Title of Invention

ANTI-CD16A ANTIBODY

Abstract

This invention relates to an anti-CD16A antibody comprising a VH domain comprising complementarity determining regions (CDRs) a CDRI having the amino acid sequence ofSEQ ill NO:35, a CDR2 having the amino acid sequence ofSEQ ill NO:39,' and a CDR3 having the amino acid sequence of SEQ ill NO:59 and a V L domain comprising a CDRI having the amino acid sequence of SEQ ill NO:67, a CDR2 having the amino acid sequence of SEQ ill NO:75, and a CDR3 having the amino acid sequence of SEQ ill NO:88, wherein at least one of said CDRs comprises at least one amino acid substitution selected from the group consisting of, in the V H domain, M34Y in CDRl, H50L in CDR2, W52F in CDR2, D54N in CDR2, N60S in CDR2, A62S in CDR2, W99Y in CDR3, AIOID in CDR3, and, in the VL domain, K24R in CDRl, A25S in CDRI, F32Y in CDRl, M33L in CDRl, N34A in CDRl, T50A, T50W, or T50S in CDR2, T51A in CDR2, N53S in CDR2, E55A or E55Q in CDR2, S56T in CDR2, N92Y in CDR3, N93S in CDR3, and D92T in CDR3, which positions are according to the Kabat numbering scheme.

Full Text

CD16A BINDING PROTEINS AND USE FOR THE TREATMENT OF IMMUNE DISORDERS FIELD OF THE INVENTION t The invention relates to CD16A binding proteins and methods for treatment of immune disorders. The invention finds application in the fields of biomedicine and immunology. BACKGROUND - Fey receptors (FcyR) are cell surface receptors that bind the Fc region of immunoglobulin G (IgG) molecules. Among other functions, these receptors couple the formation of antibody-antigen complexes to efiector cell responses. For example, cross-linking of activating Fey receptors by immune complexes can result in the phagocytosis of pathogens, killing of foreign and transformed cells by direct cytotoxicity, the clearance of toxic substances, and the initiation of an inflammatory response. Notably, the Fey recqators play a key role in autoimmunity. Autoantibody binding to activating Fc receptors triggers the pathogenic sequalae of autoimmune diseases such as idiopathic thrombocytopenic purpura, arthritis, systemic lupus erythrematosus, autoimmune hemolytic anemia, and others. In humans and rodents there are three classes of Fey receptors, designated FcyRI, FcyRII, and FcyRIlI (see, Ravetch and Holland, 2001 Annual Rev. Immunol 19:275-90; and Ravetch and Kinet, 1991, Annua! Rev. Immunol. 9:457-92). FcyRI sites are generally occupied by monomelic IgG, while RII and Rin receptors are generally unoccupied and available to interact with immune complexes. FcyRI, also called CD64, binds monomelic IgG wth high affinity, and is present on monocytes and macrophages. FcyRII, also called CD32, binds to multimeric IgG (immune complexes or aggregated IgG) with moderate affinity, and is present on a variety of cell types, including B cells, platelets, neutrophils. macrophages and monocytes. FcyRIII, also called CD16, binds to multimeric IgG with moderate affinity and is the predominant activating FcyR on myeloid cells. FcyRIII is found in two forms. FcyRIIIA (CD16A), a transmembrane signaling form (50-65 kDa), is expressed by NK cells, monocytes, macrophages, and certain T cells. Fc^RinS (CD16B), a glycosyl-phosphatidyl-inositol anchored form (48 kDa) form, is expressed by human neutrophils. See, e.g., Scallon et al., 1989, Proc. Natl. Acad. Set U.S.A. 86:5079-83 and Ravetch et al., 1989, J. Exp. Med. 170:481-97. Protein and nucleic acid sequences for CD16A are reported in Genbank as accession numbers P0S637 (protein) and X52645 (nucleic acid) and in SWISS-PROT as accession number CAA36870. Protein and nucleic acid sequences for CD16B are reported in Genbank as accession nximbers 075015 (protein) and X16863 (nucleic acid) and in SWISS-PROT as CAA34753. SUMMARY OF THE INVENTION [0004] In one aspect, the invention provides a CD 16A binding protein that may be used for treatment of an individual with an autoimmune disease. CD16A binding proteins of the invention are other than mouse antibodies, and include chimeric, human and humanized anti-CD16A monoclonal antibodies, fragments thereof, single chain antibodies, and other binding proteins comprising a VH domain and/or a VL domain. [0005] In one aspect the CD16A binding protein comprises a Fc region derived from a human IgG heavy chain (e.g., a Fc region derived from human IgGi) where the Fc region lacks effector function and/or is modified to reduce binding to a Fc effector Kgand. In one embodiment, the CD 16A binding protein is not glycosylated, for example, due to a substitution at residue 297 of the Fc ch"-"-"^ region. [0006] In one aspect, the CD 16A binding protein is a humanized 3 G8 antibody with a VH domain comprising three complementarity determining regions (CDRs) derived from the VH domain of mouse monoclonal antibody 3G8. In one embodiment, the VH domain has the sequence of the VH domain of Hu3G8VH-l. In one embodiment, the CDRs of the binding protein have the sequence of tihe mouse CDRs. La some versions, the VH domain CDRs differ from those of 3G8 at least by one or more ofthe following substitutions: Valat position 34 in CDRl, Leu at position 50 in CDR2, Phe at position 52 in CDR2, Asn at position 54 in CDR2, Ser at position 60 in CDR2, Ser at position 62 in CDR2, Tyr at position 99 in CDRS, and Asp at position 101 of CDRS. hi one embodiment, the VH domain has the sequence of the VH domain of HuSGSVH- 22. In one embodiment VH domain comprises an FR3 domain having the sequence of SEQE)N0:51. The VH domain may be linked to an antibody heavy chain constant domain, for example the human Cyl constant domain. [0007] In some versions the CD16A binding protein has a VH domain having a sequence set forth in Table 3. In some versions the CD16A binding protein has a VH domain tliat differs from the sequence of Hu3G8VH-l by one or more of the substitutions shown in Table 1. [0008] In one aspect, the CD16A binding protein is a hxmianized 3G8 antibody with a VL domain comprising three complementarity determining regions (CDRs) derived from the VL domain of mouse monoclonal antibody 3G8. In one embodiment, the CDRs of the binding protein have the sequence of the mouse CDRs. In some versions, the VL domain CDRs differ from those of 3G8 at least by one or more of the following substitutions: Arg at position 24 in CDRl; Ser at position 25 in CDRl; Tyr at position 32 inCDRl; Leu at position 33 in CDRl; Ala at position 34 in CDRl; Asp, Trp or Ser at position 50 in CDR2; Ala at position 51 in CDR2; Ser at position 53 in CDR2; Ala or Gin at position 55 in CDR2; TTir at position 56 in CDR2; Tyr at position 92 in CDR3; Ser at position 93 in CDR3; and Thr at position 94 in CDRS. In one embodiment, the VL domahihas the sequence of the VL domain of Hu3G8VL-l, Hu3G8VL-22 or Hu3G8VL-43. The VL domain may be linked to an antibody light chain constant domain, for example the human CK constant region. [0009] In some versions the CDl 6A binding protein has a VL domain having a sequence set forth in Table 4. In some versions the CD16A binding protein has a VL domain that differs from the sequence of Hu3G8VL-l by one or more of the substimtions shown in Table 2. [0010] In one aspect, the CDl 6A binding protein comprises both a VH domain and a VL domain, as described above (which may be prepared by coe^ression of polynucleotides encoding heavy and light chains). Optionally the humanized heavy chain variable region comprises a sequence set forth in Table 3 and/or the a himianized light chain variable region comprises a sequence set forth in Table 4. For example, in exemplary embodiments, the bindii^ protein has a heavy chain vanaoie region naving me sequence oi Hin^iu iNu:ii3 ana a ugnt cnain variable region having the sequence of SEQ ID NO:96,100 or 1118. In another exemplary embodiment, the binding protein has a heavy chain variable region having the sequence of SEQ ID NO:109 and light chain variable regions having the sequence of SEQ ID NO:96. In another exemplary embodiment, the binding protein has a heavy chain variable region having the sequence of SEQ ID NO: 104 and light chain variable regions having the sequence of SEQ ID NO:96. [0011] In an embodiment, the CD16A binding protein is tetrameric antibody comprising two light chains and two heavy chains, said light chains comprising a VL domain and a light chain constant domain and said heavy chaii™ comprising a VH domain and a heavy chain constant domain. la an embodiment, the light chain constant domain is human CK and/or the heavy chain constant region is C7I. [0012] In one embodiment of the invention, the CD16A binding protein comprises an antigen binding site that binds CD 16A or sCDl 6A with a binding constant of less than 5 nM. [0013] In one embodiment, the CD16A binding protein comprises an aglycosyl Fc region that has reduced binding to at least one Fc effector hgand compared to a reference CD16A binding protein that comprises an unmodified Fc region (e.g., a human IgGi Fc domain glycosylated at position 297). The Fc effector ligand can be FcyRm or the CI q component of complement. [0014] In one embodiment, the invention provides a CD16A bmding protein that is a humanized antibody that binds to CD 16A and inhibits the binding of Fc receptor to CD16. [0015] In an aspect, the invention provides a pharmaceutical composition comprising of CD 16 A binding protein described herein and a phaimaceutically acceptable excipient [0016] In an aspect, the invention provides an isolated polynucleotide, optionally an expression vector, encoding a VH domain of a CD16A binding protein described herein. In an aspectj the invention provides an isolated nucleic acid, optionally an expression vector, encoding a VL domain of a CD16A binding protem described herein. In an aspect, the invention provides a cell, optionally a mammalian cell, comprismg a polynucleotide described herein. In an aspect, the invention provides a eel! line, optionally a mammalian cell line, expressing a CD16A binding protein described herein. [0017] The invention further provides a method of reducing an deleterious immune response (or undesired immune response) in a mammal comprising administering to a mammal a CDl 6A binding protein described herein. In an embodiment, reducing the deleterious immune response comprises protecting against antibody-mediated platelet depletion. [0018] In one aspect, the invention provides a method of treating an deleterious immune response in a mammal without inducing neutropenia in the mammal (e.g., severe neutropenia or moderate neutropenia), where the method comprises administering to the mammal a CD16A binding protein having an Fc region derived from human IgG, and where the amino acid at position 297 of the Fc region is aglycosyl. [0019] In embodiments of the above-described methods, the deleterious immune response is an inflammatory response, for example, an inflammatory response caused by an autoimmune disease. In an embodiment, the inflammatory response is caused by idiopathic thrombocytopenic purpura (ITP), rheumatoid arthritis (RA), systemic lupus erythrematosus (SLE), autoimmune hemolytic anemia (AHA), scleroderma, autoantibody triggered urticaria, pemphigus, vasculitic syndromes, systemic vasculitis, Goodpasture"s syndrome, multiple sclerosis (MS), psoriatic arthritis, ankylosing spondylitis, Sjogren"s syndrome, Reiter"s syndrome, Kowasaki"s disease, polymyositis and deimatomyositis. Other examples of diseases or conditions that can be treated according to the invention also include any diseases susceptible to treatment with intravenous immunoglobulin (TVIG) therapy (e.g., allergic asthma). The invention provides CD16A binding proteins that both protect against autoimmune diseases and do not result in significant neutrophil diminution in a mammal. In an embodiment, the CD16A binding proteins are anti-CD16A antibodies. These CD16A binding proteins are particularly advantageous for use as human therapeutics. In one aspect, the invention provides a method of treating an autoimmune disease in a mammal without neutrophil diminution or neutropenia in the mammal, by administering a CDi6A binding protein having an Fc region derived fi:om human IgG and an aglycosyl amino acid at position 297 of each of the CH2 domains of the Fc region. [0020] In yet another aspect, the invention provides a method of inhibiting the binding of IgG antibodies to FcyRIII on a cell by contacting the cell with a CD 16A binding protein under conditions in which the CD 16A binding protein binds the FcyRIII on the cell. [0021] In one aspect, the invention provides a method of maJcing a CDJ 6A binding protein with improved therapeutic efficacy in treating an deleterious immune response, comprising the following steps: i) obtaining a first CD 16A binding protein, where the first CD16A binding protein comprises an Fc region derived from IgG; and ii) modifying the Fc region of the first CD16A binding protein to produce a second CD16A binding protein that is aglycosylated at position 297 of the Fc region, where the second CD16A binding protein is more effective in treating the deleterious immune response when administered to a mammal than the first CD 16A binding protein. [0022] Inoneaspect, the invention provides a method of making a CD16A binding protein with improved therapeutic efficacy in treating an deleterious immune response, comprisii^ the following steps: i) obtaining a first CD16A binding protein, wherein the first CD 16 A binding protein comprises an Fc region derived from IgG; and ii) modifying the Fc region of the first CD16A binding protein to produce a second CD 16A binding protein that has reduced binding to an Fc effector ligand compared to the unmodified Fc region of the first CD16A binding protein, where the second CD16A binding protein is more effective in treating the deleterious immune response when administered to a mammal than the first CD16A binding protein. In one embodiment, the Fc effector ligand is FcyRm or the Clq component of complement. [0023] hi one aspect the method involves administering a CD 16 A binding protein to reduce an deleterious immune response in a subject without eliciting one or more significant deleterious effects that result from 3G8 administration, or eliciting significantly lower levels of such effects than does administration of murine 3G8. [0024] In one embodiment of the invention, the improved therapeutic efficacy in treating a deleterious immune response comprises improved effectiveness at protecting against antibody-mediated platelet depletion. The deleterious immune response is optionally due to idiopathic thrombocytopenic purpura (TTP) or the administration of murine monoclonal antibody 6A6 to a muFcyRlII-/-, huFcvRIIIA transgenic mouse. [0025] The invention provides the use of a CD16A bmdingprotem comprising an Fc region derived from a human IgG heavy chain, wherein the Fc region lacks effector function, for treatment of an immune disorder or for preparation of a medicament for treatment of an immune disorder. BRIEF DESCRIPTION OF THE FIGURES [0026] Figure 1 showsresultsfromanELISAfor binding of sCD16A by CD16A binding proteins. Hu3G8-24.43 is an antibody with the heavy chain Hu3G8VH-24, and the Ught chain Hu3G8VL-43. Hu3G8-5.1 is an antibody with the heavy chain Hu3G8VH-5, and the light chain Hu3G8VL-l. Ch3G8 is the chimeric 3G8 antibody. HulgGl is an irrelevant immunoglobulin. [0027] Figure 2 shows results of an assay for binding of humanized and chimeric antibodies to CHO-Kl cells expressing the extracellular domain of CD16A. Hu3G8-22.1 is an antibody with die heavy chain Hu3G8VH-22, and the light chamHu3G8VL-l. Hu3G8-5.1 is an antibody with the heavy chain Hu3G8VH-5, and the Ught chain Hu3G8VL-l. Hu3G8-22.43 is an antibody with the heavy chain Hu3G8VH-22, and the light chain Hu3G8VL-43. N297Q indicates the antibody is aglycosylated. [0028] Figure 3 shows results of a cell based competition assay. The aglycosylated humanized antibodies shown compete with aglycosylated chimeric antibody for binding to CHO-Kl cells expressing the extraceUular domain of CD16A [0029] Figure 4 shows mhibition of binding of sCD16A to hnmune complexes. Hu3G8-l.l isanantibody with the heavy chain Hu3G8VH-l, and the Ught chain Hu3G8VL-l. [0030] Figure 5 shows IIP protection m mice injected i.v. witii mAb 3G8 (0.5}ig/g) or human IVIG (Img/g) one hour before ch6A6 i.p injection. [0031] Figure 6 shows ITP protection in mice injected i.v. with mAb 3G8 (0.5jig/g) or human IVIG (Img/g) one hour before ch6A6 i.v injection. [0032] Figure? shows the absence oflTP protection inmice injected i.v. with ch3G8 (O.S^ig/g) one hour before 6A6 i.p. injection. [0033] Figure 8 shows protection from ITP in mice injected i.v. with ch3G8 N297Q one hour before ch6A6 i.p injection . [0034] Figure 9 shows protection from ITP in mice injected i.v. with ch3G8 N297Q one hour before ch6A6 i.v injection. [0035] Figure 10 shows the results of FACS scans of neutrophils following administration of CDI6A binding protein or controls. The x-axis shows labeling with antibody to CD16, and the y-axis shows labeling with antibody to the Gr-1 antigen. The upper right quadrant shows neutrophils; the upper left quadrant shows other granulocytes and neutrophils that no longer stain with 3G8-FITC. [0036] Figure 11 showspreventionofAIHA with a humanized anti-CD 16 antibody. [0037] Figure 12 shows inhibition of ch4D5 mediated ADCC by humanized 3G8 antibodies. [0038] Figure 13 shows inhibition of ch4-4-20 mediated ADCC by mouse 3GS (Figure 13A) and humanized 3G8 antibodies (Figure 13B). [0039] Figure 14 shows protection of FcyRIII-/-, hCD16A, hCD32A mice against ITP by administration of hu3G8-5.1. [0040] Figure 15 shows protection of FcyRIII-/-, hCD16A mice against ITP by administration of hu3G8-5.1 N??7Q. Figure 15(A) shows data points for each dose at indicated times. Figure 15(B) shows dose response at the 5 hour time point. [0041] Figure 16 shows the therapeutic effect of administration of aglycosylated humanized antibody subsequent to mice in which thrombocytopenia has been induced. Figure 16(A) shows administration of Hu3G8-5.1-N297Q. Figure 16(B) shows administration of Hu3G8-22.1-N297Q and Hu3G8-22.43- N297Q. [0042] Figure 17 shows the tiierapeutic effect of a humanized anti-CD16A antibody in treatment of autoimmxme hemolytic anemia. DETAILED DESCRIPTION 1. Definitions [0043] Unless otherwise defined, all terms of art, notations and other scientific terms or terminology used herein are intended to have the meanings commonly understood by those of skill in the art to which this invention pertains. In some cases, terms with commonly understood meanings are defined herein for clarity and/or for ready reference, and the inclusion of such definitions herein should not necessarily be construed to represent a substantial difference over what is generally understood in the art. The practice of the present invention will employ, unless otherwise indicated, conventional techniques of molecular biology (including recombinant techniques), microbiology, cell biology, biochemistry, nucleic acid chemistry, and immunology, which are within the skill of the art. Such techniques are explained fiilly in the literature, such as. Current Protocols in Immunology (J.E. Coligan et al., eds., 1999, including supplements through2001)-, Current Protocols in Molecular Biology (F.M. Ausubel et al., eds., 1987, including supplements through 2001); Molecular Cloning: A Laboratory Manual, third edition (Sambrook and Russel, 2001); PCIt: The Polymerase Chain Reaction, (Midlis et al., eds., 1994); The Immunoassay Handbook (D. Wild, ed., Stockton Press NY, 1994); Bioconjugate Techniques (Greg T. Hermanson, ed.. Academic Press, 1996); Methods of Immunological Analysis (R. Masseyeff, W.H. Albert, and N.A. Stames, eds., Weinheim: VCH Verlags gesellschaft mbH, 1993), Harlow and Lane Using Antibodies: A Laboratory Manual Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, 1999; and Beaucage et al. eds.. Current Protocols in Nucleic Acid Chemistry Johii Wiley & Sons, Inc., New York, 2000). [0044] The terms "heavy chain," "light chain," "variable region," "fi-amework region," "constant domain," and the like, have their ordinary meaning in the immunology art and refer to domains in naturally occurring nmnunoglobxilins and the corresponding domains of synthetic (e.g., recombinant) binding proteins (e.g., humanized antibodies). The basic structural unit of naturally occurring unmunoglobulins (e.g., IgG) is atetramer having two light chains and two heavy chains. Usually naturally occurring immunoglobulin is expressed as a glycoprotein of about 150,000 daltons, although, as described below, IgG can also be produced in a nonglycosylated form. The amino-terminal ("N") portion of each chain includes a variable region of about 100 to 110 or more amino acids primarily responsible for antigen recognition. The carboxy-tenninal ("c") portion of each chain defines a constent region, with light chaiiB having a single coi^tant domain and heavy chains usually having three constant domains and a hinge region. Thus, the structure of the light chams of an IgG molecule is N-VL-CL-C and the structure of IgG heavy chains is N-VH-CHI-H-CH2-CH3-C (where H is the hinge region). The variable regions of an IgG molecule consists of the complementarity determining regions (CDRs), which contain the residues in contact with antigen and non-CDR segments, referred to as framework segments, which maintain the structure and determine the positioning of the CDR loops. Thus, the Vtand VHdomains have the structure N-FRl, CDRl, FR2, CDR2, FR3, CDR3, FR4-C. [0045] As used herein, the tenns "CD16A bmding protein," "CD16A antibody," and "anti-CD16A antibody," are used interchangeably and refer to a variety of immunoglobulin-like or immunoglobuUn-derived proteins. "CD 16A binding proteins" bind CD16A via an interaction with VL and/or VH domains (as distinct from Fc-mediated binding). Examples of CD16A binding proteins includes chimeric, humanized and human antibodies (e.g., comprising 2 heavy and 2 light chains), fragments thereof (e.g., Fab, Fab", F(ab")2, and Fv fragments), bifunctional or multifimctional antibodies (see, e.g., Lanzavecchia et al., 1987, Eur. J. Immunol. 17:105), single chain antibodies (see, e.g.. Bird et al., 1988, Science 242:423-26), fusion proteins (e.g., phage display fusion proteins), "minibodies" (see, e.g., U.S. pat. no. 5,837,821) and other antigen binding proteins comprising a VL and/or VH domain or fragment thereof In one aspect, the CD16A binding protein is a "tetrameric antibody" i.e., having generally the structure of a naturally occurring IgG and comprising both variable and constant domains, (i.e., two light chains comprising a VL domain and a light chain constant domain, such as human CK and two heavy chains comprising a VH domain and a heavy chain hinge and constant domains, such as human Cyl). Except as expressly noted, the mouse antibodxSGSJs. specifically excluded from the definition of GDI6A binding protein. [0046] When referring to binding proteins or antibodies (as broadly defined herein) the assignment of amino acids to each domain is in accordance with the definitions of Kabat, SEQUENCES OF PROTEINS OF IMMUNOLOGICAL INTEREST (National Institutes of Health, Bethesda, Md., 1987 and 1991). Amino acids fixjm the variable regions of the mature heavy and light chains of immunoglobulins are designated by the position of an amino acid in the chain. Kabat described numerous amino acid sequences for antibodies, identified an amino acid consensus sequence for each subgroup, and assigned a residue number to each amino acid. Kabat"s numbering scheme is extendible to antibodies not included in his compendium by aligning the antibody in question with one of the consensus sequences in Kabat by reference to conserved amino acids. This method for assigning residue numbers has become standard in the field and readily identifies amino acids at equivalent positions in different antibodies, including chimeric or humanized variants. For example, an amino acid at position 50 of a human antibody Ught chain occupies the equivalent position to an amino acid at position 50 of a mouse antibody tight chain. Thus, as used herein m the context of chimeric or humanized antibodies, a reference such as "at position 297 of the Fc region" refers to the amino acid position in an immunoglobulin chain, region of an a immunoglobulin chain, or region of a polypeptide derived &om an immunogiobulin chain, that corresponds to position 297 of the corresponding human immunoglobulin. [0047] The "Fc region" of immunoglobulins refers to the C-terminal region of an irmnunoglobulin heavy chain. Although the boundaries of the Fc region may vary somewhat, usually the Fc region is from about position 226-230 extending to the carboxy terminus of the polypeptide (and encompassing the CH2 and CH3 domains). Sequences ofhumanFc regions are found in Kabat, rapra. In addition, a variety of allelic variants are known to exist. [0048] An "Fc effector ligand" is a ligand that binds to the Fc region of an IgG antibody, thereby activating eifector mechanisms resulting in the clearance and destruction of pathogens. Fc effector ligands include three cellular Fc receptors types - FcRyl, FcRyll, and FCRYIII. The multiple isoforms of each of the three Fc receptor types are also included. Accordingly, the term "Tc effector ligand" includes both FcRylllA CCD16A) and FcRylllB (CD16B). The term "Fc effector Ugand" also includes the neonatal Fc receptor (Fcyn) and the C1 q component of complement. Binding of IgG to the Fc receptors triggers a variety of biological processes including antibody-dependent cell-^nediated cytotoxicity (ADCC), release of inflammatory mediators, control of antibody production, clearance of immune complexes and destruction of antibody-coated particles. Binding of the Clq component of complement to IgG activates the complement system. Activation of complement plays important roles in opsonization, lysis of cell pathogens, and inflammatory responses. 10049] As used herein, an Fc region that "lacks effector function" does not 3ind the Fc receptor and/or does not bind the Clq component of complement and xigger the biological responses characteristic of such binding. [0050] The term "glycosylation site" refers to an amino acid residue that is recognized by a mammalian cell as a location for the attacliment of sugar residues, ^mino acid residues to which carbohydrates, such as oligosaccharides, are Attached are usually asparagine (N-linkage), serine (0-linkage), and threonine (O-Jnkage) residues. The specific sites of attachment usually have a characteristic jequence of amino acids, referred to as a "glycosylation site sequence." The glycosylation site sequence for N-linked glycosylation is : -Asn-X-Ser- or -Asn-K-Thr-, where X can be any of the conventional amino acids, other than proline. [he Fc region of human IgG has two glycosylation sites, one in each of the dfi iomains. The glycosylation that occurs at the glycosylation site in the CH2 iomain of human IgG is N-linked glycosylation at the asparagine at position 297 :Asn297). ;0051] The term "chimeric," when referring to antibodies, has the ordinary neaning in the art and refers to an antibody in which a portion of a heavy and/or ight chain is identical to or homologous with an antibody from one species (e.g., nouse) while the remaining portion is identical to or homologous with an mtibody of another species (e.g., human). [0052] As used herein, the term "humanized" has its usual meaning in the art. ^ [n general terms, humanization of a non-himian antibody involves substituting tiie ^ CDR sequences from non-human immunoglobulin VL and VH regions into human / 5-amework regions. Further, as used herein, "himianized" antibodies may comprise additional substitutions and mutations in the CDR and/or framework regions introduced to increase affinity or for other purposes. For example, substitution of nonhuman framework residues in the human sequence can increas ifGnity. See, e.g., Jones et al., 1986, Nature 321:522-25; Queen et al., 1989, Froc. Natl. Acad Sci. U.S.A. 86:10029-33; Foote and Winter, 1992, JMol. Biol. 224:487-99; Chothiaet al., 1989, Nature 342:877-83; Riechmann et al., 1988, Nature 332:323-27; Co et al., 1991, Proc. Natl. Acad Sci. U.S.A. 88:2869-73; Padlan, 1991, Mol Immunol 28:489-98. The resulting variable domains have non-human CDR seqxiences and framework sequences derived from human antibody framework sequence(s) or a human consensus sequence (e.g., as disclosed in Kabat, supra). A variety of different human framework regions may be used singly or in combination as a basis for the humanized monoclonal antibodies of the present invention. The framework sequences of a humanized antibody are "substantially human," by which is meant that at least about 70% of the human antibody sequence, usually at least about 80% human, and most often at least about 90% of the framework sequence is from human antibody sequence. hi some embodiments, the substantially human framework comprises a serine at position 113 of the VH FR4 domain (e.g., SEQ ID NO: 64). As used herein, a "humanized antibody" includes, in addition to tettameric antibodies, single chain antibodies, antibody fragments and the like that comprise CDRs derived from a non-human antibody and framework sequences derived from human framework regions. " [0053] As used herein, "mammals" include humans, non-human primates, rodents, such as, mice and rats, and other mammals. [0054] As used herein, "neufropenia" has its ordinary meaning, and refers to a state in which the number of neutrophils circulating in the blood is abnormally low. The normal level of neutrophils in human blood varies slightly by age and race. The average adult level is about 1500 cells/mm^ of blood. Neutrophil counts less than 500 cells/mm^ result in great risk of severe infection. Generally, in humans, severe neufropenia is defined by a blood neutrophil count less thari about 500 cells/mm^, and moderate neutropenia is characterized by a blood neufrophil count from about 500-1000 cells/mm . [0055] As used herein, "treatment" refers to clinical intervention in an attempt to alter the disease course of the individual or cell being treated, and can be performed eiliier for prophylaxis or during the course of cluiical pathology. Therapeutic effects of freatment include without limitation, preventing occurrence or recurrence of disease, alleviation of symptoms, diminishment of any direct or indirect pathological consequences of the disease, decreasing the rate of disease progression, amelioration or palliation of the disease state, and remission or improved prognosis. [0056] An "effective amount" is an amount sufficient to effect a beneficial or desired clinical result upon freatment. An effective amount can be administered to a patient in one or more doses. A "therapeutically effective amount" is an amoxmt that is sufficient to palliate, ameliorate, stabilize, reverse or slow the progression of the disease, or otherwise reduce the pathological consequences of the disease, or reduce the symptoms of the disease. The amelioration or reduction need not be, and usually is not, permanent, but may be for a period of time ranging from at least one hour, at least one day, or at least on week or more. The effective amount is generally determined by the physician on a case-by-case basis and is within the skill of one in the art. Several factors are typically taken into account when determining an appropriate dosage to achieve an effective amount. These factors include age, sex and weight of the patient, fhie condition being treated, the severity of the condition and the form and effective concentration of the binding protein administered. An "inflammation reducing amount" is an amount that reduces inflammation in a subject. A reduction in inflammation can be assessed by art known criteria, including decreased C-reactive protein levels, decreased consumption of complement, reduced immune complex deposition at sites of inflammation (e.g., joints in subjects with RA, kidney in subjects with liq)us, myelin sheath, etc.), reduced cytokine release, migration of macrophages and neutrophils, and the like. [0057] "Substantial sequence identity," as used herein, refers to two or more sequences or subsequences (e.g., domains) that have at least about 80% amino acid residue identity, preferably at least about 90%, or at least about 95% identity when compared and aligned for maximum correspondence. Sequence identity between two similar sequences (e.g., antibody variable regions) can be measured by (1) aligning the sequences to maximize the total number of identities across the entire length of the sequences, or across the entire length of the shorter of the two sequences, if of different lengths (and where the length of the aligned sequences or shorter of the aligned sequences is "L" residues); (2) counting tiie number of positions (not including the number "E" residues designated as excluded from the comparison) at which there is an amino acid identity, where the number of identities is designated "N"; (3) and dividing the N by the "L" minus "E." For example, in a comparison of two sequences each of length 80 residues, in which 6 specific residues are excluded from the comparison and for which there are 63 identities in the remaining 74 positions, the sequence identity would be N/(L-E) or 65/(80-6) or 87.8%. (Residues might be specified as "excluded" from the calculation when, for illustration but not limitation, they are in a non-antibody domain of fiision protein.) Alternatively, optimal alignment and sequence identity can be calculated by computerized implementations of algorithms described in Smith & Waterman, 19BI, Adv. Appl Math. 2:482 [local homology algorithm], Needleman & Wunsch, 1970, J. Mol. Biol. 48:443 [homology alignment algorithm], Pearson & Lipman, 1988, Proc. Natl. Acad. Sci. USA 85:2444 [search for similarity method], or Altschul et al., 1990, J. Mol. Biol. 215:403-10 [BLAST algorithm ]. See Ausubel et al., supra and GAP, BESTFIT, FASTA, and TFASTA in the Wisconsin Genetics Software Package, Genetics Computer Group, 575 Science Dr., Madison, WI). When using any of the aforementioned algorithms, the default parameters (for Window length, gap penalty, etc.) are used. An amino acid or nucleic acid sequence is "substantially similar to" a second sequence when the degree of sequence identity is at least about 70% identical, preferably at least about 80%, or at least about 90%, or even at least about 95%, identical. Sequences that are substantially identical are also substantially similar. [0058] As used herein, a polypeptide, polypeptide domain or region, or amino acid sequence is "derived from" another when the two sequences are identical or substantially similar and have a similar biological fimction. For example, in a humanized mouse monoclonal antibody the complementary determining regions (CDRs) are "derived from" the corresponding CDRs of the mouse monoclonal antibody, and the variable domain framework regions can be "derived from" framework sequences of the corresponding human antibody. It will be apparent that one domain, etc., can be derived from a parental domain, etc., even though the two differ in sequence due to, for example, the introduction of mutations that affect, or alternatively do not change, binding affinity or other properties of the protein in which the domain, etc., is contained, such as those described herein. It wiU also be understood that normally a domain, etc., "derived from" a parental domain, etc., is made, produced or designed using materials (e.g. genetic material) or information (e.g., nucleotide or amino acid sequence) from the parental molecule. [0059] Standard abbreviations are used for amino acids: alanine, Ala (A); serine, Ser (S); threonine, Thr (T); aspartic acid. Asp (D); glutamic acid, Glu (E); asparagine, Asn (N); glutamine, Gin (Q); arginine, Arg (R); lysine, Lys (K); isoleucine, He (I); leucine, Leu (L); methionine. Met (M); valine, Val (V); phenylalanine, Phe (F); tyrosine, Tyr 00; tryptophan, Trp (W); glycine, Gly (G); bistidine. His (H): proline. Pro (P); and cysteine, Cys (C). 2. Introduction [0060] The FcyRIIIA receptor, CD16A, plays a role in coupling cytotoxic and immune complex antibodies to effector responses. It is believed that the interaction of the FcyRIIIA receptor and immimoglobuiin aggregates (e.g. immune complexes) present in autoimmune diseases and other pathogenic conditions results in a deleterious inflammatory response in subjects. Without intending to be bound by a specific mechanism, it is believed that reducing the interaction of the FcyRIIIA receptor (generally referred to herein as "CD16A" or "the CD16A receptor" and immunoglobulin aggregates will alleviate this inflammatory response. Also without intending to be bound by a specific mechanism, it is believed that one method for reducing the interaction of CD16A and immimoglobuiin aggregates is by use of anti-CD16A antibodies, or other CD16A binding proteins, to block the interaction. [0061] Monoclonal antibody 3G8 ("mAb 3G8") is a mouse monoclonal antibody that binds the Fc-binding domain of human CD16A and B with a Ka of 1x10^ VT^ (Fleit et al., 1982, Proc. Natl. Acad. Set U.S.A 79:3275-79). 3G8 blocks the binding of human IgGi immune complexes to isolated human NK cells, monocytes and neutrophils, as well as to CD16A-transfected 293 cells. Ejcperiments in which mAb 3G8 has been administered to human patients for treatment of idiopathic thrombocytopenic purpura (ITP) have been conducted (Clarkson et al., 1986, N. Engl. J Med 314:1236-39; Soubrane, et al., 1993, Blood 81:15-19). Administration of the 3G8 antibody was reported to result in increased platelet levels and was accompanied by one or more significant side effects, including a HAMA response, cytokine release syndrome, and/or pronounced neutropenia. [0062] TTie present invention provides novel GDI 6A binding proteins, including humanized and/or aglycosylated monoclonal antibodies, and methods for reducing an deleterious immime response in a subject by administering the proteins. Administration of these binding proteins is shown to be protective ia well established models for two distinct autoimmune diseases: autoimmune hemolytic anemia (AHA) and idiopathic thrombocytopenic purpura. These results are indicative of efficacy of this treatment for other autoimmune diseases as well. Moreover, the inventors have discovered that, unexpectedly, administration of anti-CD 16A antibodies with altered effector function (e.g., aglycosylated antibodies) protects against the deleterious immune reponses characteristic of autoimmune disorders without inducing acute severe neutropenia. Thus, the invention provides new reagents and methods for antibody-mediated effected treatment of autoimmune conditions without pronounced side-effects observed using alternative treatments. 3. CDMA Binding Froteins [0063] A variety of CD16A binding proteins may be used in the methods of the invention. Suitable CD16A binding proteins include human or humanized monoclonal antibodies as well as CD16A binding antibody fragments (e.g., scFv or single chain antibodies. Fab fragments, minibodies) and anotiier antibody-lilce proteins that bind to CD 16A via an interaction with a light chain variable region domain, a heavy chain variable region domain, or both. (0064] In some embodiments, die CDl 6A binding protein for use according to the invention comprises a VL and/or VH domain that has one or more CDRs with sequences derived from a non-human anti-CD 16A antibody, such as a mouse anti-CD 16A antibody, and one or more framework regions with derived from framework sequences of one or more human immunoglobulins. A number of non-human anti-CD16A monoclonal antibodies, from which CDR and olher sequences may be obtained, are known (see, e.g., Tamm and Schmidt, 1996, J. Imm. 157:1576-81; Flehetal., 1989,p.l59; LEUKOCYTE TYPING II: HUMAN MYELOID AND HEMATOPOIETIC CELLS, Reinherz et al., eds. New York: Springer-Verlag; 1986; LEUCOCYTE TYPING III: WHITE CELL DIFFERENTIATION ANTIGENS McMichael AJ, ed., Oxford: Oxford University Press, 1986); LEUKOCYTE TYPING rV: WHITE CELL DIFFERENTIATION ANTIGENS, Kapp et al., eds. Oxford Univ. Press, Oxford; LEUKOCYTE TYPING V: WHITE CELL DIFFERENTIATION ANTIGENS, Schlossman et al., eds. Oxford Univ. Press, Oxford; LEUKOCYTE TYPING VI: WHITE CELL DIFFERENTIATION ANTIGENS, Kishimoto, ed. Taylor & Francis. In addition, as shown in the Examples, new CDl 6A binding proteins that recognize human CDI6A expressed on cells can be obtained using well known methods for production and selection of monoclonal antibodies or related binding proteins (e.g., hybridoma technology, phage display, and the like). See, for example. O"Connel et al., 2002, J. Mol Biol. 321:49-56; Hoogenboom and Chames, 2000, Imm. Today 21:371078; Krebs et al., 2001, J: Imm. Methods 254:67-84; and other references cited herein. Monoclonal antibodies from a non-human species can be chimerized or humanized using techniques using techniques of antibody humanization known in the art. [0065] Alternatively, fully human antibodies against CD16A can be produced using transgenic animals having elements of a human immune system {see, e.g,, U.S. Patent Nos. 5,569,825 and 5,545,806), using human peripheral blood cells (Casali et al., 1986, Science 234:476), by screening a DNA library from human B cells according to the general protocol outlined by Huse et al., 1989, Science 246:1275, and by other methods. [0066] It is contemplated that, for some purposes, it may be advantageous to use CD16A binding proteins that bind the CD16A receptor at the same epitope bound by 3G8, or at least sufficiently close to this epitope to block binding by 3G8. Methods for epitope mapping and competitive binding experiments to identify binding proteins with the desired binding properties are well known to those skilled in the art of experimental immunology. See, for example, Harlow and Lane, cited iupm; Stahletal., 19%3, Methods in Enzytnology 9:242-53; Kirkland et al., 1986, J. Immunol. 137:3614-19; Morel et al., 1988, Molec. Immunol. 25:7-15; Cheung et al., 1990, Virology 176:546-52; and Moldenhauer et al., 1990, Scand J. Immunol. 32:77-82. Also see Examples and §3G(i), infra. For instance, it is possible to determine if two antibodies bind to the same site by using one of the antibodies to capture liie antigen on an ELISA plate and then measuring the ability of the second antibody to bind to die captured antigen. Epitope comparison can also be achieved by labeling a first antibody, directly or indirectly, with an enzyme, radionuclide or fluorophore, and measuring the ability of an unlabeled second antibody to inhibit the binding of the first antibody to the antigen on cells, in solution, or on a solid phase. [0067] It is also possible to measure the ability of antibodies to block the binding of the CD16A receptor to immune complexes formed on ELISA plates. Such immune complexes are formed by first coating the plate with an antigen such as fluorescein, then applying a specific anti-fluorescein antibody to the plate. This immune complex then serves as the ligand for soluble Fc receptors such as sFcREEIa. Alternatively a soluble iTmnuTie comnlex mav be formed and labeled. directly or indirectly, with an enzyme radionuclide or fluorophore. The ability of antibodies to inhibit the binding of these labeled immune complexes to Fc receptors on cells, in solution or on a solid phase can then be measured. [0068] GDI 6A binding proteins of the invention may or may not comprise a human immunoglobulin Fc region. Fc regions are not present, for example, in scFv binding proteins. Fc regions are present, for example, in human or humanized tetrameric monoclonal IgG antibodies. As described in detail below, in some embodiments of the present invention, the CD16A binding protein includes an Fc region that has an altered effector function, e.g., reduced affinity for an effector ligand such as an Fc receptor or CI component of complement compared to the xmaltered Fc region (e.g., Fc of naturally occurring IgGi proteins), hi one embodiment the Fc region is not glycosylated at the Fc region amino acid corresponding to position 297. Such antibodies lack Fc effector function. [0069] Thm, in some embodiments of the invention, the GD16A binding protein does not exhibit Fc-mediated binding to an effector ligand such as an Fc receptor or Ihe CI component of complement due to the absence of fee Fc domain in the binding protein while, in other cases, the lack of binding or effector function is due to an alteration in the constant region of the antibody. 4. CD16A Binding Proteins Comprising CDR Sequences Similar to A mAh 3G8 CDR Sequences. [0070] GD16A bmding protems that can be used in the practice of the invention include proteins comprising a CDR sequence derived from (i.e., having a sequence ihe same as or similar to) the CDRs of the mouse monoclonal antibody 3G8. Complementary cDNAs encoding the heavy chain and light chain variable regions of the mouse 3G8 monoclonal antibody, including the CDR encodii^ sequences, were cloned and sequenced as described in the Examples. The nucleic acid and protein sequences of 3G8 are provided below and are designated SEQ ID N0:1 and 2 (VL) and SEQ ID N0:3 and 4 (VH). Usmg the mouse variable region and CDR sequences, a large number of chimeric and humanized monoclonal antibodies, comprising complementary determining regions derived from 3G8 CDRs were produced and their properties analyzed. To identify humanized antibodies that bind GDI 6A with high afBnity and have other desirable properties. antibody heavy chains comprising a VH region with CDRs derived from 3G8 were produced and combined (by coexpression) with antibody light chains comprising a VL region with CDRs derived from 3G8 to produce a tetrameric antibody for analysis. Properties of the resulting tetrameric antibodies were determined as described below. As described below, CD16A binding proteins comprising 308 CDRs, such as the humanized antibody proteins described hereinbelow, may be used according to the invention to reduce an deleterious immune response. A. VH Region [0071] In one aspect, the CD16A binding protein of the invention may comprise a heavy chain variable domain in which at least one CDR (and usually three CDRs) have the sequence of a CDR (and more typically all three CDRs) of the mouse monoclonal antibody 3G8 heavy chain and for which the remaining portions of the binding protein are substantially himian (derived from and substantially similar to, the heavy chain variable region of a human antibody or antibodies). [0072] In an aspect, the invention provides a humanized 3G8 antibody or antibody fragment containing CDRs derived from tie 3G8 antibody in a substantially human framework, but in which at least one of the CDRs of the heavy chain variable domain differs in sequence from the corresponding mouse antibody 3G8 heavy chain CDR. For example, in one embodiment, the CDR(s) differs from the 3G8 CDR sequence at least by having one or more CDR substitutions shown in Table 1 (e.g., valine at position 34 in CDRl, leucine at position 50 in CDR2, phenylalanine at position 52 in CDR2, tyrosine at position 52 in CDR2, aspartic acid at position 52 in CDR2, asparagine at position 54 in CDR2, serine at position 60 in CDR2, serine at position 62 in CDR2, tyrosine at position 99 in CDR3, and/or aspartic acid at position 101 of CDR3). Suitable CD16 binding proteins may comprise 0,1,2, 3, or 4, or more of tiiese substitutions (and often have from 1 to 4 of these substitutions) and optionally can have additional substitutions as well. [0073] In one embodiment, a CD 16 A bindii^ protein may comprise a heavy chain variable domain sequence that is the same as, or similar to, the VH domain of the Hu3G8VH-l constmct, the sequence of which is provided in Table 3. For example, the invention provides a GDI 6A binding protein comprismg a VH domain with a sequence that (1) differs from the VH domain of Hu3G8VH-l by zero, one, or more than one of the CDR substitutions set forth in Table 1; (2) differs from the VH domain of Hu3G8VH-l by zero, one or more than one of the framework substitutions set forth in Table 1; and (3) is at least about 80% identical, often at least about 90%, and sometimes at least about 95% identical, or even at least about 98% identical to the Hu3G8VH-l VH sequence at the remaining positions. [0074] Exemplary VH domains of CD16 binding proteins of the invention have the sequence of Hu3G8VH-5 and Hu3G8VH-22, as shown in Tables 3 and 6. [0075] The VH domam may have a sequence that differs from that of Hu3G8VH-l (Table 3) by at least one, at least two, at least three, at least four 4, at least five, or at least six of the substitutions shown in Table 1. Tliese substitutions are believed to result in increased affinity for CD16A and/or reduce the immunogenicity of a GDI 6A binding protein when admmistered to humans. In certain embodiments, the degree of sequence identity with the Hu3G8VH-l VH domain at the remaining positions is at least about 80%, at least about 90%, at least about 95% or at least about 98%. Table 1 VH Domain Substitutions No. Kabat Position Region Substitutions 1 2 FRl He 2 5 FRl Lys 3 10 FRl Thr 4 30 FRl Arg 5 34 CDRl Val 6 50 CDR2 Leu 7 52 CDR2 Phe or Tyr or Asp 8 54 CDR2 Asn 9 60 CDR2 Ser 10 62 CDR2 Ser 11 70 FR3 Thr 12 94 FR3 Gin or Lys or Ala or His 13 99 CDR3 Tyr 14 101 CDR3 Asp [0076] For illustration and not limitation, the sequences of a number ot CD16A binding protein VH domains is shown in Table 3. As described in the Examples, infra, heavy chains comprising these sequences fused to a human Cy\ constant region were coexpressed with the hu3G8VL-l light chain (described below) to form tetramenc antibodies, and binding of the antibodies to CD16A was measured to assess the effect of amino acid substitutions compared to the hu3G8VH-l VH domain. Constructs in which the VH domain has a sequence of hu3G8VH-l, 2, 3,4, 5, 8,12,14,16,17,18,19, 20,22,23,24, 25,26,27,28,29, 30, 31, 32, 33, 34,35, 36, 37, 42,43,44, and 45 showed high affinity bmdmg, with hu3G8VH- 6 and -40 VH domains showing intermediate binding. CD16A binding proteins comprising the VH domains of hu3G8VH-5 and hu3G8VH-22 are considered to have particularly favorable binding properties. B. VL Region [0077] Similar studies were conducted to identify Ught chain variable domain sequences with favorable binding properties. In one aspect, the invention provides a CD16A binding protein containing a li^t chain variable domain in which at least one CDR (and usually three CDRs) has the sequence of a CDR (and more typically all three CDRs) cf ihe mouse monoclonal antibody 3G8 Ught chain and for which the remaining portions of the binding protein are substantially human (derived fiom and substantially similar to, the heavy chain variable region of a human antibody or antibodies). [0078] In one aspect, the invention provides a humanized 3G8 antibody or antibody fragment contaming CDRs derived from the 3G8 antibody in a substantially human framework, but in which at least one of the CDRs of the light chain variable domain differs in sequence from the mouse monoclonal antibody 3G8 light cham CDR. hi one embodunent, the CDR(s) differs from tiie 3G8 sequence at least by having one or more amino acid substitutions in a CDR, such as, one or more substitutions shown in Table 2 (e.g., arginine at position 24 in CDRl; serine at position 25 in CDRl; tyrosine at position 32 in CDRl; leucine at position 33 in CDRl; aspartic acid, tryptophan or serine at position 50 in CDR2; s^ine at position 53 in CDR2; alanine or glutamine at position 55 in CDR2; threonine at position 56 in CDR2; serine at position 93 inCDR3; and/or threonine at position 94 in CDR3). In various embodiments, the variable domain can have 0,1,2, 3, 4, 5, or more of these substitutions (and often have &om 1 to 4 of these substitutions) and optionally, can have additional substitutions as well. [0079] In one embodiment, a suitable CD16A binding protein may comprise a light chain variable domain sequence that is the same as, or similar to, the VL domain of the Hu3G8VL-l construct, the sequence of which is provided in Table 4. For example, the invention provides a CD16A binding protein comprising a VL domain with a sequence that (1) differs from the VL domain of Hu3G8VL-l by zero, one, or more of the CDR substitutions set forth in Table 2; (2) differs from the VL domain of Hu3G8VL-l by zero, one or more of the framework substitutions set forth in Table 2; and (3) is at least about 80% identical, often at least about 90%, and sometimes at least about 95% identical, or even at least about 98% identical to the Hu3G8VL-l VL sequence at the remaining positions. [0080] Exemplary VLdomainsofCD16bindingproteins of the invention have the sequence of Hu3G8VL-l or Hu3G8VL-43, as shown in Tables 4 and 6. [0081] The VL domain may have a sequence that differs from that of Hu3G8VL-l (Table 4) by zero, one, at least two, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, or at least 9 of the substitutions shovm in Table 2. These substitutions are believed to result in increased affinity for CD16A and/or reduce the immunogenicity of a CD16A binding, protein when administered to humans. In certain embodiments, the degree of sequence identity at the remaining positions is at least about 80%, at least about 90%, at least about 95% or at least about 98%. Table 2 VL Domain Substitutions No. Kabat Position Region Substitutions 1 24 CDRl Arg 2 25 CDRl Ser 3 32 CDRl Tyr 4 33 CDRl Leu 5 50 CDR2 Asp or Trp or Ser 6 51 CDR2 Ala 7 53 CDR2" Ser 8 55 CDR2 Ala or Gin 9 56 CDR2 Thr 10 93 CDR3 Ser 11 94 CDR3 Thr [00S2] For illustration and not limitation, tiie sequences of a number of CD16A binding protein VL domains is shown in Table 4. As described in the Examples, infra, light chains comprising these sequences fused to a human CK constant domain were coexpressed with the Hu3G8VH-l heavy chain (described above) to form tetrameric antibodies, and the binding of the antibodies to CD16A was measured to assess the effect of amino acid substitutions compared to the Hii^GSVL-l VL domain. Cohstmcts in which the VL, domain has a sequence of hu3G8VL-l, 2,3,4, 5,10,16,18,19,21,22,24,27,28,32,33, 34,35,36, 37, and 42 showed high affinity binding and hu3G8VL-15,17,20,23, 25,26,2% 30, 31, 38, 39,40 and 41 showed intermediate binding. CD 16A binding proteins comprising the VL domains of hu3G8VL-l, hu3G8VL-22, and hu3G8VL-43 are considered to have particularly favorable binding properties. C. Combinations ofVi and/or VH domains [0083] As is known in the art and described elsewhere herein, immunoglobulin light and heavy chains can be recombinantly expressed under conditions in which they associate to produce a tetrameric antibody, or can be so combined in vitro. Similarly, combinations of VL and/or VH domains can be expressed in the form of single chain antibodies, and still other CD16A binding proteins that comprise a VL and/or VH domain can be expressed by known methods. It will thus be appreciated ttiat a 3G8-derived VL-domain described herein can be combined with a 3G8-derived VH-domain described herein to produce a CD16A binding protein, and all such combinations are contemplated. [0084] For illustration and not for limitation, examples of useful CD16A binding proteins are those comprising at least one VH domain and at ]east one VL domain, where the VH domain is from hu3G8VH-l, hu3G8VH-22 or hu3G8VH-5 and the VL domain is &om hu3G8VL-l, hu3GSVL-22 orhu3G8VL-43. In particular, humanized antibodies that comprise hu3G8VH-22 and either hu3G8VL-l, hu3GSVL-22 orliu3GSVL-43, orhu3G8VH-5 andhu3GSVL-l have favorable properties. [0085] It will be appreciated by those of skill that the sequences of VL and VH domains described here can be further modified by art-known methods such as affinity maturation (see Schier et al., 1996, J. Mol. Biol 263:551-67; Daugherty et al., 1998, Protein Eng. 11:825-32; Boder et al.. 1997, Nat. Biotechnol. 15:553-57; Boder et al., 2000, Proc. Natl Acad. Sci. U.S.A 97:10701-705; Hudson and Souriau, 2003, Nature Medicine 9:129-39). For example, the CD16A binding proteins can be modified using affinity maturation techniques to identify proteiia with increased affinity for CD 16A and/or decreased affinity for CD 16B. D. Constant Domains and Fc Region [0086] As noted above, the CDl 6A binding protein of the invention may contain light chain and/or heavy chain constant regions (including the hinge region connecting the CHI and CH2 domains in IgG molecules). It is contemplated that a constant domain from any type (e.g., IgM, IgG, IgD, IgA and IgE) of immunoglobulin may be used. The constant domain for the light chain can be lambda or kappa. The constant domain for the heavy chain can be any isotype {e.g., IgGi, IgG2, IgGa and IgG4). Chimeric constant domains, portions of constant domains, and variants of naturally occurring human antibody constant domains (containing deletions, insertions or substitutions of amino acid residues) may be used. For instance, a change in the amino acid sequence of a constant domain can be modified to provide additional or different properties, such as altered jmmunogenicity or half-life of the resultant polypeptide. The changes range fi:om insertion, deletion or substitution of a small number (e.g.,, less than ten, e.g., one, two, three or more) amino acid residues to substantial modifications of the constant region domain. Changes contemplated include those that affect tbe interaction with membrane receptors, complement fixation, persistence in circulation, and other effector functions. For example, the hinge or other regions can be modified as described in U.S. pat. no. 6,277,375. Inparticxdar, it will often be advantageous to delete or alter amino acids of the Fc region. For example, the Fc region can be modified to reduce or elimmate binding to Fc effector hgands such as FcyRIII and the Clq component of complement, such that the antibodies lack (or have substantially reduced) effector fimction. Antibodies having such modified Fc regions induce little or no antibody dependent cellular cytotoxicity (ADCC) and/or complement mediated lysis when administered to a mammal, compared to unmodified antibodies. Assays to identify antibodies lacking effector function are known in the art. See, e.g., U.S. Pat. Nos. 6,194,551; 6,528,624; and 5,624,821, European Pat. No. EP 0 753 065 Bl, and PCX publication WO 00/42072. [0087] The CD16A binding protein of the invention may include a himian IgGi Fc domain comprising one or more amino acid substitutions or deletions (relative to the parental naturally occurring IgGi) that result in a reduced interaction between the Fc domain of the binding protein and FcyRIIA and/or FcyRIHA (e.g., to minimize potential activation of macrophages and/or minimize neutrophil diminution) and/or increased bindii^ of the Fc region to Fc/RIIB (e.g., to increase FcyRIIB-mediated inhibition of effector cell activation; see BoUand andRavetch, \999, Adv. in Immunol. 72:149). Specific mutations effecting the desired changes in binding can be identified by selection using display of mutant Fc libraries expressed on the surfece of microorganisms, viruses or mammalian cells, and screening such libraries for mxrtaut Fc variants having the desired property or properties. In addition, the literature reports that particular residues or regions of the Fc are involved in Fey interactions such that deletion or mutation of tiiese residues would be expected to result in reduced FcR binding. The binding site on human antibodies for FcyR was reported to be the residues 233-239 (Canfield et al„ 1991, J Exp Med 173:1483-91; Woof et al, 1986, Mol Imm. 23:319-30; Duncan et al., 1988, Nature 332:563). The crystal stiructure of FcyRIII complexed with human IgGI Fc revealed potential contacts between the receptor and its ligand and also revealed that a single FcyRin monomer binds to both subunits of the Fc homodimer in an asymmetric fashion. Alanine-scanning mutagenesis of the Fc region confirmed the importance of most of the predicted contact residues (Shields et al., 2001, J Biol. Chem. 276:6591-6604), [00881 Exemplary Fc region mutations uiclude, for example, L235E, L234A, L235A, and D265A, which have been shown to have low affinity for all FcR, mto CY-1 (Clynes et al., 2000, Nat. Med. 6:443-46; Alegre et al., 1992, J Immunol 148:3461-68; Xu et al., 2000, Cell Immunol 200:16-26). Additional Fc region modifications purported to affect FcR binding are described in WO 00/42072 (e.g., "class 4" Fc region variants) and WO 02/061090. [0089] Fc binding to FcyRJIA and FcylllA or other proteins can be measured by any of a number of methods, including ELIS A to measure binding to isolated recombinant FcyR and RIA or FACS to measure binding to cells. Immune complexes and heat aggregated or chemically crosslinked Fc or IgG can be used to test affinity for FcRs in such assays. In one embodiment, immune complexes are produced by expressing an Fc m the context of an Fab with afiinity for an antigen (such as fluorescein) and mixing the antibody and antigen to form an immune complex. E. Fc Regions with Reduce^ Binding to Fc Effector Ligands Due to Aglycosylation or Changes in Glycosylation [00901 As discussed above, in CD16A binding proteins of the invention that comprise Fc domains (e.g., anti-CD16A monoclonal antibodies) the Fc domain can be modified to achieve desired properties. In a particular aspect, the invention provides a CD16A binding protein, such as a human or humanized anti-CD16A monoclonal antibody, comprising an Fc region that is not glycosylated. As demonstrated in Example 10, infra, the inventors have discovered that, unexpectedly, administration of anti-CD16A antibodies with altered effector fimction (aglycosylated antibodies) protects against autoimmune disorders without inducing acute severe neutropenia. On the basis of this discovery, therapeutic anti-CD16A antibodies can be designed to protect against autoimmune diseases without inducing dangerous side effects. [00911 ^ one embodunent, the invention provides a CD 16A binding protein comprising an Fc region derived from human IgGi, where the amino acids corresponding to position 297 of the CH2 domains of the Fc region are aglycosyl. The terms "aglycosyl" or "aglycosylated," when referring to an Fc region in its entirety, or a specific amino acid residue in the Fc region, mean that no carbohydrate residues are attached to the specified region or residue. [0092] Human IgG antibodies that are aglycosylated show decreased binding to Fc effector ligands such as Fc receptors and CI q (see, e.g., Jefferisetal., 1995, Immunology Letters 44:111-17; Tao, 1989,^ of Immunology, 143:2595-2601; Friend et al., 1999, Transplantation 68:1632-37; Radaev and Sun, 2001, ^ of Biological Chemistry 276:16478-83; Shields et al, 2001, J! of Biological Chemistry 276:6591-6604, and U.S. Patent 5,624,821). Without intending to be bound by a particular mechanism, it Is believed that the aglycosylation of the amino acid at position 297 of the Fc domains of CD16A binding proteins described herein results in reduced binding to CD16A and the Clq component of complement Such aglycosylated antibodies lack effector function. [0093] In human IgGi constant regions, the residue at position 297 is asparagine. In one embodiment of the present invention, the residue at, or corresponding to, position 297 of the Fc region of the CD16A binding protein is other than asparagine. Substitution of another amino acid residue in the place of asparagine eliminates the N-glycosylation site at position 297. Substitution of any amino acid residues which will not result in glycosylation upon expression of the CDl 6A binding protein in a mammalian cell is appropriate for this embodiment. For instance, in some embodiments of the invention, the amino acid residue at position 297 is glutamine or alanine. In some embodiments, the amino acid residue at position 297 is cysteine, which is optionally linked to PEG. [0094] In other embodiments of the invention, the residue at position 297 may or may not be asparagine, but is not glycosylated. This can be accompHshed in a variety of ways. For example, amino acid residues other than the asparagine at position 297 are known to be important for N-linked glycosylation at position 297 (see Jefferis and Lund, 1997, Chem. Immunol. 65:111-28), and the substitution of residues at positions other than position 297 of the CH2 domain can result in a CD 16A binding protein aglycosylated at residue 297. For illustration and not limitation, a residue at position 299 in ttie CH2 domain that is other than threonine or serine will result in an antibody that is aglycosylated at position 297. Similarly, substitution of the amino acid at position 298 with proline will prodiwe an antibody with an aglycosylated amino acid at position 297. In other embodiments of the invention, Fc domains of IgG2 or IgG4 are used rather than IgGi domains. [009SJ Modification of the amino acid residues of CD16A binding proteins is well within the ability of the ordinarily skilled practitioner, and can be achieved by mutation of a polynucleotide encoding the binding protein or portion thereof. The CDI6A binding protein comprising an IgG-derived Fc region need not necessarily be mutated at the amino acid level to be aglycosylated. Binding proteins aglycosylated at position 297 of the IgG-derived Fc region can be produced by expressing the CD16A binding protein in certain cells (e.g., E. coli; see PCT publication WO 02061090A2), cell lines or under certain cell culture growth conditions where glycosylation at Asn 297 does not take place. Alternatively, carbohydrate groups may be removed from a CD16A binding protein following expression of the protein, e.g., enzymatically. Methods for removing or modifying carbohydrate groups on proteins are known and include use of endoglycosidases and peptide:N-glycosidases. [0096] It wll be apparent that a variety of methods can be used to modify the Fc region of a CD16A binding protein to change its properties. Accordingly, unless otherwise specified, as used herein fhs term "modifyTDg" in the context of modifying the Fc region of a CD16A binding protein includes modifying the protein itself directly, modifying the polynucleotide that encodes the protein and/or modifying or selecting a stiitable expression system production of the protein. [0097] In addition to CDI6A binding proteins that are agfycosylated at the position corresponding to arginine 297, variants with reduced binding to Fc effector ligands due to only partial removal, or modification, of the carbohydrate at that position may be used in the present inventioiL For example, the Fc region can be modified to include a non-naturally occurring carbohydrate that does not bestow binding protein with effector iunction. As used herein, a "modified Fc region" is an Fc region that has been derived from a parent Fc region, but which differs in glycosylation pattern from the parent Fc region. F. Production ofCDJ6A Binding Proteins [0098] CD16A binding proteins of the invention can be produced using a variety of methods well known in the art. including de novo nrotein synthesis and recombinant expression of nucleic acids encoding the binding proteins. The desired nucleic acid sequences can be produced by recombinant methods (e.g., PCR mutagenesis of an earlier prepared variant of the desired polynucleotide) or by solid-phase DNA synthesis. Usually recombinant e3q)ressJon methods are used. In one aspect, the invention provides a polynucleotide that comprises a sequence encoding a CD16A binding protein disclosed herein or a CD16A binding fragment thereof, for example a sequence encoding a VL or VH described herein, or antibody heavy chain or light chain described herein. Because of the degeneracy of the genetic code, a variety of nucleic acid sequences encode each immunoglobulin amino acid sequence, and the present invention includes all nucleic acids encoding the binding proteins described herein. (0099] Recombinant expression of antibodies is well known in the art and can be carried out, for example, by inserting nucleic acids encoding light and heavy chain variable regions, optionally linked to constant regions, into expression vectors. Expression vectors typically include control sequences such as a promoter, an enhancer, and a transcription termination sequence to which DNA segments encoding polypeptides (e.g., immunoglobulin chains) are operably linked to ensure the expression of immunoglobulin polypeptides. Eiqiression vectors are typically replicable in the host organisms either as episomes or as an integral part of the host cbromosomal DNA. The light and heavy chains can be cloned in the same or different expression vectors. [0100] Immunoglobulin light and heavy chains are exjsressed usir^ standard methods. A multiple polypeptide chain antibody or antibody fragment species can be made in a single host cell expression system wherein the host cell produces each chain of the antibody or antibody fragment and assembles the polypeptide chains into a multimeric structure to form the antibody or antibody fragment in vivo. Seee.g.,Lucasetal.,1996,A/uc/eic^ci£fei!e5.,24:1774-79. Whenheavy and light chains are cloned on separate expression vectors, the vectors are co-transfected to obtain expression and assembly of intact immimoglobulins. Alternatively, recombinant production of antibody heavy and light chains in separate expression hosts followed by assembly of antibody from separate heavy and light chains in vitro is known. See, e.g., U.S. Pat. No. 4,816,567 and Carter et al., 1992, Bio/Tecknology 10:163-67. [0101] The CD16A binding proteins are conveniently expressed in procaryotic or eukaryotic ceils. Useful hosts for antibody expression include bacteria (see, e.g., PCT publication WO 02/061090), yeast (e.g., Saccharomyces), insect cell culture (Putlitz et al., 1990, Bio/Technology 8:651-54), plants and plant cell cultures (Larrick and Fry, \99l. Hum. Antibodies Hybridomas 2:\12-Z9), and mammalian cells. Methods for expression are well known in the art. For example, in E. coli, vectors using the lac promoter to drive expression of heavy fd and light chains fused to various prokaryotic secretion signal sequences such as pelB have resulted in successful secretion of scFv and Fab fragments into the periplasmic space or into the culture medium (Barbas et al., 1991, Proc.NatLAcadSci. U.S.A 88:7978-82). A vector derived from pET25b in which the lac promoter has been inserted in place of the T7 promoter may be used, [0102] Mammalian cells are especially usefiil for producing CD16A binding proteins, including tetrameric antibodies or fragments thereof. A number of suitable host cell lines capable of secreting intact heterologous proteins are known, and mclude CHO cell lines, COS cell lines, HeLa cells, L cells and myeloma cell lines. Expression vectors for mammalian cells can include expression control sequences, such as an origin of replication, a promoter, an enhancer, ribosome binding sites, RNA splice sites, polyadenylation sites, and transcriptional terminator sequences. Examples of expression control sequences are promoters derived from endogenous genes, cytomegalovirus, SV40, adenovirus, bovine papillomavirus, and the like. In one embodiment, binding proteins are expressed using the CMV immediate early enhancer/promoter in the vector pCDNA3.1 or a similar vector. To facilitate secretion, the genes can be fused to a gene cassette containing the signal sequence of a mouse VH gene described by Orlandi et al., 1989, Proc. Natl. Acad Sci. U.S.A 86:3833-37, vdiich has been widely used for high-level secretion of immimoglobulins. [0103] The vectors containing the DNA segments encoding the polypeptides of interest can be transferred into the host cell using routine, depending on the type of cellular host. For example, calcium chloride transfection is commonly utilized for prokaryotic cells, whereas calcium phosphate treatment, electroporation, lipofection, biolistics or viral-based transfection may be used for othCT celliUar hosts. Other mettiods used to transfonn mammalian cells include the use of polybrene, protoplast fusion, hposomes, electroporation, and microinjection (see generally, Sambrook et al., supra). For transient expression, cells, e.g., HEK293, can be co-transfected with separate heavy and light chain expression vectors using a cationic lipid (e.g., Lipofectamine 2000, Invitrogen). This method can achieve expression levels of 10-20 mg/1 of IgG in conditioned medium after 3 days. The cells can then be re-fed and similar quantities harvested after 3 more days. It will be appreciated that, for some uses, the cells expressing CD16A binding proteins can be maintained in medium containing FBS screened for very low levels of bovine IgG, or, alternatively, in serum-free medium. 10104} In addition to expression of tetrameric antibodies, single chain antibodies, antibody fragments, and other CD16A binding proteins can be prepared. For example, immunoglobulin fragments can be prepared by proteolytic digestion of tetrameric antibodies, or more often, by recombinant expression of truncated antibody constructs. Usually, single chain V region ("scFv") constructs are made by linking VL and/or VH domain using a short hnking peptide (see, e.g.. Bird et al., 1988, Science 242:423-26; Pat. Nos. 4,946,778; 5,455,030; 6,103,889; and 6,207,804). [0105] Once expressed, the binding proteins can be purified using procedures well known in the art, including ammonium sulfate precipitation, afSnity chromatography, gel electrophoresis and the like (see, generally, Harris and Angal, 1990, PROTEIN PURIFICATION APPLICATIONS, A PRACTICAL APPROACH Oxford University Press, Oxford, UK; and Coligan et al., supra). In one embodiment, purification is accomplished by capturing the antibody usmg a high flow rate protein A resin such as Poros A (Perseptive Biosystems, Inc), and elution at low pH, followed by size exclusion chromatography to remove any traces of aggregate present. Since FcyRUIA binds preferentially to aggregated IgG, removal of aggregates will be desirable for certain applications. The binding proteins can be purified to substantial purity if desired, e.g., at least about 80% pure, often at least about 90% pure, more often least about 95%, or at least about 98% pure. In this context, the percent purity is calculated as a weight percent of the total protein content of the preparation, and does not include constituents which are deliberately added to the comnosition after the binding protein is purified. [0106] CD16A binding proteins can be modified after expression. For example, derivation of antibodies with polyethylene glycol ("pegylation") is reported to increase residence time (half-life and stability) and reduce immunogenicity in vivo without alteration of biological activity. See, e.g., Leong et al., 2001, Cytokine 16:106-19; Koumenis et al., 2000, IntJPharm 198:83-95; U.S. Pat. No. 6,025,158. CD16A binding proteins can be conjugated to a detectable label or ligand (e.g., a radioisotope or biotin). Other modifications are well known in the art and are also contemplated. G. Properties ofCD16A Binding Proteins [0107] In certain embodiments, CDI6A bindii^ proteins having properties as described below are used in the methods of the invention. i) Binding Affinity [0108] CD16A binding proteins can be described by reference to their binding properties and biological activity. In various embodiments, the binding constant for the interaction of a CD 16A binding protein of the invention and CD 16A is between 0.1 and 5 nM, less than about 2.5 nM, less than about 1 nM, or less ttian about 0.5 nM. Usually the binding protein binds CDl 6A with an affinity that is within 4-fold, optioi^Uy within 2-foId, of the binding affinity exhibited imder similar conditions by 3G8 or the chimeric antibody comprising the heavy chain Ch3G8VH and the light chain Ch3G8VL as described herein below. In an embodiment, the binding af&nity for CD16A is greater than that of 3G8. In an alternative embodiment, the binding affinity for CD16B is no greater than, and preferably less than, 3G8 or the chimeric antibody Ch3G8. [0109] Binding can be measured using a variety of methods, including ELISA, biosensor (Tdnetic analysis), and radioimmunoassay (RIA). ELISA is well known (see, Harlow and Lane, supra, and Ausubel et al., supra) and can be carried out using conditioned medium containing binding proteins or, alternatively, with purified antibodies. The concentration of antibody that results in 50% apparent maximal binding provides an estimate of antibody Kd. [0110] Binding can also be detected using a biosensor assay, which also provides information on the kinetic and equilibrium properties of antibody binding to FcyRIUA. An exemplary biosensor assay uses the BIAcore system (Mahnqvist et al., 1997, Curr. Opin. Chem. Biol 1:378-83). The BIAcore system relies on passing analyte over a sensor chip onto which the ligand (e.g., CD16A) is immobilized. The bindii^ of the analyte can be measured by following surface plasmon resonance (SPR) signal, which changes in direct proportion to the mass boimd to the chip. A fixed concentration of analyte is passed over the chip for a specific amount of time, allowing for the measurement of the association rate, k(on). Following this phase, buffer alone is passed over the chip and the rate at which the analyte dissociates from the surface, k(off) can be measured. The equihbrium dissociation constant can be calculated from the ratio of the kinetic constants; Kd = k(on)/k(off). [0111] A radioimmunoassay (RIA) can be used to measure the affinity of antibodies for FcyRIII-bearing cells, and to measure inhibition of IgG complexes to cells by these antibodies. In an exemplary assay, "^^I labeled binding protein is prepared and specific radioactivity of the protein determined. Labeled binding protein and cells are mixed for several hours, the cells and bound material are separated from the unbound material by centrifugation, and the radioactivity in both compartments is determined. A direct binding format is used to determine the Kd of, and the number of binding sites for, iodinated binding protein using Scatchard analysis of the binding data. Controls containing an excess of cold (unlabeled) binding protein competitor can be included to ensure the results reflect specific interactions. Examples of suitable cells include (1) NK cells or macrophages derived from normal human peripheral blood lymphocytes; (2) Cells obtamed from huCD16A transgenic mice (Li, 1996 J. Exp. Med. 183:1259-63); (3) mammalian cell lines expressing the extracellular portion of CD16A ilised to the transmembrane and intracellular domain of RII or another receptor (such as CDS or LFA-3); (4) mammalian cell Unes (e.g., CHO, HEK-293, COS) transfected transiently or stably with CDl 6A expression vectors (and optionally coexpressing gamma chain for optimal expression receptor expression). [0112] Examples of expression vectors useful for ejq)ression of CD16A and other polypeptides for use in binding assays include mammalian expression vectors (e.g., pCDNA 3.1 or pCI-neo) that contam a strong promoter/enhancer sequence (e.g., CMV immediate early) and a polyadenylation/transcription termination site flanking apolylinker region into which the CD16A gene is introduced. Usually the vector includes a selectable marker such as a neomycin resistance gene. [0113] In one embodiment, the CD16A expressed for use in assays has the sequence: MWQLLLPTALLLLVSAGMRTEDLPKAWFLEPQWYRVLEKDSVTLKCQGAYSPED NSTQWFHNESLISSQASSYFIDAATVDDSGEYRCQTNLSTLSDPVQLEVHIGWLL LQAPRWVFKEEDPIHLRCHSWKNTALHKVTYLQNGKGRKYFHHNSDFYIPKATLK DSGSYFCRGLFGSKNVSSETVNITITQGLAVSTISSFFPPGYQVSFCLVMVLLFA VDTGLYFSVKTNIRSSTRDWKDHKFKWRKDPQDK (SEQ IDN0:116). CD16A with the sequence: MWQLLLPTALLLLVSAGMRTEDLPKAWFLEPQWYRVLEKDSVTLKCQGAYSPED NSTQWFHNESLISSQASSYFIDAATVDDSGEYRCQTNLSTLSDPVQLEVHIGMLL LQAPRWVFKEEDPIKLRCHSWKNTALHKVTYLQNGKGRKYFHHNSDFYIPKATLK DSGSYFCRGLVGSKNVSSETVNITITQGLAVSTISSFFPPGYQVSFCLVMVLLFA VDTGLYFSVKTNIRSSTRDWKDHKFKWRKDPQDK (SEQ IDN0:117) can also be used. Additional CD16A variants and substitutes will be known to, or readily discernible firom the scientific literature by, the ordinarily skilled reader. [0114] Competitive assay formats can be xised to measure the ability of a CD16A binding protein to inhibit binding of another molecule to the receptor. For example, in one competitive assay format a fixed amount of labeled 3G8 is mixed with varying amounts of either unlabeled 3G8, CD16A binding protein or an irrelevant IgG (control) and added to FcyRHIA expressing cells. After incubation and separation of the cell-bound material fix)m the material free in solution, the amount of bound labeled 3G8 (and/or optionally also the unbound labeled 3G8) is determined. The concentration of unlabeled mAb which results in a 50% decrease in the binding of labeled 3G8 (IC50) is then determined from this data. ii. Blocking Immune Complex Binding to FCYRIHA [0115] Another characteristic of the CD16A binding proteins of the invention is the ability to inhibit binding of immune complexes to CD 16 A ("IC Blocking Activity"). Usually the binding protein has IC Blocking Activity that is within 4- fold, preferably within 2-fold, of the activity exhibited under similar conditions by 3G8 or the chimeric antibody, Ch3G8, described herein. [0116] Assays for measuring ability of an antibody to block binding of compiexed IgG to CD16 are known. See, e.g., Knapp etai, 1989, LEUKOCYTE TYPING IV, Oxford University Press, Oxford, p.574-97; and Edberg and Khnberty, 1991, J Immunol 159:3849-57. One suitable assay is an RIA assay with the format described above for the competitive assay, but substituting "^^I-labeled aggregated irrelevant human IgGi for the "^^I-labeled 3G8 used in the competitive assay described above. [0117] The invention provides a method of inhibiting the binding of IgG antibodies to CD 16 on a cell by contacting the cell with a CD 16A binding protein under conditions in which the CD16A binding protein binds the FcyRIII on the cell. The contacting can be in vivo (e.g., by administering the binding pwtem in a mammal) or i« vitro (e.g., by addition of antibodies to cultured cells expressing the FcyRin). IgG antibodies that are inhibited from binding the FcyRIU can be administered to the animal or added to a culture medium before or after addition or administration of the binding protein, or may be present in an animal normally or in re^jonse to a disease state. In one embodiment, the CD16 on the surface of the cell is CD 16A. iii. In Vivo Protection Against Platelet Depletion [0118] The ability of the CD16A binding proteins of the invention to reduce deleterious immune responses can be assessed in a variety of animal models. An exemplary model system is a mouse model for idiopathic fhrombocytopenic purpura (IT?) (see, Oyaizu et al., 1988, JExp.Med. 167:2017-22; Mizutani et al, 1993, Blood 82:837-44). See"Esamph9,in/ra. Other suitable modeis are known in HK art. Other animal models include rodent models of inflammatory diseases described in, for example. Current Protocols in Immunology (m some cases modified by using animals transgenic for human CD16A). Transgenic mice can be made using routine methods or can be purchased from commercial sources (e.g., Taconic Inc., German Town New York). [0119] A example of a procedure suitable for assessing the ability of a GDI 6A binding protein to provide protection from platelet depletion in a mouse model is described in Example 8, infi-a. CD16A bindmg proteins can be administered to muFcyRIII-/-, huFcYRIUA transgenic mice at a variety of concentrations, and ITP subsequently induced in the mice (e.g., by administering the 6A6 or chimeric 6A6 antibody) to the mice. At timed intervals after the administration of 6A6/ch6A6, the mice are bled and the platelet counts are determined. Optionally, the ICso for each molecule is then determined at the time point where maximal platelet depletion is observed in the negative control group. Based on the results of Example 8 and on prior studies, maximum depletion occurred 2-6 hr after 6A6 administration. IC50S are determined graphically, using a curve-fitting program such as the four-parameter fit provided in the SigmaPlot program. Statistically significant inhibition of depletion of platelets after administration of 6A6 in the treatment group compared to the untreated group and a group administered an identical formulation of an irrelevant, isotype matched mAb is indicative of the desired biological activity. [0120] Experiments in which protection by CD16A binding proteins was assayed are described in the Examples, infra. Preparations of recombinant mouse 3G8 produced in HEK-293 cells, chimeric 3G8 witii human IgGl or IgG2 constant domains (ch3G8-7l produced in HEK-293 and CHO-Kl cells, and ch3G8-72 produced in HEK-293 cells), and a ch3G8-yl variant (cli3G8-7l D265A) did not provide significant protection. Murine 3G8, produced firom the hybridoma, and a chimeric version of 3G8 containing an aglycosylated human Gl constant region (Ch3G8-Gl N297Q), produced in HEK-293 cells, were able to protect animals from platelet depletion in the mouse model. As shown in Example 10,11 and 1547, infra, Ch3G8 N297Q and aglycosylated humanized antibodies protected against platelet depletion in the ITP mouse model. Although not intending to be bound by a particular theory, one possibility is that since ch3G8 N297Q is largely devoid of effector fimction, it is more efficient than ch3G8 in protecting mice against ITP. Thus, these data suggest that anti-CD16A antibodies without effector function (e.g., aglycosylated antibodies) have advantages compared to some glycosylated (e.g., glycosylated recombinant) antibodies. Further, as described in the examples, administration of aglycosylated anti-CDl 6A antibody to muFcgRIH-/-, huFcRHIB transgenic mice did not result in neufrophil depletion in the blood, spleen, and bone marrow. Without intending to be boimd by a particular theory, there are several possible explanations for these unexpected results. Protein glycosylation is known to vary in different cell lines, especially those from different species. A difference in the nature of the carbohydrate attached to the antibody constant region as a consequence of expression in different cell types may be responsible for the difference in activity, i.e., if the lack of activity results in part from effector cell activation caused by ch3G8 binding to Fc receptors (or complement) via the antibody Fc region in a glycosylation-dependent manner. Alternatively, recombinant murine and ch3G8 may contain other post-translational modifications that affect activity and which can be eliminated by using different cell lines to express the CD16A binding proteins. It is possible that a combination of isotype and/or isotype containing mutations to eliminate effector function may provide similar protective effects as elimination of the carbohydrate on the Fc. 5. Methods of Treatment [0121] A number of diseases and conditions characterized by an deleterious immune response can be treated using the binding proteins of the invention a CD16A binding protein as described herein (e.g., comprising a VL and/or VH sequence as disclosed herein and, optionally, a Fc region modified as disclosed herein to have a reduced effector function). In one embodiment, the binding protein is administered to a subject with an autoimmune disease (i.e., a disease characterized by the production of autoantibodies). It is believed that pathogenic IgG antibodies observed in autoimmune diseases are either the pathogenic triggers for these diseases or contribute to disease progression and mediate disease through the inappropriate activation of cellular Fc receptors. Aggregated autoantibodies and/or autoantibodies complexed viith self antigens (immune complexes) bind to activating FcRs, thereby triggering the pathogenic sequelae of numerous autoimmune diseases (which occur in part because of immunologicaliy mediated inflammation against self tissues). Without intending to be bound by a particular mechanism of action, the CD16A binding proteins described herein interfere with and reduce the interaction of the autoimmune antibodies and FcyRin receptors. [0122] Examples of autoimmune diseases that can be treated include, without limitation, idiopathic thrombocytopenic purpura (FTP), rheumatoid arthritis (RA), systemic lupus erythrematosus (SLE), autoimmune hemolytic anemia (AHA), sclerodeima, autoantibody triggered urticaria, pemphigus, vascuUtic syndromes. systemic vasculitis, Goodpasture"s syndrome, multiple sclerosis (MS), psoriatic arthritis, anlcylosing spondylitis, Sjogren"s syndrome, Reiter"s syndrome, Kowasaki"s disease, polymyositis and dermatomyositis. Other examples of diseases or conditions that can be treated according to the invention also include any diseases susceptible to treatment with intravenous immunoglobulin (TVIG) therapy (e.g., allergic asthma). Thus, the treatment of autoimmune diseases heretofore treated by IVIG therapy (in one embodiment, a condition other than ITP) is contemplated. While detailed luiderstanding of the mechanism of action of IVIG has not been established, it is proposed that modulating the activity of cellular FcyRs plays a role in its in vivo efficacy. The protective activity of IVIG may rely on the small percentage of dimeric or polymeric IgG present in the preparation. The specificity of the FcyRIII pathway in coupling cytotoxic and immune complex antibodies to effector responses and the ability to directly block this pathway with a mAb strongly suggests that an anti-FcYRin antibody will have enhanced activity relative to IVIG. [0123] A reduction in a deleterious immune response can be detected as a reduction in inflammation. Alternatively, a reduction in a deleterious immune response can be detected as a reduction in symptoms characteristic of the condition being treated (e.g., a reduction in symptoms exhibited by a subject suffering fi:om an autoimmune condition), or by other criteria that will be easily recognized by physicians and experimentahsts inthefieldof automimmunity. ft will be apparent that, in many cases, specific indicia of reduction will depend on the specific condition beii^ treated. For example, for illustration and not limitation, a reduction in a deleterious immune response in a subject with ITP can be detected as a rise in platelet levels in the subject. Similarly, a reduction in a deleterious immime response in a subject with anemia can be detected as a rise in RBC levels in the subject. A clinician will recognize significant changes in platelet or RBC levels, or other reponses following treatment. [0124] The deleterious immune response is optionally due to idiopathic thrombocytopenic purpura resulting from the administration of an antiplatelet antibody, optionally murine monoclonal antibody 6A6, to a muFcyRIII-/-, huFcyRIHA transgenic mouse. [0125] In one aspect, the invention provides a method for treating an autoimmune disease, such as ITP, by administering a CD16A binding protein that is largely devoid of effector function. In an embodiment, the CDI6A binding protein comprises Fc regions derived from human IgG. In an embodiment, the Fc regions are aglycosyl. In an embodiment, the position 297 of each of the CH2 domains is a residue of than asparagine or proline. In one aspect, the binding protein comprises a variable region sequence as described elsewhere herein. However, as discussed hereiu, the compositions and treatment methods of the invention are not limited to specific CD 16A binding proteins derived from murine mAb 3G8, but are applicable to CD16A binding proteins in general. In an embodiment, the CD16A binding protein is a teframeric antibody protein having two light chains and two heavy chains. [0126] In a related aspect, the invention provides methods of reducing an deleterious immune response in a mammal without significantly reducing neutrophil levels or inducing neutropenia (e.g., severe neutropenia or moderate neutropenia) by administering to the mammal a tierapeuticaUy effective amount of a phaimaceutical composition comprising a CD 16 A binding proteiu described herein. In an embodiment, the mammal is human. In an embodiment, the mammal is a nonhuman mammal (e.g., mouse) comprising one or more human fransgenes. [0127] For therapeutic applications, the binding proteins of the invention are formulated with a phaimaceutically acceptable excipient or carrier, e.g., an aqueous carrier such as water, buffered water, 0.4% saline, 0.3% glycine and the like, optionally including other substances to increase stability, shelf-life or to approximate physiological conditions (sodium acetate, sodium chloride, potassium chloride, calcium chloride, sodium lactate, histidine and arginine). For administration to an individual, the composition is preferably sterile, and fii^e of pyrogens and other contaminants. The concenfration of binding protein can vary widely, e.g., from less than about 0.01%, usually at least about 0.1% to as much as 5% by weight. Methods for preparing parentally administerable compositions are known or apparent to those skilled in the art and are described in more detail in, for example, Remii^ton, THE ScffiNCE OF PRACTICE AND PHARMACY, 20th Edition Mack Publishing Company, Easton, Pa., 2001). The pharmaceutical compositions of the invention are typically administered by a parenteral route, most typically intravenous, subcutaneous, intramuscular, but other roxrtes of administration can be med (e.g., mucosal, epidermal, intraperitoneal, oral. intranasal, and intrapulmonary). Although not required, phannaceutical compositions are preferably supplied in unit dosage form suitable for administration of a precise amount. In one embodiment, CD16A binding proteins can be administered in a form, formulation or apparatus for sustained release (e.g., release over a period of several weeks or months). [0128] In one embodiment, polynucleotides encoding CD16A binding proteins (e.g., CD16A binding protein expression vectors) are administered to a patient. Following administration, the CD 16A binding protein is expressed in the patient. Vectors usefiil in administration of CD16A binding proteins can be viral (e.g., derived fi-om adenovirus) or nonvira]. Usually the vector will comprise a promoter and, optionally, an enhancer that serve to drive transcription of a protein or proteins. Such therapeutic vectors can be introduced into cells or tissues in vivo, in vitro or ex vivo. For ex vivo therapy, vectors may be introduced into cells, e.g., stem cells, taken from the patient and clonally propagated for autologous transplant back into the sanae patient (see^ e.g., U.S. Patent Nos. 5,399,493 and 5,437,994). [0129] The compositions can be administered for prophylactic and/or therapeutic treatments. In prophylactic applications, compositions are administered to a patient prior to an expected or potential deleterious immune response. For example, idiopathic thrombocytopenic purpura and systemic lupus erythrematosus are conditions in which an deleterious immune response can be exacerbated by administration of certain medications. The CD16A binding compositions of the invention can be administered in anticipation of such medication-induced responses to reduce the magnitude of the response. In therapeutic applications, compositions are administered to a patient already suffering from an deleterious iromune response in an amount sufficient to at least partially ameliorate the condition and its complications. An amount adequate to accomphsh this may be a "therapeutically efiective amoimt" or "therapeutically effective dose." Amounts effective for these uses depend upon the severity of the condition and the general state of the patient"s own unmune system, but generally range from about 0.01 to about 100 mg of antibody per dose, with dosages of from 0.1 to 50 mg and 1 to 10 mg per patient being more commonly used. An "inflammation reducing amoimt" of the binding protem can also be administered to a mammal to reduce a deleterious immune response. [0130] The administration of the CDl 6A binding proteins can be administered according to the judgement of the treating physician, e.g., daily, weekly, biweekly or at any other suitable interval, depending upon such factors, for example, as the nature of the ailment, the condition of the patient and half-Ufe of the binding protein. [0131] CD16A bmding proteins can be administered in combination other treatments directed to alleviation of the deleterious immune response or its symptoms or sequalae. Thus, the binding proteins can be administered as part of a therapeutic regimen that includes co-administration of another agent or agents, e.g., a chemotherapeutic agent such as a non-steroidal anti-inflammatory drug (e.g., aspirin, ibuprofen), steroids (e.g., a corticosteroid, prednisone), immunosuppressants (e.g., cyclosporin A, methotrexate Cytoxan), and antibodies (e.g., in conjunction with IVIG). 6. Increasing the Therapeutic Efficacy of a CDJ6A Binding Protein [0132] In a related aspect, the invention provides a method for increasing the therapeutic efficacy of a CD16A binding protein comprising one or more Fc domains (e.g,, anti-CD16A antibodies comprising two Fc domains) by modifying the protein f o that it has Fc region(s) with reduced binding to at least one Fc efifector ligand compared to the original (i.e., xmmodified) Fc region. For example, the Fc region can be modified so that the Fc region is not glycosylated. As described above, modification of the Fc region can be accomplished in several ways (e.g., by genetic mutation, by choice of expression system to change the Fc glycosylation pattem, and the like), hi one embodiment, the Fc effector Ugand is FcyRJU. In one embodiment, the Fc effector ligand is the Clq component of complement. As used in this context, a subject CD16A bmding protem has increased "therapeutic efficacy" compared to a reference binding protein that induces neutropenia when administered if the subject CDI6A binding protein does not induce neutropenia (or results in less severe neutropenia). For example, a CD16A bmding protein that reduces the severity of an deleterious hnmune response (e.g., ITP or experimentally induced ITT in a mammal) and reduces neutrophil levels in the animal by x% has greater "therapeutic efBcacy" than a CDl 6A binding protein that reduces the severity of an deleterious immune response and reduces neutrophil levels in the animal by y%, if y is greater than x. e.g. two-fold greater. In one embodiment, the protein is modified by mutation such that the modified proleio is aglycosylated. [0133] For example, the invention provides methods for producing a modified CD) 6 binding protein comprising a modified immunoglobulin heavy chain, the modified CD16 binding protein having greater therapeutic efficacy than a parent CD 16 binding protein comprising a parent immunoglobulin heavy chain, by (i) intioducing at least one mutation into a parent polynucleotide that encodes the parent immunoglobulin heavy chain to produce a modified polynucleotide that encodes the modified immunoglobulin heavy chain, the mutation introducing into the modified immunoglobulin heavy chain an amino acid substitution that changes, reduces or eliininates glycosylation in the CH2 domain of the parent immunoglobulin heavy chain; and (ii) expressing the modified polynucleotide in a cell as the modified immunoglobulin heavy chain so as to produce the modified CD 16 binding protein heavy chain. Optionally, the heavy chain is produced under conditions of co-expression with a light chain to produce a tetrameric antibody. 7. Examples Example 1: Mouse 3G8 VH and VL and Chimeric Molecules Generated Therefi-om A) Mouse 3G8 VH and VL [0134] The cDNA encoding the mouse 3G8 antibody light chain was cloned. The sequence of the 3G8 antibody heavy chain was provided by Dr. Jeffiy Ravetch. The amino acid sequences of the 3G8 VH and Vi. are provided in Tables 1 and 3, infra. Nucleic acid sequences encoding the variable regions are: SEOn)NO:l (3G8VH) CAGGTTACTCTGAAAGAGTCTGGCCCTGGGATATTGCAGCCCTCCCAGACCCTCA GTCTGACTTGTTCTTTCTCTGGGTTTTCACTGAGGACTTCTGGTATGGGTGTAGG CTGGATTCGTCAGCCTTCAGGGAAGGGTCTAGAGTGGCTGGCACACATTTGGTGG GATGATGACAAGCGCTATAATCCAGCCCTGAAGAGCCGACTGACAATCTCCAAGG ATACCTCCAGCAACCAGGTATTCCTCAAAATCGCCAGTGTGGACACTGCAGATAC TGCCACATACTACTGTGCTCAAATAAACCCCGCCTGGTTTGCTTACTGGGGCCAA GGGACTCTGGTCACTGTCTCTGCA SE0IDN0-.3 {3G8YL) GACACTGTGCTGACCCAATCTCCAGCTTCTTTGGCTGTGTCTCTAGGGCAGAGGG CCACCATCTCCTGCAAGGCCAGCCAAAGTGTTGATTTTGATGGTGATAGTTTTAT GAACT GGTACC AACAGAAACCAGGACAGCCACCCAAACTC CT CAT CTATACTACA TCCAATCTAGAATCTGGGATCCCAGCCAGGTTTAGTGCCAGTGGGTCTGGGACAG ACTTCACCCTCAACATCCATCCTGTGGAGGAGGAGGATACTGCAACCTATTACTG TCAGCAAAGTAATGAGGATCCGTACACGTTCGGAGGGGGGACCAAGCTGGAAATA AAA B) Chimeric Heavy Chain [0135] To create a chimeric gene coding for expression of the mouse 3GS VH fused to a human constant domain, the nucleic acid encoding the 3G8 VH was fiised to sequences encoding a signal peptide (see Orlandi et al., 1989, Proc. Natl. Acad Sci. U.S.A 86:3833-37; in lower case underline below) and a hiunan Cyl constant region (in lower case below) using standard techniques (including overlapping PCR amplification). To facilitate cloning, a Sad site was introduced, resulting in a single residue change in VH FR4 (ala -^ ser). This change in FR4 does not affect binding to CDl 6. The resulting nucleic acid had the sequence shown beiow. The regions encoding the VH domain is in iqiper case. SE0IDN0:5 {chSGSVH} gctagcqtttaaacttaagcttgttqactagtgagatcacagttctctctacagt tactgagcacacaggacctcaccatgggatggaqctgtatcatcctcttcttqqt agcaacagctacaggtaagqggctcacagtagcaqqcttgaggtctqgacatata tatgggtgacaatgacatccactttgcctttctctccacaggtgtccactccCAG GTTACCCTGAAAGAGTCTGGCCCTGGGATATTGCAGCCCTCCCAGACCCTCAGTC TGACTTGTTCTTTCTCTGGGTTTTCACTGAGGACTTCTGGTATGGGTGTAGGCTG GATTCGTCAGCCTTCAGGGAAGGGTCTAGAGTGGCTGGCACACATTTGGTGGGAT GATGACAAGCGCTATAAT CCAGC CCT GAAGAGCC GACT GACAAT CT CCAAGGATA CCTCCAGCAACCAGGTATTCCTCAAAATCGCCAGTGTGGACACTGCAGATACTGC GACATACTACTGTGCTCAAATAAACCCCGCCTGGTTTGCTTACTGGGGCCAAGGG ACTCTGGTCACTGTGAGCTCAgcctccaccaagggcccatcggtcttccccctgg caccctcctccaagagcacctctgggggcacagcggccctgggctgcctggtcaa ggactacttccccgaaccggtgacggtgtcgtggaactcaggcgccctgaccagc ggcgtgcacaccttcccggctgtcctacagtcctcaggactctactccctcagca gcgtggtgaccgtgccctccagcagcttgggcacccagacctacatctgcaacgt gaatcacaagcccagcaacaccaaggtggacaagagagttgagcccaaatcttgt gacaaaactcacacatgcccaccgtgcccagcacctgaactcctggggggaccgt cagtcttcctcttccccccaaaacccaaggacaccctcatgatctcccggacccc tgaggtcacatgcgtggtggtggacgtgagccacgaagaccctgaggtcaagttc aactggtacgtggacggcgtggaggtgcataatgccaagacaaagccgcgggagg agcagtacaacagcacgtaccgtgtggtcagcgtcctcaccgtcctgcaccagga ctggctgaatggcaaggagtacaagtgcaaggtctccaacaaagccctcccagcc cccatcgagaaaaccatctccaaagccaaagggcagccccgagaaccacaggtgt acaccctgcccccatcccgggatgagctgaccaagaaccaggtcagcctgacctg cctggtcaaaggcttctatcccagcgacatcgccgtggagtgggagagcaatggg cagccggagaacaactacaagaccacgcctcccgtgctggactccgacggctcct tcttcctctacagcaagctcaccgtggacaagagcaggtggcagcaggggaacgt cttctcatgctccgtgatgcatgaggctctgcacaaccactacacgcagaagagc ctctccctgtctccgggtaaatgagtgcggccgcgaattc [0136] This construct was inserted into tlie pCI-Neo (Promega Biotech) at the Nhel-EcoRl sites of the poIyUnker for use for expression of the chimeric heavy chain in cells. C) Chimeric Light Chain [0137] To create a chimeric gene coding for the mouse 3G8 VL fused to a human constant domain, this 3G8 VL segment was fused to a signal sequence (as for the VH above; (lower case underlined) and a human CK constant region (lower case) cDNA using standard techniques, resulting in a nucleic acid with the sequence shown below: SE0IDNO:6 ch3G8VL gctagctqaqatcacagttctctctacagttactgagcacacagqacctcaccat cjggatggaqctqtatcatcctcttcttggtagcaacagctacaggtaaqqqqctc acagtagcaggcttgaggtctggacatatatatgqqtqacaatqacatccacttt gcctttctctccacaggtgtccactccGACACTGTGCTGACCCAATCTCCAGCTT CTTTGGCTGTGTCTCTAGGGCAGAGGGCCACCATCTCCTGCAAGGCCAGCCAAAG TGTTGATTTTGATGGTGATAGTTTTATGAACTGGTACCAACAGAAACCAGGACAG CCACCCAAACTCCTCATCTATACTACATCCAATnTAGAATCTGGGATCCCAGCCA GGTTTAGTGCCAGTGGGTCTGGGACAGACTTCACCCTCAACATCCATCCTGTGGA GGAGGAGGATACTGCAACCTATTACTGTCAGCAAAGTAATGAGGATCCGTACACG TTCGGAGGGGGGACCAAGCTTGAGATCAAAcgaactgtggctgcaccatcggtct tcatcttcccgccatctgatgagcagttgaaatctggaactgcctctgttgtgtg cctgctgaataacttctatcccagagaggccaaagtacagtggaaggtggataac gccctccaatcgggtaactcccaggagagtgtcacagagcaggacagcaaggaca gcacctacagcctcagcagcaccctgacgctgagcaaagcagactacgagaaaca caaagtctacgcctgcgaagtcacccatcagggcctgagctcgcccgtcacaaag agcttcaacaggggagagtgttagttctagagtcgactctagaggatccccgggt accgagctcgaattc [0138] This construct was inserted into pCI-Neo (Promega Biotech) at the Nhel-EcoRl sites of the polylinker for use for expression of the chimeric light chain in cells. D) Expression [0139] The ch3G8VH and ch3G8VL chimeric proteins described above can be co-expressed to form a chimeric antibody, referred to as ch3G8. The chimeric antibody cli3G8 can be expressed either in a myeloma or in other mammalian cells (e.g., CHO, HEK-293). An example of a procedure for expression of CD16A binding proteins such as ch3G8 and variants is provided in Example 4, infra. Example 2: Humanized anti-CD 16A binding proteins A) Hmnanized Heavy Chain [0140] CDR encoding sequences &om the mouse 3G8 VH clone were fused to firamework sequences derived from the human germline VH sequence VH2-70 to create a polynucleotide encoding a VH designated Hu3G8VH. The polynucleotide was generated by an overlapping PCR procedure. In a first step, using the primers and strategy shown below and tbe mouse 3G8 VH polynucleotide (SEQ ID NO: 1) as template. [0143] For expression in mammalian cells (HEK-293), the Hu3G8VH-l sequence was cloned into the pCI-Neo polylinker at the Nhel-EcoRl sites, following intervenmg cloning into pUC and pCDNA3.1. B) Humanized Light Chain [0144] CDR encoding sequences from the mouse 3G8 VL clone were fiised to framework sequences derived from the human B3 germline V-K gene. The polynucleotide was generated by an overlapping PCR procedure using the primers and strategy shown below and the mouse 3G8 VL polynucleotide (SEQ ID NO: 2) as template. H024 -*H023 ce primer Length .e^i^. H026 NO: w%- -m^i 71 71 30 H023 H024 H025 H026 H027 H009 63 aCTCTTTGGCTGTGTCTCTAGGGGAGAGGGCCACCATCAACTGCiia 3GCCAGCCAAAGTGTTG 65 "TCTCCACAGGTGTCCACTCCGACATCGTGATGACCCAATCTCCAG RCTCTTTGGCTGTGTCTCTA JGTGAGGGTGAAGTCTGRCCCAGACCCACTGCCACTAAACCTGTCT ;GGACCCCAGATTCTAGATTGGATG 67 rGACAGTAATAAACTGCCACATCCTCAGCCTGCAGGCTGCTGATGG TGAGGGTGAAGTCTGTCCCAG gcggcAAGCTTGGTCCCCTGTCCGAACGTGTACGGATCCTCATTAC TTTGCTGACAGTAATAAACTGCCAC ::GAGCTAGCTGAGAT CACAGTTCTCTCTAC 19 20 21 22 23 [0145] The resulting polynucleotide had the sequence SEOIDNQ:25 {hu3G8VL} GACACTGTGCTGACCCAATCTCCAGCTTCTTTGGCTGTGTCTCTAGGGCAGAGGG CCACCATCTCCTGCAAGGCCAGCCAAAGTGTTGATTTTGATGGTGATAGTTTTAT GAACTGGTACCAACAGAAACCAGGACAGCCACCCAAACTCCTCATCTATACTACA TCCAATCTAGAATCTGGGATCCCAGCCAGGTTTAGTGCCAGTGGGTCTGGGACAG ACTTCACCCTCAACATCCATCCTGTGGAGGAGGAGGATACTGCAACCTATTACTG TCAGCAAAGTAATGAGGATCCGTACACGTTCGGAGGGGGGACCAAGCTTGAGATC AAA [0146] The Hu3G8 VL gene segment was combined with a signal sequence (as described above, lower case, underline) and a human C-K constant region (lower case) cDNA xising standard techniques resulting in a product with the sequence below: 5EQIDNO:26 {hu3G8VL-l} gctagctgagatcacagttctctctacagttactgagcacacaggacctcaccat gggatggagctgtatcatcctcttcttggtagcaacagctacaggtaagggqctc acagtagcaggcttgaggtctg^acatatatatgggtgacaatgacatccacttt gcctttctctccacaggtgtccactccGACACTGTGCTGACCCAATCTCCAGCTT CTTTGGCTGTGTCTCTAGGGCAGAGGGCCACCATCTCCTGCAAGGCCAGCCAAAG T GTT GAT TTT GATGGTGATAGT TT T ATGAACT GGTACCAACAGAAACCAGGAC AG CCACCCAAACTCCTCATCTATACTACATCCAATCTAGAATCTGGGATCCCAGCCA GGTTTAGTGCCAGTGGGTCTGGGACAGACTTCACCCTCAACATCCATCCTGTGGA GGAGGAGGATACTGCAACCTATTACTGTCAGCAAAGTAATGAGGATCCGTACACG TTCGGAGGGGGGACCAAGCTTGAGATCAAAcgaactgtggctgcaccatcggtct tcatcttcccgccatctgatgagcagttgaaatctggaactgcctctgttgtgtg cctgctgaataacttctatcccagagaggccaaagtacagtggaaggtggataac gccctccaatcgggtaactcccaggagagtgtcacagagcaggacagcaaggaca gcacctacagcctcagcagcaccctgacgctgagcaaagcagactacgagaaaca caaagtctacgcctgcgaagtcacccatcagggcctgagctcgcccgtcacaaag agcttcaacaggggagagtgttagttctagagtcgactctagaggatccccgggt accgagctcgaattc [01471 This construct was inserted into pCI-Neo for expression in mammalian cells. Example 3: Variant CD16A binding proteins [0148] Additional expression constructs were made in which sequence changes were introduced in the VL or VH domains by site directed mutagenesis. A typical mutagenesis reaction contained 10 ng plasmid DNA (isolated from a methylation competent strain of £. coif), 125 ngeachof aforwardandreverse primer, each containing the mutation of interest, reaction buffer, and dNTPs in 0.05 ml volume. 2.5 units of PfuTurbo DNA polymerase (Stratagene) was added and the reaction was subjected to 15 cycles of 95°, 30 sec; 55°, 1 min; 68°, 12 min. The product of the PCR was then digested witii Dpnl endonuclease and the restricted DNA was used to transform £. co/i strain XL-10 gold. Because £>pnl only digests methylated DNA it will digest the parental, non-mutated, plasmid leaving the newly synthesized non-methylated product, containing the mutation of interest, as the predominant species. 10149] The sequences of the variant VH domains are shown in Table 3. The sequences of the variant VL domains are shown in Table 4. Example 4: Expression in Mammalian cells [0150] Various combinations of heavy and light chain expression plasmids (e.g., comprising the chimeric, humanized and variant VL and VH domains fused to human Cyl and CK constant domains as described above) were co-transfected into HEK-293 cells for transient expression of recombinant tetrameric antibodies (i.e., comprising 2 heavy chains and 2 light chains), sometimes referred to herein as "recombinant antibodies." Transfection was carried out using Lipofectamine-2000 (Invitrogen) in 6 well plates according to the manirfacturer"s instmctions. [0151] Recombinant antibodies were prepared by cotransfection of a heavy chain expression plasmid (i.e., encoding a heavy chain comprising a VH and constant domains) and hght chain expression plasmids (i.e., encoding a light chain comprising a VL and constant domains) together into HEK-293 cells for transient ej^ression of recombinant antibodies. [0152] Hu3G8VH variants listed in Table 3 were coexpressed with the hu3G8VL-l hght chain. For reference, most assays included (i) recombinant antibodies produced by coexpression of chBGSVH and ch3G8VL ("ch3G8VH/ch3G8VL") and (ii) recombinant antibodies produced by coexpression of hu3G8VH-l andhu3G8VL-l ("hu3G8VH-l/hu3G8VL-I"). [0153] Hu3G8VL variants listed in Table 4 were coexpressed with Hie ch3G8VH heavy chain. For reference, most assays included (i) recombinant antibodies produced by coexpression of ch3G8VH and ch3G8VL ("ch3G8VH/ch3G8VL") and (ii) recombinant antibodies produced by coexpression of ch3G8VH and hu3G8VL-l ("ch3G8VH/hu3G8VL-l"). [0154] After three days, the levels of recombinant antibodies in the conditioned media were determined by ELISA, and the recombinant antibodies were analyzed by ELISA for binding to captured sCD16A as described in Examples 5. Selected antibodies were assayed for cell binding to cells e?q)ressing the extracellular domain of CD16A, as shown in Example 6. Example 5: ELISA Determination of Binding to CD16A [0155] Sandwich ELISA was performed to detect binding of antibodies to a soluble form of CD16A. Soluble human CD16A [0156] A soluble form of human CD16A was expressed from HEK-293 cells using a pcDNA3.1-derived expression vector containing the CD16A gene truncated just prior to the transmembrane region. To create the vector, cDNA encoding CD16A was amplified using the primers 3Aieft [gttggatcctccaactgctctgctacttctagttt] (SEQ ID NO:27) and 3Aright [gaaaagcttaaagaatgatgagatggttgacact] (SEQ ID NO:28) digested With BamHI and Hindlll, and cloned into the vector pcDNAS.l (Novagen) at the Bam/Hindin site of the polylinker. The construct was used to transiently transfect HEK-293 cells. For some assays, the secreted product was purified from conditioned medium using affinity chromatography on a human IgG Sepharose column. In some assays, the amoimt of sCD16A in conditioned medium was quantitated and unpurified sCD 16A was used. Purification was not required since the ELISA c^ture antibody (LNK16 mAb) specifically bound the antigen, allovnng removal of contaminants in washing steps. [0157] The amino acid sequence of the sCD16 construct is shown below. (Tie signal sequence, underlined, is cleaved off during expression. Note the last seven residues are derived from the vector pCDNA3.1 rather than from the CD 16A gene): MWQLLLPTALLLLV5AGMRTEDLPKAWFLEPQWYRVLEKDSVTLKCQGAYSPED NSTQWFHNESLISSQASSYFIDAATVDDSGEYRCQTNLSTLSDPVQLEVHIGWLL LQAPRWVFKEEDPIHLRCHSWKNTALHKVTYLQNGKGRKYFHHNSDFYIPKAl"LK DSGSYFCRGLFGSKNVSSETVNITITQGLAVSTISSFFECLAAARV (SEQ ID NO:29) ELISAformat [0158] Plates were first coated with 100 ng/well of the anti-CD16 mAb LNK-16 (Advanced Immunochemical, Long Beach CA; see 5th Hiunan Lymphocyte Differentiation Antigens Workshop) in carbonate buffer at room temperature for 2 hrs. Any anti-sCD16A antibody that does not block binding by 3G8 can be used. After blocking for 30 minutes with PBS-T-BSA, sCD16A conditioned medium was added at a dilution of 1/10 and incubated at room temperature for 16 hrs. Alternatively, when purified sCD16 was used, it was diluted to a concentration of 50 ng/ml in PBS-T-BSA. 0.05 ml was added to each well and incubated for at least 2 hrs at room temperature. [0159] The plate was washed and dilutions of recombmant antibodies starting at 0.5 |J,g/ml in PBS-T-BSA were then added and incubated for 1 hr at room temp. Binding of recombinant antibodies to the captured sCDI6A was then measured using an anti-human IgG-HRP conjugate and TMB substrate. After stopping color development using dilute sulfuric acid, the plate was read at 450 nM. Results of Binding Assays [0160] This example shows that the binding properties of humanized anti-CD16A antibodies for binding to CD 16A are the same or similar to the properties of the chimeric 3G8 antibody. [0161] Based on the comparative binding studies, the recombinant antibodies were classified as binding with high, intermediate, or low affinity. Antibodies viith high and intermediate binding affinity are discussed above in section 4. The recombinant antibodies with a VH domain of hu3G8VH- 9,10,11, 13,15, 21,38, 39,or41 showed little orno binding to SCD16A. From these data it appears certain substitutions (or combinations of substitutions) are generally detrimental to binding. For example, substitution of tyrosine or aspartic acid at VH position 52 (i.e., 52Y and 52D) or threonine at position 94 (94T) are detrimental to binding. Similarly, the combination leucine at position 50 with aspartic acid at position 54 (50L+54N) is detrimental to binding, as is the combination arginine at 94 and aspartic acid at 101 (94R+101D). However, aspartic acid at 101 is tolerated when position 94 is glutamine, lysine, histidine or alanine (but not arginine). Further 34V+94R4-101D has intermediate activity. This indicates a relationship between positions 34, 94 and 101 in maintaining higji affinity binding, and suggests that 34V may be an especially important residue. Likewise, recombinantantibodies witha VLdomainofhu3G8VL-6,7, 8,9,11,12,13,and 14 showed little or no binding to sCD 16A. From these data it appears certain substitutions (or combinations of substitutions) are generally detrimental to binding. For example, substitution of alanine at position 34 (34A) or tyrosine at position 92 (92Y) is generally detrimental to binding. [0162] Results of an exemplary binding assay are shown in Figure 1. Example 6: Antibody Binding to Cells Expressing CDI6A [0163] The ability of selected humanized antibodies to bind to CD16A expressed by CHO-Kl cells as assayed by direct binding competition assays. [0164] CHO-Kl cells expressing extracellular domain of FcRIIIa fiised to the transmembrane and intracellular domain of FcRIIb were used for cell binding assays. Cells were plated at 40,000 cells per well in 96 well flat bottom tissue culture plates (FALCON MICROTEST Tissue Culnire plate, 96 weU) and incubated at 37"^C CO2 incubator for approximately 24hr. The plate was then gently washed three times with 25 mM Hepes, 75 uM EDTA, 11.5 mM KCI, 115 mM NaCl, 6 mM MgS04,1.8 mM CaC12,0.25% BSA (binding buffer). [0165] For indirect binding assays, lOOjil of a serial dilution of anti-CD 16 Mab (fmal concentration: lug/ml, 0.5 , 0.25,0.125,0.0625, 0.03,0.015, 0 ug/ml) was then added to wells in binding buffer. The plate was then incubated at 23""C for 1 hr and washed three times with binding buffer. SOjil/weU of Europium (EU)-Iabeled -anti-human-IgG (lOOng/ml) was then added to each well and the piate was incubated at 23""C for 30 minutes then washed three times with bindii^ buffer. Finally, 100|J1 Delfia enhancement solution (PerkmElmer/Wallac) was added. After incubating with shaking for 15 minutes, the plate was read for time resolved fluorescence (excitation 340mn; emission 615nm) in a Victor2 instrument (PerkinElmer/Wallac). The results of the assay are shown in Figure 2. [0166] The CHO-Kl cells described above were also used in competition assays. After washing with binding buffer as described above, varying amounts of purified unlabeled Mab (1.2 - 75 nM final concentration) were mixed with a fixed concentration of Eu-Ch3G8-N297Q (final concentration 2.5 nM). The plate was then incubated at 23*"C for 1 hr and washed three times with binding buffer. lOOpl Delfia enhancement solution (PerkinElmer/Wallac) was the added and after incubating with shaking for 15 minutes, the plate was read for time resolved fluorescence (excitation 340nm; emission 615nm) in a Victor2 instrument (PerkinElmer/Wallac). The results of the assay are shown in Figure 3. [0167] These assays demonstrate that the humanized anti CD16A monoclonal antibodies bind with high affinity to CD16A on the surface of transfected cells. Hu3G8-22.1-N297Q binds to CDMA bearing cells with higher affinity than Ch3G8-N297Q. Example 7: Inhibition ofbinding of sCD16A to Immune Complexes Assay of 4-4-20 binding to FITC-BSA [0168] The binding of ch4^-20 or ch4-4-20 (D265A) to FITC-BSA was assessed by ELISA. (Ch4-4-20 is identical to Ch3G8 except that it contains the respective VH and VL regions of 4-4-20 instead of those of 3G8. Thus it retains high affinity and specificity for the hapten fluorescein. 4-4-20 is described in Bedzylcetal., \9B9, J Biol Chem 264:1565-9.) FITC-BSAO ug/ml-50 ng/well) was coated onto Nunc maxisorb immunoplates in carbonate buffer and allowed to bind for approximately 16 hr. Following blocking with BSA, dilutions of ch4-4-20 were added to the wells and allowed to bind for 1 hr at RT. After washing out imbound Mab, HRP-conjugated goat anti-human Ig secondary was added. One hour later the secondary antibody was removed, washed and developed with TMB substrate. Following addition of an acidic stop solution the plate was read at 450nm. Botii ch4-4-20 and ch4-4-20(D265A) bound to the FITC-BSA with high affinity (data not shown). Assay qfsFcR binding to ch4-4-20/FITC-BSA immune complexes [0169] The same format was used to assay binding of sFcRs to immune complexes (IC) formed on the ELISA plate between ch4-4-20 and FITC-BSA. In this case we have used either biotinylated sFcR or biotinylated anti-human G2 Mab as a secondary reagent, followed by streptavidin-HRP detection. Inhibition ofsFcR binding to IC by murine, chimeric and humanized 3G8 [0170] The concentrations of ch4-4-20 and sFcR were fixed to give approximately 90 percent maximal signal in the assay. sCD 16A was premixed with serial dilutions of murine, chimeric or humanized 3G8 and incubated for one hour prior to adding to the plate containing the immune complex. Serial dilutions of humanized or chimeric 3G8 were incubated with sCD16A-G2-biotin for one hour. The mixtures were then added to ELISA wells containing an immune complex between a human IgGl chimeric form of the anti-fluorescein Mab 4-4-20 and FITC-BSA. After one hoxjr, binding of the soluble receptor to the IC was detected using streptavidin-HRP conjugate and TMB development. The results are shown in Figure 4. This assay indicates that humanized anti-CDl 6A antibodies are potent inhibitors of CD16A binding to IgG in immune complexes. Example 8: Analysis of anti-CD 16A Monoclonal Antibody Panel {0171] A panel of hybridomas was generated following immunizing and boosting mice with sCD16A using standard methods. Eight 96-well plates were screened by ELISA for binding activity on plates coated directly with sCD16A. Ninety-three of these gave a positive signal and were expanded fiirther. Of these, 37 were positive for binding to human blood cells by FACS. These supematants were then analyzed for their ability to block tiie interaction of CD16A with immune complexes and for the similarify of the binding site (epitope) to that of 308. Assays included capture ELISA using chimeric 3G8 down and inhibition of immune complex binding to sRIIIa-Ig. Based on these assays antibodies with biading and inhibitory properties similar to 3G8 were isolated, as well as Mabs with binding and/or inhibitory properties distinct fi-om 3G8. [0172J DJ130C (DAKO) and 3G8 were used as controls in the assays. Mab DJ130c is a commercially available Mab which bmds CD16 at an epitope distinct fixDm 3G8 (Tamm and Schmidt). This Mab does not block FcRIIIa-immune complex binding (Tamm and Schmidt). In an ELISA-based inhibition assay, DJ130c enhances rather than inhibits binding. [0173] The data indicate that the panel contains antibodies which bind to the same epitope as Ch3G8 and block sCD16A binding to immime complexes. The panel of Mabs also contains antibodies which do not bind to the same epitope as Ch3G8. Most of these latter antibodies do not block the interaction of sCD16a witii IgG in immime complexes. Table Assay Effect on sCD 16a Binding to Immime Complexes Binding to sCD16 Captured by Cb3G8 Result Inhibition Enhancement No Effect Positive 2 5(+DJ-I30c) 17 Negative n(+3G8) 0 2 Example 9: Induction of Platelet Depletion In Vivo [0174] The in vivo activity of a CDMA binding protein for blocldng hum^ Fc-FcyRIII interactions induced by autoantibodies can be evaluated using animal models of autoimmune diseases. One suitable model is the "passive mouse model" of ITP and the anti-platelet mAb 6A6 (see, Oyaizu et al., 1988, J Exp.Med. 167:2017-22; Mizutanietal, 1993,5/00^82:837-44). 6A6isanIgG2a isotypemAb derived from a NZWxBSXBFl individual. Administration of 6A6 depletes platelets in muFcyRIH -/-, huFcyRUIA transgenic mice but not in muFcYRIII -/-mice without the human transgene. See Samuelssonetal.,2001, Science 291:484-86. Other anti-platelet monoclonal antibodies can be used m place of 6A6 in the model. Alternatively, a polyclonal anti-platelet antibody can be used. [0175] CD16A binding proteins that confer the greatest degree of protection from platelet depletion can be identified by administrating CD 16A bindbag protems to a muFcyRIII -/-, hiiFcYRIIIA transgenic mouse and measuring any reduction in mAb 6A6 induced platelet depletion. [0176] A related assay can be carried out using a chimeric human IgGiK chimeric derivative of 6A6 in place of the mouse mAb in the protocol provided above, so that the depleting mAb had a human isotype. To conduct this assay, a chimeric 6A6 monocJona] antibody (ch6A6) was prepared by fusing the cDNA segments encoding the murine anti-plateiet monoclonal antibody 6A6 VH and VL regions to the human Cyl and CK CDNA segments, respectively. The resulting genes were co-expressed in 293 cells and chimeric 6A6 was purified by protein A affinity chromatography followed by size exclusion chromatography. [0177] To demonstrate that the chimeric 6A6 antibody induces platelet depletion, to and ch6A6 was administered to muFcyRIir"", huFcyRIIIA transgenic mice. The ch6A6 was administered to each animal either i.v. or intraperitoneally (i.p.) (O.lfig/g). Animals were bled 2 hrs, 5 hrs, 24 hrs and 48 hrs after administration of ch6A6, and plasma platelet counts were determined using a Coulter Z2 particle counter and size analyzer equipped with a 70 jmi aperture. Particles between 1.5 and 4 |im in size (correspondmg to platelets) were counted and the data were analyzed by plotting the platelet count versus time for each concentration. [0178] Two hours after injection of 0.1|ag/g ch6A6 i.p., approximately 75% of the platelets were depleted. The number of platelets remained low for 5 hours after ch6A6 injection tlien progressively increased to return to normal 72 hours after ch6A6 injection. [0179] Two hours after mjection of 0.1 ^g/g ch6A6 i.v., approxhnately 60% of the platelets were depleted. The number of platelets remained low for 6 hours after ch6A6 injection then progressively increased to retum to normal 48 hours after ch6A6 injection. Example 10. Analysis of the Ability of CD 16 Binding Antibodies to Protect Mice from Platelet Depletion [0180] The ability of CD16A binding proteins to reduce platelet depletion in experimental ITP can be assayed as described below. CD16A binding proteins were administered intravenously (i.v.) to groups of muFcyRIir"", huFcyRIIIA transgenic mice at concentrations of 0.5,1,2 or 5 |ig/g in phosphate buffered saline (PBS). Controls were PBS alone or an irrelevant human IgGl (negative control) or human intravenous immunoglobxUin (TVIG; positive control). One hour after administration of the CD16A binding protein or control, ITP was induced by administering 0.1 ij.g/g ch6A6 to each animal either intravenously or intraperitoneally. Animals were bled 2 hrs, 5 hrs, 24 hrs and 48 hrs after administration of ch6A6. Plasma platelet counts were determined using the Coulter Z2 particle counter and size analyzer as described above and the data were analyzed by plotting the platelet count versus time for each concentration of administered binding protein. [0181] When muFcyRIIT"", huFcyRIHA transgenic mice were injected with murine 3G8 (D.5^g/g) one hour before i.p. injection of ch6A6,33% of the platelets were depleted at the 2 hours time point (Figure 5). The number of platelets then progressively increased to retum to normal 24 hours after ch6A6 injection. When muFcyRIir"", huFcyRIHA transgenic mice were injected with murine 3G8 (0.5jj,g/g) one hour before i. v. injection of ch6A6, 30% of the platelets were depleted at the 2 hours time point (Figure 6). The number of platelets then rapidly increased to retum to nonnal 5 hours after ch6A6 injection. [0182] These results were similar to the protection seen when human IVIG is administered. When muFcyRni""", huFcyRIIIA transgenic mice were injected with human IVIG (Img/g) one hour before i.p. mjection of ch6A6,33% of die platelets were depleted at the 2 hours lime point (Figure 5). The number of platelets then progressively increased to return to normal 24 hours after ch6A6 injection. When muFcyRIII ", huFcyRIIIA transgenic mice were injected with human IVIG (Img/g) one hour before i.v. injection of ch6A6,20% of the platelets were depleted at the 2 hours time point (Figure 6). The number of platelets then rapidly increased to return to normal 5 hours after ch6A6 injection. [0183] The results shown in Figures 5 and 6 show that m3G8 protects mice from ch6A6-mediated platelet depletion, and that the level of protection was similar to the protection conferred by IVIG. [0184] Preparations of recombinant mouse 3G8 produced m HEK-293 cells, chimeric 3G8 with human IgGl or IgG2 constant domains (ch3G8-yI produced in HEK-293 and CHO-Kl cells, and ch3G8-y2 produced m HEK-293 ceUs), and a ch3G8-yl variant (ch3G8-yl D265A) did not provide significant protection in this experiment. When muFc7RIII""", huFcyRIIIA transgenic mice were injected with ch3G8yl ory2 (0.5^g/g) one hour before i.p. injection of 6A6, approximately 60% of the platelets were depleted at the 5 hour time point (Figure 7). The number of platelets then progressively returned to normal. Although depletion was not as severe as in mice that received no anti-CD 16A binding protein, these chimeric antibodies provided significantly less protection, if any, than murine 3G8. A ch3G8 variant in which aspartic acid 265 was changed to alanine showed similar results. Interestmgly, as is shown in Example 11, modification of the ch3G8 to produce an aglycosylated variant increased ihs protective effect of the antibody. Example 11: Ch3G8 N2970 Protects Mice fi-om ch6A6-Mediated Platelet pepletioiL [0185] An aglycosylated version of chSGS-yl was prepared by mutating the expression polynucleotide encoding ch3G8-yl so that residue 297 was changed fi-om asparagine (N) to glutamine acid (Q), and ejqjressing the encoded antibody. Residue 297 lies in an N-iinked glycosylation site, and this mutation prevents glycosylation of the Fc domain at this site. This aglycosylated antibody, ch3G8 N297Q, was produced in HEK-293 cells as described for ch3G8-yl (see Example 4, supra). The ability of ch3G8-N297Q to protect against ch6A6-mediated platelet depletion was tested using the protocol described above. 10186] When muFcyRIIf"", huFcyRIIIA transgenic mice were injected with l|j.g/g of the aglycosyl form of ch3G8 (ch3G8 N297Q) one hour before i.p. injection of ch6A6, approximately 75% of the platelets were depleted at the 2-hotir time point (Figure 8). Platelet levels increased faster than in the absence of ch3G8 N297Q, and returned to nomial by 24 hours after ch6A6 iojection. [0187] When muFcyRIir"", huFcyRIEA transgenic mice were injected with lia,g/gch3G8N297Qone hour before i.v. injectionofch6A6, approximately 60% of the platelets were depleted at the 2 hours time point (Figure 9). Platelet levels increased faster than in the absence of ch3G8 N297Q, and returned to normal by 48 hours after ch6A6 injection. [0188] When muFcyRIII""", huFcyRHIA transgenic mice were injected with ch3G8 N297Q (2^g/g) one hour before i.v. mjection of ch6A6, only 40% of the platelets were depleted at the 2 hours time point (Figure 9). Platelet levels increased faster than in the absence of ch3G8 N297Q, and returned to normal by 5 hours after ch6A6 injection. [0189] Thus, ch3G8-N297Q was consistently able to significantly improve platelet cotmts. Binding of 3G8 to human CD16 on effector cells blocks the ability of GDI 6 to interact with immune complexes and trigger effector functions such as ADCC or phagocytosis. Chimeric and mouse 3G8 molecules have similar ability to bind CD16 and are similar in their ability to inMbit the binding of sCD16 to immtme complexes in vitro. Without intending to be bound by a particular mechanism, the binding (and thus) the blocking activity of the mAb is thought to be confined to the Fab portion of the antibody and blocking of huCD16 is believed to be the mechanism of protection in the transgenic mouse ITP model. The data above suggest that the glycosylation state of the Fc domain can affect the in vivo protective capacity of anti-CD16A antibodies. Ablation of Fc domain glycosylation (e.g., with D265A or N297Q mutations, or by using a human gamma2 Fc domain) reduces or eliminates Fc binding to FcR. hi the case of the aglycosyl (N297Q) variant, complement fixation is also abolished. Example 12: Neutrophil Levels following Administration of Aglvcosvl CD16A Binding Proteins [01901 The effect of an aglycosylated CD16A binding protein on neutrophil levels was tested and compared to that of glycosylated CDI6A binding proteins. CD16A binding proteins, or the controls such as irrelevant human IgGl (negative control) or murine RB6-8C5 (positive control), were administered to groups of muFcyRIIr", huFcyRIIIB transgenic mice at a concentration of 5 |j.g/g in phosphate buffered saline (PBS). Another negative control was administered PBS alone. Twenty four hours later, mice were euthanized and blood, spleen and bone marrow are collected. Neutrophils were analyzed by FACS. Staining experiments were performed in RPMI containing 3%FCS. Murine cells were stained using FITC-conjugated 3G8 (PharMingen) and R-PE-conjugated RB6-8C5 (PharMingen). Samples were analyzed by flow-cytometry using a FACSCalibur (Becton Dickinson). [0191] Intraperitoneal injection of 5 \ig/g ch3G8 (prepared as described above) resulted in murine neutrophil depletion in the blood and spleen (Figure 10; upper right quadrant). Similar results were seen following administration of murine 3G8 (results not shown). In the bone marrow of ch3G8 treated animals, neutrophils stained weakly for CD16, which could indicate receptor occupancy by the chimeric antibody or shedding (Figure 10; see shift from the upper right quadrant to the upper left quadrant). In contrast, intraperitoneal injection of 5 (ig/g cli3G8 N297Q did not result in murine neutrophil depletion in the blood, spleen or bone marrow (Figure 10). In additional experiments, humanized glycosylated 3G8 antibodies showed substantially more depletion of circulating blood neutrophils compared to aglycosylated forms of the same antibodies. Example 13: Autoimmune Hemolytic Anemia Model [0192] This example demonstrates that administration of CD16A binding protein prevents red blood cell depletion in a model of autoimmune hemolytic anemia, [0193] The abihty of the Hu3G8-5.1-N297Q monoclonal antibody to prevent antibody-dependent red blood cell depletion in muFcRIU-/-, huFcRnia+ mice was evaluated. Hu3G8-5.1-N297Q is an aglycosy antibody with the heavy cham Hu3GSVH-5 and the light chain Hu3GSVH-l and the indicated substitution of asparagine 297. Mice were bled on day 0 and RBC levels were determined using a Coulter Z2 particle analyzer. The next day groups of 3 animals each were then injected intravenously with either 0.5 mg/kg Hu3G8-5.1-N297Q or PBS. One group of mice did not receive any compound. One hour later, RBC depletion was induced in the first two groups by administering mouse anti-RBC IgG2a Mab 34-3C to each animal intraperitoneally (i.p.) (2.5 mg/kg). Ajiimals were bled 2 hrs, 5 hrs, 24 hre and 48 hrs after administration of 34-3C and RBC counts were determined. Data was analyzed by plotting RBC count vereus. The data, depicted in Figure 11, demonstrate the abUity of Hu3G8-5.1-N297Q to prevent RBC depletion in this model. Example 14: Inhibition of Antibody-Dependent Cellular Cytotoxicity fADCC) [0194] This example demonstrates that humanized 3G8 variants inhibit ADCC IK vitro and with an activity similar to that of mouse 3G8. [0195] Methods: The protocol for assessment of antibody dependent cellular cytotoxicity (ADCC) is similar to that previously described in (Ding et al., 1998, Immunity 8:403-11). Briefly, target cells from the HER2-overexpressing breast cancer cell line SK-BR-3 were labeled with the europium chelate bis(acetoxymethyl) 2,2":6",2"-terpyridme-6,6"-dicarboxylate (DELFIA BATDA Reagent, Perkin Ehner/Wallac). The labeled target cells were then opsonized (coated) with either chimeric anti-HER2 (ch4D5, lOOng/ml) or chimeric anti-fluorescein (ch4-4-20, lug/ml) antibodies. In the case of the anti-fluorescein antibody, SK-BR-3 cells were coated wilii the fluorescem hapten prior to antibody opsonization. Peripheral blood mononuclear cells (PBMC), isolated by Ficoll-Paque (Amersham Pharmacia) gradient centrifugation, were used as effector cells (Effector:Target ratio: ch4D5 = (37.5:1) and ch4-4-20 = (75:1)). Following a 3.5 hour incubation at 37**Ci 5%C02, cell supematants were harvested and added to an acidic europium solution (DELFIA Europium Sohition, Perkin ElmerAVallac). The fluorescence of the Europium-TDA chelates formed was quantitated m a time-resolved fluorometer (Victor2 1420, Perkin ElmerAVallac). Maximal release (MR) and spontaneous release (SR) were determined by incubation of target cells with 2% TX-l 00 and media alone, respectively. Antibody independent cellular cytotoxicity (AICC) was measured by incubation of target and effector cells in the absence of antibody. Each assay is performed in triplicate. The mean percentage specific lysis was calculated as: (ADCC - AICC)/(MR-SR) x 100. [0196] Results: Addition of anti-CD 16 variants inhibited ADCC mediated through antibodies directed against the HER2/neu protein (ch4D5) (Figure 12), or the hapten, fluorescein (ch4-4-20) (Figure 13). Inhibition of the ch4D5 mediated ADCC was greater than 50% at 300ng/ml for all 3G8 variants tested while isotype control antibodies had no effect in the assay. In the case of the anti-fluorescein antibody, inhibition was approximately 50% at concentrations above lug/ml for murine 3G8 (Figure 13A) and humanized 308 variants (Figure 13B), while isotype control antibodies and chimeric 3G8 had little effect. Example 15: Administration of Hu3G8-5.1-N2970 Prevents Immune Thrombocytopenia aTPI in huFcRIIa+. huFcRIIIa+ mice [0197] This example shows that that administration of anti-CD16A antibodies protects against FTP mediated by CD32A. As in FcyRin-/-, hCD16A mice, administration of the ch6A6 antibody induces ITP in FcyRHI-/-, hCD32A transgenic mice. Five hours after injection of O.ljig/g ch6A6 i.p., approjcimately 80% of the platelets are depleted (not shown). The number of platelets remained low for 24 hours after ch6A6 injection, and then progressively increased lo return to normal 48 hours after ch6A6 injection. As expected, the i.v. injection of hu3G8-5.3 (0.5^g) one hour prior to ch6A6 injection did not protect FcyRIH-/-, hCD32A mice agamst ITP (not shown). [0198] As in smgle transgenic mice, ch6A6 mduces ITP in FcyRIH-/-, hCD16A, hCD32A double transgenic mice. Five hours after injection of 0.1p.g/g ch6A6 i.p., approximately 80% of the platelets were depleted (Figure 14). The number of platelets remained low for 24 hours after ch6A6 injection, and then progressively increased to return to norma! 48 hours after ch6A6 injection. [0199] In contrast to FcyRHI-/-, hCD32A mice, FcyRIH-/-, hCDl 6A, hCD32A mice were protected against ITP by administration of hu3G8-5.1. Complete protection was observed when l|.^g h3G8 5.1 is injected one hour prior to ch6a6 ip injection; and partial protection resulted from administration of or 0.75 \ig/g or O.S^ig/g of h3G8 5.1 are used. (Figure 14). Thus, the data indicate that although CD32A can mediate ITP, the mjection of Ijig/g of h3G8 5.1 completely and unexpectedly protects mice against platelet depletion. Example 16. Prevention of Platelet Depletion Using Hu3G8-5.1-N2970 Produced m CHO-S Cell Line [02001 Hu3G8-5.1-N297Q was produced in a CHO-S ceUline. The ability of this antibody to protect against ITP in FcyRIII-/-, hCD 16A single transgenic mice was determined using the procedure described in Example 13. As is shown in Figure 15, admitiistration of 0.5mg/kg or more Hu3G8-5.1-N297Q produced in CHO-S cells one hour prior to ch6A6 i.p.injection completely protects mice against ITP. Example 17. Therapeutic Effect of Aglvcosylated Himianized Antibodies [0201] ITP was induced in mice as described above, by i.p. injection of O.lug/g ch6A6 at time 0. Two hours later, the number of platelets in the plasma was determined to confirm the presence of ITP. Three hours after i.p. injection of ch6A6, mice were injected i.v. with hu3G8-5.1-N297Q at different concentration (arrow). The results (Figure 16A) indicate that the number of platelets rapidly returns to normal after Hu3G8-5.1-N297Q injection whereas the number of platelets remains low in non-treated mice. These results demonstrate that administration of the hu3G8-5.1-N297Q antibody can be used to cure ITP in the mouse model. [0202] In this experiment, ITP was induced by i.p. injection of O.lug/g ch6A6 at time 0. Two hours later, the number of platelets in the plasma was determined to confirm the presence of ITP. Three hoin"s after i.p. injection of ch6A6, mice were injected i.v. with hu3G8-22.1-N297Q or hu3G8-22.43-N297Q at 0.5ug/g (arrow). The results indicate that the number of platelets rapidly returns to normal after H\i3G8-22.1-N297Q injection whereas the number of platelets remains low in non-treated mice and in mice treated with Hu3G8-22.43-N297Q (Figure 16B). These data indicate that hu3G8-22.1-N297Q can be used to cure ITP m the moxise model. Example 18: TherapeuticEffect of Hu3G8-22.1-N2970in AHAinmuFcvR^I-/-■ huFcYRIIIA transgenic mice [0203] In this experiment, AHA was induced by i.p. injection of 50 ug mouse anti-RBC IgG2a Mab 34-3C at day 0. On day 1, the number of RBC m the blood was determined to confirm the presence of AHA. Two hours later, mice were injected i.v. with Hu3G8-22.1-N297Q at various concentrations (arrow). The results indicate that the number of RBC remained stable after Hii3G8-22.I-N297Q injection whereas the number of RBC continued to drop in non-treated mice (Figure 17). The optimal concentration of Hu3G8-22.1-N297Q is 0.5ug/g. The number of RBC returned to normal in all mice at day 7. Control mice were bled every day but not injected in order to determine the effect of repeated bleedings on the number of RBC. These results in the mouse model indicate that Hu3G8-22.1-N297Q can be used to cure AHA. Hu3G8-22.1-N297Q prevents further RBC depletion by autoantibodies and therefore protects mice against anemia. [0204] It is understood that the examples and embodiments described herein are for illustrative purposes only and that various modifications or changes in light thereof will be suggested to persons skilled in the art and are to be mcluded within the spirit and purview of this appUcation and scope of the appended claims. AU publications (including sequence accession numbers and corresponding annotations), patents and patent applications cited herein are hereby incorporated by reference in their entirety for all purposes to Ihe same extent as if each individual publication, patent or patent application were specifically and individually indicated to be so incorporated by reference. WE CLAIM: 1. An anti-CD16A antibody comprising a VH domain comprising complementarity determining regions (CDRs) a CDRl having the amino acid sequence of SEQ ID NO:35, a CDR2 having the amino acid sequence of SEQ ID NO:39, and a CDRS having the amino acid sequence of SEQ ID NO:59 and a VL domain comprising a CDRl having the amino acid sequence of SEQ ID NO:67, a CDR2 having the amino acid sequence of SEQ ID NO:75, and a CDR3 having the amino acid sequence of SEQ ID NO:88, wherein at least one of said CDRs comprises at least one amino acid substitution selected from the group consisting of, in the VH domain, M34V in CDRl, H50L in CDR2, W52F in CDR2, D54N in CDR2, N60S in CDR2, A62S in CDR2, W99Y in CDR3, AIOID in CDR3, and, in the Vu domain, K24R in CDRl, A25S in CDRl, F32Y in CDRl, M33L in CDRl, N34A in CDRl, T50A, T50W, or T50S in CDR2, T51A in CDR2, N53S in CDR2, E55A or E55Q in CDR2, S56T in CDR2, N92Y in CDR3, N93S in CDR3, and D92T in CDR3, which positions are according to the Kabat numbering scheme. 2. The antibody as claimed in claim 1 that has been humanized. 3. The antibody as claimed in claim 2 that comprises a VH domain having the sequence ofSEQ ID NO: 104, 109, or 113. 4. The antibody as claimed in claim 2 or 3 comprising a VL domain having the sequence of SEQ ID NO:96, 100, or 118. 5. The antibody as claimed in claim 2, that comprises a V^ domain having the sequence of SEQ ID NO: 113 and a Vi domain having the sequence of SEQ ID NO:96, 100, or 118. 6. The antibody as claimed in claim 2 that comprises a VH domain comprising an FR3 domam having the sequence of SEQ ID N0:51. 7. A humanized anti-CD16A antibody comprising a VL domain having the amino acid sequence of SEQ ID NO:96. 8. The humanized antibody as claimed in claim 7 further comprising a VH CDRl having the amino acid sequence of SEQ ID NO:35, a VH CDR2 having the amino acid sequence of SEQ ID NO:39, and a VH CDR3 havmg the amino acid sequence of SEQ ID NO:59. 9. The humanized antibody as claimed in claim 7 further comprising a VH domain having the amino acid sequence of SEQ ID NO;104 or 109. 10. The antibody as claimed in any one of claims 1 to 9, wherein said antibody lacks an Fc region or comprises an Fc region that lacks effector function. 11. The antibody as claimed in claim 10 comprising an Fc region that lacks effector function, which Fc region is derived from a human IgGl Fc region. 12. The antibody as claimed in claim 10 comprising an Fc region that lacks effector function, wherein the amino acid corresponding to residue 297 of the Fc region according to the Kabat numbering scheme is not glycosylated. 13. The antibody as claimed in claim 12, wherein the amino acid corresponding to residue 297 of the Fc region according to the Kabat numbering scheme is not asparagine. 14. The antibody as claimed in any of claims 7 to 9, wherein said antibody comprises a light chain having the amino acid sequence SEQ ID NO:98. 15. The antibody as claimed in claim 7 further comprising a heavy cham having the amino acid sequence of SEQ ID NO: 111. 16. A humanized anti-CD16 antibody comprising a light chain having the amino acid sequence of SEQ ID NO:98 and a heavy chain having the amino acid sequence of SEQ ID NO: 111. 17. The antibody as claimed in any one of claims 1 to 16 that is a single chain antibody. 18. The antibody as claimed in any one of claims 1 to 17 that is a tetrameric antibody.

Documents:

2956-chenp-2004 abstract.pdf

2956-chenp-2004 claims duplicate.pdf

2956-chenp-2004 claims.pdf

2956-chenp-2004 correspondence others.pdf

2956-chenp-2004 description (complete) duplicate.pdf

2956-chenp-2004 description (complete).pdf

2956-chenp-2004 drawings duplicate.pdf

2956-chenp-2004 drawings.pdf

2956-chenp-2004 form-1.pdf

2956-chenp-2004 form-18.pdf

2956-chenp-2004 form-26.pdf

2956-chenp-2004 form-3.pdf

2956-chenp-2004 form-5.pdf

2956-chenp-2004 pct.pdf

2956-chenp-2004 petiiton.pdf