispecific les.

The anti-fiuorescein antibody, 420 (“mAb 420”) may be employed as a source of FITC-specific binding domains (Gruber, M. et al. (1994) “Efiicient Tumor Cell Lysis MediatedBy A Bispecific Single Chain Antibody Expressed In Escherichia coli,” J. Immunol. 152(11): 5368-5374).

Amino Acid Sequence Of The Heavy Chain Variable Domain Of mAb 420 (SEQ ID NO:65) (CDRH residues are underlined): EVKLDETGGG LVQPGRPMKL SCVASGFTFS DYWMNWVRQS BilKGTmZWVAQ IRNKPYNYET YYSDSVKGR:' T-.SRDDSKSS VYLQMNNLRV *ZDMG YYC'l'G YWGQ GTSVTVSS -llO- Amino Acid Sequence Of The Light Chain Variable Domain Of mAb 420 (SEQ ID NO:66) (CDRL residues are underlined): DVVMTQTPFS DQAS :SCRSSQSLV HSNGNTYLRW YLQKPGQSPK VLIYKVSNRF FSGS GSGTDFTLKI SRVEAEDLGV THVP ETFGGGTKLE IK ] Any of the bispecific formats provided herein may be utilized (see, e.g., Tables 1, 2, 3, and 4). Preferred bispecific molecules comprise only one hapten (e.g., fluorescein) binding site and will bind a single hapten-labeled antibody, thereby exhibiting a 1:1 ratio of universal adaptor bispecific molecule to hapten-labeled antibody in the resulting complexes.

Such universal bispeciflc adaptor molecules may be constructed using, for example, the VL and VH Domains of an anti-PD-l antibody and an anti-fluorescein antibody. Preferably, such a universal bispeciflc adaptor molecule is ntly bonded diabody or a trivalent binding molecule sing two, three, four, five, or more polypeptide chains. Representative universal bispeciflc adaptor molecules which may be ucted are provided below.

A. UBA 1 ] One universal ific adaptor molecule that may be generated is a covalently bonded diabody composed of two polypeptide chains comprising one PD—l epitope-binding site and one fluorescein binding site (“UBA 1”).

The first polypeptide chain of UBA 1 comprises, in the inal to inal direction, an N-terminus, the VL Domain of mAb 420 (SEQ ID NO:66), an intervening spacer peptide (Linker 1, GGGSGGGG (SEQ ID NO:5)), the VH Domain ofPD-l mAb 6 (SEQ ID NO:57, wherein X1 is I)), an intervening spacer peptide (Linker 2, GGCGGG (SEQ ID NO:6)), the E—coil Heterodimer—Promoting Domain: EVAALHK—+:VAAL4:K—b1VAALr:K— +1VAAT. 41K (SEQ ID NO:18)), and a C-terminus.

Thus, the amino acid sequence of the first polypeptide chain of UBA 1 is (SEQ ID NO:67): DVVMTQTPFS LPVSLGDQAS :SCQSSQSLV HSNGNTYLRW YLQKPGQSPK VLIYKVSNRF SGVPDRFSGS GSGTDFTLKI SRVEAEDLGV THVP WTFGGGT<LE IKGGGSGGGG QVQLVQSGAE VKKPGASVKV SCKASGYSFT V?QA.PGQGLEWIGV liPSDSLTWL DQKFKDQVTI TVDKSTSTAY MELSSLRSED TAVYYCAREH YGTSPFAYWG QGTLVTVSSG GCGGGEVAAL FKfiVAALfiKfi VAALjKfiVAA EEK -lll- The second polypeptide chain of UBA 1 comprises, in the N—terminal to C- al direction, an N—terminus, a VL Domain of PD-l mAb 6 (SEQ ID NO:58, wherein X1 is S and X2 is Q)), an intervening spacer e r 1, GG (SEQ ID NO:5)), the VH Domain of mAb 4-4—20 (SEQ ID NO:65)), an intervening spacer peptide (Linker 2, GGCGGG (SEQ ID NO:6)), the K-coil Heterodimer—Promoting Domain: KVAALKE— KVAALKE—KVAALKE—KVAALKE (SEQ ID NO:19)) and a C-terminus.

Thus, the amino acid ce of the second polypeptide chain ofUBA 1 is (SEQ ID NO:68): EIVLTQSPAT LSLS?GERAT LSCRASESVD NYGMSFMNWF QQKPGQPPKL LIHAASNQGS GVPSRFSGSG SGTDFTLTTS SLEPEDEAVY EVPY TFGGGTKVEI KGGGGSGGGG EVKLDETGGG LVQPGRPMKL SCVASGFTFS DYWMNWVQQS PEKGLEWVAQ :QNKPYNYET YYSDSVKGRF TISRDDSKSS VYLQMNNWRV EUMGTYYCTG SYYGMDYWGQ SSGG CGGGKVAALK EKVAALKEKV AALKEKVAAL.L K? B. UBA 2 As provided above, incorporating an IgG CH2-CH3 s onto one polypeptide chain of a diabody such as UBA 1 will permit a more complex four-chain bispecific Fc Region-containing diabody to form. Thus a second universal bispecific adaptor molecule that may be generated is a covalently bonded diabody composed of four polypeptide chains comprising two PD-l epitope-binding sites, two fluorescein binding sites, and an Fc Region (“UBA 2”). It will be noted that UBA 2 may bind two fluorescein labeled molecules via the two fluorescein binding sites.

The first and third polypeptide chains of UBA 2 comprises, in the N—terminal to inal direction, an N-terminus, a VL Domain of a mAb 420 (SEQ ID NO:66), an intervening spacer peptide r 1, GGGSGGGG (SEQ ID NO:5)),the VH Domain of PD-l mAb 6 (SEQ ID NO:57, n X1 is I)), an intervening spacer peptide (Linker 2, GGCGGG (SEQ ID NO:7)), the E—coil Heterodimer-Promoting Domain: 41VAAL+:K—PIVAALJ«:K— 41VAAT.4'.K—F.VAA’.F.K (SEQ ID NO:18)), an intervening spacer peptide (Linker 3, GGGDKTHTCPPC? (SEQ ID NO:31)), an IgGl Fc Region comprising substitutions L234A/LZ35A (SEQ ID NO:43), wherein X is K), and a C-terminus.

Thus, the amino acid sequence of the first and third polypeptide chains of UBA 2 is (SEQ ID NO:69): DVVMTQTPFS LPVSLGDQAS :SCRSSQSLV HSNGNTYLRW YLQKPGQSPK VLIYKVSNRF FSGS GSGTDFTLKI SRVEAEDLGV YFCSQSTHVP WTFGGGT<LE IKGGGSGGGG QVQLVQSGAE VKKPGASVKV SCKASGYSFT SYWMNWVRQA PGQGLEWHGV liPSDSjTWL DQKFKDQVTI TVDKSTSTAY MEWSSLRSED AREH YGTSPFAYWG QGTLVTVSSG GCGGGEVAAL FKfiVAALfi<fi VAALJKfiVAA.L LEKGGGD<TH TCPPCPAPEA AGGPSVFLFP PKPKDTLWIS CVVV DVSHEDPZVKJ. FNWYVDGVFV HNAKTKPREE QYWSTYQVVS VLTVLHQDWL NGKEYKC<VS NKALPAP E T SKAKGQPR LPPS REEMTKNQVS LTCLVKGFYP SDIAVEWESN GQPENNYKTT PPVLDSDGSF FLYSKLTVDK SRWQQGNVFS CSVMHEALHN HYTQKSLSLS The second and fourth polypeptide chains of UBA 2 are identical to the second polypeptide chain of UBA 1. Thus, the second and fourth polypeptide chains of UBA 2 each have the amino acid sequence of SEQ ID NO:68.

C. UBA 3 A third universal bispecific adaptor molecule that may be generated is a covalently bonded diabody composed of three polypeptide chains sing one PD-l e-binding site, one fluorescein binding site, and an Fc Region (“UBA 3”).

The first ptide chain of UBA 3 ses, in the inal to C-terminal direction, an N—terminus, the VL Domain of mAb 420 (SEQ ID NO:66), an ening spacer peptide r 1, GGGSGGGG (SEQ ID NO:5)), the VH Domain ofPD-l mAb 6 (SEQ ID NO:57, wherein X1 is I)), an intervening spacer peptide (Linker 2, GGCGGG (SEQ ID NO:6)), the E-coil Heterodimer—Promoting Domain: EVAALHK—+:VAAL*:K—b1VAALr:K— +1VAAT~1K (SEQ ID NO:18)), an intervening spacer peptide (Linker 3, GGGDKTHTCPPCP (SEQ ID NO:31)), a “knob-bearing” IgGl Fc Region comprising substitutions L234A/L235A (SEQ ID NO:44, wherein X is K)), and a C-terminus.

Thus, the amino acid sequence of the first polypeptide chain of UBA 3 is (SEQ ID NO:70): DVVMTQTPFS LPVSLGDQAS :SCRSSQSLV HSNGNTYLRW YLQKPGQSPK VLIYKVSNRF SGVPDRFSGS GSGTDFTLKI SRVEAEDLGV YFCSQSTHVP WTEGGGTKLfi KGGGSGGGG QVQLVQSGAE VKKPGASVKV SCKASGYSFT SYWMNWVRQA WLGV iHPSDSETWL DQKEKDRVTI TVDKSTSTAY -ll3- MELSSLRSED AREH YGTSPFAYWG QGTLVTVSSG GCGGGEVAAL FKfiVAALH<fi VAALEKEVAA LEKGGGDKTH TCBPCBAPS§_§GGPSVFLFP PK?KDTLMIS RTPEVTCVVV DVSHED?EVK FNWYVDGVEV HNAKTKPREE QYWSTYRVVS VLTVLHQDWL NGKEYKCKVS NKALBAP 5K T PR EPQVYTLPPS REEMTKNQVS TWCTVKGFYP SDIAVEWESN GQPENWYKTT PPVLDSDGSF FLYSKLTVDK SRWQQGNVFS CSVMHEALHN HYTQKSLSLS The second polypeptide chain of UBA 3 may be identical to the second polypeptide chain ofUBA 1. Thus, the second polypeptide chain ofUBA 3 has the amino acid sequence of SEQ ID NO:68.

The third polypeptide chains ofUBA 3 comprises, in the N—terminal to inal direction, an N—terminus, a spacer peptide r 3, DKTHTCPPCP (SEQ ID NO:26)), a bearing” IgGl Fc Region comprising substitutions L235A (SEQ ID NO:45, wherein X is K)), and a C-terminus.

Thus, the amino acid sequence of the third polypeptide chain of UBA 3 is (SEQ ID NO:71): DKTHTCBPCP APLAAGGBSV PKDT RM SRTBfiVT CVVVDVSHED PEVKFNWYVD GVEVHNAKTK PREEQYNSTY RVVSVLTVLH QDWLWGKEYK CKVSNKALBA.P fiKTTSKAK GQPREPQVYT LEBSRZfiWTK CAV<.L GEYBSDLAV: WESNGQPENN YKTTPPVLDS DGSFFLVSKL TVDKSRWQQG NVFSCSVMHE ALTNRYTQKS LSLSPGK D. UBA 4 A fourth sal bispecific adaptor molecule that may be generated is a covalently bonded trivalent binding molecule composed offour polypeptide chains comprising two PD-1 epitope-binding sites, one fluorescein binding site, and an Fc Region (“UBA 4”).

The first polypeptide chain of UBA 4 is identical to the first polypeptide chain of UBA 3. Thus, the first polypeptide chains of UBA 4 has the amino acid sequence of SEQ ID NO:70.

The second polypeptide chain of UBA 4 is identical to the second polypeptide chain of UBA 1. Thus, the second polypeptide chain of UBA 4 has the amino acid sequence of SEQ ID NO:68. —114— The third polypeptide chain ofUBA 4 comprises, in the inal to C-terminal direction, the VH Domain of PD-l mAb 6 (SEQ ID NO:57, wherein X1 is 1)), an IgGl CH1 Domain (SEQ ID NO:40), an IgGl Hinge Region (SEQ ID NO:33), a bearing” IgGl Fc Region comprising substitutions L234A/L235A (SEQ ID NO:45, wherein X is K)), and a C-terminus.

Thus, the amino acid sequence of the third ptide chain of UBA 4 is (SEQ ID NO:72): QVQLVQSGAE VKKPGASVKV SC<ASGYSFT SYWMNWVRQA PGQGLEWIGV ETWL DQKFKDRVTI TVDKSTSTAY MELSSLRSED TAVYYCAREH YGTSPFAYWG QGTLVTVSSA ST {GPSVFPL APSSKSTSGG TAALGCLVKD YFPEPVTVSW NSGALTSGVH TFPAVLQSSG LYSLSSVVTV PSSSLGTQTY ICWVNHKPSN T<VDKRVEPK SC PP CPAPEAAGGP SVFLFPPKP< DTLMISRTPE VTCVVVDVSH EDBEVKENWY VDGVEVHNAK TKPREEQYNS TYRVVSVLTV LHQDWLNGKE YKCKVSN<AL PAPIEKTISK AKGQPREPQV YTLPBSRfifiM TKNQVSLSCA.VKGEYPSDLA GQPE PPVL DSDGSFFTVS KWTVDKSRWQ QGNVFSCSVM HEALHNRYTQ KSLSLS?GK The fourth polypeptide chain of UBA 4 comprises, in the N-terminal to C- terminal direction, the VL Domain of PD-l mAb 6 (SEQ ID NO:58, n X1 is S and X2 is Q)), a CL Domain (e.g., an IgG Kappa Domain (SEQ ID N0:38), and a C-terminus.

] Thus, the amino acid sequence of the fourth polypeptide chain of UBA 4 is (SEQ ID NO:73): SPAT ERAT LSCRASESVD NYGMSFMNWF QQKPGQ?PKL LIHAASNQGS GVPSRFSGSG SGTDFTLT S SLfiPfiDhAVY ECQQSKEVPY TFGGGTKVE“ PSVE EPBS)HQLK SGTASVVCLL WNFYPREAKV QWKVDNALQS GNSQESVTEQ DSKDSTYSLS STLTWSKADY EKHKVYACEV THQGLSSPVT KSFNRGEC E. UBA 5 A fifth universal bispeciflc adaptor molecule that may be generated is a covalently bonded trivalent binding molecule composed of three polypeptide chains comprising two PD- 1 epitope-binding sites, one fluorescein binding site, and an Fc Region (“UBA 4”) (see, e.g., Figure 6C-6D).

The first polypeptide chain of UBA 5 is identical to the first polypeptide chain of UBA 3. Thus, the first polypeptide chains of UBA 5 has the amino acid sequence of SEQ ID NO:70.

The second polypeptide chain of UBA 5 is identical to the second ptide chain of UBA 1. Thus, the second polypeptide chain ofUBA 5 has the amino acid sequence of SEQ ID NO:68.

The third polypeptide chain ofUBA 5 comprises, in the N—terminal to C-terminal direction, the VL Domain of PD-l mAb 6 (SEQ ID NO:58, wherein X1 is S and X2 is Q)), an ening spacer peptide (Linker 4, GGGGSGGGGSGGGGS (SEQ ID NO:37)), the VH Domain of PD-l mAb 6 (SEQ ID NO:57, n X1 is 1)), an IgGl CH1 Domain (SEQ ID NO:40), an IgGl Hinge Region (SEQ ID NO:33), a “hole-bearing” IgGl Fc Region sing substitutions L234A/L235A (SEQ ID NO:45, wherein X is K)), and a C-terminus.

] Thus, the amino acid sequence of the third polypeptide chain of UBA 5 is (SEQ ID NO:74): EIVLTQSPAT LSLSPGERAT LSCRASESVD NYGMSFMNWF PPKL LIHAASNQGS GVPSRFSGSG SGTDFTLTTS SLEPEDEAVY ECQQSKEVPY TFGGGTKVZ.L KGGGGSGGGG SGGGGSQVQL VQSGAEVKKP GASVKVSCKA SGYSFTSYWM NWVRQAPGQG LfiWlGV HBS DSLTWLDQKF KDRVTITVDK STSTAYMELS SLRSEDTAVY YGTS PFAYWGQGTL VTVSSASTKG PSVFPLAPSS KSTSGGTAAL GCLVKDYFPE PVTVSWWSGA LTSGV TFPA VLQSSGLYSL SSVVTVPSSS LGTQTYICNV NHKPSNTKVD KRVEPKSCDK THTCPPCPAP EAAGGPSVFL FPPKPKDTTM SQTPfiVTCV VVDVS fiDPfi VKFWWYVDGV EVHNAKTKPR EEQYNSTYRV WLNGKEYKCK VSN<ALPAPI EKTISKAKGQ PQEPQVYTLP VSLSCAVKGF YPSDLAVfiWfi SNGQPENNYK TTPPVLDSDG SFFLVSKLTV DKSRWQQGNV HEAL HNRYTQKSLS LS?GK Using tional s, TLA-4 antibodies may be labeled with fluorescein. When such labeled molecules are incubated in the presence of a universal bispeciflc adaptor molecule provided above having an epitope-binding site that binds to PD-l and an epitope-binding site that binds to fluorescein, they form a PD-l X CTLA—4 bispecific molecule, which may be assayed as described below.

It will be appreciated in view of the teachings provided herein that different VH Domains, VL Domains, linkers, heterodimer promoting domains, and/or IgG Constant Domains could be utilized to generate alternative universal bispecific adaptor molecules. For e, the VH and VL Domains of an anti-CTLA-4 antibody and/or a different anti-PD-l antibody could be used in place of the VH and VL Domains of the employed D-l antibody to generate alternative or lent universal bispecific adaptor molecules. -ll6- Alternatively, the VH and VL Domains of an anti-CTLA-4 antibody may be used in place of the VH and VL Domains of the anti-fluorescein antibody to generate PD-l X CTLA—4 bispecific les having the general structure of Variations I, II, III, V and VI provided above, Such PD-l X CTLA—4 bispecific molecules may be used directly in the assays described below.

Example 4 Assays The PD-l X CTLA-4 bispecif1c molecules of the t invention may be characterized in any of a variety of ways. In particular, PD-l X CTLA-4 bispecific molecules of the invention may be d for their ability to immunospecifically bind to the PD-l and CTLA-4 molecules (e.g., as present on a cell surface, etc), and/or the binding kinetics of the interactions with antigen may be determined. Where the bispecific molecules comprise an Fc region (or portion thereof), their ability to exhibit R interactions, e.g., specific binding of an Fc region (or portion thereof) to an FcyR, ion of effector function, signal transduction, etc, may be assayed. The immunomodulatory activity and/or in vivo anti-tumor efficacy of the PD-l X CTLA-4 bispecific molecules of the invention may be assayed using in vitro and in vivo assays known in the art.

A. Preparation of Immune Cells and Cell Expressing PD-l and/or CTLA-4 1. Isolation of PBMCs and Immune Cell ulations from Human Whole Blood PBMCs from healthy human donors are isolated from whole blood, for example, using Ficoll gradient centrifugation. Briefly, whole blood is diluted 1:1 with sterile phosphate buffered saline (PBS). The diluted blood (35 mL) is d onto 15 mL ofFicoll-PaqueTM Plus in a 50 mL tube and the tubes are centrifuged at 400 x g (1320 rpm) for 30 minutes with the brake off. The buffy—coat layer between the two phases is collected into 50 mL tubes and centrifuged at 600 x g (1620 rpm) for 5 minutes. The supernatant is discarded and the cell pellet is washed 3 times with PBS (e.g., by centrifuging the tubes at 600 x g (1620 rpm) for minutes). Viable cell count is determined using Trypan Blue dye. The PBMCs are ended in complete culture medium (e.g., RPMI 1640, 10% PBS, 1% pen/strep) and incubated at 37°C with 5% C02 overnight or are r processed to isolate a desired immune cell ulation such as T cells, (e.g., T regs, CD8, CD4), NK cells, dendritic cells and monocytes as described below. -ll7- ] Particular immune cell ulations are readily isolated from PBMCs using a commercial preparation kit (e.g., the UntouchedTM human T cell isolation kits for isolation of s, CD4 T-cells, CD8 s, tes, Dendritic Cells (Life Technologies/ThermoFisher Scientific); the DYNABEADS® Regulatory CD4+/CD35+ T Cell Kit for isolation of T regulatory cells (CD4+/CD25+) (ThermoFisher), etc), according to the manufacturer’s instructions. After isolation, the immune cell subpopulation (e.g., T cells) are resuspended in the appropriate complete culture medium (e.g., RPMI 1640, 10% PBS, 1% penicillin/ streptomycin, which may be supplemented with cytokines (e.g., IL-2, GM-CF, IL- 4, TNF-a, etc.) and incubated at 37°C with 5% C02 overnight. As provided herein such purified subpopulations are useful to evaluate cell e expression of PD-l and/or CTLA-4 and for evaluation of the immune atory activity of the PD-1 X CTLA-4 bispecific molecules of the invention. 2. Isolation Of PBMCs From Cynomolgus Monkey Or Rhesus Monkey Whole Blood PMBCs from Cynomolgus monkey or Rhesus monkey are ed from whole blood, for example using Ficoll gradient centrifugation. Briefly, whole blood is diluted 1:3 with sterile PBS. Diluted blood (35 mL) is layered onto 15 mL of 90% -PaqueTM Plus (90 mL Ficoll + 10 mL PB S) in a 50 mL polypropylene centrifuge tube and centrifuged at 931 X g (2000 rpm) for 30 minutes at room temperature with the brake off. The buffy—coat layer between the two phases is collected and transferred to a clean 50 mL tube and washed with 45 mL PBS by centrifuging the tubes at 600 x g (1620 rpm) for 5 minutes. The supernatant is discarded and the pellet is rinsed 3X with PBS. Cynomolgus or Rhesus monkey PBMCs are then resuspended in 30 mL of te culture medium and viable cell count is determined by Trypan Blue dye exclusion.

Particular immune cell ulations are readily isolated from non-human primate PBMCs using a commercial preparation kit (e.g., Pan T-cell, CD4+ T-Cell, and CD4+/CD25+ Treg isolation kits (Miltenyl Biotech)), according to the manufacturer’s instructions. Alternatively, flow cytometric sorting using non-human primate specific or cross- reactive mAbs can be used for sorting. 3. Generation Of Human Immature Or Mature d-Derived Dendritic Cells (mDC) Cells From Isolated Human Monocytes Human monocytes are isolated from donor derived purified PBMCs using a commercial preparation kit (e.g., the UntouchedTM human monocyte kit (Life Technologies/ThermoFisher ific) according to manufacturer’s instructions. Isolated human monocytes are induced to differentiate into human immature mDCs by culturing monocytes (e.g., in alpha m Essential Media with nucleosides (OLMEM) media + 2% human AB-negative serum + 1% penicillin/streptomycin) for 5-7 days in the presence of recombinant human granulocyte macrophage-colony stimulating factor (e.g., hGM-CSF, Peprotech, 100 ng/ml) and recombinant human interleukin-4 (hIL-4, Peprotech, 40 ng/ml).

Immature mDCs are harvested and washed with PBS by centrifuging the tubes at 600 x g (1620 rpm) for 5 s for use as stimulator cells in allogeneic mixed lymphocyte reaction (allo- MLR) assays, such as those detailed below.

In certain allo—MLR experiments immature mDCs are induced to differentiate by adding TNFOL or a cocktail of additional cytokines (TFNv, E-IB) and mitogens (LPS) for two additional days of culture (see, e.g., Han, T. (2009) “Evaluation of 3 al Dendritic Cell Maturation Protocols Containing LPS and IFN-gamma, ” J Immunother 321399). The purity, maturation and activation of mDCs may be ted by flow cytometry using one or more of the ing antibodies: D14, D80, anti-CD83, anti-CD86, anti-HLA-DR; and the appropriate isotype ls. The flow cytometric data from such evaluations may be acquired on a FACSCalibur/Fortessa n Dickinson/BD Biosciences) and analyzed using FlowJo software (TreeStar). 4. sion of PD-l and CTLA-4 Cells expressing PD-l and/or CTLA-4 may be generated using methods known in the art. For example, cells (e.g., NSO, Jurkat, CHO, etc.) may be engineered to express PD- 1 and/or CTLA-4 using retroviral vectors containing the appropriate gene (e.g., human PD-l gene). Alternatively, immune cells may be stimulated to induce or increase the expression of PD-l and/or . Briefly, purified immune cells (e.g., PBMCs, T-cells, dendritic cells, etc.) isolated as described above are cultured for 2-6 days in the presence or absence of a mitogen and the expression of PD-l and/or CTLA-4 is examined on the untreated (Naive) and stimulated cells, for e using flow cytometry. Commercial anti-PD-l and anti-CTLA-4 antibodies can be used for preliminary evaluation of the expression patterns on Naive cells and -ll9- in response to mitogen stimulation. Additionally, or optionally the PD-l X CTLA-4 bispeciflc les of the invention may be used.

Mitogens which may be utilized for such studies are well known in the art and include, but are not limited to: CD3/CD28 beads, lipopolysaccharides (LPS), Staphylococcus aureus enterotoxin types A-E (e.g., SEB), phorbol myristate acetate (PMA), phytohemagglutinin (PHA), concanavalin A (conA), pokeweed n (PWM), etc.

Mitogen(s) identified as inducing/enhancing the expression of PD-l and/or CTLA-4 may be used in functional assays to evaluate the stimulatory activity of the PD-l X CTLA—4 bispecif1c molecules of the present invention. See for example the “SEB”, and “MLR” assays described herein.

B. Binding Assays Immunoassays that can be used to analyze speciflc binding to PD-l or CTLA-4 molecules, binding cross-reactivity, or Fc-chR interactions include, but are not limited to, competitive and non-competitive assay systems using techniques such as western blots, radioimmunoassays, ELISA (enzyme linked immunosorbent , “sandwich” immunoassays, precipitation assays, precipitin reactions, gel diffusion precipitin reactions, immunochromatographic assays, immunodiffusion assays, agglutination assays, complement—fixation assays, radiometiic assays, fluorescent immunoassays, and protein A immunoassays, etc. (see, e.g., Ausubel et al., 2008, t ols in Molecular Biology). Binding affinity for a target n is typically measured or determined by standard antibody-antigen assays, such as Biacore itive assays, saturation assays, or immunoassays such as ELISA or RIA. scence activated cell sorting (FACS), using any of the techniques known to those skilled in the art, is used for immunological or functional based assays to characterize the PD-l X CTLA-4 ific molecules of the invention.

For example, PBMCs may be prepared as described above. Where desired immune cell subsets (e.g., T regulatory, T helper, APCs, etc.) may be isolated from the purified PBMC. The isolated cells are then examined for PD-l and CTLA—4 expression on various cell subsets (e.g, T tory, T helper, APCs, etc.) by co-staining and FACS analysis as described below. 1. Cell Surface Binding (Saturation Assay) ] The ability of PD-1 X CTLA-4 bispecific les to bind to PD-l and/or CTLA-4 expressed on the cell surface may be measured in saturation/dilution based assays using a cell that expresses PD-l and/or CTLA-4 (target cells). Such cells may be immune cells stimulated to expressed PD—l and/or CTLA-4, or a cell line (e.g., NSO cells) engineered to stably over-express PD-l and/or CTLA-4 molecules. Briefly, cultured s cells (e.g., NSO cell ered to s PD1+) are harvested and resuspended (e.g., about 5x106 cells/ml) in blocking buffer (e.g., FACS buffer + 10% human AB Serum). Starting at equal molar concentrations (e.g., 20nM in total of 200 pl) a PD—l X CTLA-4 bispecific molecule, an anti- PD-l antibody, an anti-CTLA-4 or a combination of anti-PD-l and anti-CTLA-4 antibodies are prepared for dilution in a separate microtiter plate and then serially diluted (e.g., 1:4, 1:5, 1:10, etc.) 5-12 times to generate a 5-12 point curve. The highest starting concentration in all experiments is determined empirically. The same volume (e.g., 50 pl) of each dilution is added to a new microtiter plate and target cells are added to each well (e.g., 0.25x106 cells/well) and incubated (e.g., at 4-25°C for 30-120 minutes). The cells are washed 1-3 times (e.g., the microtiter plate is spun at 600 x g (1620 rpm) for 5 minutes and then washed with blocking buffer and spun again) and resuspended in blocking buffer. For ary staining, the appropriate secondary regent is selected, for example a goat anti-Human Fc-APC may be used to detect human primary antibodies, while a goat Anti-Mouse IgG Fc Alexa Fluor 647 is used to detect mouse primary antibodies. The selected secondary reagent is d in blocking buffer and based on the tration of the individual secondary, a stock solution is made and the same volume/well of the secondary mixture is aliquoted to individual wells and incubated (e.g, at 4-25°C 30-120 minutes). The cells are washed as described above and resuspended in blocking buffer. The stained cells are ed by flow try. The flow cytometric data may be acquired on a FACSCalibur/Fortessa (Becton Dickinson/Fortessa), ed as mean fluorescent intensity using FlowJo software (TreeStar), and plotted and fitted using the log(agonist) vs. response —Variable slope (four parameter) function in Prism6 software (Graphpad). 2. Receptor/Ligand Binding and Signaling Assays Assays that can be used to analyze the ability of the PD-l X CTLA—4 bispecific molecules of the invention to modulate (e.g., block, inhibit, stimulate, etc.) ligand g and signaling are provided in more detail below. a. PD-l Receptor/Ligand Binding The y of PD-l X CTLA-4 bispecific molecules to inhibit PD-l from binding PD-Ll and/or PD-L2 may be evaluated using cells that express PD-l (target cells). Such cells may be immune cells ated to s PD-l, or a cell line engineered to express PD-l molecule, for example NSO—cells retrovirally transduced with the human PD-l gene. Briefly, PD-l expressing cells (e.g., NSO/PDCDl (NSO-PD1+)) are harvested and resuspended (e.g., about 1.5x106 cells/ml) in blocking buffer (e.g., FACS buffer + 10% Human Ab Serum) and plated in a microtiter plate (e.g., 0.25x106 cells/well). Starting at equal molar trations (e.g, 20nM in total of 200 pl) of a PD-1 X CTLA-4 bispecific molecule, an anti-PD—l antibody, an anti-CTLA-4, or a combination of anti-PD-l and TLA-4 dies are ed for on in a separate microtiter plate and serially diluted (e.g., 1:4, 1:5, 1:10, etc.) 5-12 times to generate a 5-12 point curve. The highest starting concentration in all experiments is determined empirically. The same volume (e.g., 50 ul) of each dilution is added to each well of the microtiter plate containing the target cells. To evaluate the inhibition of PD-Ll binding a soluble PD-Ll fusion n (e.g., hPD-Ll (B7Hl) TEV-hIgGl—Fc-biotin (Ancell))is added to each well with the exception of unstained negative control wells and incubated (e.g., at 4- °C for 30-120 minutes). To evaluate the inhibition ofPD—L2 binding a soluble PD-L2 fusion protein (e.g, CD273 ) mngG/biotin (Ancell)) is added to each well with the exception of unstained negative control wells and incubated (e.g., at 4-25°C for 30-120 minutes). The cells are washed 1-3 times (e.g., the microtiter plate is spun at 600 x g (1620 rpm) for 5 minutes and then washed with blocking buffer and spun again). The cells are ended in blocking buffer. With the exception of unstained negative control wells, the appropriate secondary reagent for detection of the PD-Ll or PD-LZ fusion protein (e.g., streptavidin-PE labeled secondary (eBiosciences)) is added and incubated (e. g., at 4-25°C for 15—120 s). The cells are washed as described above and resuspended in blocking buffer. The stained cells may be analyzed by flow cytometry. The flow tric data may be acquired on a FACSCalibur/Fortessa (Becton Dickinson/Fortessa), and analyzed for the loss mean fluorescent intensity of labeled sPD-Ll or sPD-L2 in the presence of a PD-l X CTLA-4 bispecific molecule, an anti—PD-l antibody, an anti-CTLA-4, or a combination of anti-PD-l and anti-CTLA-4 antibodies using FlowJo software (TreeStar), and plotted and fitted using the log(agonist) vs. response —variable slope (four parameter) function in Prism6 re (Graphpad). b. CTLA-4 Receptor/Ligand Binding The ability of PD-1 x CTLA-4 bispecific molecules to inhibit CTLA-4 from binding CD80 and/or CD86 may be evaluated using cells that express CTLA-4 t cells).

Such cells may be immune cells stimulated to express CTLA-4, or a cell line engineered to express CTLA-4, for example lls retrovirally transduced with the human CTLA-4 gene.

Briefly, CTLA-4 expressing cells are harvested and resuspended in blocking buffer (e.g., FACS buffer + 10% Human Ab Serum) and plated in a microtiter plate (e.g., 0.25x106-10x106 cells/well). ng at equal molar trations (e.g., 20nM in total of 200 pl) of a PD-l X CTLA-4 bispecific molecule, an anti-PD-l antibody, an anti-CTLA-4, or a ation of anti- PD-l and anti-CTLA-4 antibodies are prepared for dilution in a separate microtiter plate and serially diluted (e.g., 1:4, 1:5, 1:10, etc.) 5-12 times to generate a 5-12 point curve. The highest starting concentration in all experiments is determined empirically. The same volume (e.g., 50 ul) of each dilution is added to each well of the microtiter plate containing the target cells. To evaluate the inhibition of CD80 binding a soluble CD80 fusion protein (e.g., hCD80-mng- biotin (ADIPOGEN®)) is added to each well with the exception of unstained negative control wells and incubated (e. g., at 4-25°C for 30-120 minutes). To evaluate the inhibition of CD86 binding a e CD86 fusion protein (e.g., hCD86-mng—biotin (ADIPOGEN®)) is added to each well with the exception of unstained negative control wells and incubated (e.g., at 4-25°C for 30—120 minutes). The cells are washed 1-3 times (e.g., the microtiter plate is spun at 600 x g (1620 rpm) for 5 minutes and then washed with blocking buffer and spun again). The cells are resuspended in blocking buffer. With the exception of unstained ve control wells, the appropriate secondary reagent for detection of the CD80 or CD86 fusion protein (e.g., streptavidin-PE d secondary (eBiosciences)) is added and ted (e.g., at 4-25°C for -120 s). The cells are washed as described above and resuspended in ng buffer.

The stained cells may be analyzed by flow cytometry. The flow cytometric data may be acquired on a libur/Fortessa (Becton Dickinson/Fortessa), and analyzed for the loss mean cent intensity of labeled CD86 or CD80 in the presence of a PD-l X CTLA-4 bispecific molecule, an anti-PD-l antibody, an anti-CTLA-4, or a combination of anti-PD-l and anti-CTLA-4 antibodies using FlowJo software (TreeStar), and plotted and fitted using the log(agonist) vs. se —variable slope (four parameter) function in Prism6 software (Graphpad).

C. Reporter Assays ] The functional activity of PD—l X CTLA-4 bispecific molecules in blocking the interaction ofPD-l with PD—Ll may be assessed using a commercial reporter system developed by a according to the manufacturer’s direction. Briefly, two cell lines engineered to function as either a stimulator line or reporter cell line are used. The ator line was engineered from a CHO-parental line to express the PD-L1 molecule and a T cell activator, which is a membrane bound anti-CD3 agonist mAb [CHO/PDLl cells]. The reporter cell line was engineered from a CD3-positive Jurkat parental line to express a luciferase reporter construct under the transcription control of nuclear factor of activated T-cells (NFAT) [NFAT- luc2/PD-1 Jurkat cells]. When cultured together, the anti-CD3 agonist expressed on the CH0- PDLl cell line drives rase expression by the NFAT signal transduction pathway mediated by the engagement of the TCR/CD3 signaling complex present on the Jurkat-NFAT-luc/PD-l cell line. In the absence of anti-PD-l or anti-PD—Ll dies, rase is sed at a level relative to TCR/CD3 signaling but down-modulated or inhibited by the presence of the PD-l/PD-Ll inhibitory axis, which functions as a brake. In the ce of molecules which inhibit PD—l/PD-Ll signaling (e.g., anti—PD-l or anti-PD-Ll antibodies), this inhibitory axis or “brake” is released, permitting ed luciferase expression that can be measured.

Accordingly, the PD-1 tory activity of PD-l X CTLA-4 bispecific molecules may be evaluated by culturing Ll with ch/PDl Jurkat (3H-D5). Briefly, CHO- PDL1 are plated into a microtiter plate (e.g., at 4.0x104 cells/well) and cultured overnight (e.g., in RPMI media ning 10% FBS + lOOug/mL Hygromycin B + 500ug/mL G418). The next day, assay buffer (6.9., RPMI + 2% FBS is prepared along with a 5-12 point serial dilution of a PD-l X CTLA-4 bispecific molecule, or an D-l antibody in assay buffer with highest dilution point at equal molar equivalence (e. g., 100-200 nM) and 5-12 serial dilutions (e.g., 1:4, 1:5, 1:10, etc.) are prepared. In the following order, a portion of cell the culture media is removed from the microtiter plate containing adherent CHO/PDLl cells and aliquots of each dilution are added to the CHO/PDLl cells. Cultured NFAT-lch/PD-l Jurkat cells are harvested and resuspended in assay buffer and added (e.g., 5.0x104 cells/well in 40u1/well) to the Ll cells. The co-culture is incubated (e.g., for 6 hours at 37°C). At the end of the incubation, Bio-G10 substrate (Promega) is reconstituted and added to the ambient temperature equilibrated iter plate. Following incubation (e. g., 5—10 minutes) the optical density of each well is read on a VICTORTM X4 Multilabel Plate Reader (Perkin Elmer #2030-0040) at 450nm with luminescence relative light unit (RLU) as the readout. The data may then be plotted —124— and fitted using the log(agonist) vs. response —variable slope (four parameter) function in Pri sm6 software (Graphpad).

Similar reporter assays are available for CTLA-4 ing (e.g., CTLA-4 Blockade Bioassay Kit (Promega)) and/or may be readily generated to analyze the functional activity of PD-l X CTLA-4 bispecific molecules in ng the ction CTLA-4 with its respective ligand(s).

D. Immunomodulatory Assays Assays that can be used to analyze the immunomodulatory activity of the PD-l X CTLA-4 ifrc molecules of the invention include n stimulation assays such as the “SEB” assay detailed above, and Mixed Lymphocyte Reaction (MLR) assays such as those provided in more detail below. The ability of the PD-l X CTLA-4 bispecifrc molecules of the invention to modulate both the PD-l and the CTLA-4 inhibition pathways is expected to provide enhanced stimulation in assays as compared to anti-PDl and anti— CTLA-4 antibodies alone or the combination of such antibodies.

PBMCs or T cells are ed from the blood of the same (autologous) or unrelated (allogeneic) patient(s) healthy donor(s) blood by centrifugation over a Ficoll- PaqueTM gradient as described above and resuspended in te culture medium. For allo- MLR assays that employ mDCs, monocytes are purified and d as be above. For one-way (unidirectional) allo-MLR assays der cells (e.g., PBMCs) are co-cultured with stimulating cells in a microtiter plate. Depending on the context, stimulating cells are DCs, autologous PBMCs (for auto-MLR, i.e., negative control), or allogeneic PBMCs (for allo- MLR, i.e., positive control). The ratio of responderzstimulating cells is typically 1:1 or 2: 1, but may be varied. The co-cultures are performed in the presence of equal molar amounts of serial (e.g., 1:4 1:5, 1:10, etc.) dilutions of a PD-l x CTLA-4 bispecifrc molecule, an anti-PD-l antibody, an anti-CTLA-4, a combination of anti-PD-l and anti-CTLA-4 antibodies, or the corresponding isotype mAbs. Serial antibody dilutions may be prepared as described above.

In addition, single cell populations ls ated with or without anti-CD3 +/- anti-CD28 mAbs may be used as controls in such experiments. ating cells (stimulators) are pre- irradiated (e.g., at 45 grays[Gy] (4500 rads) using a Gammacell® 3000 Elan Blood/Cell Irradiator (Theratronics)) to prevent eration ofthe stimulator cells and allow measurement of only the proliferation of the responding cell (responders). After 5 -7 days (the time will be adjusted to ensure expression of PD-1 and CTLA-4 during the assay), [3H]-thymidine (e.g., 1 uCi/well (Perkin Elmer)) is added for r 18-48 hours. The radioactivity incorporated into DNA is measured in (e.g. in a TOPCount NXT B-scintillation counter n Elmer)). Results are expressed as either mean counts per minute (cpm) or expressed as stimulation index (SI) allowing the comparison of results from different donors. SI is calculated as follows: mean counts per minute (cpm) from stimulated cells divided by mean cpm from non-stimulated cells.

MLR responses are considered positive when SI was 23 for PBMC-induced ation and $12 6 for DC-induced stimulation. Alternatively, proliferation may be measured, using a CEFSE-based proliferation assay (Boks, M.A., et al. (2010) “An optimized CFSE based T-cell suppression assay to evaluate the suppressive ty ofregulatory T—cells induced by human tolerogenic dendritic ” Scand J Immunol 72: 158—168).

Additional MLR assays which may be used to evaluate the immune stimulatory activity of the PD-l x CTLA-4 bispecific molecules of the invention are known in the art. See, for example,Davies, IK. et al. (2011) “Induction of alloantigen—specific anergy in human peripheral blood mononuclear cells by tigen ation with co-stimulatory signal blockade,” Journal of Visualized Experiments: JoVE, (49), 2673; Kruisbeek, A.M., et al. (2004) “Proliferative Assays for T cell Function,” CURRENT PROTOCOLS IN IMMUNOLOGY, 60:III:3.12.1—3.12.20; Wallgren, AC. et al. (2006) “The Direct Pathway OfHuman T-Cell Allorecognition Is Not Tolerized By Stimulation With Allogeneic Peripheral Blood Mononuclear Cells Irradiates With High-Dose Ultraviolet,” Ba. Scand J of Immunol 63:90- 96, Levitsky, J. et al. (2009) “The Human 'Treg MLR' Immune Monitoring for Foxp3+ T regulatory cell generation, Transplantation 88: 1303-1 1.

E. In Vivo Anti-Tumor Assays The anti-tumor activity of the PD-l X CTLA-4 bispeciflc les of the ion may be evaluated in various animal models known in the art. Treatment with the PD-1 X CTLA-4 bispecific molecules of the invention is ed to inhibit tumor establishment and/or tumor growth to a greater extent than treatment with anti-PD1 and anti- CTLA-4 antibodies alone or the combination of such antibodies.

Murine xenograph tumor models are particularly . Briefly, mice are implanted with a cancer cell line, or tumor cells of interest and are d with (i) a PD-1 X CTLA-4 bispecific molecule (ii) an anti-PD-l antibody (iii) an anti-CTLA-4 antibody (iv) a combination of anti-PD-1 and anti-CTLA-4 antibody, and (vi) no-treatment control which may be vehicle alone and/or an irrelevant antibody. Treatment may begin prior to implantation (e.g., 1 day before (i.e., day -1)); on the same day as implantation (i.e., day 0), or after ishment of a tumor (e. g., day 7). The animals may receive a single treatment or may receive multiple treatments (e.g., weekly post tation). The animals are monitored over time to determine the in vivo effect of these molecules on tumor establishment and/or growth.

Growth of tumors may be monitored my measuring the tumors and determining the tumor volume (height X width X length). Treated animals which show complete tumor regression can be used to examine tumor-specific immunity by lenge using the same or tumor cells and irrelevant tumor cells as a control. In addition, these models may be modified to include combination treatment with standard of care treatments such as chemotherapy, radiation, etc.

Numerous transplantable cancer cell lines which may be utilized in such xenograph models are known in the art and include, but are not limited to: MDSTS, SW480 and SW620 colorectal cancer cells; AGS gastric cancer cells, UACC-62, A2058, and LOX IMVI melanoma cells, 22rv prostate cancer cells, AsPC-l and BXPc-3 pancreatic cancer cells, Caki-l, A498 and 786-0 renal cancer cells; HT—1197 Bladder cancer cells; 4T1, MDA—l\/fl3- 231, mammary cancer cells, A549, WX322 Lung cancer cells, HT1080 Fibrosarcoma cells, HBL-2 human mantle cell lymphoma cells, Raji Burkitt’s lymphoma cells. Particularly preferred are Patient—Derived Xenograft (PDX) models. Such cancer cell lines, or patientderived tumors are engrafted into immunocompromised mice strains (e.g., Nude mice, Scid mice, NOD mice, Rag 1 null mice, etc. (see, e.g., Belizario, J.E., (2009) “Immunodeficient Mouse Models: An Overview,” m Open 1874-2262/09) or humanized mice such as enic human HLA-A2 mice (see, e. g., Shultz, L.D., et al. (2012) ized mice for immune system investigation: progress, promise and challenges,” Nature Rev Immunol 12:786-798) as described above. In addition, for evaluation of molecules which te immune checkpoint immune—deficient mice may be engrafted with human immune system components (e.g., reconstituted with human PBMCs, stem cells, immune progenitor cells, etc.) prior to or concurrently with implantation of the d tumor cells and treatment as detailed above.

Example 5 PD-l x CTLA-4 Bispecific Molecules g Studies ] Several PD-l X CTLA-4 bispeciflc molecules were generated, including Fc Region-containing diabodies and Fc-Region-containing trivalent les comprising four polypeptides chains. Three diabodies having four polypeptide chains and comprising E/K—coil Heterodimer—Promoting Domains were generated and accorded the designations “DART B,” “DART C,” and “DART D.” One diabody having four chains and sing CHI/CL Domains was generated and accorded the designation “DART E.” Two trivalent binding molecules having four chains and comprising E/K-coil dimer-Promoting Domains and CHI/CL Domains were ted and accorded the designations “TRIDENT A,” and “TRIDENT B.” In on, several antibodies having specificity for PD—l or CTLA-4 were generated. One antibody specific for PD-l was generated and accorded the designation “PD- 1 mAb 6 G4P.” Three antibodies specific for CTLA-4 were generated and ed the designations “CTLA-4 mAb 1,” “CTLA-4 mAb 3 GlAA,” and “CTLA-4 mAb 3 G4P.” The structure and amino acid sequences of these PD-l X CTLA—4 bispecific molecules, anti-PD-l antibodies, anti-CTLA-4 antibodies are provided above and are summarized in Table 9 below.

Table 9 . . . SEQ ID Other l 95 CTLA-4 mAb l IgG4 2 96 DART B E/K-Colls, see PD-l mAb 6-ISQ (YTE) 3 95 F1gure 3B 4 96 1 97 CTLA-4 mAb 3 2 98 E/K-Coils; see DART C IgG4 PD-l mAb 6-ISQ 3 97 Figure 3B 4 98 l 99 PD-l mAb 6-ISQ IgG4 2 100 DART D E/K-Cofls, see CTLA-4 mAb 3 (YTE) 3 99 F1gure 3B 4 100 CTLA-4 mAb 3 IgG4 2 CL/CHl, see DART E PD-l mAb 6-ISQ (YTE) 3 Figure 3c 1 101 PD-l mAb 6-ISQ IgGl 2 100 E/K-Coils; see DART F CTLA-4 mAb 3 (AA/YTE) 3 101 Figure 3B 4 100 -l28- Table 9 . . . SEQ ID Other 1 104 E/K-Coils and PD-l mAb 6-ISQ IgG4 2 105 TRIDENT A .

CTLA-4 mAb 3 (YTE) 3 106 C1?CHl’ see1gure 6A 4 107 1 108 E/K-Coils and PD—1 mAb 6-ISQ IgGl 2 105 TRIDENT B ' CTLA-4 mAb 3 (AA/YTE) 3 109 C1?CHl’ see1gure 6A 4 107 1 88 PD-l mAb 6 2 89 natural antibody PD‘I mAb 6_ISQ IgG4 G4P 3 88 structure 4 89 CTLA-4 811111413343?b 1 l antibody I G1 4 M p g mAb l . structure repllca) 1 92 CTLA 4 2 94 natural amlbOdy mAb 3 CTLA-4 mAb 3 IgGl (AA) 3 92 ure 4 94 1 93 CTLA-4 2 94 l antibody CTLA'4 mAb 3 IgG4 mAb 3 G4P 3 93 structure 4 94 J: les incorporating IgG4 Fc regions also incorporate a stabilized IgG4 hinge region. ** the same amino acid sequence as ipilimumab (see, e.g., IMGT 3D and 2D Structural Database Accession Nos. 8568_H and 8568_L) ] Additional PD—l X CTLA-4 ific molecules comprising alternative PD-l and/or CTLA-4 epitope-binding sites may be readily generated by orating different VH and VL Domains. Similarly, molecules comprising alternative linkers, Fc Regions, and/or having alternative structures may be generated as provided herein (see, e.g., Table 8).

A. ELISA Binding Studies ELISA studies were conducted to measure the binding of serially diluted binding molecules (antibody CTLA—4 mAb 3 G4P, DART D, T A or DARTB) to soluble hCTLAAvi-His (1 ug/mL) or hPD-l-His (l ug/mL) that had been coated onto support plates. Goat anti—human-Fc-HRP (1:10,000) was employed as the secondary detection molecule to detect binding. The results of such studies are shown in Table 10 and in Figures 8A-8B. The data shows that PD-l x CTLA-4 bispecific molecules having two binding sites for PD-l and CTLA-4 (e.g, DART D and DART B) ted binding to PD-l and CTLA-4 that was similar to that of their respective parental anti-PD-l and anti—CTLA-4 antibodies. PD- 1 x CTLA-4 bispeciflc les having two binding sites for PD-l and one binding site for CTLA-4 (e.g, TRIDENT A) exhibited binding to PD-1 that was similar to that of the parental anti-PD-l antibody and exhibited reduced binding to CTLA-4 (relative to that of the parental antibody) due to the d avidity of the trivalent molecule, which comprises only a single binding site for CTLA-4. Similar g results were observed for DART F and T B having IgGl CH1 and/or IgGl (AA/YTE) Fc regions.

Table 10 Construct ECso of CTLA-4 Bindin_ nM ECso of PD-l Bindin_ nM CTLA-4 mAb 3 G4P O N PD-l mAb 6 G4P DART D 0 TRIDENT A ——l.0 DART B ——0.4 The effect of altering orientations and binding domains on binding was igated by incubating PD-1 x CTLA-4 bispeciflc molecules comprising the CTLA-4 binding domains of CTLA-4 mAb l (e.g, DART B) and CTLA-4 mAb 3 (e. g., DART C and DART D) in the presence of soluble human PD-1 (Figure 8C), or soluble human CTLA Avi-His (Figure 8D), that had been coated onto support plates. Goat anti-human-Fcy-HRP was employed as the secondary detection le to detect binding using PICO chemiluminescent substrate. The results indicate that PD-l x CTLA-4 bispecific molecules sing the CTLA-4 binding domains of CTLA-4 mAb 1 (e.g., DART B) and CTLA-4 mAb 3 (e.g, DART C and DART D) exhibit r binding to CTLA-4. The orientation of the g domains (i.e., location on first or second chain) was not found to significantly alter binding to PD-l or CTLA-4 (compare binding ofDART C and DART D).

B. ELISA Blocking Studies A series of ELISA assays were conducted to evaluate the ability of bispeciflc molecules of the invention to block ligand binding to PD-l and , alone and in combination. Blockade of PD-Ll binding to PD-l was evaluated in the presence of equal amounts of an irrelevant antigen and in the presence of equal amounts of CTLA-4. Plates were coated with a 1:1 mix of His-tagged soluble human PD-1 (shPD-l) and a His-tagged irrelevant antigen (irrAg) (2 ug/ml each), or a 1:1 mix of shPD—l and a His-tagged soluble human CTLA- -l30- 4 (shCTLA-4) (2 ug/ml each). PD-l mAb 6 G4P, DART D, TRIDENT A or a CONTROL TRIDENT (having two binding sites for RSV and one binding site for CTLA—4) at the indicated concentrations were premixed for 5 mins with 6 ug/ml biotin-labeled PD-Ll and added to the plates. PD—Ll binding was detected using streptavidin HRP 00). The results of this evaluation are presented in s 9A-9B. All of the PD-l binding molecules tested were found to be able to inhibit PD-Ll binding to PD-l.

Blockade of of B7-1 binding to CTLA-4 was evaluated in the ce of equal amounts of an irrelevant antigen and in the presence of equal amounts of, or four-fold more PD-l. Plates were coated with a 1:1 mix of shCTLA-4 and irrAg (2 ug/ml each), a 1:1 mix of shCTLA-4 shPD-l (2 ug/ml each), or a 1:4 mix of shCTLA-4 (0.8 ug/ml) and shPD-l (3.2 . PD-1 mAb 6 G4P, DART D, T A, CTLA-4 mAb 3 G4P, or CONTROL TRIDENT at indicated concentrations were premixed for 5 mins with 0.2 ug/ml biotin-labeled B7-1 and added to the plates. B7-1 binding was detected using streptavidin HRP (13,000).

The results of this evaluation are presented in Figure 9C-9E. All of the CTLA-4 binding molecules tested were found to be able to inhibit B7-1 g to CTLA-4. TRIDENT A blocking of B7-1 binding was found to be enhanced by the interaction of its PD-1 binding arm interacting with immobilized PD-l (compare to CONTROL TRIDENT which does not bind PD-l) e 9D). Moreover, under the 1:4 CTLA-4:PD-1 condition, which better mimics the relative expression levels seen on stimulated cells (see, Figure 19A), TRIDENT A blocking of B7-1 binding was found to be further enhanced , the T A curve was further shifted compared to the curve of the CONTROL TRIDENT, which does not bind PD-l) (Figure 9E).

The results of these ELISA studies demonstrate that all of the PD-l binding molecules tested were able to inhibit PD-Ll from binding to the PD-l (Figures 9A-9B). All such molecules are nt for PD-l and exhibited similar inhibition profiles. All of the CTLA-4 binding molecules tested were able to inhibit B7-1 from binding to immobilized CTLA-4 (Figure 9C-9E) with les comprising two PD-l binding sites and one CTLA-4 binding site exhibiting stronger inhibition in the presence of PD-l (Figure 9D-9E). Thus, the trivalent les comprising a single CTLA-4 binding site exhibit a PD-l biased blockade of CTLA-4 ligands, trating that the CTLA-4 interaction can be tailored by adjusting the valency. -l3l- C. E® Studies The binding affinity of DART A, TRIDENT A, and CTLA-4 mAb 1 to human CTLA-4 and cynomolgus monkey CTLA-4 was investigated using BIACORE® analysis.

Briefly, His-tagged soluble CTLA—4 (an extracellular portion ofhuman or cynomolgus monkey CTLA-4 fused to a histidine-containing peptide) was captured on immobilized anti-PentaHis and then ent concentrations 200 nM) ofthe CTLA-4 binding molecules were passed over the lized CTLA-4 proteins. The kinetics of binding were determined Via BIACORE® analysis (affinity by 1:1 ir binding model (simultaneous ka/kd), or avidity by separate ka/kd 1:1 fit). The calculated ka, kd and KD from these studies are presented in Table 11.

Table 11 Human CTLA-4 C no CTLA-4 Molecule ka kd KD ka kd Kl) x105 xio'4 (nM) x105 x10'3 (11M) —m“——-m * avidity by separate ka/kd 1:1 fit 1 y by 1:1 Langmuir binding model ] DART D is bivalent for CTLA-4 and exhibits binding affinities to human and cynomolgus monkey CTLA—4 that are within about 2 to 4—fold that of the CTLA—4 mAb 1.

TRIDENT A is monovalent for CTLA-4 exhibits lower affinity for both human and cynomolgus monkey CTLA-4 as expected in view of its reduced avidity.

The binding affinity of DART A, T A, PD-l mAb 6 G4P, and CTLA-4 mAb 3 GlAA to human PD-l was investigated using BIACORE® analysis. The binding molecules were captured on immobilized F(ab)2 goat anti-human Fc and then different concentrations (6.25-100 nM) of His-tagged soluble human PD-l were passed over the immobilized binding molecules, and the kinetics of binding was determined via BIACORE® analysis (Langmuir 1:1 binding fit). The ated ka, kd and KD from these studies are presented in Table 12 (n.d., not detectable).

Table 12 Human PD-l DART A, TRIDENT A, PD-l mAb 6 G4P are each bivalent for PD—l and exhibit able g affinities. As expected, CTLA-4 mAb 3 GlAA did not exhibit any detectable binding for human PD-l.

D. CTLA-4 Cell Based Assays DART B, DART D, TRIDENT A, the anti-CTLA-4 antibodies CTLA-4 mAb 1, CTLA-4 mAb 3 G4P, and an hIgG control dy were evaluated for binding to CHO cells expressing cynomolgus monkey CTLA-4 (cynoCTLA-4) or human CTLA-4 (huCTLA-4).

The results of this evaluation are shown in Figures B. The binding molecules were incubated in the presence of CHO cells that were sing either cynomolgus monkey CTLA-4 (Figure 10A) or human CTLA-4 (Figure 10B). Binding to such cells was detected using an anti-human Fc secondary antibody. The results show that all the molecules tested were able to bind human and cynomolgus monkey CTLA—4 expressed on the surface of the CHO cells. The anti-CTLA-4 antibodies exhibited similar binding profiles to huCTLA-4; the bivalent, bispecific molecules DART B and DART D exhibited slightly reduced binding, and the trivalent binding molecule. TRIDENT A, which is lent for CTLA-4 exhibited lower binding than the molecules having higher valency for CTLA-4. The control antibody did not bind. Similar results were seen for binding to cynoCTLA-4.

DART C, DART D, DART E, TRIDENT A, the anti-CTLA-4 antibodies CTLA- 4 mAb 1, CTLA-4 mAb 3 G1AA, and the anti-PD-l antibody PD-l mAb 6 G4P were ted for binding to Jurkat cells which express huCTLA-4 but not PD-l on their surface. Binding of the DART and TRIDENT molecules to human CTLA-4 was detected using anti-human FC secondary Ab (FACS). The results of the tion are shown in Table 13 and Figure 11A (DART C, DART D, DART E, CTLA-4 mAb 1, CTLA-4 mAb 3 G1AA, and PD-l mAb 6 G4P) and Figure 11B (CTLA-4 mAb l, CTLA—4 mAb 3 G1AA, PD-l mAb 6 G4P and TRIDENT A). As shown in Figures 11A-11B, the PD-l antibody did not bind CTLA-4, but all the CTLA-4 binding les tested were able to bind huCTLA-4 expressed on the surface -l33- of Jurkat cells. The anti-CTLA-4 antibodies exhibited r binding profiles; the bivalent, bispeciflc molecules DART C, DART D, and DART E exhibited slightly d binding to Jurkat cells and the trivalent binding molecule. TRIDENT A, which is monovalent for CTLA- 4 exhibited lower binding than the molecules having higher valency for CTLA—4.

Molecule EC50 (nM) CTLA-4 mAb 1 0.4215 PD-1 mAb 6 G4P 6.557 CTLA-4 mAb 3 G1AA 0.3728 DART E 1.269 DART C 0.7575 DART D 08829 TRIDENT A 4.638 DART D, TRIDENT A and the anti-CTLA-4 antibodies CTLA-4 mAb l, CTLA- 4 mAb 3 GlAA were evaluated for their ability to block the CTLA-4 ligands B7-l and B7-2.

His-tagged derivatives of B7-1 and B7-2 were incubated in the presence of CTLA-4 Jurkat cells. Binding of -1 and -2 was detected using an anti-His antibody. The results of this evaluation are shown in Figure 12A (His-B7-1) and Figure 12B (His-B7—2). All the molecules tested were found to be able to t B7-1 and B7-2 from binding CTLA-4 expressed on the surface of the Jurkat cells. The anti-CTLA-4 dies exhibited similar inhibition profiles, the bivalent, bispeciflc molecule DART D was slight less potent an inhibitor and the trivalent binding molecule. TRIDENT A, which is monovalent for CTLA-4 was less potent than any of the molecules having higher valency for CTLA-4. The control antibody did not t at all. The ELISA studies described above suggest that TRIDENT A, and r les having two PD-l binding sites and one CTLA-4 binding site would be more potent inhibitors in the presence of PD-l.

An lL-2/Luc Jurkat cell CTLA-4 reporter assay was used to evaluate the y ofDART C, DART D, TRIDENT A, CTLA-4 mAb 3 GlAA and PD-l mAb 6 G4P to reverse CTLA-4 immune checkpoint inhibitory signal as demonstrated by increased luciferase expression. IL-2/Luc-Jurkat-CTLA-4 cells were therefore incubated in the presence of such molecules (R:S= l : 0.3) for 30 min at 37 CC, after which time artificial antigen presenting Raji cells were added and the incubation continued for 6 hours. The artificial antigen presenting cells activate the TCR/CD3 complex on the Jurkat reporter cells. The results of the tion are shown in Figure 13. All of the CTLA-4 binding molecules tested were able reverse the —134— CTLA-4 immune checkpoint inhibitory signal as determined by the luciferase assay.

TRIDENT A, which is monovalent for CTLA-4 was less potent in this assay than any of the molecules having higher valency for CTLA-4. The control antibody did not inhibit at all. The ELISA studies described above suggest that TRIDENT A, and similar molecules having two PD-l binding sites and one CTLA-4 binding site would be more potent in the presence of PD- E. PD-l Cell Based Assays DART D, TRIDENT A, PD-l mAb 6 G4P, and CTLA-4 mAb 3 GlAA were evaluated for their ability to bind NSO cells expressing PD-l but not CTLA-4. Binding molecules were incubated in the presence of the cells and the mean fluorescence index of the cells was measured. The s of this tion are presented in Figure 14. As expected, the CTLA-4 antibody did not bind, all the bispeciflc binding les were found to be able to bind PD-l expressed on the e ofN80 cells. All the iflc molecules are bivalent for PD-l and exhibited similar binding to NSO cells.

DART D, TRIDENT A, PD—l mAb 6 G4P, and CTLA-4 mAb 3 GlAA were ted for their ability to block binding between PD-l expressed on the cell surface and its ligands PD-Ll and PD-L2. PD-Ll-PE or PD-L2-PE was incubated in the presence of such binding molecules and their ability to bind to NSO-PD-l cells was evaluated using FACS. The results of this evaluation are presented in Figure 15A ) and Figure 15B (PD-L2). As expected, the CTLA-4 antibody did not t, all of the PD-l binding molecules tested were able to inhibit both PD-Ll (Figure 15A) and PD-L2 (Figure 15B) from binding to the PD-l expressed on the surface of the NSC cells. All the PD-l binding molecules are bivalent for PD-l and exhibited similar inhibition profiles.

DART D, TRIDENT A, CTLA-4 mAb 3 GlAA, and PD-l mAb 6 G4P were also evaluated in a PD-l blockade reporter assay. Such binding molecules were ted in the presence of PD-Ll+ CH0 and Jurkat effector cells, and the ability of the binding molecules to block immune inhibition (by blocking the PD-l / PD-Ll interaction) was assessed by following the extent of CD3-mediated activation (as trated by increased luciferase expression in the NFAT-luc/PD-l Jurkat assay; Promega). The results of this evaluation are presented in Figure 16. All of the PD-l binding molecules tested were able to reverse the PD-l immune checkpoint inhibitory signal as trated by increased luciferase expression. All the PD-l -l35- binding les are bivalent for PD-1 and exhibited similar y to inhibit PD-1 blockade of T cell signaling. The CTLA-4 antibody did not inhibit at all in this system.

F. CTLA-4/PD-1 Cell Based Assays DART D, TRIDENT A, and a ve control antibody were examined for their ability to ate PD-1 and CTLA-4 in an enzyme-fragment complementation assay by DiscoverX. In brief, aliquots of the U208 CTLA—4(1-195)—PK PD-1(1-199)-EA cell line #9 were plated in quadruplicate at 5,000 cells / well in DiscoverX CP5 plating media on 384-well plates. Cells were allowed to attach for 4 hours at 37 oC / 5% C02. 11 point, 1:3 dilution series of each of the binding molecules were then added to the PD-1 — CTLA-4 cells. The plates were incubated overnight (16 hrs) at 37 °C / 5% C02. PathHunter detection reagent was added to the wells, which were then ted for 1 hour at room temperature in the dark, and the plate was then read on an Envision luminometer. The results of this tion are presented in Table 14 and Figure 17 (U208 CTLA-4(1—195)—PK PD-1(1-199)—EA cell line #9). Both the bispecific DART D and TRIDENT A molecules show comparable co-engagement ofPD-l and CTLA-4 in cells that co-express both ors, as shown by enzyme-fragment complementation, ting that the bispecific molecules of the invention are capable of simultaneous binding of PD-l and CTLA-4, and further indicating that anchoring through PD- 1 compensates for the decreased CTLA-4 avidity of the T molecule when both target receptors are expressed. This finding is consistant with the ELISA inhibition s described above. The negative control elicited no significant increase in signal in the PD1-CTLA4 cell line. Incubation with higher concentrations of TRIDENT A elicited a robust signal increase in the U208 PD1-CTLA4 zation cell line (S:B=12.7). The response with DART D in dose-response testing in the PD-1 — CTLA-4 cell line was smaller in magnitude (S:B=9.2) but the EC50 values were similar for both these molecules (EC50=20 pM).

Table 14 —Negative Control TRIDENT A DART D HillSlope ~15.99 1.103 0.8095 EC50 (nM) ~6.883 x 10-10 2.123 x 10-11 2.090 x 10-11 The ability ofDART D, TRIDENT A, CTLA-4 mAb 3 G1AA, PD—l mAb 6 G4P and the combinations of CTLA-4 mAb 3 G1AA/PD-1 mAb 6 G4P (Ab Combo 1) to enhance the response of a Mixed Lymphocyte Reaction (MLR) was evaluated. Monocyte-derived dendritic cells were generated by ng CD14+ monocytes (isolated from PBMCs using Miltenyi positive selection kit) with GM-CSF (100 ng/ml) and IL-4 (10 ng/ml) and then culturing the cells for 7 days. On day 7, cells were harvested and plated into 96—well plates and ed for 24 h. On day 8, CD4+ T-cells (isolated by negative selection using Myltenyi kit) at 200,000 cells/well and test articles were added and cultured for 3 days. IFN—g levels in culture supernatants were then measured using using human DuoSet ELISA Kits for IFN—y (R&D Systems) according to the manufacturer’s instructions. When antibodies were used in ation, each antibody was added at the indicated concentration so that the total concentration of antibody added is d, The e y is plotted in Figure 18. Both the bispecific DART D and TRIDENT A molecules were found to enhance the MLR response to the same extent or slightly better than the combination of individual parental antibodies.

The ability ofDART D, T A, CTLA-4 mAb 3 GlAA, PD-l mAb 6 G4P and the combination of CTLA-4 mAb 1/PD-1 mAb 1 (Ab Combo 1) to enhance cytokine release through checkpoint inhibition was also evaluated in a Staphylococcus aureus enterotoxin type B (SEB) re-stimulation assay. In general, PBMCs were d from whole blood (e.g., using the Ficoll-Paque Plus density nt centrifugation method (GE Healthcare) according to manufacturer’s instructions) from y donors. Purified PBMCs were cultured in RPMI-media + 10% heat inactivated FBS + 1% Penicillin/Streptomycin in T- bulk flasks for 2-3 days alone or with SEB (e.g., Sigma-Aldrich) at 0.5 ng/mL (primary stimulation). At the end of the first round of SEB-stimulation, PBMCs are washed twice with PBS and immediately plated in 96—well tissue culture plates at a concentration of 1-5 X 105 cells/well in media alone, media with a control or a test article, media with SEB at 0.5 ng/mL (secondary stimulation) and no antibody, or media with SEB and a control IgG or a test e, and were cultured for an additional 2-3 days. At the end ofthe second stimulation, supernatants were harvested to measure cytokine secretion (e.g, using human DuoSet ELISA Kits for IFNy, IL-2, TNFOL, IL-10, and IL-4 (R&D Systems) ing to the manufacturer’s instructions).

Figures 19A-19B show fluorescence-activated cell sorting (FACS) dot plots of the expression of PD-l vs. CTLA-l by such PBMCs in the absence (Figure 19A) or presence (Figure 19B) of SEB stimulation. Figure 19C shows the effect of the SEB stimulation on IFN—y secretion. PBMCs were stimulated with Staphylococcus aureus enterotoxin type B (SEB) at 0.5 ng/ml for 48 hours. Cells were then harvested, washed and re-plated in 96 well plates with antibodies at various concentrations with fresh SEB for an additional 48 hours. The supernatant was then harvested and ed by flow try ELISA for IFN—y production. -l37- Both the bispecific DART and the TRIDENT protein showed an increase in IFN—y response that recapitulated the response observed with the combination of the dual parental mAbs.

Similar results were seen in a SEB Stimulation assay in which the PBMCs were cultured with a high concentration (500 ng/mL) of SEB for 72 hours. To further investigate the affect of PD] x CTLA-4 bispeciflc molecules on the T-cell response, PBMCs were stimulated with 0.5 ng/ml SEB for 48 hours, harvested, washed and re-plated in 96-well plates with fresh SEB and either DART D, TRIDENT A, CTLA-4 mAb 3 GlAA, PD—l mAb 6 G4P or the ation of CTLA-4 mAb 3 GlAA / PD-l mAb 6 G4P (Ab Combo 1) for an additional 48 hours, and the released IL—2 was measured (Figure 19D). Figures D show that the administration of PDl X CTLA-4 bispecific molecules significantly enhanced T-cell responses. When antibodies were used in combination, each antibody was added at the indicated concentration so that the total concentration of antibody added is d.

Example 6 In Vivo Studies A. Activity of a PD-l x CTLA-4 Bispecific Molecule in GVHD Murine Model The activity of a representative PDl X CTLA-4 bispeciflc bivalent molecule, DART D was assessed in a PBMC implanted NOG murine model of Graft Versus Host Disease (GVHD). The study design is presented in Table 15.

Table 15 Group N/sex Treatment Dose Route/ Cell Implant(s) (ug/kg) Schedule DARTD IV/Q7Dx7 PBMC (1P, 1E7) DART D IV/Q7D x 7 PBMC (IP, 1E7) DART D IV/Q7D x 7 PBMC (IP, 1E7) ] CD3+ T cell counts were med via FACS on study day 14 and are plotted in Figure 20A. Survival was monitored over the course of the study and is plotted as percent survival in Figure 20B. sed T cell expansion and accelerated GVHD was seen in animal treated with 500 ug/kg DART D, consistence with enhancement of T cell immune responses. -l38- B. Toxicology and cokinetic Study of PD-l x CTLA-4 Bispecific Molecules The safety profile of a representative PDl X CTLA-4 bispecific bivalent molecule, DART D, and a representative PDl X CTLA-4 ific trivalent molecule, TRIDENT A, was assessed in a non-GLP (Good Laboratory Practice) dosing study in cynomolgus monkeys. In on, l markers pharmacodynamics activity were ed.

In this study the potential toxicity of the PD-l x CTLA—4 bispecific molecules, when administered by multiple intravenous infusions was evaluated. The study design is presented in Table 16.

Table 16 Group Test Article Dose (mg/kg) Number of Animals 1M 1F 3M 3F —3M 3F TRIDENTA —— 2M 1F A 2-week interval was thus provided between the 50 mg/kg dose and escalation to 75 mg/kg, The following parameters and nts were evaluated in this study: clinical signs, body s, food consumption, body temperature, clinical pathology ters (coagulation, clinical chemistry and hematology pre-dose and 23 hours post-dose for Groups 1-3, out to day 22 for Group 4), bioanalysis and toxicokinetic parameters, flow cytometry (pre- dose and 23 hours post dose), cytokines (2, 6, 22 hours post-dose). Anti-Drug-Antibodies were evaluated for Group 4 only on days 8, 15 and 22. Necropsy was performed 48 hours after the 3r01 dose for Groups 1-3 only. The in vivo binding and activity of the PD-l x CTLA-4 bispeciflc molecules was also examined as described below.

All animals survived until scheduled euthanasia. No adverse clinical observations in animals receiving 3 doses up to 75 mg/kg/week. In ular, no diarrhea was observed.

The histopathology was also rkable. Increases in globulin levels were observed in the treatment groups and the organ weight of the spleen and thymus were observed to increase in Groups 2—3 (see Table 17, Group 4 was not necropsied), as would be expected upon stimulation of the immune system. The serum concentration-time profiles for each of the treatment groups are shown in Figures 21A-21C and are tent with molecules comprising human Fc regions in cynomolgus monkeys. -l39- Table 17 SleenzBod Weiht Th muszBod Wei_ht -—-_0.080mean,n=2 0.035 mean,n=2 DARTD 0.239 mean, n=6 0.088 mean, n=6 DARTD 0.225 mean, n=6 0.084 mean, n=6 It has been reported that ses in absolute lymphocyte count (ALC) after treatment with the anti-CTLA-4 antibody ipilimumab appear to correlate with clinical benefit and overall survival (see, e. g., Ku, G.Y., et al. (2010) “Single-Institution ence With Ipilimumab In AdvancedMelanoma Patients In The Compassionate Use Setting: Lymphocyte Count After 2 Doses Correlates With Survival” Cancer 1 16(7): 1767-1775) indicating that ALC may be a useful pharmacodynamic (PD) endpoint. The ALC counts were ed in each of the above-described groups pre—treatment and post-treatment on days 2, 8, 9, 15 and 16.

Occupancy of DART D or TRIDENT A binding sites on PD-1+ T cells was determined by measuring the mean fluorescent intensity (MFI) of anti-human IgG4 Alexa 488+ events in the CD4+/PD-1+ and CD8+/PD—1+ T cell tions under two conditions for each monkey blood sample. Under one ion, the MFI values obtained in the presence of excess DART D or TRIDENT A were used to determine the maximal DART D or TRIDENT A binding intensity on PD-1+ cells within each cell population. Under the second condition, the MFI values obtained in the presence of excess negative control were used to determine the binding intensity of PD-1+ cells within each cell population exhibited in the DART D or TRIDENT A- treated animal at the time of sample collection. The difference n the two conditions was used to calculate % occupancy of DART D or TRIDENT A binding sites on PD-1+ T cell subsets in DART D or TRIDENT A-treated animals as follows: MFI of Anti-HulgG4+ Events in % Occupancy of DART D or TRIDENT A {the Presence of Excess AEX1367] _ x100 Binding Sites On PD-l+ T Cell Subsets [ ce of Excess DART D or TRIDENT A]MFI ofAnti-HngG4+ Events in the The absolute counts, and the percent change normalized to Day 1 are plotted in Figure 22A (in thousands of cells /ul (th/ul)) and in Figure 22B (percent change in the ALC normalized to Day 1 (D1)). Each of the DART D treatment groups exhibited an l drop in ALC counts immediately after treatment followed by an se in ALC to levels well above ne. A similar trend was observed for the TRIDENT A treatment group, which only received only one lower dose. —140— In addition, CD4+ T cell eration and PD-l occupancy on T cells were examined for the above-described Groups 1—3. Bn'efly, CD3+/PD-l+ T cells were analyzed by FACS to evaluate the t cells bound by DART D. Forty microliters ofthe negative control molecule (respiratory syncytial virus (RSV) X fluorescein IgG4,K Fc DART) or test article (DART D or TRIDENT A) at 35 ug/mL were added to a 96 ell plate. One hundred microliters of well-mixed anticoagulated whole blood were then added into each well, thoroughly mixed using a pipette, and incubated in the dark for 45 to 75 minutes at ambient temperature. One thousand microliters of 1x BD FACS Lysing solution were then added to each well and mixed using a pipette; the plate was then incubated in the dark for an additional to 20 minutes at ambient temperature. The plate was then centrifuged at 400 x g for 5 minutes and the supernatant was discarded. One thousand microliters of FACS buffer were added in each well and mixed as a washing step. The plate was then fuged at 400 x g for s and the supernatant was discarded. The cell pellet was resuspended with twenty microliters of Panel 1 antibody mix and incubated for 30 to 60 minutes at ambient temperature.

The plate was washed as in us wash steps. At the end of incubation, the plate was washed again and the cell pellet was finally resuspended in three-hundred microliters of FACS buffer and the samples were analyzed with a BD FACSCanto 11 cell analyzer. The results of the analysis are shown in Figures 23A-23B.

As shown in Figure 23A (for DART D administered at 50 mg/kg) and Figure 23B (for DART D administered at 75 mg/kg), PD-l ncy (i.e., binding by DART D) was maximal hout the duration of treatment for Groups 2 and 3. Proliferation CD4+ T cells were ted by FACS for co-expression of Ki-67 (a cellular marker for proliferation).

Twenty microliters of an antibody mixture A (containing antibodies that bind cell surface markers: CD45, CD3, CD4, and CD8) were added into a 96 deep-well plate. Fifty microliters of ixed anticoagulated whole blood were then added into each well, mixed ghly using a pipette, and incubated in the dark for 15 to 45 minutes at ambient temperature. Five hundred microliters of 1x BD FACS Lysing solution were then added to each well and mixed using a pipette, the plate was then incubated in the dark for an additional to 20 minutes at ambient temperature. The plate was centrifuged at 1200 rpm for 5 minutes and the supernatant was discarded. Five hundred microliters of FACS buffer were then added in each well and mixed as a washing step. The plate was then centrifuged at 1200 rpm for 5 minutes and the supernatant was discarded. The cell pellet was resuspend in antibody mixture —141— B ining antibodies that bind the ellular marker, Ki 67) or were resuspended in an iso antibody preparation (containing isotype ls for the intracellular marker) and incubated in the dark for 15 to 45 minutes. After washing, the cell pellet was resuspended in three hundred iters of FACS buffer and the samples were analyzed with a BD FACSCanto II cell analyzer. From a T Cell Intracellular ng Panel, the percentage of CD4+ and CD8+ cells was determined as the fraction of total CD45+ leukocyte gated cells.

The cellular events of Ki 67+ in gated CD4+ cells were counted and the percentage of CD4+/Ki 67+ T cells (proliferative CD4 T cells) was determined as the fraction of total CD4+ cells. In a similar manner, the percentage of CD8+/Ki 67+ T cells (proliferative CD8 T cells) was determined as the on of total CD8+ cells. The results of the analysis are shown in Figures 24A-24B.

As shown in Figures 24A-24B, proliferation of CD4+ T cells was markedly ed in treatment Groups 2 and 3 throughout the duration of treatment. The results of this study indicate that administration of PD1 x CTLA-4 ific molecules is well tolerated in cynomolgus monkeys at trations of up to 75 mg/kg. Well above the 5 mg/kg dosage where adverse events have been reported for cynomolgus monkeys treated with Ipilimumab.

The molecules exhibited a favorable pharmacokinetic profile and a number of markers pharmacodynamics activity were observed including increased lymphocyte count, increased in levels, increased spleen and thymus organ weights, increased T cell proliferation (both T cell counts and sion of Ki-67) and maximal PD-1 occupancy on T cells.

All ations 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. While the invention has been described in connection with specific embodiments thereof, it will be understood that it is capable of further modifications and this application is intended to cover any ions, uses, or adaptations of the invention following, in general, the principles of the invention and ing such departures from the present disclosure as come within known or customary practice within the art to which the invention pertains and as may be applied to the essential features hereinbefore set forth.

Other aspects of the invention as described herein are defined in the following paragraphs: 1. A bispecific molecule possessing both one or more e-binding sites capable of immunospecific binding to (an) epitope(s) of PD-1 and one or more epitope-binding sites capable of immunospecific binding to (an) epitope(s) of CTLA-4, wherein said molecule comprises: (A) a Heavy Chain Variable Domain and a Light Chain Variable Domain of an antibody that binds PD-1; and (B) a Heavy Chain Variable Domain and a Light Chain Variable Domain of an antibody that binds CTLA-4; wherein said molecule is: (i) a diabody, said y being a covalently bonded complex that ses two, three, four or five polypeptide chains; or (ii) a trivalent binding molecule, said trivalent binding molecule being a covalently bonded complex that comprises three, four, five, or more polypeptide chains. 2. The bispecific molecule of paragraph 1, wherein said le exhibits an activity that is enhanced ve to such activity ted by two monospecific molecules one of which possesses said Heavy Chain Variable Domain and said Light Chain Variable Domain of said antibody that binds PD-1 and the other of which possesses said Heavy Chain Variable Domain and said Light Chain Variable Domain of said antibody that binds . 3. The ific molecule of paragraph 1 or 2, wherein said molecule elicits fewer immune-related adverse events (irAEs) when administered to a subject in need thereof relative to such iREs elicited by the administration of a monospecific antibody that binds CTLA-4. 4. The bispecific molecule of any one of paragraphs 1-3, wherein said molecule comprises an Fc Region.

. The bispecific molecule of aph 4, wherein said Fc Region is a variant Fc Region that comprises: (A) one or more amino acid modifications that reduces the ty of the variant Fc Region for an FcγR; and/or (B) one or more amino acid modifications that enhances the serum half-life of the variant Fc Region. 6. The bispecific molecule of paragraph 5, wherein said modifications that reduces the affinity of the variant Fc Region for an FcγR comprise the substitution of L234A; L235A; or L234A and L235A, wherein said numbering is that of the EU index as in Kabat. 7. The bispecific molecule of paragraph 5 or 6, wherein said modifications that that enhances the serum ife of the variant Fc Region comprise the substitution of M252Y; M252Y and S254T; M252Y and T256E; M252Y, S254T and T256E; or K288D and H435K, wherein said numbering is that of the EU index as in Kabat. 8. The bispecific molecule of any one of paragraphs 1-7, n said molecule is said y and comprises two epitope-binding sites capable of immunospecific binding to an epitope of PD-1 and two epitope-binding sites e of immunospecific binding to an epitope of CTLA-4. 9. The bispecific molecule of any one of paragraphs 1-7, wherein said molecule is said trivalent binding molecule and comprises two epitope-binding sites capable of immunospecific binding to an epitope of PD-1 and one epitope-binding site capable of immunospecific binding to an epitope of CTLA-4.

. The ific molecule of any one of aphs 1-9, wherein said molecule is capable of binding to PD-1 and CTLA-4 molecules present on the cell surface. 11. The bispecific molecule of any one of paragraphs 1-10, wherein said molecule is capable of simultaneously binding to PD-1 and CTLA-4. 12. The bispecific molecule of any one of paragraphs 1-11, n said molecule es the stimulation of immune cells. 13. The bispecific le of aph 12, wherein said stimulation of immune cells results in: (A) immune cell proliferation; and/or (B) immune cell production and/or release of at least one cytokine; and/or (C) immune cell production and/or release of at least one lytic molecule; and/or (D) immune cell expression of at least one activation marker. 14. The bispecific molecule of paragraph 12 or 13, wherein said immune cell is a T-lymphocyte or an NK-cell.

. The ific molecule of any one of aphs 1-14, wherein said epitopebinding sites capable of immunospecific binding to an epitope of PD-1 (A) the VH Domain of PD-1 mAb 1 (SEQ ID NO:47) and the VL Domain of PD-1 mAb 1 (SEQ ID NO:48); or (B) the VH Domain of PD-1 mAb 2 (SEQ ID NO:49) and the VL Domain of PD-1 mAb 2 (SEQ ID NO:50); or (C) the VH Domain of PD-1 mAb 3 (SEQ ID NO:51) and the VL Domain of PD-1 mAb 3 (SEQ ID NO:52); or (D) the VH Domain of PD-1 mAb 4 (SEQ ID NO:53) and the VL Domain of PD-1 mAb 4 (SEQ ID NO:54); or (E) the VH Domain of PD-1 mAb 5 (SEQ ID NO:55) and the VL Domain of PD-1 mAb 5 (SEQ ID NO:56); or (F) the VH Domain of PD-1 mAb 6 (SEQ ID NO:57) and the VL Domain of PD-1 mAb 6 (SEQ ID NO:58); or (G) the VH Domain of PD-1 mAb 6-I VH (SEQ ID NO:86) and the VL Domain of PD-1 mAb 6-SQ VL (SEQ ID NO:87); or (H) the VH Domain of PD-1 mAb 7 (SEQ ID NO:59) and the VL Domain of PD-1 mAb 7 (SEQ ID NO:60); or (I) the VH Domain of PD-1 mAb 8 (SEQ ID NO:61) and the VL Domain of PD-1 mAb 8 (SEQ ID . 16. The bispecific molecule of any one of paragraphs 1-15, wherein said epitopebinding site(s) capable of immunospecific binding to an epitope of CTLA-4 comprise: (A) the VH Domain of CTLA-4 mAb 1 (SEQ ID NO:76) and the VL Domain of CTLA-4 mAb 1 (SEQ ID NO:77); or (B) the VH Domain of CTLA-4 mAb 2 (SEQ ID NO:78) and the VL Domain of CTLA-4 mAb 2 (SEQ ID NO:79); or (C) the VH Domain of CTLA-4 mAb 3 (SEQ ID NO:90) and the VL Domain of CTLA-4 mAb 3 (SEQ ID . 17. The bispecific molecule of paragraph 16, wherein: (A) said epitope-binding sites capable of immunospecific binding to an epitope of PD-1 comprise the VH Domain of PD-1 mAb 6-I VH (SEQ ID NO:86) and the VL Domain of PD-1 mAb 6-SQ (SEQ ID NO:87); (B) said epitope-binding site(s) e of immunospecific binding to an epitope of CTLA-4 comprise(s) the VH Domain of CTLA-4 mAb 3 (SEQ ID NO:90) and the VL Domain of CTLA-4 mAb 3 (SEQ ID NO:91). 18. The bispecific molecule of any one of paragraphs 1-17, wherein said molecule comprises: (A) two polypeptide chains having SEQ ID NO:95, and two ptide chain having SEQ ID NO:96; or (B) two polypeptide chains having SEQ ID NO:97, and two polypeptide chain having SEQ ID NO:98; or (C) two polypeptide chains having SEQ ID NO:99, and two polypeptide chain having SEQ ID ; or (D) two polypeptide chains having SEQ ID NO:102, and two polypeptide chain having SEQ ID NO:103; or (E) two polypeptide chains having SEQ ID NO:101, and two polypeptide chain having SEQ ID NO:100; or (F) one polypeptide chains having SEQ ID NO:104, one polypeptide chain having SEQ ID NO:105, one polypeptide chain having SEQ ID NO:106, and one polypeptide chain having SEQ ID NO:107; or (G) one polypeptide chains having SEQ ID NO:108, one polypeptide chain having SEQ ID NO:105, one polypeptide chain having SEQ ID NO:109, and one polypeptide chain having SEQ ID NO:107. 19. A pharmaceutical composition that comprises an effective amount of the bispecific molecule of any of paragraphs 1-18 and a pharmaceutically able carrier.

. The bispecific le of any one of paragraphs 1-18, wherein said molecule is used to promote stimulation of an immune-mediated response of a subject in need thereof. 21. The bispecific molecule of any one of paragraphs 1-18, wherein said molecule is used in the treatment of a disease or condition associated with a suppressed immune system. 22. The bispecific le of paragraph 21, wherein the disease or condition is cancer or an infection. 23. The ific molecule of paragraph 22, wherein said cancer is characterized by the presence of a cancer cell selected from the group consisting of a cell of: an adrenal gland tumor, an AIDS-associated cancer, an alveolar soft part sarcoma, an astrocytic tumor, bladder , bone cancer, a brain and spinal cord cancer, a metastatic brain tumor, a breast cancer, a carotid body tumors, a cervical cancer, a chondrosarcoma, a ma, a chromophobe renal cell carcinoma, a clear cell carcinoma, a colon cancer, a colorectal cancer, a cutaneous benign fibrous histiocytoma, a desmoplastic small round cell tumor, an ependymoma, a Ewing’s tumor, an extraskeletal myxoid chondrosarcoma, a fibrogenesis imperfecta ossium, a fibrous sia of the bone, a gallbladder or bile duct , c cancer, a gestational trophoblastic disease, a germ cell tumor, a head and neck cancer, hepatocellular oma, an islet cell tumor, a Kaposi’s Sarcoma, a kidney cancer, a leukemia, a lipoma/benign tous tumor, a liposarcoma/malignant lipomatous tumor, a liver cancer, a lymphoma, a lung cancer, a medulloblastoma, a melanoma, a meningioma, a multiple ine neoplasia, a multiple myeloma, a myelodysplastic syndrome, a neuroblastoma, a neuroendocrine tumors, an ovarian cancer, a pancreatic cancer, a papillary thyroid carcinoma, a parathyroid tumor, a pediatric cancer, a eral nerve sheath tumor, a phaeochromocytoma, a pituitary tumor, a prostate cancer, a posterious uveal melanoma, a rare hematologic disorder, a renal metastatic cancer, a rhabdoid tumor, a rhabdomysarcoma, a sarcoma, a skin , a soft-tissue sarcoma, a squamous cell cancer, a stomach , a synovial a, a testicular cancer, a thymic carcinoma, a thymoma, a thyroid metastatic cancer, and a uterine cancer. 24. The bispecific molecule of paragraph 22, wherein said infection is terized by the presence of a bacterial, fungal, viral or protozoan pathogen.

Still further embodiments are within the scope of the following claims.

Claims (18)

Claims

1. A bispecific molecule, comprising: (A) one or more epitope-binding sites e of immunospecific binding to an epitope of PD-1 comprising the VH Domain of PD-1 mAb 6-I VH (SEQ ID NO:86) and the VL Domain of PD-1 mAb 6-SQ (SEQ ID NO:87); and (B) one or more epitope-binding sites capable of immunospecific binding to an epitope of CTLA-4 comprising the VH Domain of CTLA-4 mAb 3 (SEQ ID NO:90) and the VL Domain of CTLA-4 mAb 3 (SEQ ID NO:91).

2. The bispecific molecule of claim 1, wherein said le comprises an Fc .

3. The bispecific molecule of claim 2, wherein said Fc Region is a t Fc Region that comprises: (A) one or more amino acid cations that reduces the affinity of the variant Fc Region for an FcγR; and/or (B) one or more amino acid cations that enhances the serum half-life of the variant Fc Region.

4. The bispecific molecule of claim 3, wherein said modifications that reduces the affinity of the variant Fc Region for an FcγR comprise the substitution of L234A; L235A; or L234A and L235A, wherein said numbering is that of the EU index as in Kabat.

5. The bispecific molecule of claim 3 or claim 4, wherein said modifications that that enhances the serum half-life of the variant Fc Region comprise the substitution of M252Y; M252Y and S254T; M252Y and T256E; M252Y, S254T and T256E; or K288D and H435K, wherein said numbering is that of the EU index as in Kabat.

6. The bispecific molecule of any one of claims 1-5, which is an antibody, antibody fragment, singlechain binding molecule, diabody, or bispecific T-cell engaging antibody.

7. The bispecific molecule of claim 1, comprising two polypeptide chains each comprising SEQ ID NO:99 and two polypeptide chains each comprising SEQ ID .

8. The bispecific molecule of claim 1, comprising one polypeptide chain comprising SEQ ID NO:104, one polypeptide chain comprising SEQ ID NO:105, one polypeptide chain comprising SEQ ID NO:106, and one polypeptide chain comprising SEQ ID NO:107.

9. The bispecific molecule of claim 1, comprising two polypeptide chains each comprising SEQ ID NO:101 and two polypeptide chains each comprising SEQ ID .

10. The bispecific molecule of claim 1, comprising one polypeptide chain comprising SEQ ID NO:108, one polypeptide chain comprising SEQ ID NO:105, one polypeptide chain sing SEQ ID , and one polypeptide chain comprising SEQ ID NO:107.

11. A pharmaceutical composition that comprises an effective amount of the bispecific molecule of any one of claims 1-10 and a pharmaceutically acceptable carrier.

12. Use of the bispecific molecule of any one of claims 1-10, or the pharmaceutical composition of claim 11 for the manufacture of a medicament for promoting stimulation of an immune-mediated response or for treating a disease or condition associated with a suppressed immune .

13. The use of claim 12, wherein the disease or condition is cancer or an infection.

14. The use of claim 13, n said cancer is characterized by the presence of a cancer cell selected from the group consisting of a cell of: an adrenal gland tumor, an AIDS-associated cancer, an alveolar soft part sarcoma, an astrocytic tumor, bladder cancer, bone cancer, a brain and spinal cord cancer, a metastatic brain tumor, a breast cancer, a carotid body tumors, a cervical cancer, a chondrosarcoma, a chordoma, a chromophobe renal cell carcinoma, a clear cell carcinoma, a colon cancer, a colorectal cancer, a cutaneous benign fibrous histiocytoma, a desmoplastic small round cell tumor, an ependymoma, a s tumor, an extraskeletal myxoid osarcoma, a fibrogenesis imperfecta ossium, a fibrous dysplasia of the bone, a gallbladder or bile duct cancer, gastric cancer, a ional trophoblastic disease, a germ cell tumor, a head and neck , hepatocellular carcinoma, an islet cell tumor, a Kaposi’s Sarcoma, a kidney cancer, a leukemia, a lipoma/benign lipomatous tumor, a liposarcoma/malignant lipomatous tumor, a liver cancer, a lymphoma, a lung cancer, a medulloblastoma, a melanoma, a meningioma, a multiple endocrine neoplasia, a multiple a, a myelodysplastic syndrome, a neuroblastoma, a neuroendocrine tumors, an ovarian cancer, a pancreatic , a papillary thyroid carcinoma, a parathyroid tumor, a ric , a peripheral nerve sheath tumor, a phaeochromocytoma, a pituitary tumor, a prostate cancer, a posterior uveal melanoma, a rare logic disorder, a renal metastatic cancer, a rhabdoid tumor, a rhabdomyosarcoma, a sarcoma, a skin cancer, a issue sarcoma, a squamous cell cancer, a stomach cancer, a al sarcoma, a testicular cancer, a thymic carcinoma, a thymoma, a thyroid metastatic cancer, and a uterine cancer.

15. The use of claim 13, wherein the cancer is characterized by the presence of a cancer cell selected from the group consisting of a cell of: a colorectal cancer, a lung cancer, a al , a head and neck cancer, a prostate cancer, a sarcoma, and a thymoma.

16. The use of claim 13, wherein the cancer is terized by the ce of a cancer cell selected from the group consisting of a cell of: a colorectal cancer, a hepatocellular carcinoma, a glioma, a kidney cancer, a breast cancer, a multiple myeloma, a bladder cancer, a neuroblastoma, a sarcoma, a non-Hodgkin’s lymphoma, a non-small cell lung cancer, an ovarian cancer, a pancreatic cancer and a rectal cancer.

17. The use of claim 13, n the cancer is characterized by the presence of a cancer cell selected from the group ting of a cell of: a colorectal cancer, a gastric cancer, a melanoma, a prostate cancer, a pancreatic cancer, a renal cancer, a bladder , a mammary cancer, a lung cancer, a fibrosarcoma, a human mantle cell lymphoma, a Raji Burkitt’s lymphoma.

18. The use of claim 13, wherein said infection is characterized by the presence of a bacterial, fungal, viral or protozoan pathogen. K—coil {or E-coii) Polypeptide Chain 1 COOHWC Linker 22 Linker 2 Polypeptide Chain 2 COOH WI: C ------/S