HETERODIMERIC ANTIBODIES INCLUDING BINDING TO CD8

[0001] This application claims priority under 35 U.S.C. §119(e) to U.S. Provisional Patent Application No. 62/084,741, filed Nov. 26, 2014, U.S. Provisional Patent Application No. 62/084,750, filed Nov. 26, 2014, U.S. Provisional Patent Application No. 62/085,003, filed Nov. 26, 2014, U.S. Provisional Patent Application No. 62/084,757, filed Nov. 26, 2014, U.S. Provisional Patent Application No. 62/251,005, filed Nov. 4, 2015 and U.S. Provisional Patent Application No. 62/250,971, filed Nov. 4, 2015, all of which are expressly incorporated herein by reference in their entirety, with particular reference to the figures, legends and claims therein.

[0002] Antibody-based therapeutics have been used successfully to treat a variety of diseases, including cancer and autoimmune/inflammatory disorders. Yet improvements to this class of drugs are still needed, particularly with respect to enhancing their clinical efficacy. One avenue being explored is the engineering of additional and novel antigen binding sites into antibody-based drugs such that a single immunoglobulin molecule co-engages two different antigens. Such non-native or alternate antibody formats that engage two different antigens are often referred to as bispecifics. Because the considerable diversity of the antibody variable region (Fv) makes it possible to produce an Fv that recognizes virtually any molecule, the typical approach to bispecific generation is the introduction of new variable regions into the antibody.

[0003] A number of alternate antibody formats have been explored for bispecific targeting (Chames & Baty, 2009, mAbs 1 [6]:1-9; Holliger & Hudson, 2005, Nature Biotechnology 23[9]:1126-1136; Kontermann, mAbs 4(2):182 (2012), all of which are expressly incorporated herein by reference). Initially, bispecific antibodies were made by fusing two cell lines that each produced a single monoclonal antibody (Milstein et al., 1983, Nature 305:537-540). Although the resulting hybrid hybridoma or quadroma did produce bispecific antibodies, they were only a minor population, and extensive purification was required to isolate the desired antibody. An engineering solution to this was the use of antibody fragments to make bispecifics. Because such fragments lack the complex quaternary structure of a full length antibody, variable light and heavy chains can be linked in single genetic constructs. Antibody fragments of many different forms have been generated, including diabodies, single chain diabodies, tandem scFv's, and Fab₂ bispecifics (Chames & Baty, 2009, mAbs 1[6]:1-9; Holliger & Hudson, 2005, Nature Biotechnology 23[9]:1126-1136; expressly incorporated herein by reference). While these formats can be expressed at high levels in bacteria and may have favorable penetration benefits due to their small size, they clear rapidly in vivo and can present manufacturing obstacles related to their production and stability. A principal cause of these drawbacks is that antibody fragments typically lack the constant region of the antibody with its associated functional properties, including larger size, high stability, and binding to various Fc receptors and ligands that maintain long half-life in serum (i.e. the neonatal Fc receptor FcRn) or serve as binding sites for purification (i.e. protein A and protein G).

[0004] More recent work has attempted to address the shortcomings of fragment-based bispecifics by engineering dual binding into full length antibody-like formats (Wu et al., 2007, Nature Biotechnology 25[11]:1290-1297; U.S. Ser. No. 12/477,711; Michaelson et al., 2009, mAbs 1[2]:128-141; PCT/US2008/074693; Zuo et al., 2000, Protein Engineering 13[5]:361-367; U.S. Ser. No. 09/865,198; Shen et al., 2006, J Biol Chem 281[16]:10706-10714; Lu et al., 2005, J Biol Chem 280[20]:19665-19672; PCT/US2005/025472; expressly incorporated herein by reference). These formats overcome some of the obstacles of the antibody fragment bispecifics, principally because they contain an Fc region. One significant drawback of these formats is that, because they build new antigen binding sites on top of the homodimeric constant chains, binding to the new antigen is always bivalent.

[0005] For many antigens that are attractive as co-targets in a therapeutic bispecific format, the desired binding is monovalent rather than bivalent. For many immune receptors, cellular activation is accomplished by cross-linking of a monovalent binding interaction. The mechanism of cross-linking is typically mediated by antibody/antigen immune complexes, or via effector cell to target cell engagement. For example, the low affinity Fc gamma receptors (FcγRs) such as FcγRIIa, FcγRIIb, and FcγRIIIa bind monovalently to the antibody Fc region. Monovalent binding does not activate cells expressing these FcγRs; however, upon immune complexation or cell-to-cell contact, receptors are cross-linked and clustered on the cell surface, leading to activation. For receptors responsible for mediating cellular killing, for example FcγRIIIa on natural killer (NK) cells, receptor cross-linking and cellular activation occurs when the effector cell engages the target cell in a highly avid format (Bowles & Weiner, 2005, J Immunol Methods 304:88-99, expressly incorporated by reference).. Similarly, on B cells the inhibitory receptor FcγRIIb downregulates B cell activation only when it engages into an immune complex with the cell surface B-cell receptor (BCR), a mechanism that is mediated by immune complexation of soluble IgG's with the same antigen that is recognized by the BCR (Heyman 2003, Immunol Lett 88[2]:157-161; Smith and Clatworthy, 2010, Nature Reviews Immunology 10:328-343; expressly incorporated by reference). As another example, CD3 activation of T-cells occurs only when its associated T-cell receptor (TCR) engages antigen-loaded MHC on antigen presenting cells in a highly avid cell-to-cell synapse (Kuhns et al., 2006, Immunity 24:133-139). Indeed nonspecific bivalent cross-linking of CD3 using an anti-CD3 antibody elicits a cytokine storm and toxicity (Perruche et al., 2009, J Immunol 183[2]:953-61; Chatenoud & Bluestone, 2007, Nature Reviews Immunology 7:622-632; expressly incorporated by reference). Thus for practical clinical use, the preferred mode of CD3 co-engagement for redirected killing of targets cells is monovalent binding that results in activation only upon engagement with the co-engaged target.

[0006] Thus while bispecifics generated from antibody fragments suffer biophysical and pharmacokinetic hurdles, a drawback of those built with full length antibody-like formats is that they engage co-target antigens multivalently in the absence of the primary target antigen, leading to nonspecific activation and potentially toxicity. The present invention solves this problem by introducing a novel set of bispecific formats that enable the multivalent co-engagement of distinct target antigens. In addition, the present invention provides novel heterodimerization variants that allow for better formation and purification of heterodimeric proteins, including antibodies.

[0007] The present invention provides in some embodiments a heterodimeric antibody comprising:

• • [0008] a) a first monomer comprising:
    • [0009] i) a first Fc domain;
    • [0010] ii) an scFv comprising a scFv variable light domain, an scFv linker and a scFv variable heavy domain; wherein said scFv is covalently attached to the N-terminus of said Fc domain using a domain linker;
  • [0011] b) a second monomer comprising a heavy chain comprising:
    • [0012] i) a heavy variable domain; and
    • [0013] ii) a heavy chain constant domain comprising a second Fc domain; and
  • [0014] c) a light chain comprising a variable light domain and a variable light constant domain;
    • [0015] wherein said variable light domain and said variable domain form an antigen binding domain, and wherein one of said scFv and said antigen binding domain bind to CD8 and the other to a target tumor antigen (TTA).

[0016] In some embodiments, the scFv binds to CD8 and the antigen binding domain binds to said TTA.

[0017] In some embodiments, the scFv binds to the TTA and the antigen binding domain binds to CD8.

[0018] In some embodiments, the first and said second Fc domains have a set of amino acid substitutions selected from the group consisting of S364K/E357Q:L368D/K370S; L368D/K370S:S364K; L368E/K370S:S364K; T411T/E360E/Q362E:D401K; L368D/K370S:S364K/E357L and K370S:S364K/E357Q.

[0019] In some embodiments, the CD8 antigen binding domain is selected from the group consisting of OKT8_H1L1, OKT8_H2L1, 51.1_H1L1 and 51.1_H1L2.

[0020] The present invention provides in some embodiments a heterodimeric antibody comprising:

• • [0021] a) a first monomer comprising:
    • [0022] i) a first heavy chain comprising:
      • [0023] 1) a first variable heavy domain;
      • [0024] 2) a first constant heavy domain comprising a first Fc domain; and
      • [0025] 3) a first variable light domain, wherein said first variable light domain is covalently attached to the C-terminus of said first Fc domain using a domain linker;
  • [0026] b) a second monomer comprising:
    • [0027] i) a second variable heavy domain;
    • [0028] ii) a second constant heavy domain comprising a second Fc domain; and
    • [0029] iii) a third variable heavy domain, wherein said second variable heavy domain is covalently attached to the C-terminus of said second Fc domain using a domain linker;
  • [0030] c) a common light chain comprising a variable light domain and a constant light domain;
    • [0031] wherein said first and said second Fc domain have a set of amino acid substitutions selected from the group consisting of S364K/E357Q:L368D/K370S; L368D/K370S:S364K; L368E/K370S:S364K; T411T/E360E/Q362E:D401K; L368D/K370S:S364K/E357L and K370S:S364K/E357Q, wherein said first variable heavy domain and said variable light domain bind a first antigen, said second variable heavy domain and said variable light domain bind said first antigen, and said second variable light domain and said third variable heavy domain binds a second antigen, wherein one of said first and second antigens is CD8 and the other is a TTA.

[0032] In some embodiments, the first antigen is CD8.

[0033] In some embodiments, the second antigen is CD8.

[0034] The present invention provides in some embodiments a heterodimeric antibody comprising:

• • [0035] a) a first monomer comprising:
    • [0036] i) a first heavy chain comprising:
      • [0037] 1) a first variable heavy domain;
      • [0038] 2) a first constant heavy chain comprising a first Fc domain;
      • [0039] 3) a scFv comprising a scFv variable light domain, an scFv linker and a scFv variable heavy domain; wherein said scFv is covalently attached to the C-terminus of said Fc domain using a domain linker;
  • [0040] b) a second monomer comprising a second heavy chain comprising a second variable heavy domain and a second constant heavy chain comprising a second Fc domain; and
  • [0041] c) a common light chain comprising a variable light domain and a constant light domain;
    • [0042] wherein said first and said second Fc domains have a set of amino acid substitutions selected from the group consisting of S364K/E357Q:L368D/K370S; L368D/K370S:S364K; L368E/K370S:S364K; T411T/E360E/Q362E:D401K; L368D/K370S:S364K/E357L and K370S:S364K/E357Q, wherein said first variable heavy domain and said variable light domain bind a first antigen, said second variable heavy domain and said variable light domain bind said first antigen, and said second variable light domain and said third variable heavy domain binds a second antigen, wherein one of said first and second antigens is CD8 and the other is a TTA.

[0043] In some embodiments, the first antigen is CD8.

[0044] In some embodiments, the second antigen is CD8.

[0045] The present invention provides in some embodiments a heterodimeric antibody comprising:

• • [0046] a) a first monomer comprising:
    • [0047] i) a first heavy chain comprising:
      • [0048] 1) a first variable heavy domain;
      • [0049] 2) a first constant heavy chain comprising a first CH1 domain and a first Fc domain;
      • [0050] 3) a scFv comprising a scFv variable light domain, an scFv linker and a scFv variable heavy domain; wherein said scFv is covalently attached between the C-terminus of said CH1 domain and the N-terminus of said first Fc domain using domain linkers;
  • [0051] b) a second monomer comprising a second heavy chain comprising a second variable heavy domain and a second constant heavy chain comprising a second Fc domain; and
  • [0052] c) a common light chain comprising a variable light domain and a constant light domain;
    • [0053] wherein said first and said second Fc domain have a set of amino acid substitutions selected from the group consisting of S364K/E357Q:L368D/K370S; L368D/K370S:S364K; L368E/K370S:S364K; T411T/E360E/Q362E:D401K; L368D/K370S:S364K/E357L and K370S:S364K/E357Q, wherein said first variable heavy domain and said variable light domain bind a first antigen, said second variable heavy domain and said variable light domain bind said first antigen, and said scFv binds a second antigen, wherein one of said first and second antigens is CD8 and the other is a TTA.

[0054] In some embodiments, the first antigen is CD8.

[0055] In some embodiments, the second antigen is CD8.

[0056] The present invention provides in some embodiments a heterodimeric antibody comprising:

• • [0057] a) a first monomer comprising:
    • [0058] i) a first heavy chain comprising:
      • [0059] 1) a first variable heavy domain;
      • [0060] 2) a first constant heavy chain comprising a first CH1 domain and a first Fc domain;
      • [0061] 3) a scFv comprising a scFv variable light domain, an scFv linker and a scFv variable heavy domain; wherein said scFv is covalently attached between the C-terminus of said CH1 domain and the N-terminus of said first Fc domain using domain linkers;
  • [0062] b) a second monomer comprising a second Fc domain; and
  • [0063] c) a light chain comprising a variable light domain and a constant light domain;
    • [0064] wherein said first and said second Fc domain have a set of amino acid substitutions selected from the group consisting of S364K/E357Q:L368D/K370S; L368D/K370S:S364K; L368E/K370S:S364K; T411T/E360E/Q362E:D401K; L368D/K370S:S364K/E357L and K370S:S364K/E357Q, wherein said first variable heavy domain and said variable light domain bind a first antigen, said scFv binds a second antigen, and wherein one of said first and second antigens is CD8 and the other is a TTA.

[0065] In some embodiments, the first antigen is CD8.

[0066] In some embodiments, the second antigen is CD8.

[0067] The present invention provides in some embodiments a heterodimeric antibody comprising:

• • [0068] a) a first monomer comprising:
    • [0069] i) a first heavy chain comprising:
      • [0070] 1) a first variable heavy domain;
      • [0071] 2) a first constant heavy domain comprising a first Fc domain; and
      • [0072] 3) a first variable light domain, wherein said second variable light domain is covalently attached between the C-terminus of the CH1 domain of said first constant heavy domain and the N-terminus of said first Fc domain using domain linkers;
  • [0073] b) a second monomer comprising:
    • [0074] i) a second variable heavy domain;
    • [0075] ii) a second constant heavy domain comprising a second Fc domain; and
    • [0076] iii) a third variable heavy domain, wherein said second variable heavy domain is covalently attached to the C-terminus of said second Fc domain using a domain linker;
  • [0077] c) a common light chain comprising a variable light domain and a constant light domain;
    • [0078] wherein said first and said second Fc domains have a set of amino acid substitutions selected from the group consisting of S364K/E357Q:L368D/K370S; L368D/K370S:S364K; L368E/K370S:S364K; T411T/E360E/Q362E:D401K; L368D/K370S:S364K/E357L and K370S:S364K/E357Q, wherein said first variable heavy domain and said variable light domain bind a first antigen, said second variable heavy domain and said variable light domain bind said first antigen, and said second variable light domain and said third variable heavy domain binds a second antigen;
    • [0079] wherein one of said first and second antigens is CD8 and the other is a TTA.

[0080] In some embodiments, the TTA is selected from the group consisting of CD19, CD20, CD38 and CD123.

[0081] In some embodiments, the first antigen is CD8.

[0082] In some embodiments, the second antigen is CD8.

[0083] The present invention provides in some embodiments a nucleic acid composition comprising:

• • [0084] a) a first nucleic acid encoding the first monomer as described above and herein, respectively;
  • [0085] b) a second nucleic acid encoding said first monomer as described above and herein, respectively;
  • [0086] c) a third nucleic acid encoding said first monomer as described above and herein, respectively.

[0087] The present invention provides in some embodiments an expression vector composition comprising:

• • [0088] a) a first expression vector comprising said first nucleic as described above and herein, respectively;
  • [0089] b) a second expression vector comprising said second nucleic acid as described above and herein, respectively;
  • [0090] c) a third expression vector comprising said second nucleic acid as described above and herein, respectively.

[0091] In some embodiments, the host cell comprises the nucleic acid composition as described above and herein or the expression vector composition as described above and herein.

[0092] The present invention provides in some embodiments a method of making a heterodimeric antibody as described above and herein, the method comprising culturing said cells under conditions wherein said heterodimeric antibody is produced and recovering said antibody.

[0093] The present invention provides in some embodiments a method of treating comprising administering a heterodimeric antibody as described above and herein.

[0094] The present invention provides in some embodiments a bispecific antibody comprising:

• • [0095] a) a heavy chain comprising:
    • [0096] i) a heavy constant domain comprising an Fc domain;
    • [0097] ii) a heavy chain variable domain; and
    • [0098] iii) a scFv;
  • [0099] b) a light chain comprising:
    • [0100] i) a light constant domain comprising an Fc domain;
    • [0101] ii) a light chain variable domain; and
    • [0102] iii) a scFv; and
    • [0103] wherein said heavy and light variable domains form an antigen binding domain, and wherein one of said antigen binding domain and said scFv binds to CD8 and the other binds to a target tumor antigen.

[0104] In some embodiments, the scFv comprises a charged scFv linker.

[0105] In some embodiments, the scFv is covalently attached at the C-terminus of said heavy chain using a domain linker.

[0106] In some embodiments, the scFv is covalently attached at the N-terminus of said heavy chain using a domain linker.

[0107] In some embodiments, the scFv is covalently attached between said Fc domain and said heavy chain variable region using a domain linker at each end.

[0108] The present invention provides in some embodiments a nucleic acid composition comprising:

• • [0109] a) a first nucleic acid encoding a heavy chain of an antibody as described above and herein; and
  • [0110] b) a second nucleic acid encoding a light chain of an antibody as described above and herein; respectively.

[0111] The present invention provides in some embodiments an expression vector composition comprising:

• • [0112] a) a first expression vector comprising the first nucleic acid as described above and herein;
  • [0113] b) a second expression vector comprising the second nucleic acid as described above and herein; and
  • [0114] c) a third expression vector comprising the third nucleic acid as described above and herein.

[0115] In some embodiments, the host cell comprises the nucleic acid composition as described above and herein.

[0116] In some embodiments, the host cell comprises the expression vector composition as described above and herein.

[0117] The present invention provides in some embodiments a method of making the bispecific antibody as described above and herein comprising culturing said host cell under conditions wherein the bispecific antibody is made and recovering said antibody.

[0118] The present invention provides in some embodiments a method of treating a patient in need thereof by administering a bispecific antibody as described above and herein.

[0165] In order that the application may be more completely understood, several definitions are set forth below. Such definitions are meant to encompass grammatical equivalents.

[0166] By “ablation” herein is meant a decrease or removal of activity. Thus for example, “ablating FcγR binding” means the Fc region amino acid variant has less than 50% starting binding as compared to an Fc region not containing the specific variant, with less than 70-80-90-95-98% loss of activity being preferred, and in general, with the activity being below the level of detectable binding in a Biacore assay. Of particular use in the ablation of FcγR binding are those shown in FIG. 19.

[0167] By “ADCC” or “antibody dependent cell-mediated cytotoxicity” as used herein is meant the cell-mediated reaction wherein nonspecific cytotoxic cells that express FcγRs recognize bound antibody on a target cell and subsequently cause lysis of the target cell. ADCC is correlated with binding to FcγRIIIa; increased binding to FcγRIIIa leads to an increase in ADCC activity.

[0168] By “ADCP” or antibody dependent cell-mediated phagocytosis as used herein is meant the cell-mediated reaction wherein nonspecific cytotoxic cells that express FcγRs recognize bound antibody on a target cell and subsequently cause phagocytosis of the target cell.

[0169] By “modification” herein is meant an amino acid substitution, insertion, and/or deletion in a polypeptide sequence or an alteration to a moiety chemically linked to a protein. For example, a modification may be an altered carbohydrate or PEG structure attached to a protein. By “amino acid modification” herein is meant an amino acid substitution, insertion, and/or deletion in a polypeptide sequence. For clarity, unless otherwise noted, the amino acid modification is always to an amino acid coded for by DNA, e.g. the 20 amino acids that have codons in DNA and RNA.

[0170] By “amino acid substitution” or “substitution” herein is meant the replacement of an amino acid at a particular position in a parent polypeptide sequence with a different amino acid. In particular, in some embodiments, the substitution is to an amino acid that is not naturally occurring at the particular position, either not naturally occurring within the organism or in any organism. For example, the substitution E272Y refers to a variant polypeptide, in this case an Fc variant, in which the glutamic acid at position 272 is replaced with tyrosine. For clarity, a protein which has been engineered to change the nucleic acid coding sequence but not change the starting amino acid (for example exchanging CGG (encoding arginine) to CGA (still encoding arginine) to increase host organism expression levels) is not an “amino acid substitution”; that is, despite the creation of a new gene encoding the same protein, if the protein has the same amino acid at the particular position that it started with, it is not an amino acid substitution.

[0171] By “amino acid insertion” or “insertion” as used herein is meant the addition of an amino acid sequence at a particular position in a parent polypeptide sequence. For example, −233E or 233E designates an insertion of glutamic acid after position 233 and before position 234. Additionally, −233ADE or A233ADE designates an insertion of AlaAspGlu after position 233 and before position 234.

[0172] By “amino acid deletion” or “deletion” as used herein is meant the removal of an amino acid sequence at a particular position in a parent polypeptide sequence. For example, E233- or E233# or E233( ) designates a deletion of glutamic acid at position 233. Additionally, EDA233- or EDA233# designates a deletion of the sequence GluAspAla that begins at position 233.

[0173] By “variant protein” or “protein variant”, or “variant” as used herein is meant a protein that differs from that of a parent protein by virtue of at least one amino acid modification. Protein variant may refer to the protein itself, a composition comprising the protein, or the amino sequence that encodes it. Preferably, the protein variant has at least one amino acid modification compared to the parent protein, e.g. from about one to about seventy amino acid modifications, and preferably from about one to about five amino acid modifications compared to the parent. As described below, in some embodiments the parent polypeptide, for example an Fc parent polypeptide, is a human wild type sequence, such as the Fc region from IgG1, IgG2, IgG3 or IgG4, although human sequences with variants can also serve as “parent polypeptides”, for example the IgG1/2 hybrid of FIG. 19. The protein variant sequence herein will preferably possess at least about 80% identity with a parent protein sequence, and most preferably at least about 90% identity, more preferably at least about 95-98-99% identity. Variant protein can refer to the variant protein itself, compositions comprising the protein variant, or the DNA sequence that encodes it. Accordingly, by “antibody variant” or “variant antibody” as used herein is meant an antibody that differs from a parent antibody by virtue of at least one amino acid modification, “IgG variant” or “variant IgG” as used herein is meant an antibody that differs from a parent IgG (again, in many cases, from a human IgG sequence) by virtue of at least one amino acid modification, and “immunoglobulin variant” or “variant immunoglobulin” as used herein is meant an immunoglobulin sequence that differs from that of a parent immunoglobulin sequence by virtue of at least one amino acid modification. “Fc variant” or “variant Fc” as used herein is meant a protein comprising an amino acid modification in an Fc domain. The Fc variants of the present invention are defined according to the amino acid modifications that compose them. Thus, for example, N434S or 434S is an Fc variant with the substitution serine at position 434 relative to the parent Fc polypeptide, wherein the numbering is according to the EU index. Likewise, M428L/N434S defines an Fc variant with the substitutions M428L and N434S relative to the parent Fc polypeptide. The identity of the WT amino acid may be unspecified, in which case the aforementioned variant is referred to as 428L/434S. It is noted that the order in which substitutions are provided is arbitrary, that is to say that, for example, 428L/434S is the same Fc variant as M428L/N434S, and so on. For all positions discussed in the present invention that relate to antibodies, unless otherwise noted, amino acid position numbering is according to the EU index. The EU index or EU index as in Kabat or EU numbering scheme refers to the numbering of the EU antibody (Edelman et al., 1969, Proc Natl Acad Sci USA 63:78-85, hereby entirely incorporated by reference.) The modification can be an addition, deletion, or substitution. Substitutions can include naturally occurring amino acids and, in some cases, synthetic amino acids. Examples include U.S. Pat. No. 6,586,207; WO 98/48032; WO 03/073238; US2004-0214988A1; WO 05/35727A2; WO 05/74524A2; J. W. Chin et al., (2002), Journal of the American Chemical Society 124:9026-9027; J. W. Chin, & P. G. Schultz, (2002), ChemBioChem 11:1135-1137; J. W. Chin, et al., (2002), PICAS United States of America 99:11020-11024; and, L. Wang, & P. G. Schultz, (2002), Chem. 1-10, all entirely incorporated by reference.

[0174] As used herein, “protein” herein is meant at least two covalently attached amino acids, which includes proteins, polypeptides, oligopeptides and peptides. The peptidyl group may comprise naturally occurring amino acids and peptide bonds, or synthetic peptidomimetic structures, i.e. “analogs”, such as peptoids (see Simon et al., PNAS USA 89(20):9367 (1992), entirely incorporated by reference). The amino acids may either be naturally occurring or synthetic (e.g. not an amino acid that is coded for by DNA); as will be appreciated by those in the art. For example, homo-phenylalanine, citrulline, ornithine and noreleucine are considered synthetic amino acids for the purposes of the invention, and both D- and L-(R or S) configured amino acids may be utilized. The variants of the present invention may comprise modifications that include the use of synthetic amino acids incorporated using, for example, the technologies developed by Schultz and colleagues, including but not limited to methods described by Cropp & Shultz, 2004, Trends Genet. 20(12):625-30, Anderson et al., 2004, Proc Natl Acad Sci USA 101 (2):7566-71, Zhang et al., 2003, 303(5656):371-3, and Chin et al., 2003, Science 301(5635):964-7, all entirely incorporated by reference. In addition, polypeptides may include synthetic derivatization of one or more side chains or termini, glycosylation, PEGylation, circular permutation, cyclization, linkers to other molecules, fusion to proteins or protein domains, and addition of peptide tags or labels.

[0175] By “residue” as used herein is meant a position in a protein and its associated amino acid identity. For example, Asparagine 297 (also referred to as Asn297 or N297) is a residue at position 297 in the human antibody IgG1.

[0176] By “Fab” or “Fab region” as used herein is meant the polypeptide that comprises the VH, CH1, VL, and CL immunoglobulin domains. Fab may refer to this region in isolation, or this region in the context of a full length antibody, antibody fragment or Fab fusion protein. By “Fv” or “Fv fragment” or “Fv region” as used herein is meant a polypeptide that comprises the VL and VH domains of a single antibody. As will be appreciated by those in the art, these generally are made up of two chains.

[0177] By “IgG subclass modification” or “isotype modification” as used herein is meant an amino acid modification that converts one amino acid of one IgG isotype to the corresponding amino acid in a different, aligned IgG isotype. For example, because IgG1 comprises a tyrosine and IgG2 a phenylalanine at EU position 296, a F296Y substitution in IgG2 is considered an IgG subclass modification.

[0178] By “non-naturally occurring modification” as used herein is meant an amino acid modification that is not isotypic. For example, because none of the IgGs comprise a serine at position 434, the substitution 434S in IgG1, IgG2, IgG3, or IgG4 (or hybrids thereof) is considered a non-naturally occurring modification.

[0179] By “amino acid” and “amino acid identity” as used herein is meant one of the 20 naturally occurring amino acids that are coded for by DNA and RNA.

[0180] By “effector function” as used herein is meant a biochemical event that results from the interaction of an antibody Fc region with an Fc receptor or ligand. Effector functions include but are not limited to ADCC, ADCP, and CDC.

[0181] By “IgG Fc ligand” as used herein is meant a molecule, preferably a polypeptide, from any organism that binds to the Fc region of an IgG antibody to form an Fc/Fc ligand complex. Fc ligands include but are not limited to FcγRIs, FcγRIIs, FcγRIIIs, FcRn, C1q, C3, mannan binding lectin, mannose receptor, staphylococcal protein A, streptococcal protein G, and viral FcγR. Fc ligands also include Fc receptor homologs (FcRH), which are a family of Fc receptors that are homologous to the FcγRs (Davis et al., 2002, Immunological Reviews 190:123-136, entirely incorporated by reference). Fc ligands may include undiscovered molecules that bind Fc. Particular IgG Fc ligands are FcRn and Fc gamma receptors. By “Fc ligand” as used herein is meant a molecule, preferably a polypeptide, from any organism that binds to the Fc region of an antibody to form an Fc/Fc ligand complex.

[0182] By “Fc gamma receptor”, “FcγR” or “FcqammaR” as used herein is meant any member of the family of proteins that bind the IgG antibody Fc region and is encoded by an FcγR gene. In humans this family includes but is not limited to FcγRI (CD64), including isoforms FcγRIa, FcγRIb, and FcγRIc; FcγRII (CD32), including isoforms FcγRIIa (including allotypes H131 and R131), FcγRIIb (including FcγRIIb-1 and FcγRIIb-2), and FcγRIIc; and FcγRIII (CD16), including isoforms FcγRIIIa (including allotypes V158 and F158) and FcγRIIIb (including allotypes FcγRIIb-NA1 and FcγRIIb-NA2) (Jefferis et al., 2002, Immunol Lett 82:57-65, entirely incorporated by reference), as well as any undiscovered human FcγRs or FcγR isoforms or allotypes. An FcγR may be from any organism, including but not limited to humans, mice, rats, rabbits, and monkeys. Mouse FcγRs include but are not limited to FcγRI (CD64), FcγRII (CD32), FcγRIII (CD16), and FcγRIII-2 (CD16-2), as well as any undiscovered mouse FcγRs or FcγR isoforms or allotypes.

[0183] By “FcRn” or “neonatal Fc Receptor” as used herein is meant a protein that binds the IgG antibody Fc region and is encoded at least in part by an FcRn gene. The FcRn may be from any organism, including but not limited to humans, mice, rats, rabbits, and monkeys. As is known in the art, the functional FcRn protein comprises two polypeptides, often referred to as the heavy chain and light chain. The light chain is beta-2-microglobulin and the heavy chain is encoded by the FcRn gene. Unless otherwise noted herein, FcRn or an FcRn protein refers to the complex of FcRn heavy chain with beta-2-microglobulin. A variety of FcRn variants used to increase binding to the FcRn receptor, and in some cases, to increase serum half-life, are shown in the Figure Legend of FIG. 83.

[0184] By “parent polypeptide” as used herein is meant a starting polypeptide that is subsequently modified to generate a variant. The parent polypeptide may be a naturally occurring polypeptide, or a variant or engineered version of a naturally occurring polypeptide. Parent polypeptide may refer to the polypeptide itself, compositions that comprise the parent polypeptide, or the amino acid sequence that encodes it. Accordingly, by “parent immunoglobulin” as used herein is meant an unmodified immunoglobulin polypeptide that is modified to generate a variant, and by “parent antibody” as used herein is meant an unmodified antibody that is modified to generate a variant antibody. It should be noted that “parent antibody” includes known commercial, recombinantly produced antibodies as outlined below.

[0185] By “Fc” or “Fc region” or “Fc domain” as used herein is meant the polypeptide comprising the constant region of an antibody excluding the first constant region immunoglobulin domain and in some cases, part of the hinge. Thus Fc refers to the last two constant region immunoglobulin domains of IgA, IgD, and IgG, the last three constant region immunoglobulin domains of IgE and IgM, and the flexible hinge N-terminal to these domains. For IgA and IgM, Fc may include the J chain. For IgG, the Fc domain comprises immunoglobulin domains Cγ2 and Cγ3 (Cγ2 and Cγ3) and the lower hinge region between Cγ1 (Cγ1) and Cγ2 (Cγ2). Although the boundaries of the Fc region may vary, the human IgG heavy chain Fc region is usually defined to include residues C226 or P230 to its carboxyl-terminus, wherein the numbering is according to the EU index as in Kabat. In some embodiments, as is more fully described below, amino acid modifications are made to the Fc region, for example to alter binding to one or more FcγR receptors or to the FcRn receptor.

[0186] By “heavy constant region” herein is meant the CH1-hinge-CH2-CH3 portion of an antibody.

[0187] By “Fc fusion protein” or “immunoadhesin” herein is meant a protein comprising an Fc region, generally linked (optionally through a linker moiety, as described herein) to a different protein, such as a binding moiety to a target protein, as described herein. In some cases, one monomer of the heterodimeric antibody comprises an antibody heavy chain (either including an scFv or further including a light chain) and the other monomer is a Fc fusion, comprising a variant Fc domain and a ligand. In some embodiments, these “half antibody-half fusion proteins” are referred to as “Fusionbodies”.

[0188] By “position” as used herein is meant a location in the sequence of a protein. Positions may be numbered sequentially, or according to an established format, for example the EU index for antibody numbering.

[0189] By “target antigen” as used herein is meant the molecule that is bound specifically by the variable region of a given antibody. A target antigen may be a protein, carbohydrate, lipid, or other chemical compound. A wide number of suitable target antigens are described below.

[0190] By “strandedness” in the context of the monomers of the heterodimeric antibodies of the invention herein is meant that, similar to the two strands of DNA that “match”, heterodimerization variants are incorporated into each monomer so as to preserve the ability to “match” to form heterodimers. For example, if some pI variants are engineered into monomer A (e.g. making the pI higher) then steric variants that are “charge pairs” that can be utilized as well do not interfere with the pI variants, e.g. the charge variants that make a pI higher are put on the same “strand” or “monomer” to preserve both functionalities. Similarly, for “skew” variants that come in pairs of a set as more fully outlined below, the skilled artisan will consider pI in deciding into which strand or monomer that incorporates one set of the pair will go, such that pI separation is maximized using the pI of the skews as well.

[0191] By “target cell” as used herein is meant a cell that expresses a target antigen.

[0192] By “variable region” as used herein is meant the region of an immunoglobulin that comprises one or more Ig domains substantially encoded by any of the V.kappa., V.lambda., and/or VH genes that make up the kappa, lambda, and heavy chain immunoglobulin genetic loci respectively.

[0193] By “wild type or WT” herein is meant an amino acid sequence or a nucleotide sequence that is found in nature, including allelic variations. A WT protein has an amino acid sequence or a nucleotide sequence that has not been intentionally modified.

[0194] The antibodies of the present invention are generally isolated or recombinant. “Isolated,” when used to describe the various polypeptides disclosed herein, means a polypeptide that has been identified and separated and/or recovered from a cell or cell culture from which it was expressed. Ordinarily, an isolated polypeptide will be prepared by at least one purification step. An “isolated antibody,” refers to an antibody which is substantially free of other antibodies having different antigenic specificities. “Recombinant” means the antibodies are generated using recombinant nucleic acid techniques in exogeneous host cells.

[0195] “Specific binding” or “specifically binds to” or is “specific for” a particular antigen or an epitope means binding that is measurably different from a non-specific interaction. Specific binding can be measured, for example, by determining binding of a molecule compared to binding of a control molecule, which generally is a molecule of similar structure that does not have binding activity. For example, specific binding can be determined by competition with a control molecule that is similar to the target.

[0196] Specific binding for a particular antigen or an epitope can be exhibited, for example, by an antibody having a KD for an antigen or epitope of at least about 10-4 M, at least about 10-5 M, at least about 10-6 M, at least about 10-7 M, at least about 10-8 M, at least about 10-9 M, alternatively at least about 10-10 M, at least about 10-11 M, at least about 10-12 M, or greater, where KD refers to a dissociation rate of a particular antibody-antigen interaction. Typically, an antibody that specifically binds an antigen will have a KD that is 20-, 50-, 100-, 500-, 1000-, 5,000-, 10,000- or more times greater for a control molecule relative to the antigen or epitope.

[0197] Also, specific binding for a particular antigen or an epitope can be exhibited, for example, by an antibody having a KA or Ka for an antigen or epitope of at least 20-, 50-, 100-, 500-, 1000-, 5,000-, 10,000- or more times greater for the epitope relative to a control, where KA or Ka refers to an association rate of a particular antibody-antigen interaction.

[0198] Bispecific antibodies that co-engage CD3 and a tumor antigen target have been designed and used to redirect T cells to attack and lyse targeted tumor cells. Examples include the BiTE and DART formats, which monovalently engage CD3 and a tumor antigen. While the CD3-targeting approach has shown considerable promise, a common side effect of such therapies is the associated production of cytokines, often leading to toxic cytokine release syndrome. Because the anti-CD3 binding domain of the bispecific antibody engages all T cells, the high cytokine-producing CD4 T cell subset is recruited. Moreover, the CD4 T cell subset includes regulatory T cells, whose recruitment and expansion can potentially lead to immune suppression and have a negative impact on long-term tumor suppression. In addition, these formats do not contain Fc domains and show very short serum half-lives in patients.

[0199] The present invention is directed to a solution of these issues, by using more selective T cell targets rather than the pan-T cell activator CD3. We demonstrate herein that bispecific antibodies designed to selectively recruit the CD8 T cell subset can target and kill tumor cells effectively. Selective CD8 recruitment leads to significantly reduced cytokine release, expanding the therapeutic window for T cell recruitment.

[0200] Accordingly, the invention provides a variety of bispecific, multivalent antibodies, using anti-CD8 as one of the antigens for Fv binding, and a target tumor antigen (TTA).

[0201] In some embodiments, the co-engagement of CD8 and the tumor target antigen can be either monovalent or bivalent, or mixed, as is generally depicted in the Figures. In some embodiments, bivalent engagement of CD8 on the T cells is used for promotion of effective and potent CD8-mediated killing.

[0202] The bispecific antibodies of the invention can also be used to target peptide/MHC complexes. For example, antibodies have been reported to act as T cell receptor (TCR) mimetics. Such antibodies have been directed against peptide MHC complexes such as NY-eso-1/HLA-A2. Other peptides include additional cancer testes antigens such as MAGE-A1 and MAGE-A3, as well as other tumor-selective peptides such as gp100 and p53. In a further embodiment, the Fab component of the bispecific antibodies can be replaced by the TCR alpha/beta chains recognizing such peptide/MHC complexes. In a preferred embodiment, the bispecific antibodies target peptides complexes with HLA-A2.

[0203] In addition, as outlined herein, the present invention can be used in the context of traditional homodimeric antibodies or heterodimeric antibodies.

[0204] The present invention relates to the generation of bispecific antibodies that bind two different antigens, e.g. CD8 and a target tumor antigen (TTA) such as CD19, CD20, CD38 and CD123, and are generally therapeutic antibodies. As is discussed below, the term “antibody” is used generally. Antibodies that find use in the present invention can take on a number of formats as described herein, including traditional antibodies as well as antibody derivatives, fragments and mimetics, described herein.

[0205] Traditional antibody structural units typically comprise a tetramer. Each tetramer is typically composed of two identical pairs of polypeptide chains, each pair having one “light” (typically having a molecular weight of about 25 kDa) and one “heavy” chain (typically having a molecular weight of about 50-70 kDa). Human light chains are classified as kappa and lambda light chains. The present invention is directed to the IgG class, which has several subclasses, including, but not limited to IgG1, IgG2, IgG3, and IgG4. Thus, “isotype” as used herein is meant any of the subclasses of immunoglobulins defined by the chemical and antigenic characteristics of their constant regions. It should be understood that therapeutic antibodies can also comprise hybrids of isotypes and/or subclasses. For example, as shown in US Publication 2009/0163699, incorporated by reference, the present invention covers pI engineering of IgG1/G2 hybrids.

[0206] The amino-terminal portion of each chain includes a variable region of about 100 to 110 or more amino acids primarily responsible for antigen recognition, generally referred to in the art and herein as the “Fv domain” or “Fv region”. In the variable region, three loops are gathered for each of the V domains of the heavy chain and light chain to form an antigen-binding site. Each of the loops is referred to as a complementarity-determining region (hereinafter referred to as a “CDR”), in which the variation in the amino acid sequence is most significant. “Variable” refers to the fact that certain segments of the variable region differ extensively in sequence among antibodies. Variability within the variable region is not evenly distributed. Instead, the V regions consist of relatively invariant stretches called framework regions (FRs) of 15-30 amino acids separated by shorter regions of extreme variability called “hypervariable regions” that are each 9-15 amino acids long or longer.

[0207] Each VH and VL is composed of three hypervariable regions (“complementary determining regions,” “CDRs”) and four FRs, arranged from amino-terminus to carboxy-terminus in the following order: FR1-CDR1-FR2-CDR2-FR3-CDR3-FR4.

[0208] The hypervariable region generally encompasses amino acid residues from about amino acid residues 24-34 (LCDR1; “L” denotes light chain), 50-56 (LCDR2) and 89-97 (LCDR3) in the light chain variable region and around about 31-35B (HCDR1; “H” denotes heavy chain), 50-65 (HCDR2), and 95-102 (HCDR3) in the heavy chain variable region; Kabat et al., SEQUENCES OF PROTEINS OF IMMUNOLOGICAL INTEREST, 5th Ed. Public Health Service, National Institutes of Health, Bethesda, Md. (1991) and/or those residues forming a hypervariable loop (e.g. residues 26-32 (LCDR1), 50-52 (LCDR2) and 91-96 (LCDR3) in the light chain variable region and 26-32 (HCDR1), 53-55 (HCDR2) and 96-101 (HCDR3) in the heavy chain variable region; Chothia and Lesk (1987) J. Mol. Biol. 196:901-917. Specific CDRs of the invention are described below.

[0209] Throughout the present specification, the Kabat numbering system is generally used when referring to a residue in the variable domain (approximately, residues 1-107 of the light chain variable region and residues 1-113 of the heavy chain variable region) and the EU numbering system for Fc regions (e.g, Kabat et al., supra (1991)).

[0210] The present invention provides a large number of different CDR sets. In this case, a “full CDR set” comprises the three variable light and three variable heavy CDRs, e.g. a vlCDR1, vlCDR2, vlCDR3, vhCDR1, vhCDR2 and vhCDR3. These can be part of a larger variable light or variable heavy domain, respectfully. In addition, as more fully outlined herein, the variable heavy and variable light domains can be on separate polypeptide chains, when a heavy and light chain is used (for example when Fabs are used), or on a single polypeptide chain the case of scFv sequences.

[0211] The CDRs contribute to the formation of the antigen-binding, or more specifically, epitope binding site of antibodies. “Epitope” refers to a determinant that interacts with a specific antigen binding site in the variable region of an antibody molecule known as a paratope. Epitopes are groupings of molecules such as amino acids or sugar side chains and usually have specific structural characteristics, as well as specific charge characteristics. A single antigen may have more than one epitope.

[0212] The epitope may comprise amino acid residues directly involved in the binding (also called immunodominant component of the epitope) and other amino acid residues, which are not directly involved in the binding, such as amino acid residues which are effectively blocked by the specifically antigen binding peptide; in other words, the amino acid residue is within the footprint of the specifically antigen binding peptide.

[0213] Epitopes may be either conformational or linear. A conformational epitope is produced by spatially juxtaposed amino acids from different segments of the linear polypeptide chain. A linear epitope is one produced by adjacent amino acid residues in a polypeptide chain. Conformational and nonconformational epitopes may be distinguished in that the binding to the former but not the latter is lost in the presence of denaturing solvents.

[0214] An epitope typically includes at least 3, and more usually, at least 5 or 8-10 amino acids in a unique spatial conformation. Antibodies that recognize the same epitope can be verified in a simple immunoassay showing the ability of one antibody to block the binding of another antibody to a target antigen, for example “binning.”

[0215] The carboxy-terminal portion of each chain defines a constant region primarily responsible for effector function. Kabat et al. collected numerous primary sequences of the variable regions of heavy chains and light chains. Based on the degree of conservation of the sequences, they classified individual primary sequences into the CDR and the framework and made a list thereof (see SEQUENCES OF IMMUNOLOGICAL INTEREST, 5th edition, NIH publication, No. 91-3242, E. A. Kabat et al., entirely incorporated by reference).

[0216] In the IgG subclass of immunoglobulins, there are several immunoglobulin domains in the heavy chain. By “immunoglobulin (Ig) domain” herein is meant a region of an immunoglobulin having a distinct tertiary structure. Of interest in the present invention are the heavy chain domains, including, the constant heavy (CH) domains and the hinge domains. In the context of IgG antibodies, the IgG isotypes each have three CH regions. Accordingly, “CH” domains in the context of IgG are as follows: “CH1” refers to positions 118-220 according to the EU index as in Kabat. “CH2” refers to positions 237-340 according to the EU index as in Kabat, and “CH3” refers to positions 341-447 according to the EU index as in Kabat. As shown herein and described below, the pI variants can be in one or more of the CH regions, as well as the hinge region, discussed below.

[0217] It should be noted that the sequences depicted herein start at the CH1 region, position 118; the variable regions are not included except as noted.

[0218] Another type of Ig domain of the heavy chain is the hinge region. By “hinge” or “hinge region” or “antibody hinge region” or “immunoglobulin hinge region” herein is meant the flexible polypeptide comprising the amino acids between the first and second constant domains of an antibody. Structurally, the IgG CH1 domain ends at EU position 220, and the IgG CH2 domain begins at residue EU position 237. Thus for IgG the antibody hinge is herein defined to include positions 221 (D221 in IgG1) to 236 (G236 in IgG1), wherein the numbering is according to the EU index as in Kabat. In some embodiments, for example in the context of an Fc region, the lower hinge is included, with the “lower hinge” generally referring to positions 226 or 230. As noted herein, pI variants can be made in the hinge region as well.

[0219] The light chain generally comprises two domains, the variable light domain (containing the light chain CDRs and together with the variable heavy domains forming the Fv region), and a constant light chain region (often referred to as CL or Cκ.

[0220] Another region of interest for additional substitutions, outlined below, is the Fc region.

[0221] Thus, the present invention provides different antibody domains. As described herein and known in the art, the heterodimeric antibodies of the invention comprise different domains within the heavy and light chains, which can be overlapping as well. These domains include, but are not limited to, the Fc domain, the CH1 domain, the CH2 domain, the CH3 domain, the hinge domain, the heavy constant domain (CH1-hinge-Fc domain or CH1-hinge-CH2-CH3), the variable heavy domain, the variable light domain, the light constant domain, FAb domains and scFv domains.

[0222] Thus, the “Fc domain” includes the —CH2-CH3 domain, and optionally a hinge domain. The heavy chain comprises a variable heavy domain and a constant domain, which includes a CH1-optional hinge-Fc domain comprising a CH2-CH3. The light chain comprises a variable light chain and the light constant domain.

[0223] Some embodiments of the invention comprise at least one scFv domain, which, while not naturally occurring, generally includes a variable heavy domain and a variable light domain, linked together by a scFv linker. As shown herein, there are a number of suitable scFv linkers that can be used, including traditional peptide bonds, generated by recombinant techniques.

[0224] The linker peptide may predominantly include the following amino acid residues: Gly, Ser, Ala, or Thr. The linker peptide should have a length that is adequate to link two molecules in such a way that they assume the correct conformation relative to one another so that they retain the desired activity. In one embodiment, the linker is from about 1 to 50 amino acids in length, preferably about 1 to 30 amino acids in length. In one embodiment, linkers of 1 to 20 amino acids in length may be used, with from about 5 to about 10 amino acids finding use in some embodiments. Useful linkers include glycine-serine polymers, including for example (GS)n, (GSGGS)n (SEQ ID NO: 238), (GGGGS)n (SEQ ID NO: 239), and (GGGS)n (SEQ ID NO: 240), where n is an integer of at least one (and generally from 3 to 4), glycine-alanine polymers, alanine-serine polymers, and other flexible linkers. Alternatively, a variety of nonproteinaceous polymers, including but not limited to polyethylene glycol (PEG), polypropylene glycol, polyoxyalkylenes, or copolymers of polyethylene glycol and polypropylene glycol, may find use as linkers, that is may find use as linkers.

[0225] Other linker sequences may include any sequence of any length of CL/CH1 domain but not all residues of CL/CH1 domain; for example the first 5-12 amino acid residues of the CL/CH1 domains. Linkers can be derived from immunoglobulin light chain, for example Cκ or Cλ. Linkers can be derived from immunoglobulin heavy chains of any isotype, including for example Cγ1, Cγ2, Cγ3, Cγ4, Cal, Cα2, Cδ, Cε, and Cμ. Linker sequences may also be derived from other proteins such as Ig-like proteins (e.g. TCR, FcR, KIR), hinge region-derived sequences, and other natural sequences from other proteins.

[0226] In some embodiments, the linker is a “domain linker”, used to link any two domains as outlined herein together. While any suitable linker can be used, many embodiments utilize a glycine-serine polymer, including for example (GS)n, (GSGGS)n (SEQ ID NO: 238), (GGGGS)n (SEQ ID NO: 239), and (GGGS)n (SEQ ID NO: 240), where n is an integer of at least one (and generally from 3 to 4 to 5) as well as any peptide sequence that allows for recombinant attachment of the two domains with sufficient length and flexibility to allow each domain to retain its biological function. In some cases, and with attention being paid to “strandedness”, as outlined below, charged domain linkers, as used in some embodiments of scFv linkers can be used.

[0227] In some embodiments, the scFv linker is a charged scFv linker, a number of which are shown in the Figures. Accordingly, the present invention further provides charged scFv linkers, to facilitate the separation in pI between a first and a second monomer. That is, by incorporating a charged scFv linker, either positive or negative (or both, in the case of scaffolds that use scFvs on different monomers), this allows the monomer comprising the charged linker to alter the pI without making further changes in the Fc domains. These charged linkers can be substituted into any scFv containing standard linkers. Again, as will be appreciated by those in the art, charged scFv linkers are used on the correct “strand” or monomer, according to the desired changes in pI. For example, as discussed herein, to make triple F format heterodimeric antibody, the original pI of the Fv region for each of the desired antigen binding domains are calculated, and one is chosen to make an scFv, and depending on the pI, either positive or negative linkers are chosen.

[0228] Charged domain linkers can also be used to increase the pI separation of the monomers of the invention as well, and thus those included in the figures an be used in any embodiment herein where a linker is utilized.

[0229] In some embodiments, the antibodies are full length. By “full length antibody” herein is meant the structure that constitutes the natural biological form of an antibody, including variable and constant regions, including one or more modifications as outlined herein, particularly in the Fc domains to allow either heterodimerization formation or the purification of heterodimers away from homodimers. Full length antibodies generally include Fab and Fc domains, and can additionally contain extra antigen binding domains such as scFvs, as is generally depicted in the Figures.

[0230] In one embodiment, the antibody is an antibody fragment, as long as it contains at least one constant domain which can be engineered to produce heterodimers, such as pI engineering. Other antibody fragments that can be used include fragments that contain one or more of the CH1, CH2, CH3, hinge and CL domains of the invention that have been pI engineered. For example, Fc fusions are fusions of the Fc region (CH2 and CH3, optionally with the hinge region) fused to another protein. A number of Fc fusions are known the art and can be improved by the addition of the heterodimerization variants of the invention. In the present case, antibody fusions can be made comprising CH1; CH1, CH2 and CH3; CH2; CH3; CH2 and CH3; CH1 and CH3, any or all of which can be made optionally with the hinge region, utilizing any combination of heterodimerization variants described herein.

[0231] In particular, the formats depicted in FIG. 1 are antibodies, usually referred to as “heterodimeric antibodies”, meaning that the protein has at least two associated Fc sequences self-assembled into a heterodimeric Fc domain. FIG. 28 depicts homodimeric bispecific antibodies as well as heterodimeric antibodies.

[0232] In some embodiments, the antibody can be a mixture from different species, e.g. a chimeric antibody and/or a humanized antibody. In general, both “chimeric antibodies” and “humanized antibodies” refer to antibodies that combine regions from more than one species. For example, “chimeric antibodies” traditionally comprise variable region(s) from a mouse (or rat, in some cases) and the constant region(s) from a human. “Humanized antibodies” generally refer to non-human antibodies that have had the variable-domain framework regions swapped for sequences found in human antibodies. Generally, in a humanized antibody, the entire antibody, except the CDRs, is encoded by a polynucleotide of human origin or is identical to such an antibody except within its CDRs. The CDRs, some or all of which are encoded by nucleic acids originating in a non-human organism, are grafted into the beta-sheet framework of a human antibody variable region to create an antibody, the specificity of which is determined by the engrafted CDRs. The creation of such antibodies is described in, e.g., WO 92/11018, Jones, 1986, Nature 321:522-525, Verhoeyen et al., 1988, Science 239:1534-1536, all entirely incorporated by reference. “Backmutation” of selected acceptor framework residues to the corresponding donor residues is often required to regain affinity that is lost in the initial grafted construct (U.S. Pat. No. 5,530,101; U.S. Pat. No. 5,585,089; U.S. Pat. No. 5,693,761; U.S. Pat. No. 5,693,762; U.S. Pat. No. 6,180,370; U.S. Pat. No. 5,859,205; U.S. Pat. No. 5,821,337; U.S. Pat. No. 6,054,297; U.S. Pat. No. 6,407,213, all entirely incorporated by reference). The humanized antibody optimally also will comprise at least a portion of an immunoglobulin constant region, typically that of a human immunoglobulin, and thus will typically comprise a human Fc region. Humanized antibodies can also be generated using mice with a genetically engineered immune system. Roque et al., 2004, Biotechnol. Prog. 20:639-654, entirely incorporated by reference. A variety of techniques and methods for humanizing and reshaping non-human antibodies are well known in the art (See Tsurushita & Vasquez, 2004, Humanization of Monoclonal Antibodies, Molecular Biology of B Cells, 533-545, Elsevier Science (USA), and references cited therein, all entirely incorporated by reference). Humanization methods include but are not limited to methods described in Jones et al., 1986, Nature 321:522-525; Riechmann et al., 1988; Nature 332:323-329; Verhoeyen et al., 1988, Science, 239:1534-1536; Queen et al., 1989, Proc Natl Acad Sci, USA 86:10029-33; He et al., 1998, J. Immunol. 160: 1029-1035; Carter et al., 1992, Proc Natl Acad Sci USA 89:4285-9, Presta et al., 1997, Cancer Res. 57(20):4593-9; Gorman et al., 1991, Proc. Natl. Acad. Sci. USA 88:4181-4185; O'Connor et al., 1998, Protein Eng 11:321-8, all entirely incorporated by reference. Humanization or other methods of reducing the immunogenicity of nonhuman antibody variable regions may include resurfacing methods, as described for example in Roguska et al., 1994, Proc. Natl. Acad. Sci. USA 91:969-973, entirely incorporated by reference. In one embodiment, the parent antibody has been affinity matured, as is known in the art. Structure-based methods may be employed for humanization and affinity maturation, for example as described in U.S. Ser. No. 11/004,590. Selection based methods may be employed to humanize and/or affinity mature antibody variable regions, including but not limited to methods described in Wu et al., 1999, J. Mol. Biol. 294:151-162; Baca et al., 1997, J. Biol. Chem. 272(16):10678-10684; Rosok et al., 1996, J. Biol. Chem. 271(37): 22611-22618; Rader et al., 1998, Proc. Natl. Acad. Sci. USA 95: 8910-8915; Krauss et al., 2003, Protein Engineering 16(10):753-759, all entirely incorporated by reference. Other humanization methods may involve the grafting of only parts of the CDRs, including but not limited to methods described in U.S. Ser. No. 09/810,510; Tan et al., 2002, J. Immunol. 169:1119-1125; De Pascalis et al., 2002, J. Immunol. 169:3076-3084, all entirely incorporated by reference.

[0233] Accordingly, in some embodiments the present invention provides heterodimeric antibodies that rely on the use of two different heavy chain variant Fc sequences, that will self-assemble to form heterodimeric Fc domains and heterodimeric antibodies.

[0234] The present invention is directed to novel constructs to provide heterodimeric antibodies that allow binding to more than one antigen or ligand, e.g. to allow for bispecific binding. The heterodimeric antibody constructs are based on the self-assembling nature of the two Fc domains of the heavy chains of antibodies, e.g. two “monomers” that assemble into a “dimer”. Heterodimeric antibodies are made by altering the amino acid sequence of each monomer as more fully discussed below. Thus, the present invention is generally directed to the creation of heterodimeric antibodies which can co-engage antigens in several ways, relying on amino acid variants in the constant regions that are different on each chain to promote heterodimeric formation and/or allow for ease of purification of heterodimers over the homodimers.

[0235] Thus, the present invention provides bispecific antibodies. An ongoing problem in antibody technologies is the desire for “bispecific” antibodies that bind to two different antigens simultaneously, in general thus allowing the different antigens to be brought into proximity and resulting in new functionalities and new therapies. In general, these antibodies are made by including genes for each heavy and light chain into the host cells. This generally results in the formation of the desired heterodimer (A-B), as well as the two homodimers (A-A and B-B (not including the light chain heterodimeric issues)). However, a major obstacle in the formation of bispecific antibodies is the difficulty in purifying the heterodimeric antibodies away from the homodimeric antibodies and/or biasing the formation of the heterodimer over the formation of the homodimers.

[0236] There are a number of mechanisms that can be used to generate the heterodimers of the present invention. In addition, as will be appreciated by those in the art, these mechanisms can be combined to ensure high heterodimerization. Thus, amino acid variants that lead to the production of heterodimers are referred to as “heterodimerization variants”. As discussed below, heterodimerization variants can include steric variants (e.g. the “knobs and holes” or “skew” variants described below and the “charge pairs” variants described below) as well as “pI variants”, which allows purification of homodimers away from heterodimers. As is generally described in WO2014/145806, hereby incorporated by reference in its entirety and specifically as below for the discussion of “heterodimerization variants”, useful mechanisms for heterodimerization include “knobs and holes” (“KIH”; sometimes herein as “skew” variants (see discussion in WO2014/145806), “electrostatic steering” or “charge pairs” as described in WO2014/145806, pI variants as described in WO2014/145806, and general additional Fc variants as outlined in WO2014/145806 and below.

[0237] In the present invention, there are several basic mechanisms that can lead to ease of purifying heterodimeric antibodies; one relies on the use of pI variants, such that each monomer has a different pI, thus allowing the isoelectric purification of A-A, A-B and B-B dimeric proteins. Alternatively, some scaffold formats, such as the “triple F” format, also allows separation on the basis of size. As is further outlined below, it is also possible to “skew” the formation of heterodimers over homodimers. Thus, a combination of steric heterodimerization variants and pI or charge pair variants find particular use in the invention.

[0238] In general, embodiments of particular use in the present invention rely on sets of variants that include skew variants, which encourage heterodimerization formation over homodimerization formation, coupled with pI variants, which increase the pI difference between the two monomers.

[0239] Additionally, as more fully outlined below, depending on the format of the heterodimer antibody, pI variants can be either contained within the constant and/or Fc domains of a monomer, or charged linkers, either domain linkers or scFv linkers, can be used. That is, scaffolds that utilize scFv(s) such as the Triple F format can include charged scFv linkers (either positive or negative), that give a further pI boost for purification purposes. As will be appreciated by those in the art, some Triple F formats are useful with just charged scFv linkers and no additional pI adjustments, although the invention does provide pI variants that are on one or both of the monomers, and/or charged domain linkers as well. In addition, additional amino acid engineering for alternative functionalities may also confer pI changes, such as Fc, FcRn and KO variants.

[0240] In the present invention that utilizes pI as a separation mechanism to allow the purification of heterodimeric proteins, amino acid variants can be introduced into one or both of the monomer polypeptides; that is, the pI of one of the monomers (referred to herein for simplicity as “monomer A”) can be engineered away from monomer B, or both monomer A and B change be changed, with the pI of monomer A increasing and the pI of monomer B decreasing. As is outlined more fully below, the pI changes of either or both monomers can be done by removing or adding a charged residue (e.g. a neutral amino acid is replaced by a positively or negatively charged amino acid residue, e.g. glycine to glutamic acid), changing a charged residue from positive or negative to the opposite charge (aspartic acid to lysine) or changing a charged residue to a neutral residue (e.g. loss of a charge; lysine to serine.). A number of these variants are shown in the Figures.

[0241] Accordingly, this embodiment of the present invention provides for creating a sufficient change in pI in at least one of the monomers such that heterodimers can be separated from homodimers. As will be appreciated by those in the art, and as discussed further below, this can be done by using a “wild type” heavy chain constant region and a variant region that has been engineered to either increase or decrease it's pI (wt A−+B or wt A−−B), or by increasing one region and decreasing the other region (A+−B− or A−B+).

[0242] Thus, in general, a component of some embodiments of the present invention are amino acid variants in the constant regions of antibodies that are directed to altering the isoelectric point (pI) of at least one, if not both, of the monomers of a dimeric protein to form “pI antibodies”) by incorporating amino acid substitutions (“pI variants” or “pI substitutions”) into one or both of the monomers. As shown herein, the separation of the heterodimers from the two homodimers can be accomplished if the pIs of the two monomers differ by as little as 0.1 pH unit, with 0.2, 0.3, 0.4 and 0.5 or greater all finding use in the present invention.

[0243] As will be appreciated by those in the art, the number of pI variants to be included on each or both monomer(s) to get good separation will depend in part on the starting pI of the components, for example in the triple F format, the starting pI of the scFv and Fab of interest. That is, to determine which monomer to engineer or in which “direction” (e.g. more positive or more negative), the Fv sequences of the two target antigens are calculated and a decision is made from there. As is known in the art, different Fvs will have different starting pIs which are exploited in the present invention. In general, as outlined herein, the pIs are engineered to result in a total pI difference of each monomer of at least about 0.1 logs, with 0.2 to 0.5 being preferred as outlined herein.

[0244] Furthermore, as will be appreciated by those in the art and outlined herein, in some embodiments, heterodimers can be separated from homodimers on the basis of size. As shown in FIGS. 1, 2 and 28 for example, several of the formats allow separation of heterodimers and homodimers on the basis of size.

[0245] In the case where pI variants are used to achieve heterodimerization, by using the constant region(s) of the heavy chain(s), a more modular approach to designing and purifying bispecific proteins, including antibodies, is provided. Thus, in some embodiments, heterodimerization variants (including skew and purification heterodimerization variants) are not included in the variable regions, such that each individual antibody must be engineered. In addition, in some embodiments, the possibility of immunogenicity resulting from the pI variants is significantly reduced by importing pI variants from different IgG isotypes such that pI is changed without introducing significant immunogenicity. Thus, an additional problem to be solved is the elucidation of low pI constant domains with high human sequence content, e.g. the minimization or avoidance of non-human residues at any particular position.

[0246] A side benefit that can occur with this pI engineering is also the extension of serum half-life and increased FcRn binding. That is, as described in U.S. Ser. No. 13/194,904 (incorporated by reference in its entirety), lowering the pI of antibody constant domains (including those found in antibodies and Fc fusions) can lead to longer serum retention in vivo. These pI variants for increased serum half life also facilitate pI changes for purification.

[0247] In addition, it should be noted that the pI variants of the heterodimerization variants give an additional benefit for the analytics and quality control process of bispecific antibodies, as the ability to either eliminate, minimize and distinguish when homodimers are present is significant. Similarly, the ability to reliably test the reproducibility of the heterodimeric antibody production is important.

[0248] The present invention provides heterodimeric proteins, including heterodimeric antibodies in a variety of formats, which utilize heterodimeric variants to allow for heterodimeric formation and/or purification away from homodimers.

[0249] There are a number of suitable pairs of sets of heterodimerization skew variants. These variants come in “pairs” of “sets”. That is, one set of the pair is incorporated into the first monomer and the other set of the pair is incorporated into the second monomer. It should be noted that these sets do not necessarily behave as “knobs in holes” variants, with a one-to-one correspondence between a residue on one monomer and a residue on the other; that is, these pairs of sets form an interface between the two monomers that encourages heterodimer formation and discourages homodimer formation, allowing the percentage of heterodimers that spontaneously form under biological conditions to be over 90%, rather than the expected 50% (25% homodimer A/A:50% heterodimer A/B:25% homodimer B/B).

[0250] In some embodiments, the formation of heterodimers can be facilitated by the addition of steric variants. That is, by changing amino acids in each heavy chain, different heavy chains are more likely to associate to form the heterodimeric structure than to form homodimers with the same Fc amino acid sequences. Suitable steric variants are included in FIG. 22 and FIG. 17 (there are also pI variants in FIG. 17).

[0251] One mechanism is generally referred to in the art as “knobs and holes”, referring to amino acid engineering that creates steric influences to favor heterodimeric formation and disfavor homodimeric formation can also optionally be used; this is sometimes referred to as “knobs and holes”, as described in U.S. Ser. No. 61/596,846, Ridgway et al., Protein Engineering 9(7):617 (1996); Atwell et al., J. Mol. Biol. 1997 270:26; U.S. Pat. No. 8,216,805, all of which are hereby incorporated by reference in their entirety. The Figures identify a number of “monomer A-monomer B” pairs that rely on “knobs and holes”. In addition, as described in Merchant et al., Nature Biotech. 16:677 (1998), these “knobs and hole” mutations can be combined with disulfide bonds to skew formation to heterodimerization.

[0252] An additional mechanism that finds use in the generation of heterodimers is sometimes referred to as “electrostatic steering” as described in Gunasekaran et al., J. Biol. Chem. 285(25):19637 (2010), hereby incorporated by reference in its entirety. This is sometimes referred to herein as “charge pairs”. In this embodiment, electrostatics are used to skew the formation towards heterodimerization. As those in the art will appreciate, these may also have an effect on pI, and thus on purification, and thus could in some cases also be considered pI variants. However, as these were generated to force heterodimerization and were not used as purification tools, they are classified as “steric variants”. These include, but are not limited to, D221E/P228E/L368E paired with D221R/P228R/K409R (e.g. these are “monomer corresponding sets) and C220E/P228E/368E paired with C220R/E224R/P228R/K409R.

[0253] Additional monomer A and monomer B variants that can be combined with other variants, optionally and independently in any amount, such as pI variants outlined herein or other steric variants that are shown in FIG. 37 of US 2012/0149876, the figure and legend and SEQ ID NOs of which are incorporated expressly by reference herein.

[0254] In some embodiments, the steric variants outlined herein can be optionally and independently incorporated with any pI variant (or other variants such as Fc variants, FcRn variants, etc.) into one or both monomers, and can be independently and optionally included or excluded from the proteins of the invention.

[0255] A list of suitable skew variants is found in FIG. 17, with FIG. 22 showing some pairs of particular utility in many embodiments. Of particular use in many embodiments are the pairs of sets including, but not limited to, S364K/E357Q:L368D/K370S; L368D/K370S:S364K; L368E/K370S:S364K; T411T/E360E/Q362E:D401K; L368D/K370S:S364K/E357L and K370S:S364K/E357Q. In terms of nomenclature, the pair “S364K/E357Q:L368D/K370S” means that one of the monomers has the double variant set S364K/E357Q and the other has the double variant set L368D/K370S.

[0256] In general, as will be appreciated by those in the art, there are two general categories of pI variants: those that increase the pI of the protein (basic changes) and those that decrease the pI of the protein (acidic changes). As described herein, all combinations of these variants can be done: one monomer may be wild type, or a variant that does not display a significantly different pI from wild-type, and the other can be either more basic or more acidic. Alternatively, each monomer is changed, one to more basic and one to more acidic.

[0257] Preferred combinations of pI variants are shown in the figures. As outlined herein and shown in the figures, these changes are shown relative to IgG1, but all isotypes can be altered this way, as well as isotype hybrids. In the case where the heavy chain constant domain is from IgG2-4, R133E and R133Q can also be used.

[0258] In the case of antibody based heterodimers, e.g. where at least one of the monomers comprises a light chain in addition to the heavy chain domain, pI variants can also be made in the light chain. Amino acid substitutions for lowering the pI of the light chain include, but are not limited to, K126E, K126Q, K145E, K145Q, N152D, S156E, K169E, S202E, K207E and adding peptide DEDE (SEQ ID NO: 241) at the c-terminus of the light chain. Changes in this category based on the constant lambda light chain include one or more substitutions at R108Q, Q124E, K126Q, N138D, K145T and Q199E. In addition, increasing the pI of the light chains can also be done.

[0259] In addition, many embodiments of the invention rely on the “importation” of pI amino acids at particular positions from one IgG isotype into another, thus reducing or eliminating the possibility of unwanted immunogenicity being introduced into the variants. A number of these are shown in FIG. 21 of US Publ. 2014/0370013, hereby incorporated by reference. That is, IgG1 is a common isotype for therapeutic antibodies for a variety of reasons, including high effector function. However, the heavy constant region of IgG1 has a higher pI than that of IgG2 (8.10 versus 7.31). By introducing IgG2 residues at particular positions into the IgG1 backbone, the pI of the resulting monomer is lowered (or increased) and additionally exhibits longer serum half-life. For example, IgG1 has a glycine (pI 5.97) at position 137, and IgG2 has a glutamic acid (pI 3.22); importing the glutamic acid will affect the pI of the resulting protein. As is described below, a number of amino acid substitutions are generally required to significant affect the pI of the variant antibody. However, it should be noted as discussed below that even changes in IgG2 molecules allow for increased serum half-life.

[0260] In other embodiments, non-isotypic amino acid changes are made, either to reduce the overall charge state of the resulting protein (e.g. by changing a higher pI amino acid to a lower pI amino acid), or to allow accommodations in structure for stability, etc. as is more further described below.

[0261] In addition, by pI engineering both the heavy and light constant domains, significant changes in each monomer of the heterodimer can be seen. As discussed herein, having the pIs of the two monomers differ by at least 0.5 can allow separation by ion exchange chromatography or isoelectric focusing, or other methods sensitive to isoelectric point.

[0262] The pI of each monomer can depend on the pI of the variant heavy chain constant domain and the pI of the total monomer, including the variant heavy chain constant domain and the fusion partner. Thus, in some embodiments, the change in pI is calculated on the basis of the variant heavy chain constant domain, using the chart in the Figure. As discussed herein, which monomer to engineer is generally decided by the inherent pI of the Fv and scaffold regions. Alternatively, the pI of each monomer can be compared.

pI Variants that Also Confer Better FcRn In Vivo Binding

[0263] In the case where the pI variant decreases the pI of the monomer, they can have the added benefit of improving serum retention in vivo.

[0264] Although still under examination, Fc regions are believed to have longer half-lives in vivo, because binding to FcRn at pH 6 in an endosome sequesters the Fc (Ghetie and Ward, 1997 Immunol Today. 18(12): 592-598, entirely incorporated by reference). The endosomal compartment then recycles the Fc to the cell surface. Once the compartment opens to the extracellular space, the higher pH, ˜7.4, induces the release of Fc back into the blood. In mice, Dall'Acqua et al. showed that Fc mutants with increased FcRn binding at pH 6 and pH 7.4 actually had reduced serum concentrations and the same half life as wild-type Fc (Dall'Acqua et al. 2002, J. Immunol. 169:5171-5180, entirely incorporated by reference). The increased affinity of Fc for FcRn at pH 7.4 is thought to forbid the release of the Fc back into the blood. Therefore, the Fc mutations that will increase Fc's half-life in vivo will ideally increase FcRn binding at the lower pH while still allowing release of Fc at higher pH. The amino acid histidine changes its charge state in the pH range of 6.0 to 7.4. Therefore, it is not surprising to find His residues at important positions in the Fc/FcRn complex.

[0265] Recently it has been suggested that antibodies with variable regions that have lower isoelectric points may also have longer serum half-lives (Igawa et al., 2010 PEDS. 23(5): 385-392, entirely incorporated by reference). However, the mechanism of this is still poorly understood. Moreover, variable regions differ from antibody to antibody. Constant region variants with reduced pI and extended half-life would provide a more modular approach to improving the pharmacokinetic properties of antibodies, as described herein.

[0266] In addition to formats that rely on the use of heterodimeric Fc domains, some bispecific formats do not rely on heterodimerization variants in the Fc or constant domains as shown in the Figures. Rather, bispecific tetravalent antibodies are generated, where each heavy chain is identical.

[0267] As will be appreciated, a number of suitable antibody formats can be used. In many cases, these constructs rely on the use of at least one of the two antigen binding domains being in scFv format, for ease of purification, although this is not required for all formats. FIG. 3 shows two such embodiments (using anti-CD20 antigen binding domains, although any target tumor antigens can be used as is shown in FIG. 28D), generally referred to as “mAb-scFv2” (FIG. 3A) and “central scFv2” (FIG. 3B). Other such formats finding use in the present invention include, but are not limited to, other IgG-scFv formats (for example, where the scFvs are attached to the C-terminus of the light chain, or where the scFvs are attached to the Fv region of the Fab portion, for example either at the N terminus of the heavy chain or the N-terminus of the light chain), DVD-Ig formats, IgG-sVD formats, sVD-IgG formats, 2-in-1 IgG formats, mAb2 formations, etc., such as those depicted in FIG. 2 of Kontermann, mAbs 4:2, 182-197 (2012) under “bispecific IgG and IgG-like molecules”, specifically incorporated by reference with the accompanying references.

[0268] In addition to pI amino acid variants, there are a number of useful Fc amino acid modification that can be made for a variety of reasons, including, but not limited to, altering binding to one or more FcγR receptors, altered binding to FcRn receptors, etc.

[0269] Accordingly, the proteins of the invention can include amino acid modifications, including the heterodimerization variants outlined herein, which includes the pI variants and steric variants. Each set of variants can be independently and optionally included or excluded from any particular heterodimeric protein.

[0270] Accordingly, there are a number of useful Fc substitutions that can be made to alter binding to one or more of the FcγR receptors. Substitutions that result in increased binding as well as decreased binding can be useful. For example, it is known that increased binding to Fc□RIIIa generally results in increased ADCC (antibody dependent cell-mediated cytotoxicity; the cell-mediated reaction wherein nonspecific cytotoxic cells that express FcγRs recognize bound antibody on a target cell and subsequently cause lysis of the target cell). Similarly, decreased binding to FcγRIIb (an inhibitory receptor) can be beneficial as well in some circumstances. Amino acid substitutions that find use in the present invention aaainclude those listed in U.S. Ser. No. 11/124,620 (particularly FIG. 41), Ser. Nos. 11/174,287, 11/396,495, 11/538,406, all of which are expressly incorporated herein by reference in their entirety and specifically for the variants disclosed therein. Particular variants that find use include, but are not limited to, 236A, 239D, 239E, 332E, 332D, 239D/332E, 267D, 267E, 328F, 267E/328F, 236A/332E, 239D/332E/330Y, 239D, 332E/330L, 243A, 243L, 264A, 264V and 299T.

[0271] In addition, there are additional Fc substitutions that find use in increased binding to the FcRn receptor and increased serum half life, as specifically disclosed in U.S. Ser. No. 12/341,769, hereby incorporated by reference in its entirety, including, but not limited to, 434S, 434A, 428L, 308F, 259I, 428L/434S, 259I/308F, 436I/428L, 436I or V/434S, 436V/428L and 259I/308F/428L.

[0272] Similarly, another category of functional variants are “FcγR ablation variants” or “Fc knock out (FcKO or KO)” variants. In these embodiments, for some therapeutic applications, it is desirable to reduce or remove the normal binding of the Fc domain to one or more or all of the Fcγ receptors (e.g. FcγR1, FcγRIIa, FcγRIIb, FcγRIIIa, etc.) to avoid additional mechanisms of action. That is, for example, in many embodiments, particularly in the use of bispecific antibodies that bind CD3 monovalently it is generally desirable to ablate FcγRIIIa binding to eliminate or significantly reduce ADCC activity. wherein one of the Fc domains comprises one or more Fcγ receptor ablation variants. These ablation variants are depicted in FIG. 19, and each can be independently and optionally included or excluded, with preferred aspects utilizing ablation variants selected from the group consisting of G236R/L328R, E233P/L234V/L235A/G236del/S239K, E233P/L234V/L235A/G236del/S267K, E233P/L234V/L235A/G236del/S239K/A327G, E233P/L234V/L235A/G236del/S267K/A327G and E233P/L234V/L235A/G236del. It should be noted that the ablation variants referenced herein ablate FcγR binding but generally not FcRn binding.

[0273] As will be appreciated by those in the art, all of the recited heterodimerization variants (including skew and/or pI variants) can be optionally and independently combined in any way, as long as they retain their “strandedness” or “monomer partition”. In addition, all of these variants can be combined into any of the heterodimerization formats.

[0274] In the case of pI variants, while embodiments finding particular use are shown in the Figures, other combinations can be generated, following the basic rule of altering the pI difference between two monomers to facilitate purification.

[0275] In addition, any of the heterodimerization variants, skew and pI, are also independently and optionally combined with Fc ablation variants, Fc variants, FcRn variants, as generally outlined herein.

[0276] As will be appreciated by those in the art and discussed more fully below, the heterodimeric fusion proteins of the present invention can take on a wide variety of configurations, as are generally depicted in FIGS. 1, 2 and 28. Some figures depict “single ended” configurations, where there is one type of specificity on one “arm” of the molecule and a different specificity on the other “arm”. Other figures depict “dual ended” configurations, where there is at least one type of specificity at the “top” of the molecule and one or more different specificities at the “bottom” of the molecule. Thus, the present invention is directed to novel immunoglobulin compositions that co-engage a different first and a second antigen.

[0277] As will be appreciated by those in the art, the heterodimeric formats of the invention can have different valencies as well as be bispecific. That is, heterodimeric antibodies of the invention can be bivalent and bispecific, wherein one target tumor antigen (e.g. CD8) is bound by one binding domain and the other target tumor antigen (e.g. CD20, CD19, CD38, CD123, etc.) is bound by a second binding domain. The heterodimeric antibodies can also be trivalent and bispecific, wherein the first antigen is bound by two binding domains and the second antigen by a second binding domain.

[0278] The present invention utilizes anti-CD8 antigen binding domains in combination with anti-target tumor antigen (TTA) antigen binding domains. As will be appreciated by those in the art, any collection of anti-CD8 CDRs, anti-CD8 variable light and variable heavy domains, Fabs and scFvs as depicted in any of the Figures can be used. Similarly, any of the anti-TTA antigen binding domains can be used, e.g. anti-CD38, anti-CD20, anti-CD19 and anti-CD123 antigen binding domains, whether CDRs, variable light and variable heavy domains, Fabs and scFvs as depicted in any of the Figures can be used, optionally and independently combined in any combination.

[0279] One heterodimeric scaffold that finds particular use in the present invention is the “triple F” or “bottle opener” scaffold format as shown in FIG. 1A, A and B. In this embodiment, one heavy chain of the antibody contains an single chain Fv (“scFv”, as defined below) and the other heavy chain is a “regular” FAb format, comprising a variable heavy chain and a light chain. This structure is sometimes referred to herein as “triple F” format (scFv-FAb-Fc) or the “bottle-opener” format, due to a rough visual similarity to a bottle-opener (see FIG. 1). The two chains are brought together by the use of amino acid variants in the constant regions (e.g. the Fc domain, the CH1 domain and/or the hinge region) that promote the formation of heterodimeric antibodies as is described more fully below.

[0280] There are several distinct advantages to the present “triple F” format. As is known in the art, antibody analogs relying on two scFv constructs often have stability and aggregation problems, which can be alleviated in the present invention by the addition of a “regular” heavy and light chain pairing. In addition, as opposed to formats that rely on two heavy chains and two light chains, there is no issue with the incorrect pairing of heavy and light chains (e.g. heavy 1 pairing with light 2, etc.).

[0281] Many of the embodiments outlined herein rely in general on the bottle opener format that comprises a first monomer comprising an scFv, comprising a variable heavy and a variable light domain, covalently attached using an scFv linker (charged, in many but not all instances), where the scFv is covalently attached to the N-terminus of a first Fc domain usually through a domain linker (which, as outlined herein can either be un-charged or charged). The second monomer of the bottle opener format is a heavy chain, and the composition further comprises a light chain.

[0282] In some embodiments, the scFv is the domain that binds to the CD8, with the Fab of the heavy and light chains binding to the other TTA. In addition, the Fc domains of the invention generally comprise skew variants (e.g. a set of amino acid substitutions as shown in FIG. 17 and FIG. 24, with particularly useful skew variants being selected from the group consisting of S364K/E357Q:L368D/K370S; L368D/K370S:S364K; L368E/K370S:S364K; T411T/E360E/Q362E:D401K; L368D/K370S:S364K/E357L and K370S:S364K/E357Q), optionally ablation variants (including those shown in FIG. 19), optionally charged scFv linkers (including those shown in Figure XX) and the heavy chain comprises pI variants (including those shown in FIG. 17).

[0283] The present invention provides bottle opener formats where the anti-CD8 sequences are as shown in FIG. 37 to FIG. 40, whether sets of 6 CDRs, variable heavy and light domain pairs, or scFv sequences, including scFv linkers whether charged or uncharged.

[0284] The present invention provides bottle opener formats with CD38 antigen binding domains wherein the anti-CD38 sequences are as shown in the Figures.

[0285] The present invention provides bottle opener formats with CD20 antigen binding domains wherein the anti-CD20 sequences are as shown in the Figures.

[0286] The present invention provides bottle opener formats with CD19 antigen binding domains wherein the anti-CD19 sequences are as shown in the Figures.

[0287] The present invention provides bottle opener formats with CD123 antigen binding domains wherein the anti-CD123 sequences are as shown in the Figures.

mAb-Fv Format

[0288] One heterodimeric scaffold that finds particular use in the present invention is the mAb-Fv format shown in FIG. 1. In this embodiment, the format relies on the use of a C-terminal attachment of an “extra” variable heavy domain to one monomer and the C-terminal attachment of an “extra” variable light domain to the other monomer, thus forming a third antigen binding domain, wherein the Fab portions of the two monomers bind a TTA and the “extra” scFv domain binds CD8, or vice versa.

[0289] In this embodiment, the first monomer comprises a first heavy chain, comprising a first variable heavy domain and a first constant heavy domain comprising a first Fc domain, with a first variable light domain covalently attached to the C-terminus of the first Fc domain using a domain linker. The second monomer comprises a second variable heavy domain of the second constant heavy domain comprising a second Fc domain, and a third variable heavy domain covalently attached to the C-terminus of the second Fc domain using a domain linker. The two C-terminally attached variable domains make up a scFv that binds CD3. This embodiment further utilizes a common light chain comprising a variable light domain and a constant light domain, that associates with the heavy chains to form two identical Fabs that bind a TTA. As for many of the embodiments herein, these constructs include skew variants, pI variants, ablation variants, additional Fc variants, etc. as desired and described herein.

[0290] The present invention provides mAb-Fv formats where the anti-CD8 scFv sequences are as shown in the Figures.

[0291] The present invention provides mAb-Fv formats wherein the anti-CD38 sequences are as shown in the Figures.

[0292] The present invention provides mAb-Fv formats with CD20 antigen binding domains wherein the anti-CD20 sequences are as shown in the Figures.

[0293] The present invention provides mAb-Fv formats with CD19 antigen binding domains wherein the anti-CD19 sequences are as shown in the Figures.

[0294] The present invention provides mAb-Fv formats with CD123 antigen binding domains wherein the anti-CD123 sequences are as shown in the Figures.

[0295] The present invention provides mAb-Fv formats comprising ablation variants as shown in FIG. 19.

[0296] The present invention provides mAb-Fv formats comprising skew variants as shown in FIGS. 17 and 22.

mAb-scFv

[0297] One heterodimeric scaffold that finds particular use in the present invention is the mAb-Fv format shown in FIG. 1. In this embodiment, the format relies on the use of a C-terminal attachment of a scFv to one of the monomers, thus forming a third antigen binding domain, wherein the Fab portions of the two monomers bind a TTA and the “extra” scFv domain binds CD8, or vice versa. Thus, the first monomer comprises a first heavy chain (comprising a variable heavy domain and a constant domain), with a C-terminally covalently attached scFv comprising a scFv variable light domain, an scFv linker and a scFv variable heavy domain. This embodiment further utilizes a common light chain comprising a variable light domain and a constant light domain, that associates with the heavy chains to form two identical Fabs that bind a TTA. As for many of the embodiments herein, these constructs include skew variants, pI variants, ablation variants, additional Fc variants, etc. as desired and described herein.

[0298] The present invention provides mAb-Fv formats where the anti-CD3 scFv sequences are as shown the Figures.

[0299] The present invention provides mAb-Fv formats wherein the anti-CD38 sequences are as shown in the Figures.

[0300] The present invention provides mAb-Fv formats with CD20 antigen binding domains wherein the anti-CD20 sequences are as shown in the Figures.

[0301] The present invention provides mAb-Fv formats with CD19 antigen binding domains wherein the anti-CD19 sequences are as shown in the Figures.

[0302] The present invention provides mAb-Fv formats with CD123 antigen binding domains wherein the anti-CD123 sequences are as shown in the Figures.

[0303] The present invention provides mAb-Fv formats comprising ablation variants as shown in FIG. 19.

[0304] The present invention provides mAb-Fv formats comprising skew variants as shown in FIGS. 17 and 22.

Central scFv

[0305] One heterodimeric scaffold that finds particular use in the present invention is the Central-scFv format shown in FIG. 1. In this embodiment, the format relies on the use of an inserted scFv domain thus forming a third antigen binding domain, wherein the Fab portions of the two monomers bind a TTA and the “extra” scFv domain binds CD8 or vice versa. The scFv domain is inserted between the Fc domain and the CH1-Fv region of one of the monomers, thus providing a third antigen binding domain.

[0306] In this embodiment, one monomer comprises a first heavy chain comprising a first variable heavy domain, a CH1 domain and Fc domain, with a scFv comprising a scFv variable light domain, an scFv linker and a scFv variable heavy domain. The scFv is covalently attached between the C-terminus of the CH1 domain of the heavy constant domain and the N-terminus of the first Fc domain using domain linkers. This embodiment further utilizes a common light chain comprising a variable light domain and a constant light domain, that associates with the heavy chains to form two identical Fabs that bind a TTA. As for many of the embodiments herein, these constructs include skew variants, pI variants, ablation variants, additional Fc variants, etc. as desired and described herein.

[0307] The present invention provides Central-scFv formats where the anti-CD8 scFv sequences are as shown in the Figures.

[0308] The present invention provides Central-scFv formats wherein the anti-CD38 sequences are as shown in the Figures.

[0309] The present invention provides Central-scFv formats with CD20 antigen binding domains wherein the anti-CD20 sequences are as shown in the Figures.

[0310] The present invention provides Central-scFv formats with CD19 antigen binding domains wherein the anti-CD19 sequences are as shown in the Figures.

[0311] The present invention provides Central-scFv formats with CD123 antigen binding domains wherein the anti-CD123 sequences are as shown in the Figures

[0312] The present invention provides Central-scFv formats comprising ablation variants as shown in FIG. 19.

[0313] The present invention provides Central-scFv formats comprising skew variants as shown in FIGS. 17 and 22.

[0314] One heterodimeric scaffold that finds particular use in the present invention is the Central-Fv format shown in FIG. 1. In this embodiment, the format relies on the use of an inserted scFv domain thus forming a third antigen binding domain, wherein the Fab portions of the two monomers bind a TTA and the “extra” scFv domain binds CD3 or vice versa. The scFv domain is inserted between the Fc domain and the CH1-Fv region of the monomers, thus providing a third antigen binding domain, wherein each monomer contains a component of the scFv (e.g. one monomer comprises a variable heavy domain and the other a variable light domain).

[0315] In this embodiment, one monomer comprises a first heavy chain comprising a first variable heavy domain, a CH1 domain and Fc domain and an additional variable light domain. The light domain is covalently attached between the C-terminus of the CH1 domain of the heavy constant domain and the N-terminus of the first Fc domain using domain linkers. The other monomer comprises a first heavy chain comprising a first variable heavy domain, a CH1 domain and Fc domain and an additional variable heavy domain. The light domain is covalently attached between the C-terminus of the CH1 domain of the heavy constant domain and the N-terminus of the first Fc domain using domain linkers.

[0316] This embodiment further utilizes a common light chain comprising a variable light domain and a constant light domain, that associates with the heavy chains to form two identical Fabs that bind a TTA. As for many of the embodiments herein, these constructs include skew variants, pI variants, ablation variants, additional Fc variants, etc. as desired and described herein.

[0317] The present invention provides Central-Fv formats where the anti-CD8 scFv sequences are as shown in the Figures.

[0318] The present invention provides Central-Fv formats wherein the anti-CD38 sequences are as shown in the Figures.

[0319] The present invention provides Central-Fv formats with CD20 antigen binding domains wherein the anti-CD20 sequences are as shown in the Figures.

[0320] The present invention provides Central-Fv formats with CD19 antigen binding domains wherein the anti-CD19 sequences are as shown in the Figures.

[0321] The present invention provides Central-Fv formats with CD123 antigen binding domains wherein the anti-CD123 sequences are as shown in the Figures.

[0322] The present invention provides Central-Fv formats comprising ablation variants as shown in FIG. 19.

[0323] The present invention provides Central-Fv formats comprising skew variants as shown in FIGS. 17 and 22.

[0324] One heterodimeric scaffold that finds particular use in the present invention is the one armed central-scFv format shown in FIG. 1. In this embodiment, one monomer comprises just an Fc domain, while the other monomer uses an inserted scFv domain thus forming the second antigen binding domain. In this format, either the Fab portion binds a TTA and the scFv binds CD8 or vice versa. The scFv domain is inserted between the Fc domain and the CH1-Fv region of one of the monomers.

[0325] In this embodiment, one monomer comprises a first heavy chain comprising a first variable heavy domain, a CH1 domain and Fc domain, with a scFv comprising a scFv variable light domain, an scFv linker and a scFv variable heavy domain. The scFv is covalently attached between the C-terminus of the CH1 domain of the heavy constant domain and the N-terminus of the first Fc domain using domain linkers. The second monomer comprises an Fc domain. This embodiment further utilizes a light chain comprising a variable light domain and a constant light domain, that associates with the heavy chain to form a Fab. As for many of the embodiments herein, these constructs include skew variants, pI variants, ablation variants, additional Fc variants, etc. as desired and described herein.

[0326] The present invention provides one armed central-scFv formats where the anti-CD8 scFv sequences are as shown in the Figures.

[0327] The present invention provides one armed central-scFv formats wherein the anti-CD38 sequences are as shown in the Figures.

[0328] The present invention provides one armed central-scFv formats with CD20 antigen binding domains wherein the anti-CD20 sequences are as shown in the Figures.

[0329] The present invention provides one armed central-scFv formats with CD19 antigen binding domains wherein the anti-CD19 sequences are as shown in the Figures.

[0330] The present invention provides one armed central-scFv formats with CD123 antigen binding domains wherein the anti-CD123 sequences are as shown in the Figures.

[0331] The present invention provides one armed central-scFv formats comprising ablation variants as shown in FIG. 19.

[0332] The present invention provides one armed central-scFv formats comprising skew variants as shown in FIGS. 17 and 22.

Dual scFv Formats

[0333] The present invention also provides dual scFv formats as are known in the art and shown in FIG. 1.

[0334] The present invention provides dual scFv formats where the anti-CD8 scFv sequences are as shown in the Figures.

[0335] The present invention provides dual scFv formats wherein the anti-CD38 sequences are as shown in the Figures.

[0336] The present invention provides dual scFv formats with CD20 antigen binding domains wherein the anti-CD20 sequences are as shown in the Figures.

[0337] The present invention provides dual scFv formats with CD19 antigen binding domains wherein the anti-CD19 sequences are as shown in the Figures.

[0338] The present invention provides dual scFv formats with CD123 antigen binding domains wherein the anti-CD123 sequences are as shown in the Figures.

[0339] The present invention provides dual scFv formats comprising ablation variants as shown in FIG. 19.

[0340] The present invention provides dual scFv formats comprising skew variants as shown in FIGS. 17 and 22.

[0341] The bispecific antibodies of the invention have two different antigen binding domains: one that binds to CD8, and one that binds to a target tumor antigen (sometimes referred to herein as “TTA”). Suitable target tumor antigens include, but are not limited to, CD20, CD38, CD123; ROR1, ROR2, BCMA; PSMA; SSTR2; SSTR5, CD19, FLT3, CD33, PSCA, ADAM 17, CEA, Her2, EGFR, EGFR-vIII, CD30, FOLR1, GD-2, CA-IX, Trop-2, CD70, CD38, mesothelin, EphA2, CD22, CD79b, GPNMB, CD56, CD138, CD52, CD74, CD30, CD123, RON, ERBB2, and EGFR.

[0342] The “triple F” format is particularly beneficial for targeting two (or more) distinct antigens. (As outlined herein, this targeting can be any combination of monovalent and divalent binding, depending on the format). Thus the immunoglobulins herein preferably co-engage two target antigens. Each monomer's specificity can be selected from the lists herein. Additional useful bispecific formats for use with an anti-CD8 binding domain are shown in FIG. 1.

[0343] Particular suitable applications of the heterodimeric antibodies herein are co-target pairs for which it is beneficial or critical to engage each target antigen monovalently. Such antigens may be, for example, immune receptors that are activated upon immune complexation. Cellular activation of many immune receptors occurs only by cross-linking, achieved typically by antibody/antigen immune complexes, or via effector cell to target cell engagement. For some immune receptors, for example the CD8 signaling receptor on T cells, activation only upon engagement with co-engaged target is critical, as nonspecific cross-linking in a clinical setting can elicit a cytokine storm and toxicity. Therapeutically, by engaging such antigens monovalently rather than multivalently, using the immunoglobulins herein, such activation occurs only in response to cross-linking only in the microenvironment of the primary target antigen. The ability to target two different antigens with different valencies is a novel and useful aspect of the present invention. Examples of target antigens for which it may be therapeutically beneficial or necessary to co-engage monovalently include but are not limited to immune activating receptors such as CD3, CD8, FcγRs, toll-like receptors (TLRs) such as TLR4 and TLR9, cytokine, chemokine, cytokine receptors, and chemokine receptors. In many embodiments, one of the antigen binding sites binds to CD3, and in some embodiments it is the scFv-containing monomer.

[0344] Virtually any antigen may be targeted by the immunoglobulins herein, including but not limited to proteins, subunits, domains, motifs, and/or epitopes belonging to the following list of target antigens, which includes both soluble factors such as cytokines and membrane-bound factors, including transmembrane receptors: 17-IA, 4-1BB, 4Dc, 6-keto-PGF1a, 8-iso-PGF2a, 8-oxo-dG, A1 Adenosine Receptor, A33, ACE, ACE-2, Activin, Activin A, Activin AB, Activin B, Activin C, Activin RIA, Activin RIA ALK-2, Activin RIB ALK-4, Activin RIIA, Activin RIIB, ADAM, ADAM10, ADAM12, ADAM15, ADAM17/TACE, ADAMS, ADAMS, ADAMTS, ADAMTS4, ADAMTS5, Addressins, aFGF, ALCAM, ALK, ALK-1, ALK-7, alpha-1-antitrypsin, alpha-V/beta-1 antagonist, ANG, Ang, APAF-1, APE, APJ, APP, APRIL, AR, ARC, ART, Artemin, anti-Id, ASPARTIC, Atrial natriuretic factor, av/b3 integrin, Axl, b2M, B7-1, B7-2, B7-H, B-lymphocyte Stimulator (BlyS), BACE, BACE-1, Bad, BAFF, BAFF-R, Bag-1, BAK, Bax, BCA-1, BCAM, Bcl, BCMA, BDNF, b-ECGF, bFGF, BID, Bik, BIM, BLC, BL-CAM, BLK, BMP, BMP-2 BMP-2a, BMP-3 Osteogenin, BMP-4 BMP-2b, BMP-5, BMP-6 Vgr-1, BMP-7 (OP-1), BMP-8 (BMP-8a, OP-2), BMPR, BMPR-IA (ALK-3), BMPR-IB (ALK-6), BRK-2, RPK-1, BMPR-II (BRK-3), BMPs, b-NGF, BOK, Bombesin, Bone-derived neurotrophic factor, BPDE, BPDE-DNA, BTC, complement factor 3 (C3), C3a, C4, C5, C5a, C10, CAl25, CAD-8, Calcitonin, cAMP, carcinoembryonic antigen (CEA), carcinoma-associated antigen, Cathepsin A, Cathepsin B, Cathepsin C/DPPI, Cathepsin D, Cathepsin E, Cathepsin H, Cathepsin L, Cathepsin O, Cathepsin S, Cathepsin V, Cathepsin X/Z/P, CBL, CCI, CCK2, CCL, CCL1, CCL11, CCL12, CCL13, CCL14, CCL15, CCL16, CCL17, CCL18, CCL19, CCL2, CCL20, CCL21, CCL22, CCL23, CCL24, CCL25, CCL26, CCL27, CCL28, CCL3, CCL4, CCL5, CCL6, CCL7, CCL8, CCL9/10, CCR, CCR1, CCR10, CCR10, CCR2, CCR3, CCR4, CCR5, CCR6, CCR7, CCR8, CCR9, CD1, CD2, CD3, CD3E, CD4, CD5, CD6, CD7, CD8, CD10, CD11a, CD11b, CD11c, CD13, CD14, CD15, CD16, CD18, CD19, CD20, CD21, CD22, CD23, CD25, CD27L, CD28, CD29, CD30, CD30L, CD32, CD33 (p67 proteins), CD34, CD38, CD40, CD40L, CD44, CD45, CD46, CD49a, CD52, CD54, CD55, CD56, CD61, CD64, CD66e, CD74, CD80 (B7-1), CD89, CD95, CD123, CD137, CD138, CD140a, CD146, CD147, CD148, CD152, CD164, CEACAMS, CFTR, cGMP, CINC, Clostridium botulinum toxin, Clostridium perfringens toxin, CKb8-1, CLC, CMV, CMV UL, CNTF, CNTN-1, COX, C-Ret, CRG-2, CT-1, CTACK, CTGF, CTLA-4, CX3CL1, CX3CR1, CXCL, CXCL1, CXCL2, CXCL3, CXCL4, CXCL5, CXCL6, CXCL7, CXCL8, CXCL9, CXCL10, CXCL11, CXCL12, CXCL13, CXCL14, CXCL15, CXCL16, CXCR, CXCR1, CXCR2, CXCR3, CXCR4, CXCR5, CXCR6, cytokeratin tumor-associated antigen, DAN, DCC, DcR3, DC-SIGN, Decay accelerating factor, des(1-3)-IGF-I (brain IGF-1), Dhh, digoxin, DNAM-1, Dnase, Dpp, DPPIV/CD26, Dtk, ECAD, EDA, EDA-A1, EDA-A2, EDAR, EGF, EGFR (ErbB-1), EMA, EMMPRIN, ENA, endothelin receptor, Enkephalinase, eNOS, Eot, eotaxinl, EpCAM, Ephrin B2/EphB4, EPO, ERCC, E-selectin, ET-1, Factor IIa, Factor VII, Factor VIIIc, Factor IX, fibroblast activation protein (FAP), Fas, FcR1, FEN-1, Ferritin, FGF, FGF-19, FGF-2, FGF3, FGF-8, FGFR, FGFR-3, Fibrin, FL, FLIP, Flt-3, Flt-4, Follicle stimulating hormone, Fractalkine, FZD1, FZD2, FZD3, FZD4, FZD5, FZD6, FZD7, FZD8, FZD9, FZD10, G250, Gas 6, GCP-2, GCSF, GD2, GD3, GDF, GDF-1, GDF-3 (Vgr-2), GDF-5 (BMP-14, CDMP-1), GDF-6 (BMP-13, CDMP-2), GDF-7 (BMP-12, CDMP-3), GDF-8 (Myostatin), GDF-9, GDF-15 (MIC-1), GDNF, GDNF, GFAP, GFRa-1, GFR-alpha1, GFR-alpha2, GFR-alpha3, GITR, Glucagon, Glut 4, glycoprotein IIb/IIIa (GP IIb/IIIa), GM-CSF, gp130, gp72, GRO, Growth hormone releasing factor, Hapten (NP-cap or NIP-cap), HB-EGF, HCC, HCMV gB envelope glycoprotein, HCMV) gH envelope glycoprotein, HCMV UL, Hemopoietic growth factor (HGF), Hep B gp120, heparanase, Her2, Her2/neu (ErbB-2), Her3 (ErbB-3), Her4 (ErbB-4), herpes simplex virus (HSV) gB glycoprotein, HSV gD glycoprotein, HGFA, High molecular weight melanoma-associated antigen (HMW-MAA), HIV gp120, HIV IIIB gp 120 V3 loop, HLA, HLA-DR, HM1.24, HMFG PEM, HRG, Hrk, human cardiac myosin, human cytomegalovirus (HCMV), human growth hormone (HGH), HVEM, 1-309, IAP, ICAM, ICAM-1, ICAM-3, ICE, ICOS, IFNg, Ig, IgA receptor, IgE, IGF, IGF binding proteins, IGF-1R, IGFBP, IGF-I, IGF-II, IL, IL-1, IL-1R, IL-2, IL-2R, IL-4, IL-4R, IL-5, IL-5R, IL-6, IL-6R, IL-8, IL-9, IL-10, IL-12, IL-13, IL-15, IL-18, IL-18R, IL-23, interferon (INF)-alpha, INF-beta, INF-gamma, Inhibin, iNOS, Insulin A-chain, Insulin B-chain, Insulin-like growth factor 1, integrin alpha2, integrin alpha3, integrin alpha4, integrin alpha4/beta1, integrin alpha4/beta7, integrin alpha5 (alphaV), integrin alpha5/beta1, integrin alpha5/beta3, integrin alpha6, integrin beta1, integrin beta2, interferon gamma, IP-10, I-TAC, JE, Kallikrein 2, Kallikrein 5, Kallikrein 6, Kallikrein 11, Kallikrein 12, Kallikrein 14, Kallikrein 15, Kallikrein L1, Kallikrein L2, Kallikrein L3, Kallikrein L4, KC, KDR, Keratinocyte Growth Factor (KGF), laminin 5, LAMP, LAP, LAP (TGF-1), Latent TGF-1, Latent TGF-1 bp1, LBP, LDGF, LECT2, Lefty, Lewis-Y antigen, Lewis-Y related antigen, LFA-1, LFA-3, Lfo, LIF, LIGHT, lipoproteins, LIX, LKN, Lptn, L-Selectin, LT-a, LT-b, LTB4, LTBP-1, Lung surfactant, Luteinizing hormone, Lymphotoxin Beta Receptor, Mac-1, MAdCAM, MAG, MAP2, MARC, MCAM, MCAM, MCK-2, MCP, M-CSF, MDC, Mer, METALLOPROTEASES, MGDF receptor, MGMT, MHC (HLA-DR), MIF, MIG, MIP, MIP-1-alpha, MK, MMAC1, MMP, MMP-1, MMP-10, MMP-11, MMP-12, MMP-13, MMP-14, MMP-15, MMP-2, MMP-24, MMP-3, MMP-7, MMP-8, MMP-9, MPIF, Mpo, MSK, MSP, mucin (Mucl), MUC18, Muellerian-inhibitin substance, Mug, MuSK, NAIP, NAP, NCAD, N-Cadherin, NCA 90, NCAM, NCAM, Neprilysin, Neurotrophin-3, -4, or -6, Neurturin, Neuronal growth factor (NGF), NGFR, NGF-beta, nNOS, NO, NOS, Npn, NRG-3, NT, NTN, OB, OGG1, OPG, OPN, OSM, OX40L, OX40R, p150, p95, PADPr, Parathyroid hormone, PARC, PARP, PBR, PBSF, PCAD, P-Cadherin, PCNA, PDGF, PDGF, PDK-1, PECAM, PEM, PF4, PGE, PGF, PGI2, PGJ2, PIN, PLA2, placental alkaline phosphatase (PLAP), P1GF, PLP, PP14, Proinsulin, Prorelaxin, Protein C, PS, PSA, PSCA, prostate specific membrane antigen (PSMA), PTEN, PTHrp, Ptk, PTN, R51, RANK, RANKL, RANTES, RANTES, Relaxin A-chain, Relaxin B-chain, renin, respiratory syncytial virus (RSV) F, RSV Fgp, Ret, Rheumatoid factors, RLIP76, RPA2, RSK, S100, SCF/KL, SDF-1, SERINE, Serum albumin, sFRP-3, Shh, SIGIRR, SK-1, SLAM, SLPI, SMAC, SMDF, SMOH, SOD, SPARC, Stat, STEAP, STEAP-II, TACE, TACI, TAG-72 (tumor-associated glycoprotein-72), TARC, TCA-3, T-cell receptors (e.g., T-cell receptor alpha/beta), TdT, TECK, TEM1, TEMS, TEM7, TEM8, TERT, testicular PLAP-like alkaline phosphatase, TfR, TGF, TGF-alpha, TGF-beta, TGF-beta Pan Specific, TGF-beta RI (ALK-5), TGF-beta RII, TGF-beta RIIb, TGF-beta RIII, TGF-beta1, TGF-beta2, TGF-beta3, TGF-beta4, TGF-beta5, Thrombin, Thymus Ck-1, Thyroid stimulating hormone, Tie, TIMP, TIQ, Tissue Factor, TMEFF2, Tmpo, TMPRSS2, TNF, TNF-alpha, TNF-alpha beta, TNF-beta2, TNFc, TNF-RI, TNF-RII, TNFRSF10A (TRAIL R1 Apo-2, DR4), TNFRSF10B (TRAIL R2 DR5, KILLER, TRICK-2A, TRICK-B), TNFRSF10C (TRAIL R3 DcR1, LIT, TRID), TNFRSF10D (TRAIL R4 DcR2, TRUNDD), TNFRSF11A (RANK ODF R, TRANCE R), TNFRSF11B (OPG OCIF, TR1), TNFRSF12 (TWEAK R FN14), TNFRSF13B (TACI), TNFRSF13C (BAFF R), TNFRSF14 (HVEM ATAR, HveA, LIGHT R, TR2), TNFRSF16 (NGFR p75NTR), TNFRSF17 (BCMA), TNFRSF18 (GITR AITR), TNFRSF19 (TROY TAJ, TRADE), TNFRSF19L (RELT), TNFRSF1A (TNF RI CD120a, p55-60), TNFRSF1B (TNF RII CD120b, p75-80), TNFRSF26 (TNFRH3), TNFRSF3 (LTbR TNF RIII, TNFC R), TNFRSF4 (OX40 ACT35, TXGP1 R), TNFRSF5 (CD40 p50), TNFRSF6 (Fas Apo-1, APT1, CD95), TNFRSF6B (DcR3 M68, TR6), TNFRSF7 (CD27), TNFRSF8 (CD30), TNFRSF9 (4-1BB CD137, ILA), TNFRSF21 (DR6), TNFRSF22 (DcTRAIL R2 TNFRH2), TNFRST23 (DcTRAIL R1 TNFRH1), TNFRSF25 (DR3 Apo-3, LARD, TR-3, TRAMP, WSL-1), TNFSF10 (TRAIL Apo-2 Ligand, TL2), TNFSF11 (TRANCE/RANK Ligand ODF, OPG Ligand), TNFSF12 (TWEAK Apo-3 Ligand, DR3 Ligand), TNFSF13 (APRIL TALL2), TNFSF13B (BAFF BLYS, TALL1, THANK, TNFSF20), TNFSF14 (LIGHT HVEM Ligand, LTg), TNFSF15 (TL1A/VEGI), TNFSF18 (GITR Ligand AITR Ligand, TL6), TNFSF1A (TNF-a Conectin, DIF, TNFSF2), TNFSF1B (TNF-b LTa, TNFSF1), TNFSF3 (LTb TNFC, p33), TNFSF4 (OX40 Ligand gp34, TXGP1), TNFSF5 (CD40 Ligand CD154, gp39, HIGM1, IMD3, TRAP), TNFSF6 (Fas Ligand Apo-1 Ligand, APT1 Ligand), TNFSF7 (CD27 Ligand CD70), TNFSF8 (CD30 Ligand CD153), TNFSF9 (4-1BB Ligand CD137 Ligand), TP-1, t-PA, Tpo, TRAIL, TRAIL R, TRAIL-R1, TRAIL-R2, TRANCE, transferring receptor, TRF, Trk, TROP-2, TSG, TSLP, tumor-associated antigen CA 125, tumor-associated antigen expressing Lewis Y related carbohydrate, TWEAK, TXB2, Ung, uPAR, uPAR-1, Urokinase, VCAM, VCAM-1, VECAD, VE-Cadherin, VE-cadherin-2, VEFGR-1 (flt-1), VEGF, VEGFR, VEGFR-3 (flt-4), VEGI, VIM, Viral antigens, VLA, VLA-1, VLA-4, VNR integrin, von Willebrands factor, WIF-1, WNT1, WNT2, WNT2B/13, WNT3, WNT3A, WNT4, WNTSA, WNTSB, WNT6, WNT7A, WNT7B, WNT8A, WNT8B, WNT9A, WNT9A, WNT9B, WNT10A, WNT10B, WNT11, WNT16, XCL1, XCL2, XCR1, XCR1, XEDAR, XIAP, XPD, and receptors for hormones and growth factors.

[0345] Exemplary antigens that may be targeted specifically by the immunoglobulins of the invention include but are not limited to: CD20, CD19, Her2, EGFR, EpCAM, CD3, FcγRIIIa (CD16), FcγRIIa (CD32a), FcγRIIb (CD32b), FcγRI (CD64), Toll-like receptors (TLRs) such as TLR4 and TLR9, cytokines such as IL-2, IL-5, IL-13, IL-12, IL-23, and TNFα, cytokine receptors such as IL-2R, chemokines, chemokine receptors, growth factors such as VEGF and HGF, and the like. To form the bispecific antibodies of the invention, antibodies to any combination of these antigens can be made; that is, each of these antigens can be optionally and independently included or excluded from a bispecific antibody according to the present invention.

[0346] Particularly preferred combinations for bispecific antibodies are an antigen-binding domain to CD8 and an antigen binding domain selected from a domain that binds CD19, CD20, CD38 and CD123, the sequences of which are shown in the Figures.

[0347] The invention further provides nucleic acid compositions encoding the bispecific antibodies of the invention. As will be appreciated by those in the art, the nucleic acid compositions will depend on the format and scaffold of the heterodimeric protein. Thus, for example, when the format requires three amino acid sequences, such as for the triple F format (e.g. a first amino acid monomer comprising an Fc domain and a scFv, a second amino acid monomer comprising a heavy chain and a light chain), three nucleic acid sequences can be incorporated into one or more expression vectors for expression. Similarly, some formats (e.g. dual scFv formats such as disclosed in FIG. 1) only two nucleic acids are needed; again, they can be put into one or two expression vectors.

[0348] As is known in the art, the nucleic acids encoding the components of the invention can be incorporated into expression vectors as is known in the art, and depending on the host cells used to produce the heterodimeric antibodies of the invention. Generally the nucleic acids are operably linked to any number of regulatory elements (promoters, origin of replication, selectable markers, ribosomal binding sites, inducers, etc.). The expression vectors can be extra-chromosomal or integrating vectors.

[0349] The nucleic acids and/or expression vectors of the invention are then transformed into any number of different types of host cells as is well known in the art, including mammalian, bacterial, yeast, insect and/or fungal cells, with mammalian cells (e.g. CHO cells), finding use in many embodiments.

[0350] In some embodiments, nucleic acids encoding each monomer and the optional nucleic acid encoding a light chain, as applicable depending on the format, are each contained within a single expression vector, generally under different or the same promoter controls. In embodiments of particular use in the present invention, each of these two or three nucleic acids are contained on a different expression vector. As shown herein and in 62/025,931, hereby incorporated by reference, different vector ratios can be used to drive heterodimer formation. That is, surprisingly, while the proteins comprise first monomer:second monomer:light chains (in the case of many of the embodiments herein that have three polypeptides comprising the heterodimeric antibody) in a 1:1:2 ratio, these are not the ratios that give the best results.

[0351] The heterodimeric antibodies of the invention are made by culturing host cells comprising the expression vector(s) as is well known in the art. Once produced, traditional antibody purification steps are done, including an ion exchange chromotography step. As discussed herein, having the pIs of the two monomers differ by at least 0.5 can allow separation by ion exchange chromatography or isoelectric focusing, or other methods sensitive to isoelectric point. That is, the inclusion of pI substitutions that alter the isoelectric point (pI) of each monomer so that such that each monomer has a different pI and the heterodimer also has a distinct pI, thus facilitating isoelectric purification of the “triple F” heterodimer (e.g., anionic exchange columns, cationic exchange columns). These substitutions also aid in the determination and monitoring of any contaminating dual scFv-Fc and mAb homodimers post-purification (e.g., IEF gels, cIEF, and analytical IEX columns).

[0352] Once made, the compositions of the invention find use in a number of applications. CD20, CD38 and CD123 are all unregulated in many hematopoeitic malignancies and in cell lines derived from various hematopoietic malignancies, accordingly, the heterodimeric antibodies of the invention find use in treating cancer, including but not limited to, all B cell lymphomas and leukemias, including but not limited to non-Hodgkin's lymphoma (NHL), Burkitt's lymphoma (BL), multiple myeloma (MM), B chronic lymphocytic leukemia (B-CLL), B and T acute lymphocytic leukemia (ALL), T cell lymphoma (TCL), acute myeloid leukemia (AML), hairy cell leukemia (HCL), Hodgkin's Lymphoma (HL), chronic lymphocytic leukemia (CLL), non-Hodgkin's lymphoma, and chronic myeloid leukemia (CML).

[0353] Accordingly, the heterodimeric compositions of the invention find use in the treatment of these cancers.

[0354] Formulations of the antibodies used in accordance with the present invention are prepared for storage by mixing an antibody having the desired degree of purity with optional pharmaceutically acceptable carriers, excipients or stabilizers (Remington's Pharmaceutical Sciences 16th edition, Osol, A. Ed. [1980]), in the form of lyophilized formulations or aqueous solutions. Acceptable carriers, excipients, or stabilizers are nontoxic to recipients at the dosages and concentrations employed, and include buffers such as phosphate, citrate, and other organic acids; antioxidants including ascorbic acid and methionine; preservatives (such as octadecyldimethylbenzyl ammonium chloride; hexamethonium chloride; benzalkonium chloride, benzethonium chloride; phenol, butyl or benzyl alcohol; alkyl parabens such as methyl or propyl paraben; catechol; resorcinol; cyclohexanol; 3-pentanol; and m-cresol); low molecular weight (less than about 10 residues) polypeptides; proteins, such as serum albumin, gelatin, or immunoglobulins; hydrophilic polymers such as polyvinylpyrrolidone; amino acids such as glycine, glutamine, asparagine, histidine, arginine, or lysine; monosaccharides, disaccharides, and other carbohydrates including glucose, mannose, or dextrins; chelating agents such as EDTA; sugars such as sucrose, mannitol, trehalose or sorbitol; salt-forming counter-ions such as sodium; metal complexes (e.g. Zn-protein complexes); and/or non-ionic surfactants such as TWEEN™, PLURONICS™ or polyethylene glycol (PEG).

[0355] The formulation herein may also contain more than one active compound as necessary for the particular indication being treated, preferably those with complementary activities that do not adversely affect each other. For example, it may be desirable to provide antibodies with other specificities. Alternatively, or in addition, the composition may comprise a cytotoxic agent, cytokine, growth inhibitory agent and/or small molecule antagonist. Such molecules are suitably present in combination in amounts that are effective for the purpose intended.

[0356] The active ingredients may also be entrapped in microcapsules prepared, for example, by coacervation techniques or by interfacial polymerization, for example, hydroxymethylcellulose or gelatin-microcapsules and poly-(methylmethacylate) microcapsules, respectively, in colloidal drug delivery systems (for example, liposomes, albumin microspheres, microemulsions, nano-particles and nanocapsules) or in macroemulsions. Such techniques are disclosed in Remington's Pharmaceutical Sciences 16th edition, Osol, A. Ed. (1980).

[0357] The formulations to be used for in vivo administration should be sterile, or nearly so. This is readily accomplished by filtration through sterile filtration membranes.

[0358] Sustained-release preparations may be prepared. Suitable examples of sustained-release preparations include semipermeable matrices of solid hydrophobic polymers containing the antibody, which matrices are in the form of shaped articles, e.g. films, or microcapsules. Examples of sustained-release matrices include polyesters, hydrogels (for example, poly(2-hydroxyethyl-methacrylate), or poly(vinylalcohol)), polylactides (U.S. Pat. No. 3,773,919), copolymers of L-glutamic acid and .gamma. ethyl-L-glutamate, non-degradable ethylene-vinyl acetate, degradable lactic acid-glycolic acid copolymers such as the LUPRON DEPOT™ (injectable microspheres composed of lactic acid-glycolic acid copolymer and leuprolide acetate), and poly-D-(−)-3-hydroxybutyric acid. While polymers such as ethylene-vinyl acetate and lactic acid-glycolic acid enable release of molecules for over 100 days, certain hydrogels release proteins for shorter time periods.

[0359] When encapsulated antibodies remain the body for a long time, they may denature or aggregate as a result of exposure to moisture at 37° C., resulting in a loss of biological activity and possible changes in immunogenicity. Rational strategies can be devised for stabilization depending on the mechanism involved. For example, if the aggregation mechanism is discovered to be intermolecular S—S bond formation through thio-disulfide interchange, stabilization may be achieved by modifying sulfhydryl residues, lyophilizing from acidic solutions, controlling moisture content, using appropriate additives, and developing specific polymer matrix compositions.

[0360] The antibodies and chemotherapeutic agents of the invention are administered to a subject, in accord with known methods, such as intravenous administration as a bolus or by continuous infusion over a period of time, by intramuscular, intraperitoneal, intracerobrospinal, subcutaneous, intra-articular, intrasynovial, intrathecal, oral, topical, or inhalation routes. Intravenous or subcutaneous administration of the antibody is preferred.

[0361] In the methods of the invention, therapy is used to provide a positive therapeutic response with respect to a disease or condition. By “positive therapeutic response” is intended an improvement in the disease or condition, and/or an improvement in the symptoms associated with the disease or condition. For example, a positive therapeutic response would refer to one or more of the following improvements in the disease: (1) a reduction in the number of neoplastic cells; (2) an increase in neoplastic cell death; (3) inhibition of neoplastic cell survival; (5) inhibition (i.e., slowing to some extent, preferably halting) of tumor growth; (6) an increased patient survival rate; and (7) some relief from one or more symptoms associated with the disease or condition.

[0362] Positive therapeutic responses in any given disease or condition can be determined by standardized response criteria specific to that disease or condition. Tumor response can be assessed for changes in tumor morphology (i.e., overall tumor burden, tumor size, and the like) using screening techniques such as magnetic resonance imaging (MRI) scan, x-radiographic imaging, computed tomographic (CT) scan, bone scan imaging, endoscopy, and tumor biopsy sampling including bone marrow aspiration (BMA) and counting of tumor cells in the circulation.

[0363] In addition to these positive therapeutic responses, the subject undergoing therapy may experience the beneficial effect of an improvement in the symptoms associated with the disease.

[0364] An improvement in the disease may be characterized as a complete response. By “complete response” is intended an absence of clinically detectable disease with normalization of any previously abnormal radiographic studies, bone marrow, and cerebrospinal fluid (CSF) or abnormal monoclonal protein the case of myeloma.

[0365] Such a response may persist for at least 4 to 8 weeks, or sometimes 6 to 8 weeks, following treatment according to the methods of the invention. Alternatively, an improvement in the disease may be categorized as being a partial response. By “partial response” is intended at least about a 50% decrease in all measurable tumor burden (i.e., the number of malignant cells present in the subject, or the measured bulk of tumor masses or the quantity of abnormal monoclonal protein) in the absence of new lesions, which may persist for 4 to 8 weeks, or 6 to 8 weeks.

[0366] Treatment according to the present invention includes a “therapeutically effective amount” of the medicaments used. A “therapeutically effective amount” refers to an amount effective, at dosages and for periods of time necessary, to achieve a desired therapeutic result.

[0367] A therapeutically effective amount may vary according to factors such as the disease state, age, sex, and weight of the individual, and the ability of the medicaments to elicit a desired response in the individual. A therapeutically effective amount is also one in which any toxic or detrimental effects of the antibody or antibody portion are outweighed by the therapeutically beneficial effects.

[0368] A “therapeutically effective amount” for tumor therapy may also be measured by its ability to stabilize the progression of disease. The ability of a compound to inhibit cancer may be evaluated in an animal model system predictive of efficacy in human tumors.

[0369] Alternatively, this property of a composition may be evaluated by examining the ability of the compound to inhibit cell growth or to induce apoptosis by in vitro assays known to the skilled practitioner. A therapeutically effective amount of a therapeutic compound may decrease tumor size, or otherwise ameliorate symptoms in a subject. One of ordinary skill in the art would be able to determine such amounts based on such factors as the subject's size, the severity of the subject's symptoms, and the particular composition or route of administration selected.

[0370] Dosage regimens are adjusted to provide the optimum desired response (e.g., a therapeutic response). For example, a single bolus may be administered, several divided doses may be administered over time or the dose may be proportionally reduced or increased as indicated by the exigencies of the therapeutic situation. Parenteral compositions may be formulated in dosage unit form for ease of administration and uniformity of dosage. Dosage unit form as used herein refers to physically discrete units suited as unitary dosages for the subjects to be treated; each unit contains a predetermined quantity of active compound calculated to produce the desired therapeutic effect in association with the required pharmaceutical carrier.

[0371] The specification for the dosage unit forms of the present invention are dictated by and directly dependent on (a) the unique characteristics of the active compound and the particular therapeutic effect to be achieved, and (b) the limitations inherent in the art of compounding such an active compound for the treatment of sensitivity in individuals.

[0372] The efficient dosages and the dosage regimens for the bispecific antibodies used in the present invention depend on the disease or condition to be treated and may be determined by the persons skilled in the art.

[0373] An exemplary, non-limiting range for a therapeutically effective amount of an bispecific antibody used in the present invention is about 0.1-100 mg/kg, such as about 0.1-50 mg/kg, for example about 0.1-20 mg/kg, such as about 0.1-10 mg/kg, for instance about 0.5, about such as 0.3, about 1, or about 3 mg/kg. In another embodiment, he antibody is administered in a dose of 1 mg/kg or more, such as a dose of from 1 to 20 mg/kg, e.g. a dose of from 5 to 20 mg/kg, e.g. a dose of 8 mg/kg.

[0374] A medical professional having ordinary skill in the art may readily determine and prescribe the effective amount of the pharmaceutical composition required. For example, a physician or a veterinarian could start doses of the medicament employed in the pharmaceutical composition at levels lower than that required in order to achieve the desired therapeutic effect and gradually increase the dosage until the desired effect is achieved.

[0375] In one embodiment, the bispecific antibody is administered by infusion in a weekly dosage of from 10 to 500 mg/kg such as of from 200 to 400 mg/kg Such administration may be repeated, e.g., 1 to 8 times, such as 3 to 5 times. The administration may be performed by continuous infusion over a period of from 2 to 24 hours, such as of from 2 to 12 hours.

[0376] In one embodiment, the bispecific antibody is administered by slow continuous infusion over a long period, such as more than 24 hours, if required to reduce side effects including toxicity.

[0377] In one embodiment the bispecific antibody is administered in a weekly dosage of from 250 mg to 2000 mg, such as for example 300 mg, 500 mg, 700 mg, 1000 mg, 1500 mg or 2000 mg, for up to 8 times, such as from 4 to 6 times. The administration may be performed by continuous infusion over a period of from 2 to 24 hours, such as of from 2 to 12 hours. Such regimen may be repeated one or more times as necessary, for example, after 6 months or 12 months. The dosage may be determined or adjusted by measuring the amount of compound of the present invention in the blood upon administration by for instance taking out a biological sample and using anti-idiotypic antibodies which target the antigen binding region of the bispecific antibody.

[0378] In a further embodiment, the bispecific antibody is administered once weekly for 2 to 12 weeks, such as for 3 to 10 weeks, such as for 4 to 8 weeks.

[0379] In one embodiment, the bispecific antibody is administered by maintenance therapy, such as, e.g., once a week for a period of 6 months or more.

[0380] In one embodiment, the bispecific antibody is administered by a regimen including one infusion of an bispecific antibody followed by an infusion of an bispecific antibody conjugated to a radioisotope. The regimen may be repeated, e.g., 7 to 9 days later.

[0381] As non-limiting examples, treatment according to the present invention may be provided as a daily dosage of an antibody in an amount of about 0.1-100 mg/kg, such as 0.5, 0.9, 1.0, 1.1, 1.5, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 40, 45, 50, 60, 70, 80, 90 or 100 mg/kg, per day, on at least one of day 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, or 40, or alternatively, at least one of week 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 or 20 after initiation of treatment, or any combination thereof, using single or divided doses of every 24, 12, 8, 6, 4, or 2 hours, or any combination thereof.

[0382] In some embodiments the bispecific antibody molecule thereof is used in combination with one or more additional therapeutic agents, e.g. a chemotherapeutic agent. Non-limiting examples of DNA damaging chemotherapeutic agents include topoisomerase I inhibitors (e.g., irinotecan, topotecan, camptothecin and analogs or metabolites thereof, and doxorubicin); topoisomerase II inhibitors (e.g., etoposide, teniposide, and daunorubicin); alkylating agents (e.g., melphalan, chlorambucil, busulfan, thiotepa, ifosfamide, carmustine, lomustine, semustine, streptozocin, decarbazine, methotrexate, mitomycin C, and cyclophosphamide); DNA intercalators (e.g., cisplatin, oxaliplatin, and carboplatin); DNA intercalators and free radical generators such as bleomycin; and nucleoside mimetics (e.g., 5-fluorouracil, capecitibine, gemcitabine, fludarabine, cytarabine, mercaptopurine, thioguanine, pentostatin, and hydroxyurea).

[0383] Chemotherapeutic agents that disrupt cell replication include: paclitaxel, docetaxel, and related analogs; vincristine, vinblastin, and related analogs; thalidomide, lenalidomide, and related analogs (e.g., CC-5013 and CC-4047); protein tyrosine kinase inhibitors (e.g., imatinib mesylate and gefitinib); proteasome inhibitors (e.g., bortezomib); NF-κB inhibitors, including inhibitors of IκB kinase; antibodies which bind to proteins overexpressed in cancers and thereby downregulate cell replication (e.g., trastuzumab, rituximab, cetuximab, and bevacizumab); and other inhibitors of proteins or enzymes known to be upregulated, over-expressed or activated in cancers, the inhibition of which downregulates cell replication.

[0384] In some embodiments, the antibodies of the invention can be used prior to, concurrent with, or after treatment with Velcade® (bortezomib).

[0385] All cited references are herein expressly incorporated by reference in their entirety.

[0386] Whereas particular embodiments of the invention have been described above for purposes of illustration, it will be appreciated by those skilled in the art that numerous variations of the details may be made without departing from the invention as described in the appended claims.

[0387] Examples are provided below to illustrate the present invention. These examples are not meant to constrain the present invention to any particular application or theory of operation. For all constant region positions discussed in the present invention, numbering is according to the EU index as in Kabat (Kabat et al., 1991, Sequences of Proteins of Immunological Interest, 5th Ed., United States Public Health Service, National Institutes of Health, Bethesda, entirely incorporated by reference). Those skilled in the art of antibodies will appreciate that this convention consists of nonsequential numbering in specific regions of an immunoglobulin sequence, enabling a normalized reference to conserved positions in immunoglobulin families. Accordingly, the positions of any given immunoglobulin as defined by the EU index will not necessarily correspond to its sequential sequence.

[0388] Antibody constant chains were modified with lower pI by engineering substitutions in the constant domains. Reduced pI can be engineered by making substitutions of basic amino acids (K or R) to acidic amino acids (D or E), which result in the largest decrease in pI. Mutations of basic amino acids to neutral amino acids and neutral amino acids to acidic amino acids will also result in a decrease in pI. A list of amino acid pK values can be found in Table 1 of Bjellqvist et al., 1994, Electrophoresis 15:529-539.

[0389] We chose to explore substitutions in the antibody CH1 (Cγ1) and CL (Ckappa or CK) regions (sequences are shown in FIG. 13of U.S. Ser. No. 14/216,705, incorporated by reference) because, unlike the Fc region, they do not interact with native ligands that impact the antibody's pharmacological properties. In deciding which positions to mutate, the surrounding environment and number of contacts the WT amino acid makes with its neighbors was taken into account such as to minimize the impact of a substitution or set of substitutions on structure and/or function. The solvent accessibility or fraction exposed of each CH1 and CK position was calculated using relevant crystal structures of antibody Fab domains. The results are shown in FIGS. 2 and 3 of U.S. Ser. No. 13/648,951 for the Cγ1 and CK respectively (Figures and accompanying legends are expressly incorporated herein by reference). Design was guided further by examining the CH1 and CL domains for positions that are isotypic between the immunoglobulin isotypes (IgG1, IgG2, IgG3, and IgG4). Because such variations occur naturally, such positions are expected to be amenable to substitution. Based on this analysis, a number of substitutions were identified that reduce pI but are predicted to have minimal impact on the biophysical properties of the domains.

[0390] As for all the bispecific antibodies herein, genes encoding the heavy and light chains of the antibodies were constructed in the mammalian expression vector pTT5. The human IgG1 constant chain gene was obtained from IMAGE clones and subcloned into the pTT5 vector. VH and VL genes encoding the anti-VEGF antibodies were synthesized commercially (Blue Heron Biotechnologies, Bothell Wash.), and subcloned into the vectors encoding the appropriate CL and IgG1 constant chains. Amino acid modifications were constructed using site-directed mutagenesis using the QuikChange® site-directed mutagenesis methods (Stratagene, La Jolla Calif.). All DNA was sequenced to confirm the fidelity of the sequences.

[0391] Plasmids containing heavy chain gene (VH-Cγ1-Cγ2-Cγ3) were co-transfected with plasmid containing light chain gene (VL-Cκ) into 293E cells using lipofectamine (Invitrogen, Carlsbad Calif.) and grown in FreeStyle 293 media (Invitrogen, Carlsbad Calif.). After 5 days of growth, the antibodies were purified from the culture supernatant by protein A affinity using the MabSelect resin (GE Healthcare). Antibody concentrations were determined by bicinchoninic acid (BCA) assay (Pierce).

[0392] The pI engineered mAbs were generally characterized by SDS PAGE on an Agilent Bioanalyzer, by size exclusion chromatography (SEC), isoelectric focusing (IEF) gel electrophoresis, binding to antigen by Biacore, and differential scanning calorimetry (DSC). All mAbs showed high purity on SDS-PAGE and SEC. IEF gels indicated that each variant had the designed isoelectric point. Generally the binding analysis on Biacore showed that pI engineered variants bound to antigen with similar affinity as the parent antibodies, indicating that the designed substitutions did not perturb the function of the mAb. DSC in the Figures of of U.S. Ser. No. 14/216,705, incorporated by reference, show which variants generally had high thermostability.

[0393] Pharmacokinetic experiments for serum half life as appropriate were performed in B6 mice that are homozygous knock-outs for murine FcRn and heterozygous knock-ins of human FcRn (mFcRn−/−, hFcRn+) (Petkova et al., 2006, Int Immunol 18(12):1759-69, entirely incorporated by reference), herein referred to as hFcRn or hFcRn+ mice.

[0394] A single, intravenous tail vein injection of antibody (2 mg/kg) was given to groups of 4-7 female mice randomized by body weight (20-30 g range). Blood (˜50 ul) was drawn from the orbital plexus at each time point, processed to serum, and stored at −80° C. until analysis. Antibody concentrations were determined using an ELISA assay. Serum concentration of antibody was measured using recombinant antigen as capture reagent, and detection was carried out with biotinylated anti-human kappa antibody and europium-labeled streptavidin. The time resolved fluorescence signal was collected. PK parameters were determined for individual mice with a non-compartmental model using WinNonLin (Pharsight Inc, Mountain View Calif.). Nominal times and dose were used with uniform weighing of points.

[0395] Reduction in the pI of a protein or antibody can be carried out using a variety of approaches. At the most basic level, residues with high pKa's (lysine, arginine, and to some extent histidine) are replaced with neutral or negative residues, and/or neutral residues are replaced with low pKa residues (aspartic acid and glutamic acid). The particular replacements may depend on a variety of factors, including location in the structure, role in function, and immunogenicity.

[0396] Because immunogenicity is a concern, efforts can be made to minimize the risk that a substitution that lowers the pI will elicit immunogenicity. One way to minimize risk is to minimize the mutational load of the variants, i.e. to reduce the pI with the fewest number of mutations. Charge swapping mutations, where a K, R, or H is replaced with a D or E, have the greatest impact on reducing pI, and so these substitutions are preferred. Another approach to minimizing the risk of immunogenicity while reducing pI is to utilize substitutions from homologous human proteins. Thus for antibody constant chains, the isotypic differences between the IgG subclasses (IgG1, IgG2, IgG3, and IgG4) provide low-risk substitutions. Because immune recognition occurs at a local sequence level, i.e. MHC II and T-cell receptors recognize epitopes typically 9 residues in length, pI-altering substitutions may be accompanied by isotypic substitutions proximal in sequence. In this way, epitopes can be extended to match a natural isotype. Such substitutions would thus make up epitopes that are present in other human IgG isotypes, and thus would be expected to be tolerized.

[0397] One approach for engineering changes in pI is to use isotype switching, as described herein.

[0398] Another approach to engineering lower pI into proteins and antibodies is to fuse negatively charged residues to the N- or C-termini. Thus for example, peptides consisting principally of aspartic acids and glutamic acid may be fused to the N-terminus or C-terminus to the antibody heavy chain, light chain or both. Because the N-termini are structurally close to the antigen binding site, the C-termini are preferred.

[0399] Based on the described engineering approaches, a number of variants were designed to alter the isoelectric point of the antibody heavy chain (Fc region generally) and in some cases the light chain.

[0400] Homology between CK and Cλ is not as high as between the IgG subclasses, however the sequence and structural homology that exists was still used to guide substitutions to create an isotypic low-pI light chain constant region. In FIG. 56 of U.S. Ser. No. 14/216,705, incorporated by reference, positions with residues contributing to a higher pI (K, R, and H) or lower pI (D and E) are highlighted in bold. Gray indicates lysine, arginines, and histidines that may be substituted, preferably with aspartic or glutatmic acids, to lower the isoelectric point. These variants, alone or in any combination, can independently and optionally be combined with all other heavy chain variants in scaffolds that have at least one light chain.

[0401] Substitutions that modify the antibody isoelectric point may be introduced into one or more chains of an antibody variant to facilitate analysis and purification. For instance, heterodimeric antibodies such as those disclosed in US2011/0054151A1 can be purified by modifying the isolectric point of one chain, so that the multiple species present after expression and Protein A purification can be purified by methods that separate proteins based on differences in charge, such as ion exchange chromatography.

[0402] As an example, the heavy chain of bevacizumab was modified by introducing substitutions to lower its isolectric point such that the difference in charges between the three species produced when WT-IgG1-HC, low-pI-HC, and WT-LC are transfected in 293E cells is large enough to facilitate purification by anion exchange chromatography. Clones were created as described above, and transfection and initial purification by Protein A chromatography is also as described above. Sequences of the three chains “Heavy chain 1 of XENP10653”, “Heavy chain 2 of XENP10653”, and “Light chain of XENP10653” in the Figures of U.S. Ser. No. 14/216,705, incorporated by reference. After Protein A purification, three species with nearly identical molecular weights, but different charges are obtained. These are the WT-IgG1-HC/WT-IgG1-HC homodimer (pI=8.12), WT-IgG1-HC/low-pI-HC heterodimer (pI=6.89), and low-pI-HC/low-pI-HC homodimer (pI=6.20). The mixture was loaded onto a GE HiTrap Q HP column in 20 mM Tris, pH 7.6 and eluted with a step-wise gradient of NaCl consisting of 50 mM, 100 mM, and finally 200 mM NaCl in the same Tris buffer. Elution was monitored by A280, and each fraction analyzed on Invitrogen pH 3-10 IEF gels with Novex running buffer and these results are shown in FIG. 40 of U.S. Ser. No. 14/216,705, incorporated by reference. WT-IgG1-HC/WT-IgG1-HC homodimer does not bind to the anion exchange column at pH 7.6 and is thus present in the flowthrough and wash (lanes 1-2). The desired heterodimer elutes with 50 mM NaCl (lane 3), while the low-pI-HC/low-pI-HC homodimer binds tightest to the column and elutes at 100 (lane 4) and 200 mM (lane 5) NaCl. Thus the desired heterodimer variant, which is difficult to purify by other means because of its similar molecular weight to the other two species, is easily purified by the introduction of low pI substitutions into one chain. This method of purifying antibodies by engineering the isoelectric point of each chain can be applied to methods of purifying various bispecific antibody constructs. The method is particulary useful when the desired species in the mixture has similar molecular weight and other properties such that normal purification techniques are not capable of separating the desired species in high yield.

[0403] The pI of antibody constant chains were altered by engineering substitutions in the constant domains. Reduced pI can be engineered by making substitutions of basic amino acids (K or R) to acidic amino acids (D or E), which result in the largest decrease in pI. Mutations of basic amino acids to neutral amino acids and neutral amino acids to acidic amino acids will also result in a decrease in pI. Conversely, increased pI can be engineered by making substitutions of acidic amino acids (D or E) to basic amino acids (K or R), which result in the largest increase in pI. Mutations of acidic amino acids to neutral amino acids and neutral amino acids to basic amino acids will also result in a increase in pI. A list of amino acid pK values can be found in Table 1 of Bjellqvist et al., 1994, Electrophoresis 15:529-539.

[0404] In deciding which positions to mutate, the surrounding environment and number of contacts the WT amino acid makes with its neighbors was taken into account such as to minimize the impact of a substitution or set of substitutions on structure and/or function. The solvent accessibility or fraction exposed of each constant region position was calculated using relevant crystal structures. Based on this analysis, a number of substitutions were identified that reduce or increase pI but are predicted to have minimal impact on the biophysical properties of the domains.

[0405] Calculation of protein pI was performed as follows. First, a count was taken of the number of D, E, C, H, K, R, and Y amino acids as well as the number of N- and C-termini present in the protein. Then, the pI was calculated by identifying the pH for which the protein has an overall charge of zero. This was done by calculating the net charge of the protein at a number of test pH values. Test pH values were set in an iterative manner, stepping up from a low pH of 0 to a high pH of 14 by increments of 0.001 until the charge of the protein reached or surpassed zero. Net charge of a protein at a given pH was calculated by the following formula:

$q_{\mathrm{protein}}\left(\mathrm{pH}\right)=\sum _{i=H,K,R,N\mathrm{termini}}\frac{N_i}{1+{10}^{pH-{\mathrm{pK}}_i}}-\sum _{i=D,E,C,Y,\mathrm{Ctermini}}\frac{N_i}{1+{10}^{{\mathrm{pK}}_i-pH}}$

where q_protein(pH) is the net charge on the protein at the given pH, is the number of amino acid i (or N- or C-termini) present in the protein, and is the pK of amino acid i (or N- or C-termini).

[0406] Variants were first purified by Protein A, and then loaded onto a GE Healthcare HiTrap SP HP cation exchange column in 50 mM MES (pH 6.0) and eluted with an NaCl gradient. Following elution, fractions from each peak were loaded onto a Lonza IsoGel IEF plate (pH range 7-11) for analysis. Separation of the middle pI heterodimer is achieved in each case, with separation improved when the heterodimer has a larger difference in pI from the homodimers.

[0407] Differential scanning fluorimetry (DSF) was used to evaluate the stability of antibodies containing isosteric pI substitutions. DSF experiments were performed using a Bio-Rad CFX Connect Real-Time PCR Detection System. Proteins were mixed with SYPRO Orange fluorescent dye and diluted to 0.25 or 0.50 mg/mL in PBS. The final concentration of SYPRO Orange was 10×. After an initial 10 minute incubation period at 25° C., proteins were heated from 25 to 95° C. using a heating rate of 1° C./min. A fluorescence measurement was taken every 30 sec. Melting temperatures were calculated using the instrument software. The results are shown in FIG. 110 of U.S. Ser. No. 14/216,705, incorporated by reference. The results indicated that isosteric(+) pI variants had lower stability. We therefore made further variants to reduce the number of substitutions on the increased pI side, but results showed that only E269Q had a small effect on stability, while E272Q and E283Q had large negative impacts on stability.

[0408] We have previously engineered the antibody constant regions of heterodimeric antibodies to have higher or lower pI using both isotypic and isosteric charge substitutions. These methods enable efficient IEX purification of heterodimeric species, but may impact stability or immunogenicity of the antibodies due to the unnatural substitutions introduced. For scFv containing heterodimeric bispecific antibodies (Examples are shown in FIG. 84 of U.S. Ser. No. 14/216,705, incorporated by reference), another region to introduce charged substitutions is the scFv linker that connects the VH and VL of scFv constructs. The most common linker used is (GGGGS)3 (SEQ ID NO: 66) or (GGGGS)4 (SEQ ID NO: 77), which has been shown to be flexible enough to allow stable scFv formation without diabody formation. These sequences are already unnatural, and contain little sequence specificity for likely immunogenic epitopes. Therefore we thought that introducing charged substitutions into scFv linkers may be a good strategy to enable IEX purification of heterodimeric bispecific species containing scFvs. Various positively and negatively charged scFv linkers were designed and are shown in FIG. 20. All linkers are novel constructs except for the “Whitlow” linker which was reported by Whitlow et al., (Whitlow M, Protein Eng. 1993 (8), 989-995.). Linkers designated as 6paxA_1 (+A) and 3hsc_2 (−A) were taken from a database of unstructured regions in human proteins obtained from PDB files and these linkers are approximately the same length as (GGGGS)3 (SEQ ID NO: 66) and contain positive or negative charges. Other linkers are based on introducing repetitive Lys or Glu residues, as well as Lys-Pro motifs designed to reduce the chance of proteolytic degradation in the positively charged linkers.

[0409] Charged linkers were first evaluated for biophysical behavior in the scFv-His format and then were later constructed in anti-CD19×CD3 Fab-scFv-Fc bispecific format. Genes encoding the scFv of engineered forms of the anti-CD3 antibody SP34 or the anti-CD19 4G7 antibody were constructed in the mammalian expression vector pTT5. For full-length constructs, the human IgG1 constant chain gene was obtained from IMAGE clones and subcloned into the pTT5 vector. scFv genes were synthesized commercially (Blue Heron Biotechnologies, Bothell Wash. Amino acid modifications were constructed using site-directed mutagenesis using the QuikChange® site-directed mutagenesis methods (Stratagene, La Jolla Calif.). All DNA was sequenced to confirm the fidelity of the sequences.

[0410] Plasmids containing scFv or heavy chain and light chain genes were transfected (or co-transfected for full-length formats) into 293E cells using lipofectamine (Invitrogen, Carlsbad Calif.) and grown in FreeStyle 293 media (Invitrogen, Carlsbad Calif.). After 5 days of growth, the antibodies were purified from the culture supernatant by protein A (full-length) using the MabSelect resin (GE Healthcare) or using Ni-NTA resing for His-tagged scFvs. Heterodimers were further purified by ion exchange chromatograpy (IEX) to assess the ability of the altered pI heavy chains to enable efficient purification. Examples of IEX purifications for an anti-CD19×CD3 bispecific containing a positively charged linker in the CD3 scFv is shown in FIG. 90 of U.S. Ser. No. 14/216,705, incorporated by reference. Antibody concentrations were determined by bicinchoninic acid (BCA) assay (Pierce).

[0411] The pI engineered scFvs or antibodies were characterized by SDS-PAGE, size exclusion chromatography (SEC), isoelectric focusing (IEF) gel electrophoresis, and/or differential scanning fluorimetry (DSF).

[0412] Anti-CD3 scFv's and anti-CD19 scFv's containing positively or negatively charged linkers, respectively, were evaluated for SEC behavior as well as for stability using DSF. Differential scanning fluorimetry (DSF) was used to evaluate the stability of scFvs containing charged linkers. DSF experiments were performed using a Bio-Rad CFX Connect Real-Time PCR Detection System. Proteins were mixed with SYPRO Orange fluorescent dye and diluted to 0.25 or 0.50 mg/mL in PBS. The final concentration of SYPRO Orange was 10×. After an initial 10 minute incubation period at 25° C., proteins were heated from 25 to 95° C. using a heating rate of 1° C./min. A fluorescence measurement was taken every 30 sec. Melting temperatures were calculated using the instrument software. Tm values for scFvs are shown in FIG. 86 of U.S. Ser. No. 14/216,705, incorporated by reference. Charged linkers had only marginal impacts on overall scFv stability as indicated by their Tm values. SEC chromatograms obtained from purified scFvs are shown in FIG. 4 of U.S. Ser. No. 14/216,705, incorporated by reference. Highly charged linkers have a longer elution time and noticeable peak tails indicating that too much charge causes the scFvs to stick to the SEC resin longer than expected. Binding results for positively charged anti-CD3 scFvs binding to CD4+ T cells (FIG. 88 of U.S. Ser. No. 14/216,705, incorporated by reference) indicated that binding of most scFvs was similar, with the exception of the very highly charged (GKGKS)4 (SEQ ID NO: 76) scFv, which showed weaker binding. No off-target binding was detected when gating for CD20+ cells in PBMCs. However, when off-target binding was tested using SP34 cells, some amount of off-target binding was seen with the highest charged linkers at high concentrations (FIG. 89 of U.S. Ser. No. 14/216,705, incorporated by reference).

[0413] Positively charged scFv linkers on the anti-CD3 scFv in an anti-CD19×CD3 Fab-scFv-Fc construct had the unexpected property of reducing the amount of high molecular weight aggregation (FIG. 91 of U.S. Ser. No. 14/216,705, incorporated by reference). SEC chromatograms of two bispecific constructs (13121—with standard (GGGGS)4 linker (SEQ ID NO: 77)) and (13124—with charged linker (GKPGS)4 (SEQ ID NO: 75)) incubated at various concentrations confirmed this phenomenon.

[0414] Activity of anti-CD19×CD3 constructs containing charged scFv linkers in the anti-CD3 scFv was evaluated using an RTCC assay with PBMCs and Fab-scFv-Fc format bispecific anti-CD19×CD3 antibodies containing different scFv linkers (FIG. 92 of U.S. Ser. No. 14/216,705, incorporated by reference). Linkers have little impact on RTCC activity, except for the highly charged linker (GKGKS)3 (SEQ ID NO: 72) which has lower activity.

[0415] Sequences for all constructs of these experiments are shown in FIG. 93 of U.S. Ser. No. 14/216,705, incorporated by reference.

[0416] Schematics for anti-CD20×anti-CD8, anti-CD19×anti-CD8, and anti-CD38×anti-CD8 Fab-scFv-Fc bispecifics are shown in FIGS. 1 to 5. Amino acid sequences for these bispecifics are listed in FIGS. 6 and 14.

[0417] FIG. 7 illustrates the ability of various anti-CD20×anti-CD8 Fab-scFv-Fc bispecifics to mediate redirected T cell cytotoxicity (RTCC). Anti-CD20×anti-CD3 Fab-scFv-Fc bispecifics are included as comparators. During the RTCC reaction, CD25 (FIG. 8) and CD69 (FIG. 9) are modulated on the surface of T cells by varying amounts.

[0418] FIG. 10 show RTCC using un-activated PBMC and anti-CD20×anti-CD8 bispecifics. FIGS. 11 and 12 show CD25 and CD69 upregulation of anti-CD20×anti-CD8 bispecifics, respectively. FIG. 13 shows IL-6 release during the above RTCC assays. Note that CD8 bispecifics release very little IL-6 compared to the anti-CD20×anti-CD3 bispecific. This lower amount of IL-6 release should correspond to less activation of CD4 T cells as well as lower expected toxicity. Amino acid sequences for these bispecifics are listed in the Figures.

[0419] All of the claims of U.S. Ser. No. 62/084,741 are expressly incorporated herein by reference.