(12) United States Patent (10) Patent No.: US 8,188,231 B2 Lazar Et Al

USOO8188231B2

(12) United States Patent (10) Patent No.: US 8,188,231 B2 LaZar et al. (45) Date of Patent: *May 29, 2012

(54) OPTIMIZED FC VARIANTS 5,633,162 A 5/1997 Keen et al. 5,648,237 A 7, 1997 Carter (75) Inventors: Gregory Alan Lazar, Los Angeles, CA 5,648,260 A 7, 1997 Winter et al. (US); Wei Dang, Pasadena, CA (US); 5,821,337 A 10, 1998 Carter et al. 5,834,597 A 11/1998 Tso et al. John R. Desjarlais, Pasadena, CA (US); 5,885,573 A 3, 1999 Bluestone et al. Sher Bahadur Karki, Pasadena, CA 6,030,613 A 2/2000 Blumberg et al. (US); Omid Vafa, Monrovia, CA (US); 6,086,875 A 7/2000 Blumberg et al. 6,121,022 A 9, 2000 Presta et al. Robert Hayes, Paoli, PA (US); Jost 6,165,745 A 12/2000 Ward et al. Vielmetter. Altadena, CA (US) 6,188,965 B1 2/2001 Mayo et al. 6,194.551 B1 2/2001 Idusogie et al. (73) Assignee: Xencor, Inc., Monrovia, CA (US) 6.242,195 B1 6/2001 Idusogie et al. (*) Notice: Subject to any disclaimer, the term of this 6,269,312 B1 7/2001 Mayo et al. 6,277.375 B1 8, 2001 Ward patent is extended or adjusted under 35 6,284,536 B1 9, 2001 Morrison et al. U.S.C. 154(b) by 802 days. 6,331.415 B1 12/2001 Cabilly et al. 6,358,733 B1 3/2002 Motwani et al. This patent is Subject to a terminal dis 6,365,161 B1 4/2002 Deo et al. claimer. 6,403,312 B1 6/2002 Dahiyat et al. 6,444,789 B1 9, 2002 Luo (21) Appl. No.: 11/124,620 6,485,726 B1 1 1/2002 Blumberg et al. (22) Filed: May 5, 2005 6,528,624 B1 3/2003 Idusogie et al. 6,538,124 B1 3/2003 Idusogie et al. (65) Prior Publication Data 6,632,927 B2 10/2003 Adair et al. 6,649,165 B2 11/2003 Schubert US 2006/OO24298 A1 Feb. 2, 2006 6,708, 120 B1 3/2004 Mayo et al. 6,719,971 B1 4/2004 Carter et al. US 2009/0053211 A9 Feb. 26, 2009 6,737,056 B1 5, 2004 Presta 6,792.356 B2 9/2004 Mayo et al. Related U.S. Application Data 6,797.492 B2 9/2004 Daugherty et al. (63) Continuation-in-part of application No. 10/822,231, 6,801,861 B2 10/2004 Mayo et al. filed on Mar. 26, 2004, now Pat. No. 7,317,091, which (Continued) is a continuation-in-part of application No. 10/672.280, filed on Sep. 26, 2003, now abandoned. FOREIGN PATENT DOCUMENTS EP O 268 636 B1 1, 1997 (60) Provisional application No. 60/589,906, filed on Jul. 20, 2004, provisional application No. 60/627,026, (Continued) filed on Nov. 9, 2004, provisional application No. OTHER PUBLICATIONS 60/627,774, filed on Nov. 12, 2004, provisional Martin et al. Molecular Cell 2001 7:867-877. application No. 60/626,991, filed on Nov. 10, 2004, Algre, et al., “A non-activating "humanized' anti-CD3 monoclonal provisional application No. 60/477,839, filed on Jun. antibody retains immunosuppressive properties in vivo. Transplan 12, 2003, provisional application No. 60/467,606, tation, 57: 1537-1543 (1994). filed on May 2, 2003, provisional application No. Armour, et al., “Recombinant human IgG molecules lacking 60/442,301, filed on Jan. 23, 2003, provisional Ficgamma receptor I binding and monocyte triggering activities.” Eur application No. 60/414.433, filed on Sep. 27, 2002. J Immunol. 29:2613-2624 (1999). Ashkenazi, et al., “Immunoadhesins as research tools and therapeutic (51) Int. Cl. agents.” Curr Opin Immunol, 9:195-200 (1997). C07K 16/00 (2006.01) Chamow, et al., “Immunoadhesins: principles and applications.” Trends Biotechnol, 14:52-60 (1996). CI2P 2/08 (2006.01) Davies, et al. “Expression of GnTIII in a recombinant anti-CD20 A 6LX39/395 (2006.01) CHO production cell line: Expression of antibodies with altered A6 IK39/40 (2006.01) glycoforms leads to an increase in ADCC through higher affinity for CO7K I/OO (2006.01) FC gamma RIII.” Biotechnol Bioeng, 74:288-294 (2001). (52) U.S. Cl...... 530/387.1:530/387.3: 530/388.1; Hutchins, et al., “Improved biodistribution, tumor targeting, and 530/388.22; 424/130.1; 424/133.1; 424/136.1: reduced immunogenicity in mice with a gamma 4 variant of 424/141.1; 424/143.1 Campath-1H.” PNAS USA,92: 11980-1 1984 (1995). (58) Field of Classification Search ...... None Jefferies, et al., Immunol Lett, 54: 101-104 (1996). See application file for complete search history. (Continued) (56) References Cited Primary Examiner — Chun Dahle (74) Attorney, Agent, or Firm — Morgan Lewis & Bockius U.S. PATENT DOCUMENTS LLP, Robin M. Silva, Esq.; Ada O. Wong, Esq. 4,816,397 A 3, 1989 BOSS et al. (57) ABSTRACT 5,225.348 A 7/1993 Nagata et al. 5,266.491 A 11/1993 Nagata et al. The present invention relates to optimized Fc variants, meth 5,328,987 A 7, 1994 Maliszewski ods for their generation, Fc polypeptides comprising opti 5,541,087 A 7, 1996 Lo et al. mized Fc variants, and methods for using optimized Fc vari 5,576, 184 A 11/1996 Better et al. antS. 5,623,053 A 4/1997 Gastinel et al. 5,624,821 A 4/1997 Winter et al. 10 Claims, 69 Drawing Sheets US 8,188,231 B2 Page 2

U.S. PATENT DOCUMENTS FOREIGN PATENT DOCUMENTS 6,804,611 B2 10/2004 Mayo et al. EP 1176 195 A1 1, 2002 6,821,505 B2 11/2004 Ward EP 1 229 125 A1 8, 2002 6,933,368 B2 8, 2005 CO et al. EP 1255 209 A2 11/2002 6,946,292 B2 9, 2005 Kanda et al. EP 1255 826 B1 11/2002 6,950,754 B2 9/2005 Mayo et al. EP O 753 065 B1 5, 2003 6,982,321 B2 1/2006 Winter EP O 805 628 B1 5, 2003 6,992.234 B2 1/2006 Roopenian EP 1323 346 A2 11/2003

7,355,008 B2 * 4/2008 Stavenhagen et al. ... 530,387.1 EP 1323 346 A3 11 2003 7,662.925 B2 * 2/2010 Lazaret al...... 530,387.1 EP O 888 125 B1 5, 2004 2001.0036459 A1 11, 2001 Ravetch EP O 904 107 B1 10, 2004 2002fOOO4587 A1 1/2002 Miller et al. EP O 383 799 B2 2, 2005 2002/004.8772 A1 4/2002 Dahiyat et al. WO WO 88,07089 A1 9, 1988 2002.0062010 A1 5/2002 Arathoon et al. WO WO91,06305 A1 5, 1991 2002/0090648 A1 7/2002 Dahiyat et al. WO WO91, 19515 A1 12, 1991 2002/01423.74 A1 10, 2002 Gallo et al. WO WO92/04053 A1 3, 1992 2002fO155537 A1 10, 2002 Carter et al. WO WO92, 16562 A1 10, 1992 2002/0164328 A1 11/2002 Shinkawa et al. WO WO92,22324 A1 12, 1992 2002/0192222 A1 12/2002 Blumberg et al. WO WO94,29351 A2 12, 1994 2003, OO12789 A1 1/2003 Blumberg et al. WO WO94,29351 A3 12?1994 2003.0049654 A1 3/2003 Dahiyat et al. WO WO95/O5468 A1 2, 1995 2003, OO73164 A1 4/2003 Simmons et al. WO WO 96,22024 A1 T 1996 2003/0078385 A1 4/2003 Arathoon et al. WO WO 97.28267 A1 8, 1997 2003. O105294 A1 6, 2003 Gilles et al. WO WO97,34631 A1 9, 1997 2003. O108548 A1 6, 2003 Bluestone et al. WO WO 98.02462 A1 1, 1998 2003.0118592 A1 6, 2003 Ledbetter et al. WO WO 98.05787 A1 2, 1998 2003. O130827 A1 7/2003 Bentzien et al. WO WO 98/23289 A1 6, 1998 2003. O133939 A1 7/2003 Ledbetter et al. WO WO98,47089 A1 11, 1998 2003.0143682 A1 7/2003 Nicolaides et al. WO WO99,04813 A1 2, 1999 2003. O157108 A1 8, 2003 Presta WO WO99,51642 A1 10, 1999 2003. O158289 A1 8, 2003 Rusin et al. WO WO99.54342 A1 10, 1999 2003. O158389 A1 8/2003 Idusogie et al. WO WO99,54484 A1 10, 1999 2003. O166868 A1 9, 2003 Presta et al. WO WOOOO9560 A2 2, 2000 2003/0175884 A1 9, 2003 Umana et al. WO WOOOO9560 A3 2, 2000 2003/01903 11 A1 10/2003 Dall’Acqua et al. WO WOOO. 23564 A2 4/2000 2003/0208054 A1 11/2003 Olsen et al. WO WOOO. 23564 A3 4/2000 2003/0224397 A1 12/2003 Lowman et al. WO WOOO,24782 A2 5, 2000 2003/0229208 Al 12/2003 Queen et al. WO WOOO,24782 A3 5, 2000 2003/0235536 Al 12/2003 Blumberg et al. WO WOOOf 42072 A2 T 2000 2004/OOO2587 A1 1/2004 Watkins et al. WO WOOOf 42072 A3 T 2000 2004.0043429 A1 3/2004 Dahiyat et al. WO WOOOf 61739 A1 10, 2000 2004.0043430 A1 3/2004 Dahiyat et al. WO WOO1/29246 A1 4/2001 2004.0062763 A1 4/2004 Mosser et al. WO WOO1/38490 A2 5, 2001 2004/0093621 A1 5, 2004 Shitara et al. WO WOO1,57088 A1 8, 2001 2004/O132101 A1* 7/2004 LaZaret al...... 435.7.1 WO WOO1, 59066 A2 8, 2001 2004/O185045 A1 9/2004 Koenig et al. WO WOO1, 59066 A3 8, 2001 20040191244 A1 9, 2004 Presta WO WOO2,30954 A1 4/2002 2004/019 1256 A1 9/2004 Raju WO WOO2A31140 A1 4/2002 2004/O192897 A2 9, 2004 Winter WO WO O2/44215 A2 6, 2002 2004/022885.6 A1 11, 2004 Presta WO WO O2/060919 A2 8, 2002 2004/0258677 A1 12/2004 Waldmann et al. WO WO O2/060919 A3 8, 2002 2004/0258682 Al 12/2004 Leung et al. WO WO O2/O61090 A3 8, 2002 2004/025915.0 A1 12/2004 Nakamura et al. WO WO 02/061093 A1 8, 2002 2005, OO14934 A1 1/2005 Hinton et al. WO WOO2/O66514 A 8, 2002 2005, OO31626 A1 2/2005 Stevenson WO WOO3,O14325 A2 2, 2003 2005, OO32114 A1 2/2005 Hinton et al. WO WOO3,O14325 A3 2, 2003 2005/0033029 A1 2, 2005 Lu WO WOO3,O16470 A2 2, 2003 2005, 0037000 A1 2/2005 Stavenhagen et al. WO WOO3,O746.79 * 3/2003 2005, 0037002 A1 2/2005 Velardi et al. WO WOO3,035.835 A2 5, 2003 2005/003861.0 A1 2/2005 Mayo et al. WO WOO3,035.835 A3 5, 2003 2005.0054.046 A1 3/2005 Presta et al. WO WOO3/O54213 A2 T 2003 2005, OO64514 A1 3/2005 Stavenhagen et al. WO WOO3,059282 A T 2003 2005, 0118174 A1 6, 2005 Presta WO WOO3,O746.79 9, 2003 2005. O152894 A1 7/2005 Krummen et al. WO WO 03/089624 A2 10, 2003 2005/0175614 A1 8, 2005 Ledbetter et al. WO WO 2004/004662 A2 1, 2004 2005/0202023 A1 9, 2005 Ledbetter et al. WO WO 2004.?004.798 A2 1, 2004 2005/0202028 A1 9, 2005 Ledbetter et al. WO WO 2004.?004.798 A3 1, 2004 2005/0202534 A1 9, 2005 Ledbetter et al. WO WO 2004/O16750 A3 2, 2004 2005/0215767 A1 9/2005 Koenig et al. WO WO 2004/O22717 A2 3, 2004 2005/0226864 A1 10, 2005 Hinton et al. WO WO 2004/O22717 A3 3, 2004 2005/0233382 A1 10, 2005 Presta WO WO 2004/024.871 A2 3, 2004 2005/0272128 A1 12/2005 Umana et al. WO WO 2004/024.889 A2 3, 2004 2005/0276799 A1 12, 2005 Hinton et al. WO WO 2004/O292O7 A 4/2004 2006, OO1931.6 A1 1/2006 Mayo et al. WO WO 2004/035752 A2 4/2004 2006/0275283 Al 12/2006 Van Vlijmen et al. WO WO 2004/056312 A2 T 2004 2007. O148164 A1 6/2007 Farrington et al. WO WO 2004/063351 A2 T 2004 2007/0224.192 A1 9, 2007 LaZaret al. WO WO 2004/063351 A3 T 2004 2007/0238665 A1 10, 2007 LaZaret al. WO WO2004/063963 A T 2004 2008/0206867 A1 8/2008 Desjarlais et al. WO WO 2004/074455 A2 9, 2004 US 8,188,231 B2 Page 3

WO WO 2004/074455 A3 9, 2004 Boruchov, A. M., et al., “Activating and inhibitory IgGFc receptors WO WO 2004/092219 A2 10, 2004 on human DCs mediate opposing functions' J. Clin. Invest, doi:10. WO WO2004/0992.49 A 11, 2004 WO WO 2004,103404 A1 12/2004 1172/JCI24772 (Sep. 16, 2005). WO WO 2004f1 10472 A2 12/2004 Bowles, J. A., et al., “CD16 polymorphisms and NK activation WO WO 2005/OOO899 A2 1, 2005 induced by monoclonal antibody-coated target cells,” Journal of WO WO 2005/001025 A2 1, 2005 Immunological Methods, pp. 1-12 (2005). WO WO 2005/007809 A2 1, 2005 Brekke, O. H., et al., “Human IgG isotype-specific amino acid resi WO WO 2005/O 11376 A2 2, 2005 dues affecting complement-mediated cell lysis and phogocytosis.” WO WO 2005/O12877 A2 2, 2005 Eur.J. Immunil., 24(10):2542-5247 (Oct. 1994). WO WO 2005/O 13090 A2 2, 2005 WO WO 2005/O18572 A2 3, 2005 Brekke, O. H., et al., “Human IgG3 can adopt the disulfide bond WO WO 2005/023866 A2 3, 2005 pattern characteristic for IgG1 without resembling it in complement WO WO 2005/027966 A2 3, 2005 mediated cell lysis.” Mol. Immunol. 30(16): 1419-1425 (Nov. 1993). WO WO 2005/037867 A1 4/2005 Bruggeman, M., et al., "Comparison of the Effector Functions of WO WO 2005/040217 A2 5, 2005 Human Immunoglobulins Using a Matched Set of Chimeric Anti WO WO 2005/047327 A2 5, 2005 bodies,” J. Exp. Med., 166:1351-1361 (Nov. 1987). WO WO 2005/0566.06 A 6, 2005 Bruggemann, M., et al., “A matched set of rat?mouse chimeric anti WO WO 2005/056759. A 6, 2005 WO WO 2005/060642 A2 7/2005 bodies. Identification and biological properties of rat H chain con WO WO 2005/063815 A2 7/2005 stant regions mu, gammal, gamma 2a, gamma 2b, gamma 2c, epsi WO WO 2005, O70963 A1 8, 2005 lon, and alpha.” J. Immunol., 142(9):3145-3150 (May 1989). WO WO 2005,116O78 A1 12/2005 Burmeister, W. P. et al., “Crystal structure of the complex of rat WO WO 2005,123780 A2 12/2005 neonatal Fc receptor with Fc” Nature, 372:379-383 (Nov. 24, 1994). WO WO 2006/O12500 A2 2, 2006 Canfield, S. M., et al., “The Binding Affinity of Human IgG for its High Affinity Fc Receptor is Determined by Multiple Amino Acids in OTHER PUBLICATIONS the C2 Domain and Is Modulated by the Hinge Region.” J. Exp. Krapp, et al., “Structural analysis of human IgG-Fc glycoforms Med., 173: 1483-1491 (Jun. 1991). reveals a correlation between glycosylation and structural integrity.” Caron, P. C., et al., “Engineered Humanized Dimeric Forms of IgG J Mol Biol, 325:979-989 (2003). Are More Effective Antibodies,” J. Exp. Med., 176: 1191-1195 (Oct. Lehrnbecher, et al., “Variant Genotypes of the Low-Affinity Fcy 1992). Receptors in Two Control Populations and a Review of Low-Affinity Caron, P.C., et al., “Murine and humanized constructs of monoclonal Fcy Receptor Polymorphisms in Control and Disease Populations.” antibody M19 (anti-CD33) for the therapy of acute myelogenous Blood,94:4220-4232 (1999). leukemia.” Cancer, 73(3 Supp):1049-1056 (Feb. 1994). Lund, et al., “Human Fc gamma RI and Fc gamma RII interact with Carpenter, P.A., et al., “Non-Fc Receptor-Binding Humanized Anti distinct but overlapping sites on human IgG.”J Immunol. 147:2657 CD3 Antibodies Induce Apoptosis of Activated Human T Cells.” 2662 (1991). Journal of Immunology, 165:6205-6213 (2000). Lund, et al., “Multiple binding sites on the CH2 domain of IgG for Carter, P. et al., “Humanization of an anti-p185' antibody for human cancer therapy” PNAS, 89:4285-4289 (May 1992). mouse Fc gamma R11.” Mol Immunol. 29:53-59 (1992). Cartron, G., et al., “Therapeutic activity of humanized anit-Cd20 Lund, et al., “Multiple interactions of IgG with its core oligosac monoclonal antibody and polymorphism in IgGFc receptor FcyRIIIa charide can modulate recognition by complement and human Fc gene.” Blood, 99(3):754-758 (Feb. 1, 2002). gamma receptor I and influence the synthesis of its oligosaccharide Chapman, P. B., "T-Cell Chauvinists Versus Antibody Advocates— chains.”J Immunol. 154:4963-4969 (1996). Can't We All Just Get Along?” J. Clin. Oncology, 22(22):4446-4448 Lund, et al., “Oligosaccharide-protein interactions in IgG can modu (Nov. 15, 2004). late recognition by Fc gamma receptors.” Faseb.J.9:115-119 (1995). Chappel, M. S., et al., “Identification of a Secondary Fcy RI Binding White, et al., “Antibody-targeted immunotherapy for treatment of Site within a Genetically Engineered Human IgG Actibody.” J. Biol. malignancy.” Annu Rev Med. 52:125-145 (2001). Chem., 268(33):25124-25131 (Nov. 1993). Aase, A. etal. “The extended hinge region of IgG3 is not required for Chappel, M. S., et al., “Identification of the Fcy receptor class I high phogocytic capacity mediated by Fc gamma receptors, but the binding site in human IgG through the use of recombinant IgG1/IgG2 heavy chains must be disulfidebonded.” EurJImmunol., 23(7): 1546 hybrid and point-mutated antibodies.” PNAS, USA, 88:9036-9040 1551 (Jul 1993). (Oct. 1991). Abadeh, S., et al., “Remodelling the oligosaccharide of human IgG Chintalacharuvu, K. R. et al., “Hybrid IgA2/IgG1 Antibodies with antibodies: effects on biological activities.” Biochem Soc Trans. Tailor-Made Effector Functions.” Clinical Immunology, 101(1):21 25(4):S661 (Nov. 1997). 31 (Oct. 2001). Akewanlop, C., et al., “Phagocytosis of Breast Cancer Cells Medi Clark, M. R. “Chemical Immunology Antibody Engineering IgG ated by Anti-MUC-1 Monoclonal antibody, DF3, and Its Bispecific Effector Mechanisms.” Dissertation submitted to Immunology Divi Antibody” Cancer Research, 61:4061-4065 (May 15, 2001). sion of Department of Pathology at Cambridge University, UK (No Date). Alegre, M., et al., “Effect of a Single Amino Acid Mutation on the Clynes, R. A., et al. “Inhibitory Fc receptors modulate in vivo Activating and Immunosuppressive Properties of a “Humanised cytoxicity against tumor targets.” Nature Medicine, 6(4):443-446 OKT3 Monoclonal Antibody.” J. Immunology, 148:3461-3468 (Jun. (Apr. 2000). 1992). Clynes, R. et al., “Modulation of Immune complex-induced Inflam Amigorena, S., et al., “Fc receptors for IgG and antigen presentation mation. In Vivo by the Coordinate Expression of Activation and on MHC class I and class II molecules' Immunology, 11:385-390 Inhibitory Fc Receptors.” J. Exp. Med., 189(1): 179-185 (Jan. 4. (1999). 1999). Armour, K. L., et al., “Differential binding to human FcyRIIa and Clynes, R., “Immune complexes as therapy for autoimmunity' J. FcyRIIb receptors by human IgG wildtype and mutant antibodies.” Clin. Invest., 115:25-27 (2005). Molecular Immunology, 40:585-593 (2003). Clynes, R., et al., “Fc receptors are required in passive and active Ashkenazi, A., et al., “Mapping the CD4 binding site for human immunity to melanoma.” PNAS USA, 95:652-656 (Jan. 1998). immunodeficiency virus by alanine-Scanning mutagenesis.” PNAS, Cohen-Sodal, J. F.G. et al., “Review: Fcy receptors' Immunology USA, 87:7150-7154 (Sep.1990). Letts, 92: 199-205 (2004). Bolland, S. “A Newly Discovered Fc Receptor tha Explains IgG Cole, M. S., et al., “Human IgG2 variants of chimeric anti-CD3 are Isotype Disparities in Effector Responses,” J. Immunity, 23:2-4 (Jul. nonmitogenic to T cells,” J. Immunol. 159(7):3613-3621 (Oct. 1, 2005). 1997). US 8,188,231 B2 Page 4

Coloma, M. J., et al., “The hinge as a spacer contributes to convalent Ghetie, V., et al., “Multiple Roles for the Major Histocompatibility assembly and is required for function of IgG. J. Immunol. Complex Class I-Related Receptor FcRn.” Annu. Rev. Immunol. 158(2):733-740 (Jan. 15, 1997). 18:739-766 (2000). D'Uscio, C. H., et al., "Cellular cytotoxicity mediated by isotype Gonzales, N. R., et al., “SDR grafting of a murine antibody using Switch variants of a monoclonal antibody to human neuroblastoma, multiple human germline templates to minimize its immunogenic Br. J. Cancer, 64(3):445-450 (Sep.1991). ity.” Molecular Immunology, 41:863-872 (2004). Da Silveira, S.A., et al., “Complement Activation Selectively Poten Greenwood, J. "Molecular Recognition in the Structure and Assem tiates the Pathogenicity of the IgG2b and IgG3 Isotypes of a High bly of Filamentous Bacteriphages.” Dissertation submitted to the Affinity Anti-Erythrocyte Autoantibody.” J. Exp. Med., 195(6):665 University of Cambridge (Oct. 1989). 672 (Mar. 18, 2002). Greenwood, J., et al. "Structural motifs involved in human IgG anti Dall'Acqua, D.F., et al., “Increasing the Affinity of a Human IgG1 for body effector functions.” Eur: J. Immunol. 23(5):1098-1104 (May the Neonatal Fc Receptor: Biological Consequences,” Journal of 1993). Immunology, 169:5171-5180 (2002). Greenwood, J., et al., “Dual Importance of Positive Charge in the Davis, R. S., et al., “Fc receptor homologs: newest members of a C-Terminal Region of Filamentous Bacteriophage Coat Protein for remarkably diverse Fc receptor gene family.” Imm. Revs, 190: 123 Membrane Insertion and DNA-Protein Interaction in Virus Assem 136 (2002). bly” Virology, 171:444-452 (1989). Davis, R.S., et al., “Identification of a family of Fc receptor homologs Greenwood, J., et al., “Effector functions of matched sets of recom with preferential B cell expression.” PNAS, USA, 98(17):9772-9777 binant human IgG subclass antibodies.” Dissertation Submitted to (Aug. 2001). Cambridge University, Cambridge, UK (Feb. 1993). Delano, W. L., et al., “Convergent Solutions to Binding at a Protein Greenwood, J., et al., “Engineering multiple-domain forms of the Protein Interface” Science, 287: 1279-1283 (Feb. 18, 2000). therapeutic antibody CAMPATH-1H: effects complement lysis.” Dhodapkar, K.M., et al., “Antitumor Monoclonal Antibodies Ther. Immunol., 1(5):247-255 (Oct. 1994). Enhance Cross-Presentation of Cellular Antigens and the Generation Groh, V., et al., “Efficient cross-priming of tumor antigen specific T of Myeloma-specific Killer T-Cells by Dendritic Cells' J. Exp Med., cells by dendritic cells sensitized with diverse anti-MICA opsonized 195(1):125-133 (Jan. 7, 2002). tumor cells' PNAS, 102(18):6461-6466 (May 3, 2005). Dhodapkar, K.M., et al., “Recruiting dendritic cells to improve anti Harrison, P.T. et al., “Domain Swap chimeras to study the binding of body therapy of cancer" PNAS, 102(18):6243-6244 (May 3, 2005). IgG by Fc gamm RI, the high affinity receptor for IgG.” Biochem Soc Dhodapkar, K.M., et al., “Selective blockade of inhibitory Fcy recep tor enables human dendritic cell maturation with IL-12p70 produc Trans. 24(1): 144S (Feb 1996). tion and immunity to antibody-coated tumor cells' PNAS, Hazenbos, W.L., et al., “Murine IgG1 complexes Trigger Immune 102(8):2910-2915 (Feb. 22, 2005). Effector Functions Predominately via FcyRIII (CD16).”.J. of Immu Dhodapkar, M.V., et al., “T cells from the tumor microenvironment nology, 161:3026-3032 (1998). of patients with progressive myeloma can generate strong, tumor Henry, A. J., et al., “Participation of the N-Terminal of Ce3 in the specific cytolytic responses to autologous, tumor-loaded dendritic Binding of Human IgE to Its High-Affinity Receptor FceRI,” Bio cells” PNAS, 99(20): 13009-13013 (Oct. 1, 2002). chemistry, 36:15568-15578 (1997). Duncan, A. R. et al., “Localization of the binding site for the human Hezareh, M., et al., “Effector Function Activities of a Panel of high-affinity Fc receptor on IgG.” Nature, 332:563-564 (Apr. 7, Mutants of a Broadly Neutralizing Antibody against Human 1988). Immunodeficiency Virus Type I.” Journal of Virology, 75(24): 12161 Duncan, A. R. et al., “The binding site for C1q on IgG.” Nature 12168 (2001). 332:738-740 (Apr. 21, 1988). Hinton, P. R. et al., “Engineered human IgG Antibodies with Longer Edelman, G. M., et al., “The Covalent Structure of an Entire YG Serum Half-Lives in Primates.” J. Biol Chem., 279(8):6213-6216 Immunoglobulin Molecule.” PNAS, 63:78-85 (1969). (Feb. 20, 2004). Ehrhardt, G. R. A., et al., “Th inhibitory potential of Fc receptor Idusogie, E. E., et al., “Engineered Antibodies with Increased Activ homolog 4 on memory B cells.” PNAS, USA, 100(23): 13489-13494 ity to Recruit Complement.” J. of Immunology, 166:2571-2575 (Nov. 2003). (2001). Ellison, J. W., et al., “The nucleotide sequence of a human Idusogie, E.E., et al., “Mapping of the C1q Binding Site on Rituxan, immunoglobulin Cygene” Nucleic Acids Research, 10(13):4071 a Chimeric Antibody with a Human IgG1 Fc.” J. of Immunology, 4079(1982). 164:41.78-4.184 (2000). Ernst, L. K., et al., “Molecular characterization of six variant Fcy Isaacs, J. D., “Improving Serotherapy with Monoclonal Antibodies' receptor class I (CD64) transcripts.” Molecular Immunology,35:943 dissertation submitted to the University of Cambridge (Mar. 1991). 954 (1998). Isaacs, J. D., et al., “From bench to bedside: discovering rules for Facchetti, F., et al., “An unusual Fc receptor-related protein expressed antibody design, and improving serotherapy with monoclonal anti in human centroblasts.” PNAS, USA, 99(6):3776-3781 (Mar. 19, bodies.” Rheumatology, 40:724-738 (2001). 2002). Issacs, J. D., et al., “Therapy with Monoclonal Antibodies, II. The Gaboriaud, C., et al., “The Crystal Structure of the Globular Head of contribution of Fcy Receptor binding and the Influenece of C1 and Complement Protein C1q Provides a Basis for Its Versatile Recogni C3 Domains on In Vivo Effector Function.” J. of Immunology, tion Properties,” J. Biol. Chem., 278(47):46974-46982 (2003). 161:3862-3869 (1998). Garman, S. C., et al., “Structure of the Fc fragment of human IgG Issacs, J. D., et al., “Therapy with Monoclonal Antibodies: an in vivo bound to its high-affinity receptor FceRIa,” Nature, 406:259-266 model for the assessment of therapeutic potential. J. Immunol. (2000). 148(10):3062-3071 (May 15, 1992). Getahun, A., et al., “IgG2a-Mediated Enhancement of Antibody and Jefferis, R. et al., “Recognition sites on human IgG for Fc gamma T Cell Responses and Its Relation to Inhibitory and Activating Fcy receptors: the role of glycosylatin.” Immunol Letters, 44(2-3): 111 Receptors.”.J. of Immunology, 172:5269-5276 (2004). 117 (Jan. 1995). Ghazizadeh, S., et al., “Physical and Functional Association of Src Jefferis, R., et al., “Interaction sites on human IgG-Fc for FcyR: related Protein Tyrosine Kinases with FcRII in Monocytic THP-1 current models.” Immunology Letts. 82:57-65 (2002). Cells,” J. Biol. Chem., 269(12):8878-8884 (Mar. 25, 1994). Jefferis, R., et al., “Modulation of FcyR and human complement Ghetie, V., et al., “FcRn: the MHC class I-related receptor that is more activation by IgG3-core oligosaccharide interactions.” Immunology than an IgG transporter” Immunology Today, 18(12):592-598 (Dec. Letters, 54: 101-104 (1996) and errata at Immunology Letters, 58:67 1997). (1997). Ghetie, V., et al., “Increasing the serum persistence of an IgG frag Jefferis, R., et al., “Molecular definition of interaction sites on human ment random mutagenesis.” Nat. Biotechol., 15(7):637-640 (Jul. IgG for Fc receptors (huFc gamma R).” Mol Immunol. 27(12): 1237 1997). 1240 (Dec. 1990). US 8,188,231 B2 Page 5

Jendeberg, L, et al., “Engineering of Fc and Fc from human Miller, I., et al., “ITRAS: a new family of immunoglobulinlike recep immunoglobulin G to analyse Subclass specificity for Staphylococcal tors differentially expressed in B cells.” Blood, 99(8):2662-2669 protein A,” Journal of Immunological Methods, 201:25-34 (1997). (Apr. 15, 2002). Junghans, R. P. et al., “The protection receptor for IgG catabolism is Mimura, Y, et al., “Role of Oligosaccharide Residues of IgG1-Fc in the f-microglobulin-containing neonatal intestinal transport recep Fcy RIIb Binding.” J. Biol. Chem. 276(49):45539-45547 (Dec. 7, tor.” PNAS, 93:5512-5516 (May 1996). 2001). Kalergis, A.M., et al., “Inducing Tumor Immunity through the Selec Morgan, A., et al., “The N-terminal end of the CH2 domain of tive Engagement of Activating Fcy Receptors on Dendritic Cells' J. chimeric human IgG1 anti-HLA-DR is necessary for C1q, Fc gamma R1 and Fc gamma RIII binding.” Immunology, 86(2):319-324 (Oct. Exp. Med. 195(12): 1653-1659 (Jun. 17, 2002). 1995). Kan, K. S., et al., “Thioether-Bonded Constructs of Faby and Fcy Nakamura, K., et al., “Dissection and optimization of immune effec Modules Utilizing Differential Reduction of Interchain Disulfide tor functions of humanized anti-ganglioside GM2 monoclonal anti Bonds,” Journal of Immunology, 166:1320-1326 (2001). body.” Molecular Immunology, 37: 1035-1046 (2000). Karassa, F. B., et al., “The role of FcyRIIA and IIIA polymorphisms Neidhardt-Berard, E., et al., “Dendritic cells loaded with killed breast in autoimmune diseases.” Biomedicine & Pharmacotherapy, 58:286 cells induce differentiation of tumor-specific cytoxic T lymphocytes' 291 (2004). Breast Cancer Res., 6R322-R328 (Apr. 30, 2004). Kim, J. et al., “Mapping the site on human IgG for binding of the Nimmerjahn, F., et al., “Divergent Immunoglobulin-G Subclass MHC class I-related receptor, FcRn' Eur:J. Immunol. 29:2819-2825 Activity Through Selective Fc Receptor Binding Science, 310: 1510 (1999). (2005). Kim, J. K., et al., “Localization of the site of the murine IgG1 Nimmerjahn, F., et al., “Fcy RIV: A Novel FcR with Distinct IgG molecule that is involved in binding to the murine intestinal Fc Subclass Specificity.” Immunity, 23:41-51 (Jul. 2005). receptor.” Eur.J Immunol., 24(10):2429-2439 (Oct. 1994). Nimmerjahn, F., et al., “Supporting Online Material for: Divergent Kim, J.K., et al., “Identifying amino acid residues that influence Immunoglobulin G Subclass Activity Through Selective Fc Receptor plasma clearance of murine IgG1 fragments by site-directed Binding Science, 310: 1510 (2005). mutagenesis.” Eur.J. Immunol. 24(3):542-548 (Mar. 1994). Niwa, R., et al., “Defucosylated Chimeric Anti-CC Chemokine Kim, T. D., et al., “Analysis of FcyRIII and IgGFc Polymorphism Receptor 4 IgG1 with Enhanced Antibody-Dependinent Cellular Reveals Functional and Evolutionary Implications of Protein-Protein cytotoxicity Shows Potent Therapeutic Activity to T-Cell Leukemia Interaction.” J. Mol. Evol., 53: 1-9 (2001). and Lymphoma.” Cancer Research, 64:2127-2133 (Mar. 15, 2004). Kurucz, I., et al., “Bacterially expressed human FcyRIIb is soluble Norderhaug, L., et al., "Chimeric mouse human IgG3 antibodies with and functionally active after in vitro refolding” Immunology Letts. an IgG4-like hinge region induce complement-mediated lysis more 75:33-40 (2000). efficiently than IgG3 with normal hinge.” Eur J immunol. Lund, J., et al., “A protein structural change in aglycosylated IgG3 21(10):2379-2384. correlates with loss of huFc gamma R1 and huFc gamma R111 O'Connor, S.J., et al., “Humanization of an antibody against human binding and/or activation.” Mol. Immunol. 27(11): 1145-1153 (Nov. protein C and calcium-dependence involving framework residues.” 1990). Protein Engineering, 11(4):321-328 (1998). Lund, J., et al., “Control of IgG/Fc glycosylation: a comparision of Ober, R. J., et al., “Differences in promiscuity for antibody-FcRn oligosaccharides from chimeric human/mouse and mouse Subclass interactions across species: implications for therapeutic antibodies.” immunoglobulin Gs.” Mol Immunol. 30(8):741-748 (Jun. 1993). International Immunology, 13(12): 1551-1559 (2001). Maenaka, K., et al., “The Human Low Affinity Fcy Receptors IIa, IIb Ober, R. J., et al., “Exocytosis of IgG as mediated by the receptor, and III Bind IgG with Fast Kinetics and Distinct Thermodynamic FcRn: An analysis at the single-molecule level” PNAS, Properties”.J. Biol. Chem. 276(48):44898-44904 (2001). 101(30): 11076-1 1081 (Jul 27, 2004). Martin, W. L., et al., “Characterization of the 2:1 Complex between Okazaki, A., et al., “Fucose Depletion from Human IgG1 Oligosac the Class IMHC-Related Fc Receptor and Its Fc Ligand in Solution.” charide Enhances Binding Enthalpy and Association Rate Between Biochemistry, 38:12639-12647 (1999). IgG1 and Fcy RIIIa.” J. Mol. Biol. 336:1239-1249 (2004). Martin, W. L., et al., “Crystal Structure at 2.8 A of an FcRn/ Parren, P. W., et al., “Characterization of IgGFcR-mediated prolif Heterodimeric Fc Complex: Mechanism of pH-Dependent Binding” eration of human T-cells induced by mouse and human anti-CD3 Molecular Cell, 7:867-877 (Apr. 2000). monoclonal antibodies. Identification of a functional polymorphism Masztalerz, A., et al., “Mechanisms of macrophage cytotoxicity in to human IgG2 anti-CD3.” J. Immunol. 148(3):695-701 (Feb. 1992). IL-2 and IL-12 mediated tumor regression.” Cancer Immunol Parren, P.W., et al., “On the interaction of IgG subclasses with the low Immunother, 52:235-242 (2003). affinity Fc gamma RIIa (CD32) on human monocytes, neutrophils, Maxwell, K.F., et al., “Crystal structure of the human leukocyte Fic and platelets. Analysis of a functional polymorphism to human receptor, FcRIIa.” Nature Structural Biology, 6(5):437-442 (May IgG2.”J Clin Invest., 90(4): 1537-1546 (Oct. 1992). 1999). Pearce, K. H., et al., “Mutational Analysis of Thrombopoietin for Mayfield, S. P. et al., “Expression and assembly of a fully active Identification of Receptor and Neutralizing Antibody Sites.” J. Biol. antibody algae.” PNAS, 100(2):438-442 (Jan. 21, 2003). Chem., 272(33):20595-20602 (1997). Mechetina, L. V., et al., “Identification of CD16-2, a novel mouse Preithner, S., et al., "High concentrations of therapeutic Iggl anti receptor homologous to CD16/FcyRIII.” Immunogenetics, 4:463 bodies are needed to compensate for inhibition of antibody 468 (2002). dependinent cellular cytotoxicity by excess endogenous Merchant, A. M. et al., “An efficient route to human bispecific IgG.” immunoglobulin G.” Molecular Immunology, (2005). Nat Biotechnol. 16(7):677-681 (1998). Presta, L.G., et al., “Engineering therapeutic antibodies for Improved Metes. D., et al., “Expression of Functional CD32 Molecules on function.” Biochemical Society Transactions, 30(part 4):487-490 Human NK Cells Is Determined by and Allelic Polymorphism of the (2002). Fcy RIIC Gene.” Blood. 91(7):2369-2380 (Apr. 1, 1998). Radaev, S., et al., “Recognition of IgG by Fcy Receptor.” J. Biol. Michaelson, T. E., et al., “Antibody Dependent Cell-Mediated Chem., 276(19): 16478-16483 (May 11, 2001). Cytotoxicity Induced by Chimeric Mouse-Human IgG Subclass and Radaev, S., et al., “Review: Recognition of immunoglobulins by Fcy IgG3 Antibodies with Altered Hinge Region.” Molecular Immunol recptors.” Molecular Immunology, 38: 1073-1083 (2001). ogy, 29(3):319-326 (1992). Radaev, S., et al., “The Structure of Human Type III Fcy Receptor in Michaelson, T. E., et al., “One disulfide bond in front of the second Complex with Fc.” J. Biol. Chem., 276(19) 16469-16477 (May 11, heavy chain constant region is necessary and Sufficient for effector 2001). functions of human IgG3 without a genetic hinge.” PNAS, 91:9243 Rafiq, K., et al., “Immune complex-mediated antigen presentation 9247 (Sep. 1994). induces tumor immunity”.J. Clin. Invest. 110:71-79 (2002). Michaelson, T. E., et al., “Primary Structure of the Hinge' Region of Raghavan, M., et al., “Fc Receptors and Their Interactions with Human IgG3.”J Biol Chem. 252(3):883-889 (Feb. 1977). Immunoglobulins' Annu. Rev. Cell Div. Biol. 12:181-220 (1996). US 8,188,231 B2 Page 6

Ravetch, J. V., et al., “IgG Fc Receptors' Annu. Rev. Immunol. Sorenson, V., et al., “Effect of the IgM and IgA secretory tailpieces on 19:275-290 (2001). polymerization and Secretion of IgM and IgG, J Immunol. Ravetch, J. V., et al., “Immune Inhibitory Receptors.” Science, 156(8):2858-2865 (Apr. 1996). 290:84-89 (Oct. 6, 2000). Steplewski, Z. et al., “Biological activity of human-mouse IgG1. Ravetch, J.V., et al., “Fc Receptors.” Annu. Rev. Immunol. 9:457-492 IgG2, IgG3, and IgG4 chimeric monoclonal antibodies with antitu (1991). mor specificity.” PNAS USA, 85:4852-4856 (Jul 1988). Reddy, P. R., et al., “Elimination of Fc Receptor-Dependent Effector Stevenson, G.T. et al., “Preparation of Fcy for addition to sulfhydryl Functions of a Modified IgG4 Monoclonal Antibody to Human CD4” expressing ligands with minimal disturbance of the hinge.” J. of J. Immunol. 164: 1925-1933 (2000). Immunological Methods, 231: 169-175 (1999). Redpath, S., et al., “The Influence of the Hinge Region Length in Tao, M., et al., “Structural Features of Human immunoglobulin G that Binding of Human IgGto Human Fcy Receptors.” Human Immunol Determine Isotype-specific Differences in Complement Activation.” ogy, 59:720-727 (1998). J. Exp. Med. 178:661-667 (Aug. 1993). Rozsnyay, Z. et al., “Distinctive role of IgG1 and IgG3 isotypes in Tao, M., et al., “The Differential Ability of Human IgG1 and IgG4 to FcR-mediated functions.” Immunology, 66(4):491-498 (Apr. 1989). Activate Complement Is Determined by the COOH-terminal Sandlie, A.A., “The extended hinge region of IgG3 is not required for Sequence of the C2 domain' J. Exp. Med. 173: 1025-1028 (Apr. high phogocytic capacity mediated by Fc gamma receptors, but the 1991). heavy chains must be disulfidebonded.” Eur.J. Immunol. 23(7): 1546 Thommesen, J. E., et al., “Lysine 322 in the human IgG3 C2 domain 1551 (Jul 1993). is crucial for antibody dependent complement activation” Molecular Sarmay, G., et al., “Mapping and Comparison of the Interaction Sites Immunology, 37:995-1014 (2000). on the Fc Region of IgG Responsible for Triggering Antibody Depen Tuijnman W. B., et al., "A flow cytometric rosetting assay for the dent Cellular Cytotoxicity (ADCC) Through Different Types of analysis of IgG-Fc receptor interactions.” J Immunol Methods, Human FCY Receptor.” Molecular Immunology, 29(5):633-639 127(2):207-214 (Mar. 1990). (1992). Uchide, J. et al., “The Innate Mononuclear Phagocyte Network Sautes-Fridman, C., et al., “Fc Gamma Receptors: A Magic Link with Depletes B Lymphocytes through Fc Receptor-dependent mecha the Outside World.” ASHI Quarterly, 148-151. (Fourth Quarter nisms during Anti-CD20 Antibody Immunotherapy” J. Exp. Med. 199(12): 1659-1669 (Jun. 21, 2004). 2003). Umana, P. et al., “Engineered glycoforms of an antineuroblastoma Sensel, M. G., et al., “Amino Acid Differences in the N-Teminus of IgG1 with optimized antibody-dependent cellular cytotoxic activity.” C2 Influence the Relative abilities of IgG2 and IgG3 to Activate Nature Biotechnology, 17:176-180 (1999). Complement” Mol. Immunol. 34(14) 1019-1029 (1997). Van Royen-Kerkhof. A. et al., “Flow cytometric determination of Fcy Shields, R. L., et al., “High Resolution Mapping of the Binding Site RIIa (CD32) polymorphism.” J. Immunol. Methods, 294:135-144 on Human IgG1 for Fcy RI. Fcy RII, Fcy RIII, and FcRn and Design (2004). of IgG1 Varients with Improved Binding to the Fcy R.J. Biol. Chem. Van Schie, R.C.A.A., et al., “Evaluation of Human Fcy RIIA (CD32) 276(9):6591-6604 (2001). and Fcy RIIIB (CD16) Polymorphisms in Caucasians and African Shields, R. L., et al., “Lack of Fucose on human IgG1 N-Linked Americans. Using Salivary DNA.” Clinical and Diagnostic Labora Oligodaccharide Improves Binding to Human Fcy RIII and Anti tory Immunology, 7(4):676-681 (Jul. 2000). body-dependent Cellular Toxicity” J. Biol. Chem., 277(30)26733 Van Sorge, N. M., et al., “Fcy R polymorphisms: Implications for 26740 (2002). function, disease and Susceptibility and immunotherapy' Tissue Anti Shinkawa, T., et al., “The Absence of Fucose but Not the Presence of gens, 63:189-202 (2003). Galactose or Bisecting N-Acetylglucosamine of Human IgG1 com Vidarte, L., et al., “Serine 132 Is the C3 Covalent Attachment Point of plex-type Oligosaccharides Shows the Critical Role of Enhancing the CH1 domain of Human IgG1” J. Biol. Chem., 276(41):38217 Antibody-dependent Cellular Cytotoxicity” J. Biol. Chem. 38223 (2001). 278(5):3466-3473 (2003). Ward, E. S., et al., “Evidence to support the cellular mechanism Shopes, B., “A genetically engineered human IgG mutant with involved in serum IgG homeostatis in humans' International Immu enhanced cytolytic activity.” J Immunol. 148(9):2918-2922 (May nology, 15(2): 187-195 (2003). 1992). Warmerdam, P. A., et al., “Interaction of a human Fc gamma RIIbl Shopes, B., et al., “Recombinant human IgG1-murine IgE chimeric (CD32) Isoform with murine and human IgG subclasses.” Int Ig. Construction, expression, and binding to human Fc gamma recep Immunol. 5(3):239-247 (Mar. 1993). tors.” J. Immunol., 145(11):3842-3848 (Dec. 1, 1990). Wawrzynczak, E. J., et al., “Recombinant mouse monoclonal anti Simmons, L. C., et al., “Expression of full-length immunoglobulins bodies with single amino acid Substitutions affecting Cld and high in Esherichia Coli; rapid and efficient production of a glycosylated affinity Fc receptor binding have identical serum half-lives in the antibodies”.J. Immunol. Methods, 263:133-147 (2002). BALB/c mouse.” Mol. Immunol., 29(2):221-227 (Feb. 1992). Smith, I. F. R. et al., “Addition of a u-Tailpiece to IgG Results in Weiner, L. M., et al., “Tunable antibodies.” Nature Biotechnology, Polymeric Antibodies with Enhanced Effector Functions Including 23(5):556-557 (May 2005). Complement-Mediated Cytolysis by IgG4.” J. Immunology, pp. Weng, W., et al., "Clinical Outcome of Lymphoma Patients. After 2226-2236 (1995). Idiotype Vaccination Is Correlated With Humoral Immune Response Smith, K.G., et al., “T cell activation by anti-T3 antibodies: compari and Immunoglobulin G Fc Receptor Genotype.” J. Clin Oncol., Son of IgG1 and IgG2b Switch variants and direct evidence for acces 22(23): 1-8 (2004). sory function of macrophage Fc receptors.” Eur J. Immunol. Weng, W., et al., “Two Immunoglobulin G Fragment C Receptor 16(5):478-486 (May 1986). Polymorphisms Independently Predict Response to Rituximab in Sonderman, P. et al., “Crystal structure of the soluble form of the Patients With Follicular Lymphoma.” Journal of Clinical Oncology, human FCy-receptor IIb: a new member of the immunoglobulin 21(21):3940-3947 (Nov. 1, 2003). superfamily at 1.7A resolution” EMBO Journal, 18(5):1095-1 103 West, A. P. et al., “Crystal Structure and immunoglobulin G Binding (1999). Properties of the Human Major Histocompatibility Complex-Related Sonderman, P. et al., “Human Fcy Receptor IIb Expressed in Fc Receptor.” Biochemistry, 39:9698-9708 (2000). Escherichia coli Reveals IgG Binding Capability” Biol. Chem. Wing, M. G. et al., “Mechanism of First-Dose Cytokine-Release 380:717-721 (Jun. 1999). Syndrome of CAMPATH 1-H:Involvement of CD16 (FcyRIII) and Sonderman, P. et al., “Molecular Basis for Immune Complex Rec CD11a/CD18 (LFA-1)on NK Cells,” J. Clin. Invest., 98(12):2819 ognition: A comparison of Fc-Receptor Structures' J. Mol. Biol. 2826 (Dec. 1996). 309:737-749 (2001). Wolff, E.A., et al., “Monoclonal antibody homodimers: enhanced Sonderman, P. et al., “The 3.2-A crystal structure of the human IgG1 antitumor activity in nude mice.” Cancer Res., 53( 11):2560-2565 Fc fragment-FcyRIII complex” Nature, 406:267-273 (Jul. 20, 2000). (Jun. 1, 1993). US 8,188,231 B2 Page 7

Wright, A., et al., “Effect of C2-Associated carbohydrate Structure Van Dijk, M., et al., “Human antibodies as next generation therapeu on IgEffector Function: Studies with Chimeric Mouse-Human IgG1 tics.” Curr: Opin. Chem. Biol. 5:368-374 (2001). Antibodies in Glycosylation Mutants of Chinese Hamster Ovary Van Sorge, N., et al., “Fcy R polymorphisms: Implications for func Cells”.J. of Immunology, 160:3393-3402 (1998). tion, disease Susceptibility and immunotherapy,” Tissue Antigens, Wright, A., et al., “In vivo trafficking and catabolism of IgG1 anti 61: 189-202 (2003). bodies with Fc associated carbohydrates of differing structure.” Vasserot, A., et al., “Optimization of protein therapeutics by directed Glycobiology, 10(12): 1347-1355 (2000). evolution.” Drug Discovery Today, 8(3):118-126 (2003). Xu, D., et al., “In Vitro Characterization of Five Humanized OKT3 Waldmann, T., et al., “Emerging Therapies: Spectrum of Application Effector Function Varient Antibodies,” Cellular Immunology, 200:16-26 (2000). of Monoclonal Antibody Therapy.” Hemotology, 394-408 (2000). Xu, M., et al., “Molecular Cloning and Characterization of SPAP1, an Bastida-Corcuera, et al., “Differential complement activation by Inhibitory Receptor.” Biochemical and Biophysical Research Com bovine IgG2 allotypes' Veterinary Immunology and Immunopathol munications, 280:768-775 (2001). ogy, 1999, vol. 71 No. 2, 115-123. Xu, Y, et al., “Residue at Position 331 in the IgG1 and IgG4 C2 Burton, et al. “Immunoglobulin G. Functional sites'. Molecular Domains Contributes to Their Differential Ability to Bind and Acti Immunology, vol. 22, No. 3, (Mar. 1985). vate Complement” J. Biol. Chem. 269(5):3469-3474 (1994). Carter, P., “Improving the Efficacy of Antibody-Based Cancer Thera Zelaschi. D., et al., “Human immunoglobulin allotypes: previously pies.” Nature Reviews, 1:118-129 (2001). unrecognized determinants and alleles defined with monoclonal anti Chan, et al. “Variable Region Domain Exchange in Human IgGs bodies.” PNAS, USA, 80:3762-3766 (Jun. 1983). promotes antibody complex formation with accompanying structural Zhou, H., et al., “DNA-based vaccines activate innate and adaptive changes and altered effector functions' Molecular Immunology antitumor immunity by engaging the NKG2D receptor" PNAS, 2004, 21:527-538. 102(31): 10846-10851 (Aug. 2, 2005). Chirino, A.J. etal. “Minimizing the immunogenicity of protein thera Zhou, J., et al., “Generation of Mutated Variants of the Human Form peutics'. Drug Discovery Today, 2004, vol. 9, No. 2, pp. 82-90. of the MHC Class I-related Receptor, FcRn, with Increased Affinity Clark, M. R. “Chemical Immunology Antibody Engineering IgG for Mouse Immunoglobulin G. J. Mol. Biol. 332(4):901-13 (Sep. Effector Mechanisms.” Dissertation submitted to Immunology Divi 2003). sion of Department of Pathology at Cambridge University, UK.J. Zhu, D., et al., “A novel human immunoglobulin Fc gamma Fc Chem. Immunol. 1997, vol. 65, pp. 88-110. epsilon bifunctional fusion protein Inhibits Fc epsilon RI-mediated Cole, M.S., et al., “HUM291, a Humanized Anti-CD3 Antibody, is degranulation.” Nat Med., 8(5):518-521 (May 2002). immunosuppressive to T cells while exhibiting reduced mitogenicity Andreakos, E., et al., “Monoclonal antibodies in immune and inflam in vitro”. Transplantation, vol. 68, No. 4, pp. 563-571 (1999). matory diseases.” Cur: Opin. Biotech., 13:615-620 (2002). Dahiyat, B. I. et al. “Protein Design Automation'. Protein Science, Chadd, H., et al., “Therapeutic antibody expression technology.” 1996, vol. 5, No. 5, p. 895-903. Curr: Opin. Biotech., 12:188-194 (2001). Dall'Acqua, W. et al. “Modulation of the Effector Functions of a Clark, M. “Antibody humanization: a case of the Emperor's new Human IgG1 through Engineering of Its Hinge Region', 2006, J. clothes?” Immunol. Today, 21 (8):397-402 (2000). Immunology, 177:1129-1138. Cragg, M., et al., "Signaling antibodies in cancer therapy,” Curr: Dall’Acqua, W., et al., "Antibody Engineering.” Curr: Opin Struc Opin. Immunol., 11:541-547 (1999). tural Biol. 8:443-450 (1998). Glennie, M., et al., “Clinical trials of antibody therapy.” Immun. Hayes R.J. et al. "Combining computational and experimental Today, 21 (8):403-410 (2000). screening for rapidoptimization of protein properties'. PNAS, 2002, Glennie, M., et al., “Renaissance of cancer therapeutic antibodies.” vol. 99...No. 25, pp. 15926-15931. Drug Discovery Today, 8(11):503–510 (2003). Kato, K. et al., “Analysis of IgG-FcgammaR interactions in Solution: Hayhurst, A., et al., “High-throughput antibody isolation.” Curr: Mapping of the FcgammaR binding site and evidence for a Opin. Chem. Biol. 5:683-689 (2001). conformational change occurring in the Fc region'. Immunology Hogarth, P., “Fc receptors are major mediators of antibody based Letters, vol. 73, No. 2-3 (2000). inflammation in autoimmunity.” Curr: Opin. Immun., 14:798-802 Morrison, et al. “Variable Region Domain Exchange Influences the (2002). Functional Properties of IgG''The Journal of Immunology 1998, Holliger, P. et al., “Antibodies come back from the brink.” Nature 160:2802-2808. Biotechnology, 16:1015-1016 (1998). Natsume, A. et al. "Engineered Antibodies of IgG1/IgG3 Mixed Hudson, P., “Recombinant antibody constructs in cancer therapy.” Isotype with Enhanced Cytotoxic Activities”, 2008, Cancer Curr: Opin. Immunology, 11:548-557 (1999). Research, 68:(10) pp. 3863-3871. Hudson, P., “Recombinant antibody fragments.” Curr: Opin in Pendley C. et al., “Immunogencity of therapeutic monoclonal anti Biotechnology, 9:395-402 (1998). bodies'. Current Opinion in Molecular Therapeutics, 2003, vol. 5, Johnson, G., et al., “Kabat Database and its applications: 30 years No. 2, pp. 172-179. after the first variability plot,” Nucleic Acids Research, 28(1):214 Shitara et al. "A new vector for the high level expression of chimeric 218 (2000). antibodies in myeloma cells' J. of Immunological Methods, 1994, Johnson, G., et al., "Kabat Database and its applications: future 167: 271-278. directions.” Nucleic Acids Research, 29(1):205-206 (2001). Tamm. A. et al., “IgG Binding Sites on Human Fcy Receptors' 1997. Maynard, J., et al., “Antibody Engineering.” Annu. Rev. Biomed. International Reviews of Immunology, 16:1.57-85. Eng., 2:339-376 (2000). Valerius, T. et al., “Fcalpha RI (CD89) as a Novel Trigger Molecule Morea, V., et al., “Antibody Modeling: Implications for Engineering for Bispecific Antibody Therapy” Blood, 1997, 90:4485-4492. and Design.” Methods, 20:267-279 (2000). WHO Review of the notation for the allotypic and related markers of Penichet, M., et al., “Antibody-cytokine fusion proteins for the human immunoglobulins. J Immunogen 1976, 3:357-362. therapy of cancer.” Journal of Immunological Methods, 248:91-1010 WHO Review of the notation for the allotypic and related markers of (2001). human immunoglobulins. 1976, Eur, J. Immunol. 6, 599-601. Reichert, J., “Monoclonal antibodies in the clinic, s Nature Wines, B.D. et al. “The IgGFc contains distinct Fc receptor (FcR) Biotechnology, 19:819-822 (2001). binding sites: the leukocyte receptors Fc gamma RIIa bind to a Thrush, G., et al., “Immunotoxins: An Update.” Ann. Rev. Immunol. region in the Fc distinct from that recognized by neonatal FeR and 14:49-71 (1996). protein A' Journal of Immunology, (2000), pp. 5313-5318. Torphy, T., et al., “Pharmaceutical biotechnology Monoclonal anti Woof, J.M. et al. “Localisation of the Monocyte-Binding Region on bodies: boundless potential, daunting challenges—Editorial Over Human Immunoglobulin G” Molecular Immunology, 1986, vol. 23. view.” Curr. Opin. Biotechnol., 13:589-591 (2002). No. 3, pp. 319-330. Trail, P. et al., “Monoclonal antibody drug conjugates in the treat Medesan, et al. “Comparative studies of rat IgG to further delineate ment of cancer Curr: Opin. Immunol. 11:584-588 (1999). the Fc:FcRn interaction site” Eur.J. Immunol. (1998) 28:2092-2100. Trikha, M., “Monoclonal antibodies as therapeutics in oncology.” Curr: Opin. Biotech., 13:609-614 (2002). * cited by examiner U.S. Patent May 29, 2012 Sheet 1 of 69 US 8,188,231 B2

Figure 1

FC ASn297

FCRn Proteins A & G U.S. Patent May 29, 2012 Sheet 2 of 69 US 8,188,231 B2

Figure 2

FoyRIIIb

U.S. Patent May 29, 2012 Sheet 7 of 69 US 8,188,231 B2

VNod (d o (o win n n & Ll LL Ll Ll Nod (d od (d w w cN cy S>>> S.-- - - szzzzre CNs () (d () (d & D cd co co -j-ul-u-u V V Oto o ou scoO () () (o COso o o c O Y \, \, \ n issss soOO o O O

wSor (z or or RsV - H -- O vso to do >->|->l N \ \ y \, 31. Llorlt w V Soo do szzzz w v N D>>> glitt i ves CN V V - - - - -C-C-C-C w V

3.v Y. Yang SuV Lu Luu 8 (dud up to 8C citi to cost s V O O. O. O. N. O 2 Z 2 2. s>>>> S2 > 22 CQC a CO N. SSS S - - - - S>>>> of CO of C/D

V3ul LL Ll LL &w Co o ud (d SYYYY v O Su- l LL So o O O S. Do D (D g g g ty. No St o 5 yr SN 2 . s5 ye SN st () () () (9 () (D Co (9 cu) () () () O O. C. O C to O C to -e is is is a cle E is - - - as U U.S. Patent May 29, 2012 Sheet 8 of 69 US 8,188,231 B2

Figure 4

FoyRIIb

FC U.S. Patent May 29, 2012 Sheet 9 of 69 US 8,188,231 B2

Figure 5

2 3 4 5 6 7, 8 9 10

Figure 6

Stained gel Western U.S. Patent May 29, 2012 Sheet 10 of 69 US 8,188,231 B2

Figure 7

191kd 97kd

39kd

PNGSe F - - t - it - it - it - t Campath #15 F6 17 i8 i22 WT

Figure 8

Campath. 1:50 (Xencor). U.S. Patent May 29, 2012 Sheet 11 of 69 US 8,188,231 B2

Figure 9

Sinply Blue Stain Western Blot s:

1 2 3 4.5 67 8 1 2 3 4 56 78. U.S. Patent May 29, 2012 Sheet 12 of 69 US 8,188,231 B2

Figure 10

100 90 80 3i 70 60 k PBS 50 V T299A s {0 V263 40 WT 30 A V266 SS 20 O W240 {0 S239D 10 O 13 log antibody (2M) U.S. Patent May 29, 2012 Sheet 13 of 69 US 8,188,231 B2

Figure 11a

100

Gid 80 PBS WT A V264 E 60 O S239D 4 S239E A 332 D is 40 o S2390/332E

S. 20

O -13 log antibody (2M)

Figure 11b

100

U 80 O. CD . E 60: C - SX 40 2 SS 2 O

12 - 1 -10 9 8 7 6 log antibody) (M) U.S. Patent May 29, 2012 Sheet 14 of 69 US 8,188,231 B2

Figure 12

1 OO

80 PBS o Y296Q SO A Y296E WT O S239E332E 4 S239Q/332E

-13 -12 -11 -10 9 -8 -7 log antibody (2M)

U.S. Patent May 29, 2012 Sheet 16 of 69 US 8,188,231 B2

Figure 14

100 ess PBS 80. O S298AVE333AVK334A 0 S239D/332E A. V.264/332E 60 O S239D/332EIA330L A A330L

40 WT

logantibody) (1.3M) U.S. Patent May 29, 2012 Sheet 17 Of 69 US 8,188,231 B2

Figure 15

G) i

s k PBS E 60 S A 332E

O W264 s WT 5 40 KX W264/332E t s SS 2O

O -12 -11 -10 -9 -8 -7 log antibody (2M) U.S. Patent May 29, 2012 Sheet 18 of 69 US 8,188,231 B2

Figure 16a

6 0

4. O

2 O

-12 -11 1 O -9 -8 -7 -6 log antibody (2M)

Figure 16b

-12 -11 -10 -9 -8 -7 -6 log antibody (2M)

U.S. Patent May 29, 2012 Sheet 20 of 69 US 8,188,231 B2

Figure 17c

100

WT 80 A S239D 6 S239DIA330LII332E 60 W S239DIS298AVI332E 0 S239DfI332E

-13 -12 - 1 -lo -9 8 7 logantibody (2M) U.S. Patent May 29, 2012 Sheet 21 of 69 US 8,188,231 B2

Figure 18

100 iss- RAfs or k &\ 3 80 A 8 PBS W E 60 0 W264 O W264,332E s E 40 X t s SS 20 V

O esses tog antibody (M) U.S. Patent May 29, 2012 Sheet 22 of 69 US 8,188,231 B2

Figure 19

60. Q: W264/1332E O S2390/332E O S239DfI332EIA330L 40

Ot-r-re --- 13 -12 -11 -10 -9 -8 -7 log antibody (8M) U.S. Patent May 29, 2012 Sheet 23 of 69 US 8,188,231 B2

Figure 20a

100

80 PBS W W264 60 S298A S298AE333AVK334A 40 332E 332EIW264

log antibody (2M)

Figure 20b

1 OO

60 v S298A S298AVE333AK334A O 332E 40 {0 332EV264

-13 -12 -11 - O -9 -8 -7 log antibody (2M) U.S. Patent May 29, 2012 Sheet 24 of 69 US 8,188,231 B2

Figure 21a

SO S298A O.

40 O S298AE333AK334A 30-332Es

20 O W264,332E 10 y = 8.5x - 19.9 ra:0.90

O 1 2 3 4. 5 6 7 AlphaScreen IC50 (nM)

Figure 21b

800 y = 41.3x - 48.6 r=0.84 WT

600

400

200 O S298AWE333AK334A 332E W264,332E O 5 10 15 20 AlphaScreen IC50 (nM) U.S. Patent May 29, 2012 Sheet 25 of 69 US 8,188,231 B2

Figure 21c

O W264,332E

S298AWE333AK334A S298A W264

y = 0.24x + 0.82 r *=0.98

O 5 10 15 Fold improvement over WT by AlphaScreen

Figure 21d

15 y = 0.84x + 1.9 r?=0.90 O W264/332E

O 0 5 O 15 Fold improvement over WT by AlphaScreen U.S. Patent May 29, 2012 Sheet 26 of 69 US 8,188,231 B2

Figure 22a

100

: 80 0 P329F 60 D265N PBS A T299 E 40 A D26.5G s o W264F SS (> 328F 20 WT

O -13 log antibody (4M)

Figure 22b

100

8C 80 c

E 60 o ad s E 40 s s t SS 20

-13 -12 -11 -10 -9 -8 -7 log antibody (4M) U.S. Patent May 29, 2012 Sheet 27 of 69 US 8,188,231 B2

Figure 23a

2.0x104

19x104

3o 1.8x104. 3. c) ge 1.7x104 O B E. - 16x10 .

e E 1.5x104. s

1.4x104.

1.3x104. --- : PBMC WT W264 332E V264/332E U.S. Patent May 29, 2012 Sheet 28 of 69 US 8,188,231 B2

Figure 23b

100

40 WT Cd V264/332E

2O O S239D/332E

log antibody (nglml) U.S. Patent May 29, 2012 Sheet 29 of 69 US 8,188,231 B2

Figure 24a

BT474 Sk-Br-3 2.1 x 104

2.0x104

1.8x104.

1.7x104.

1.5x104.

1.4x104

PBMC WT W264 W264/ PBMC WT W264 W264/ 332E 332E U.S. Patent May 29, 2012 Sheet 30 of 69 US 8,188,231 B2

Figure 24b

100 80 C) a 60 (s E s 40 WT S A 332E 20 0 332EA33OY O 332EIA330L O -1 O 1 2 log antibody (ng/ml)

Figure 24c

100

o e . resse 80 2 Y ar W O L at 60 O s G E X in s° d S298AJE333AIK334A 20 O S239D/332E/A330L

O -4 -3 2 1 0 1 2 log antibody (ng/ml) U.S. Patent May 29, 2012 Sheet 31 of 69 US 8,188,231 B2

Figure 25a

PBMC U.S. Patent May 29, 2012 Sheet 32 of 69 US 8,188,231 B2

Figure 25b

100 8O 80 g 60 Eg 40 WT SS a {X S298AWE333AWK334A 20 O W264.332E

O log antibody (nglml)

Figure 25c

100.

5 O

25. o S239DII332EIA330L

4 3 2 1 0. 2 logantibody) (ng/ml) U.S. Patent May 29, 2012 Sheet 33 of 69 US 8,188,231 B2

Figure 26a

100

80 O S239D/332E 0 S239N/I332E A 1332EIV264 WT

-1 O 1. log antibody (ng/ml)

Figure 26b

O S239DI332EIA330L. 0 S239DfI332EIA330 WT

log (antibody) (nglml) U.S. Patent May 29, 2012 Sheet 34 of 69 US 8,188,231 B2

Figure 27

100

4. O S239DI332E 20 E272H

logantibody (M) U.S. Patent May 29, 2012 Sheet 35 of 69 US 8,188,231 B2

Figure 28

100 8O 80gn g or 60 WT s A K326E 40 0, S324G SS O H268D 20. o logantibody (ug/ml) U.S. Patent May 29, 2012 Sheet 36 of 69 US 8,188,231 B2

Figure 29a

SkOV3 MCP.7 SkBr3 OVCAR3

K= Her2/neu

Figure 29b

100 Ex S239D/I332E Og so : S O & 25 & E. & :E. Ek

SkBr3 SkOV3. OVCAR3 MCF-7 U.S. Patent May 29, 2012 Sheet 37 of 69 US 8,188,231 B2

Figure 30

100.

80:

l 60 s E. . WT 40 A S239D SS v 1332E O S239DII332E 20 o S239DI332EIA330L S298AJE333AVK334A.

O 3 2 1 0 2 3 4. log antibody) (ng/ml) U.S. Patent May 29, 2012 Sheet 38 of 69 US 8,188,231 B2

Figure 31

WT 30: S239DI332EIA330L

20.

15 10

O. 1 logantibody (ng/ml) U.S. Patent May 29, 2012 Sheet 39 of 69 US 8,188,231 B2

Figure 32a

1 OO

80 k PBS 0 332E

A S239E SO O W264 V 332EIW264

WT 40

log antibody (2M) U.S. Patent May 29, 2012 Sheet 40 of 69 US 8,188,231 B2

Figure 32b

60 A W264. to W264/332E

go S239E 40

-3 -2 -1 O 1 2 3 4. log antibody (ngiml)

Figure 32c

WT 60. (> S239D/332E 8 S239DA330/332E

2 0

logantibody (ng/ml) U.S. Patent May 29, 2012 Sheet 41 of 69 US 8,188,231 B2

Figure 33a food. --WT 1.8 ug/kg -o-, WT 33.6 ug/kg Y re-S239,332E 2, ug/kg i w

120

6 O

40 / Y --WT 1.8 uglkg 20 2. -O-WT 33.6 uglkg O -e-S239D/1332E 2.1 ug/kg

5 0. 15 20 25 3. Day

Figure 33c

OO ---.S239.0332E i

0 10 20 30 40 50 Dose (ugikg) U.S. Patent May 29, 2012 Sheet 42 of 69 US 8,188,231 B2

Figure 34a

100.

80 PBS A330L S298A/E333AIK334A 60. WT A V264/1332E S239D/332E 40 S239D/332EIA330L

11 -lo -9 -8 7 -6 log antibody) (1.3M) U.S. Patent May 29, 2012 Sheet 43 of 69 US 8,188,231 B2

Figure 34b.

100

80 wT S. a v2641/1332E e V264/1332EIA330L o S239D/332ElA330L 40

log antibody (6M)

Figure 34c

100

80. PBS. 3 W 3 A S239D/A330/332E 60 0 S239DIK326T IA330/332E is 40 o S239DK326E E A330.332E s S.

log antibody (M) U.S. Patent May 29, 2012 Sheet 44 of 69 US 8,188,231 B2

Figure 35

S239D? PBMCs TX 332El A330. U.S. Patent May 29, 2012 Sheet 45 of 69 US 8,188,231 B2

Figure 36

80 k PBS WT 293 6 O O WT CHO 0 239D332E1330L 293 KX 239D 332E330L CHO 4. O

log antibody (2M) U.S. Patent May 29, 2012 Sheet 46 of 69 US 8,188,231 B2

Figure 37a

100 PBS 80. WT,293T WT, CHO V209, 293T 60. V209, CHO WT, Lec-13 40 V209, lec-13

20.

-14 -13 -12 -11 -10 -9 8 7 logantibody (M)

Figure 37b 100

WT 293T 0 WT CHO O WT Lec3 D V209:293T. (). V209 CHO O. V209 Lec13.

-2 -1 0. 1 2 log antibody (nglml) U.S. Patent May 29, 2012 Sheet 47 of 69 US 8,188,231 B2

Figure 38

100

k PBS 60 Ko N297E/D265F/332E v 2970265Y1299/.332E A N297D/D265Y/332E 40 0 N297E/332E O N29707332E W

13 -12 -11 -10 9 -8 -7 log antibody (2.3M)

U.S. Patent May 29, 2012 Sheet 50 of 69 US 8,188,231 B2

Figure 41

0.17.0.07 100. U.S. Patent May 29, 2012 Sheet 51 of 69 US 8,188,231 B2

FcRn Substitution(s) onf fold conf Fold confilaib

0.2g. 134031

108.04 083 U.S. Patent May 29, 2012 Sheet 52 of 69 US 8,188,231 B2

Figure 41 (continued) Rila C1q FcRn. Confold: Confold: Co L328O1332E ------93 9 -

383 08023.1 1098 0.28 255

to 1970 44 1971 16. 1413 U.S. Patent May 29, 2012 Sheet 53 of 69 US 8,188,231 B2

Figure 41 (continued)

conflillaith 195

24Y.243YrW262Tf 25 0.11 0.23

1450 13041 U.S. Patent May 29, 2012 Sheet 54 of 69 US 8,188,231 B2

Figure 41 (continued) FcRita IFCyrib Feyrie Ficy Rila I Substitution(s) con onffold confroid conffold conf 41 3661

843369.17.20:023.of 134 - 0.150.-- 2.06 -

1423.366 1976. 1055 236 0.39.099 U.S. Patent May 29, 2012 Sheet 55 of 69 US 8,188,231 B2

variant cor ld conf Foid confilalib

175 332T ------0.22 102

0.18 U.S. Patent May 29, 2012 Sheet 56 of 69 US 8,188,231 B2

Figure 41 (continued) Substitution(s) I context Fold conf onf Fold conffold conf illaib 225 S267t 0.93 0.140.15 0.32 1.06 1.261.01

246 N278S - 124 25.25

. . . . ; 250 N27BY 251 Y278T

w323K322H - - - 0.90 0.130.- s324 to 20i o 28 og 115 008115 ogs 237 0.36 2.15 103 0.23 1 : 0.19, 160

O56 0.172.06 U.S. Patent May 29, 2012 Sheet 57 of 69 US 8,188,231 B2

Substitution(s) 0.34780 : 0.30 188 15 009 194 007 2.25 0.360900 10. 0.98, 103 76.44 106 Q. 19. 166 Q. 19.09

33 1.19 046 0.73. 1.78 Oog 2.10 0071. ...O.30C

0.69 0.86 0.81.032,220- - - - 0.96 O.19 0.89 0.063.68- 156 0.37 0.60 0.12 3.82 004: G 15 0.25 114034,052,025,057. 9. 3 0.25 0.50 U.S. Patent May 29, 2012 Sheet 58 Of 69 US 8,188,231 B2

41 (continued) Variant Substitutions Conffold S232641A330Y,332E 2. 0.26 48.74 0.30 2756; 0.23.298 ; 641 0.2241.48 22.590.22.185

O 45 O2

0.34 0.68 0.11 0.76 0.29 0.78 0.14 (.51

0.22 192 0.136.96

0.12 136 ...... - . . m . 0.160 08 0.33 009. 336 F241E d 0.15, 190 0.11 1.18 m 0.28 033 1320-1 152 0.63

0.18337. 97.5 0.20 335 (21203

; 6.12 0.22 0.24 002. 023 Q44 0.26

O O O O 56014 U.S. Patent May 29, 2012 Sheet 59 of 69 US 8,188,231 B2

Figure 41 (continued) Fcyria FC onf Fold Conf. Fold Conffold Confallb O48 0.39

0.09 0.49

0.15 101 o. O90.

1.27 U.S. Patent May 29, 2012 Sheet 60 of 69 US 8,188,231 B2

Feyrilla

65,058 0.261,055

o37,035 0.13 0.16 0.09.039

O 0.59 0.46 1.21 0.51 129 0.54. 139

190 O19090 0.34.09 0.32082 U.S. Patent May 29, 2012 Sheet 61 of 69 US 8,188,231 B2

Figure 41 (continued) FcRia" FcRib FcRic Feyria 0.13088 0.210210.5 O410.12. d.15031 0.150.23

0.19074 "C3

0.14132 018 0.76 0.12 24 0.3 87

0.23 0.08 20. 0.66 0300.50 (36 0.29,071 8

0.14 150 U.S. Patent May 29, 2012 Sheet 62 of 69 US 8,188,231 B2

Figure 41 (continued) FCyrila II onf Fold Conf Fold 75

0.27 157 0.14 - 0.98

O 14042 0.13 0.48

- ... -- : ... -- 0.79 0.46 1. 22 O52; O42 (27 0.62 0.39 1 U.S. Patent May 29, 2012 Sheet 63 of 69 US 8,188,231 B2

"F.Rite "FRIlla onf Fold conf Fold 13 O42 O1909 0.21 102

0.39 0.26

0.140.71 O. 0.15. 104

0.25

0.12 2.92 : 0.09 Olaa O2 221

O 0.18 m 0.650.15. O30, 1830.3.2.45.13083.0.18161.

K32Ow

093043,094. C11ogo. 99 0.3 114 0.36 186 U.S. Patent May 29, 2012 Sheet 64 of 69 US 8,188,231 B2

Wariant Substitutions) 692 K322

9.309.34933

G236N G2360

0.97.007 0.45 0.08 500. O36 G2a

o680.0 U.S. Patent May 29, 2012 Sheet 65 of 69 US 8,188,231 B2

Figure 41 (continued)

FcyRIlla ---, -,------r ------w - FcRitc.nf Fold conf Fold conf Fold 0.36 U.S. Patent May 29, 2012 Sheet 66 of 69 US 8,188,231 B2

Figure 41 (continued) IFcyRila III FcRib Ferlic FcRita Ciq FcRn Substitutions) a mu-mund confifold conf Fo Conf fold conf fold conf foid confilamb 2 2

138949. 1.76 0.30 1250.24

0.18 0.50

446 : 4.230.34 450.2204.

228 to 6 0.46 0.14 0.30 0.17

4945 1895.25

0.64 0.78086

0.58, 169 0.28. 166 0.29 U.S. Patent May 29, 2012 Sheet 67 of 69 US 8,188,231 B2

Figure 41 (continued)

Cornfla: 0.83 0.31 370.89 0.350.82

124 O51 0.94 044

Of OS4020

O33052

0.9134

12 1633 13 O 0.215.20 U.S. Patent May 29, 2012 Sheet 68 of 69 US 8,188,231 B2

U.S. Patent May 29, 2012 Sheet 69 of 69 US 8,188,231 B2

Figure 42 SEQ ID NO: 8 Xaa Xala Xala Xaa Xaa Cys Xaa Xala Cys Xaa Xala Xala Xala Xala Xala Xala 221 225 23 O 235

Xaa Xala Xala Xaa Xala Lieu Xaa Xala Xala Xala Xala Lys Xaa Thir Lieu Met 24 O 245 25 O

Ile Ser Xaa Thr Pro Xaa Val Xala Cys Xaa Xala Xala Xala Xala Xaa Xaa 255 26 O 265

Xaa Xala Xala Xala Xaa Xala Xala Xala Trp Xaa Val Xaa Xala Xala Xala Xala 27 O 275 28O

Xaa Xaa Ala Xala Thr Xaa Xala Xala Xala Xaa Xala Xaa Xala Xala Xaa Xaa 285 29 O 295 3 OO

Xaa Xala Xaa Xaa Xala Lieu. Thr Val Lieu. His Glin Asp Xaa Lieu Asn Gly 305 31 O 315

Xaa Xala Tyr Xaa Cys Xaa Xala Xala Xala Xala Xala Xala Xaa Xala Xala Xaa 32O 325 33 O

Xaa Xala Xala Xala Xala 335 US 8,188,231 B2 1. 2 OPTIMIZED FC VAIRLANTS sequence variability does occur in the FR region between different antibodies. Overall, this characteristic architecture This application claims benefit under 35 U.S.C. S 119(e) to of antibodies provides a stable scaffold (the FR region) upon U.S. Provisional Application Nos. 60/589,906, filed Jul. 20, which substantial antigen binding diversity (the CDRs) can 2004; 60/627,026, filed Nov. 9, 2004; 60/627,774, filed Nov. be explored by the immune system to obtain specificity for a 12, 2004; 60/626.991, filed Nov. 10, 2004; and this applica broad array of antigens. A number of high-resolution struc tion is also a continuation-in-part of Ser. No. 10/822.231, tures are available for a variety of variable region fragments filed Mar. 26, 2004, (now issued as U.S. Pat. No. 7,317,091), from different organisms, some unbound and some in com which is a continuation-in-part of Ser. No. 10/672.280, filed plex with antigen. The sequence and structural features of Sep. 26, 2003 (now abandoned), which claims the benefit 10 antibody variable regions are well characterized (Morea et al., under 35 U.S.C. S 119(e) to U.S. Provisional Application Nos. 1997, Biophys Chem 68:9-16; Morea et al., 2000, Methods 60/477,839, filed Jun. 12, 2003, and 60/467,606, filed May 2, 20:267-279, incorporated by reference), and the conserved 2003, and 60/442,301, filed Jan. 23, 2003, and 60/414,433, features of antibodies have enabled the development of a filed Sep. 27, 2002. wealth of antibody engineering techniques (Maynard et al., 15 2000, Annu Rev. Biomed Eng 2:339-376, incorporated by FIELD OF THE INVENTION reference). For example, it is possible to graft the CDRs from one antibody, for example a murine antibody, onto the frame The present invention relates to novel optimized Fc vari work region of another antibody, for example a human anti ants, engineering methods for their generation, and their body. This process, referred to in the art as “humanization', application, particularly for therapeutic purposes. enables generation of less immunogenic antibody therapeu tics from nonhuman antibodies. Fragments comprising the BACKGROUND OF THE INVENTION variable region can exist in the absence of other regions of the antibody, including for example the antigenbinding fragment Antibodies are immunological proteins that bind a specific (Fab) comprising V-CY1 and V-C, the variable fragment antigen. In most mammals, including humans and mice, anti 25 (FV) comprising V and V, the single chain variable frag bodies are constructed from paired heavy and light polypep ment (ScPV) comprising V and V, linked together in the tide chains. Each chain is made up of individual immunoglo same chain, as well as a variety of other variable region bulin (Ig) domains, and thus the generic term fragments (Little et al., 2000, Immunol Today 21:364-370, immunoglobulin is used for Such proteins. Each chain is made incorporated by reference). up of two distinct regions, referred to as the variable and 30 The Fc region of an antibody interacts with a number of Fc constant regions. The light and heavy chain variable regions receptors and ligands, imparting an array of important func show significant sequence diversity between antibodies, and tional capabilities referred to as effector functions. For IgG are responsible for binding the target antigen. The constant the Fc region, as shown in FIG. 1, comprises Ig domains Cy2 regions show less sequence diversity, and are responsible for and Cy3 and the N-terminal hinge leading into Cy2. An impor binding a number of natural proteins to elicit important bio 35 tant family of Fc receptors for the IgG class are the Fc gamma chemical events. In humans there are five different classes of receptors (FcyRs). These receptors mediate communication antibodies including IgA (which includes Subclasses IgA1 between antibodies and the cellular arm of the immune sys and IgA2), Ig), IgE, IgG (which includes Subclasses IgG1. tem (Raghavanet al., 1996, Annu Rev Cell Dev Biol 12:181 IgG2, IgG3, and IgG4), and IgM. The distinguishing features 220; Ravetch et al., 2001, Annu Rev Immunol 19:275-290). between these antibody classes are their constant regions, 40 In humans this protein family includes FcyRI (CD64), includ although subtler differences may exist in the variable region. ing isoforms FcyRIa, FcyRIb, and FcyRIc; FcyRII (CD32), FIG. 1 shows an IgG1 antibody, used here as an example to including isoforms FcyRIIa (including allotypes H131 and describe the general structural features of immunoglobulins. R131), FcyRIIb (including FcyRIIb-1 and FcyRIIb-2), and IgG antibodies are tetrameric proteins composed of two FcyRIIc; and FcyRIII (CD16), including isoforms FcyRIIIa heavy chains and two light chains. The IgG heavy chain is 45 (including allotypes V158 and F158) and FcyRIIIb (including composed of four immunoglobulin domains linked from N allotypes FcyRIIIb-NA1 and FcyRIIIb-NA2) (Jefferis et al., to C-terminus in the order V-C1 C2-C3, referring to the 2002, Immunol Lett 82:57-65, incorporated by reference). variable heavy domain, constant heavy domain 1, constant These receptors typically have an extracellular domain that heavy domain 2, and constant heavy domain 3. The IgG C-1, mediates binding to Fc., a membrane spanning region, and an C2, and C3 domains are also referred to as constant gamma 50 intracellular domain that may mediate some signaling event 1 domain (CY1), constant gamma 2 domain (Cy2), and con within the cell. These receptors are expressed in a variety of stant gamma 3 domain (Cy3) respectively. The IgG light chain immune cells including monocytes, macrophages, neutro is composed of two immunoglobulin domains linked from N phils, dendritic cells, eosinophils, mast cells, platelets, B to C-terminus in the order V-C, referring to the light chain cells, large granular lymphocytes, Langerhans cells, natural variable domain and the light chain constant domain respec 55 killer (NK) cells, and yö T cells. Formation of the Fc/FcyR tively. complex recruits these effector cells to sites of bound antigen, The variable region of an antibody contains the antigen typically resulting in signaling events within the cells and binding determinants of the molecule, and thus determines important Subsequent immune responses Such as release of the specificity of an antibody for its target antigen. The Vari inflammation mediators, B cell activation, endocytosis, able region is so named because it is the most distinct in 60 phagocytosis, and cytotoxic attack. The ability to mediate sequence from other antibodies within the same class. The cytotoxic and phagocytic effector functions is a potential majority of sequence variability occurs in the complementa mechanism by which antibodies destroy targeted cells. The rity determining regions (CDRs). There are 6 CDRs total, cell-mediated reaction wherein nonspecific cytotoxic cells three each per heavy and light chain, designated V. CDR1. that express FcyRS recognize bound antibody on a target cell V, CDR2, V, CDR3, V, CDR1, V, CDR2, and V, CDR3. 65 and Subsequently cause lysis of the target cell is referred to as The variable region outside of the CDRs is referred to as the antibody dependent cell-mediated cytotoxicity (ADCC) framework (FR) region. Although not as diverse as the CDRs, (Raghavanet al., 1996, Annu Rev Cell Dev Biol 12:181-220; US 8,188,231 B2 3 4 Ghetie et al., 2000, Annu Rev Immunol 18:739-766: Ravetch binding mediates complement dependent cytotoxicity et al., 2001, Annu Rev Immunol 19:275-290, incorporated by (CDC). C1q forms a complex with the serine proteases C1r reference). The cell-mediated reaction wherein nonspecific and C1s to form the C1 complex. C1q is capable of binding cytotoxic cells that express FcyRs recognize bound antibody six antibodies, although binding to two IgGs is sufficient to on a target cell and Subsequently cause phagocytosis of the 5 activate the complement cascade. Similar to Fc interaction target cell is referred to as antibody dependent cell-mediated with FcyRs, different IgG subclasses have different affinity phagocytosis (ADCP). A number of structures have been for C1q, with IgG1 and IgG3 typically binding substantially solved of the extracellular domains of human FcyRs, includ better to the FcyRs than IgG2 and IgG4. There is currently no ing FcyRIIa (pdb accession code 1 H9V) (Sondermann et al., structure available for the Fc/C1q complex; however, 2001, J Mol Biol 309:737-749) (pdb accession code 1 FCG) 10 mutagenesis studies have mapped the binding site on human (Maxwell et al., 1999, Nat Struct Biol 6:437-442), FcyRIIb IgG for C1q to a region involving residues D270, K322, (pdb accession code 2FCB) (Sondermann et al., 1999, Embo K326, P329, and P331, and E333 (Idusogie et al., 2000, J J 18:1095-1103); and FcyRIIIb (pdb accession code 1 E4J) Immunol 164:41 78-4184: Idusogie et al., 2001, J Immunol (Sondermann et al., 2000, Nature 406:267-273, incorporated 166:2571-2575, incorporated by reference). by reference). All FcyRs bind the same region on Fc, at the 15 A site on Fc between the Cy2 and Cy3 domains, shown in N-terminal end of the Cy2 domain and the preceding hinge, FIG. 1, mediates interaction with the neonatal receptor FcRn, shown in FIG. 2. This interaction is well characterized struc the binding of which recycles endocytosed antibody from the turally (Sondermann et al., 2001, J Mol Biol 309:737-749 endosome back to the bloodstream (Raghavan et al., 1996, incorporated by reference), and several structures of the Annu Rev Cell Dev Biol 12:181-220; Ghetie et al., 2000, Annu human Fc bound to the extracellular domain of human Rev Immunol 18:739-766, incorporated by reference). This FcyRIIIb have been solved (pdb accession code 1 E4K)(Son process, coupled with preclusion of kidney filtration due to dermann et al., 2000, Nature 406:267-273) (pdb accession the large size of the full length molecule, results in favorable codes 1 IIS and 1 IIX) (Radaev et al., 2001, J Biol Chem antibody serum half-lives ranging from one to three weeks. 276:16469-16477, incorporated by reference), as well as has Binding of Fc to FcRn also plays a key role in antibody the structure of the human IgE Fc/Fce RIC. complex (pdb 25 transport. The binding site for FcRn on Fc is also the site at accession code 1F6A) (Garman et al., 2000, Nature 406:259 which the bacterial proteins A and G bind. The tight binding 266, incorporated by reference). by these proteins is typically exploited as a means to purify The different IgG subclasses have different affinities for antibodies by employing protein A or protein Gaffinity chro the FcyRs, with IgG1 and IgG3 typically binding substan matography during protein purification. Thus the fidelity of tially better to the receptors than IgG2 and IgG4. All FcyRs 30 this region on Fc is important for both the clinical properties bind the same region on IgGFc, yet with different affinities: of antibodies and their purification. Available structures of the the high affinity binder FcyRI has a Kd for IgG1 of 10 M', rat Fc/FcRn complex (Martin et al., 2001, Mol Cell 7:867 whereas the low affinity receptors FcyRII and FcyRIII gener 877, incorporated by reference), and of the complexes of Fc ally bind at 10 and 10 respectively. The extracellular with proteins A and G (Deisenhofer, 1981, Biochemistry domains of FcyRIIIa and FcyRIIIb are 96% identical, how 35 20:2361-2370; Sauer-Eriksson et al., 1995, Structure 3:265 ever FcyRIIIb does not have a intracellular signaling domain. 278; Tashiro et al., 1995, Curr Opin Struct Biol 5:471-481, Furthermore, whereas FcyRI. FcyRIIa/c, and FcyRIIIa are incorporated by reference) provide insight into the interaction positive regulators of immune complex-triggered activation, of Fc with these proteins. characterized by having an intracellular domain that has an A key feature of the Fc region is the conserved N-linked immunoreceptor tyrosine-based activation motif (ITAM), 40 glycosylation that occurs at N297, shown in FIG. 1. This FcyRIIb has an immunoreceptor tyrosine-based inhibition carbohydrate, or oligosaccharide as it is sometimes referred, motif (ITIM) and is therefore inhibitory. Thus the former are plays a critical structural and functional role for the antibody, referred to as activation receptors, and FcyRIIb is referred to and is one of the principle reasons that antibodies must be as an inhibitory receptor. The receptors also differ in expres produced using mammalian expression systems. While not sion pattern and levels on different immune cells. Yet another 45 wanting to be limited to one theory, it is believed that the level of complexity is the existence of a number of FcyR structural purpose of this carbohydrate may be to stabilize- or polymorphisms in the human proteome. A particularly rel solubilize Fc, determine a specific angle or level of flexibility evant polymorphism with clinical significance is V158/F 158 between the Cy3 and Cy2 domains, keep the two Cy2 domains FcyRIIIa. Human IgG1 binds with greater affinity to the V158 from aggregating with one another across the central axis, or allotype than to the F158 allotype. This difference in affinity, 50 a combination of these. Efficient Fc binding to FcyR and C1q. and presumably its effect on ADCC and/or ADCP, has been requires this modification, and alterations in the composition shown to be a significant determinant of the efficacy of the of the N297 carbohydrate or its elimination affect binding to anti-CD20 antibody rituximab (Rituxan R, a registered trade these proteins (Umafia et al., 1999, Nat Biotechnol 17: 176 mark of IDEC Pharmaceuticals Corporation). Patients with 180; Davies et al., 2001, Biotechnol Bioeng 74:288-294; the V158 allotype respond favorably to rituximab treatment; 55 Mimura et al., 2001, J Biol Chem 276:45539-45547. Radaev however, patients with the lower affinity F158 allotype et al., 2001, J Biol Chem 276:16478-16483; Shields et al., respond poorly (Cartron et al., 2002, Blood 99:754-758, 2001, J Biol Chem 276:6591-6604; Shields et al., 2002, J Biol incorporated by reference). Approximately 10-20% of Chem 277:26733-26740; Simmons et al., 2002, J Immunol humans are V158/V158 homozygous, 45% are V158/F158 Methods 263:133-147, incorporated by reference). Yet the heterozygous, and 35-45% of humans are F 158/F 158 60 carbohydrate makes little if any specific contact with FcyRs homozygous (Lehrnbecher et al., 1999, Blood 94:4220-4232; (Radaev et al., 2001, J Biol Chem 276:16469-16477, incor Cartronet al., 2002, Blood 99:754-758, incorporated by ref porated by reference), indicating that the functional role of erence). Thus 80-90% of humans are poor responders, that is the N297 carbohydrate in mediating Fc/FcyRbinding may be they have at least one allele of the F158 FcyRIIIa. via the structural role it plays in determining the Fc confor An overlapping but separate site on Fc, shown in FIG. 1, 65 mation. This is Supported by a collection of crystal structures serves as the interface for the complement protein C1q. In the of four different Fc glycoforms, which show that the compo same way that Fc/FcyR binding mediates ADCC, Fc/C1q. sition of the oligosaccharide impacts the conformation of Cy2 US 8,188,231 B2 5 6 and as a result the Fc/FcyR interface (Krapp et al., 2003, J Mol type of high (V158) or low (F158) affinity polymorphic forms Biol 325:979-989, incorporated by reference). of FcyRIIIa (Cartron et al., 2002, Blood 99:754-758, incor The features of antibodies discussed above specificity porated by reference). for target, ability to mediate immune effector mechanisms, Mutagenesis studies have been carried out on Fc towards and long half-life in serum make antibodies powerful thera various goals, with Substitutions typically made to alanine peutics. Monoclonal antibodies are used therapeutically for (referred to as alanine Scanning) or guided by sequence the treatment of a variety of conditions including cancer, homology substitutions (Duncan et al., 1988, Nature 332: inflammation, and cardiovascular disease. There are currently 563-564; Lund et al., 1991, J. Immunol 147:2657-2662: Lund over ten antibody products on the market and hundreds in et al., 1992, Mol Immunol 29:53-59; Jefferis et al., 1995, development. In addition to antibodies, an antibody-like pro 10 Immunol Lett 44:111-117; Lund et al., 1995, FasebJ 9:115 tein that is finding an expanding role in research and therapy 119; Jefferis et al., 1996, Immunol Lett 54:101-104; Lund et is the Fc fusion (Chamow et al., 1996, Trends Biotechnol al., 1996, J Immunol 157:4963-4969; Armour et al., 1999, 14:52-60; Ashkenazi et al., 1997, Curr Opin Immunol 9:195 Eur J Immunol 29:2613-2624; Shields et al., 2001, J Biol 200, incorporated by reference). An Fc fusion is a protein Chem 276:6591-6604) (U.S. Pat. No. 5,624,821; U.S. Pat. wherein one or more polypeptides is operably linked to Fc. 15 No. 5,885,573; PCT WO 00/42072; PCT WO99/58572), all An Fc fusion combines the Fc region of an antibody, and thus incorporated by reference. The majority of substitutions its favorable effector functions and pharmacokinetics, with reduce or ablate binding with FcyRs. However some success the target-binding region of a receptor, ligand, or some other has been achieved at obtaining Fc variants with higher FcyR protein or protein domain. The role of the latter is to mediate affinity. (See for example U.S. Pat. No. 5,624,821 and PCT target recognition, and thus it is functionally analogous to the WO 00/42072). For example, Winter and colleagues substi antibody variable region. Because of the structural and func tuted the human amino acid at position 235 of mouse IgG2b tional overlap of Fc fusions with antibodies, the discussion on antibody (a glutamic acid to leucine mutation) that increased antibodies in the present invention extends directly to Fc binding of the mouse antibody to human FcyRI by 100-fold fusions. (Duncan et al., 1988, Nature 332:563-564) (U.S. Pat. No. There are a number of possible mechanisms by which 25 5,624,821). Shields et al. used alanine Scanning mutagenesis antibodies destroy tumor cells, including anti-proliferation to map Fc residues important to FcyR binding, followed by via blockage of needed growth pathways, intracellular sig Substitution of select residues with non-alanine mutations naling leading to apoptosis, enhanced down regulation and/or (Shields et al., 2001, J Biol Chem 276:6591-6604: Presta et turnover of receptors, CDC, ADCC, ADCP, and promotion of al., 2002, Biochem Soc Trans 30:487-490) (PCT WO an adaptive immune response (Cragg et al., 1999, Curr Opin 30 00/42072), incorporated by reference. Immunol 11:541-547: Glennie et al., 2000, Immunol Today Enhanced affinity of Fc for FcyR has also been achieved 21:403-410. incorporated by reference). Anti-tumor efficacy using engineered glycoforms generated by expression of anti may be due to a combination of these mechanisms, and their bodies in engineered or variant cell lines (Umafia et al., 1999, relative importance in clinical therapy appears to be cancer Nat Biotechnol 17:176-180; Davies et al., 2001, Biotechnol dependent. Despite this arsenal of anti-tumor weapons, the 35 Bioeng 74:288-294; Shields et al., 2002, J Biol Chem 277: potency of antibodies as anti-cancer agents is unsatisfactory, 26733-26740; Shinkawa et al., 2003, J Biol Chem 278:3466 particularly given their high cost. Patient tumor response data 3473, incorporated by reference). This approach has gener show that monoclonal antibodies provide only a small ated enhancement of the capacity of antibodies to bind improvement in therapeutic Success over normal single-agent FcyRIIIa and to mediate ADCC. cytotoxic chemotherapeutics. For example, just half of all 40 Another major shortcoming of antibodies is their demand relapsed low-grade non-Hodgkin’s lymphoma patients ing production requirements (Garber, 2001, Nat Biotechnol respond to the anti-CD20 antibody rituximab (McLaughlin et 19:184-185; Dove, 2002, Nat Biotechnol 20:777-779, incor al., 1998, J. Clin Oncol 16:2825-2833, incorporated by refer porated by reference). Antibodies must be expressed in mam ence). Of 166 clinical patients, 6% showed a complete malian cells, and the currently marketed antibodies together response and 42% showed a partial response, with median 45 with other high-demand biotherapeutics consume essentially response duration of approximately 12 months. Trastuzumab all of the available manufacturing capacity. With hundreds of (Herceptin R, a registered trademark of Genentech), an anti biologics in development, the majority of which are antibod HER2/neu antibody for treatment of metastatic breast cancer, ies, there is an urgent need for more efficient and cheaper has less efficacy. The overall response rate using trastuzumab methods of production. The downstream effects of insuffi for the 222 patients tested was only 15%, with 8 complete and 50 cient antibody manufacturing capacity are three-fold. First, it 26 partial responses and a median response duration and dramatically raises the cost of goods to the producer, a cost survival of 9 to 13 months (Cobleigh et al., 1999, JClin Oncol that is passed on to the patient. Second, it hinders industrial 17:2639-2648, incorporated by reference). Currently for anti production of approved antibody products, limiting availabil cancer therapy, any Small improvement in mortality rate ity of high demand therapeutics to patients. Finally, because defines Success. Thus there is a significant need to enhance the 55 clinical trials require large amounts of a protein that is not yet capacity of antibodies to destroy targeted cancer cells. profitable, the insufficient Supply impedes progress of the The role of FcyR-mediated effector functions in the anti growing antibody pipeline to market. cancer activity of antibodies has been demonstrated in mice Alternative production methods have been explored in (Clynes et al., 1998, Proc Natl AcadSci USA 95:652-656: attempts at alleviating this problem. Transgenic plants and Clynes et al., 2000, Nat Med 6:443-446, incorporated by 60 animals are being pursued as potentially cheaper and higher reference), and the affinity of interaction between Fc and capacity production systems (Chadd et al., 2001, Curr Opin certain FcyRs correlates with targeted cytotoxicity in cell Biotechnol 12:188-194, incorporated by reference). Such based assays (Shields et al., 2001, J Biol Chem 276:6591 expression systems, however, can generate glycosylation pat 6604; Presta et al., 2002, Biochem Soc Trans 30:487-490; terns significantly different from human glycoproteins. This Shields et al., 2002, J Biol Chem 277:26733-26740, incorpo 65 may result in reduced or even lack of effector function rated by reference). Additionally, a correlation has been because, as discussed above, the carbohydrate structure can observed between clinical efficacy in humans and their allo significantly impact FcyRand complement binding. A poten US 8,188,231 B2 7 8 tially greater problem with nonhuman glycoforms may be immunogenicity; carbohydrates are a key source of antige nicity for the immune system, and the presence of nonhuman glycoforms has a significant chance of eliciting antibodies that neutralize the therapeutic, or worse cause adverse 5 immune reactions. Thus the efficacy and safety of antibodies produced by transgenic plants and animals remains uncertain. Bacterial expression is another attractive solution to the anti body production problem. Expression in bacteria, for example E. coli, provides a cost-effective and high capacity 10 method for producing proteins. For complex proteins such as antibodies there are a number of obstacles to bacterial expres Sion, including folding and assembly of these complex mol ecules, proper disulfide formation, and solubility, stability, and functionality in the absence of glycosylation because 15 proteins expressed in bacteria are not glycosylated. Full length unglycosylated antibodies that bind antigen have been successfully expressed in E. coli (Simmons et al., 2002, J Immunol Methods 263:133-147, incorporated by reference), and thus, folding, assembly, and proper disulfide formation of 20 bacterially expressed antibodies are possible in the absence of the eukaryotic chaperone machinery. However the ultimate utility of bacterially expressed antibodies as therapeutics remains hindered by the lack of glycosylation, which results in lack effector function and may result in poor stability and 25 solubility. This will likely be more problematic for formula tion at the high concentrations for the prolonged periods demanded by clinical use. In Summary, there is a need for antibodies with enhanced therapeutic properties. 30 SUMMARY OF THE INVENTION The present invention provides Fc variants that are opti mized for a number of therapeutically relevant properties. 35 These Fc variants are generally contained within a variant protein, that preferably comprises an antibody or a Fc fusion protein. It is an object of the present invention to provide novel Fc positions at which amino acid modifications may be made to 40 generate optimized Fc variants. Said Fc positions include 230, 240, 244, 245, 247, 262, 263,266, 273, 275,299, 302, 313, 323, 325, 328, and 332, wherein the numbering of the residues in the Fc region is that of the EU index as in Kabat. The present invention describes any amino acid modification 45 at any of said novel Fc positions in order to generate an optimized Fc variant. It is a further object of the present invention to provide Fc variants that have been characterized herein. In one embodi ment, said Fc variants comprise at least one amino acid Sub- 50 stitution at a position selected from the group consisting of 221, 222, 223, 224, 225, 227, 228, 230, 231, 232, 233,234, 235, 236, 237,238, 239, 240, 241, 243, 244, 245, 246, 247, 249, 255, 258, 260, 262, 263,264, 265, 266, 267, 268, 269, 270, 271, 272,273, 274, 275,276, 278, 280, 281, 282,283, 55 284, 285, 286, 288, 290, 291, 292, 293, 294, 295, 296, 297, 298, 299, 300, 301, 302,303,304,305,313,317, 318,320, 322,323,324, 325, 326, 327,328,329, 330, 331, 332, 333, 334, 335, 336, and 337, wherein the numbering of the resi dues in the Fc region is that of the EU index as in Kabat. In a 60 preferred embodiment, said Fc variants comprise at least one substitution selected from the group consisting of D221K. D221Y, K222E, K222Y, T223E, T223K, H224E, H224Y, T225E, T225K, T225W, P227E, P227G, P227K, P227Y, P228E, P228G, P228K, P228Y, P230A, P230E, P230G, 65 P230Y, A231E, A231G, A231 K, A231P, A231Y, P232E, P232G, P232K, P232Y, E233A, E233D, E233F, E233G,

US 8,188,231 B2 13 14 It is a further object of the present invention to provide Fc site in the variable region, and the binding sites for FcyRs, variants that mediate effector function more effectively in the FcRn, C1q, and proteins A and G in the Fc region. presence of effector cells. In one embodiment, said Fc vari FIG. 2. The Fc/FcyRIIIb complex structure 1 IIS. Fc is ants mediate ADCC that is greater than that mediated by the shown as a gray ribbon diagram, and FcyRIIb is shown as a parent Fc polypeptide. In a preferred embodiment, said Fc black ribbon. The N297 carbohydrate is shown as black variants mediate ADCC that is more than 5-fold greater than Sticks. that mediated by the parent Fc polypeptide. In a mostly pre FIGS. 3a-3b. Alignment of the amino acid sequences of the ferred embodiment, said Fc variants mediate ADCC that is human IgG immunoglobulins IgG1, IgG2, IgG3, and IgG4. between 5-fold and 1000-fold greater than that mediated by FIG. 3a provides the sequences of the CH1 (CY1) and hinge the parent Fc polypeptide. 10 domains, and FIG. 3b provides the sequences of the CH2 (Cy2) and CH3 (Cy3) domains. Positions are numbered It is a further object of the present invention to provide Fc according to the EU index of the IgG1 sequence, and differ variants that bind with weaker affinity to one or more FcyRs. ences between IgG1 and the other immunoglobulins IgG2. It is a further object of the present invention to provide Fc IgG3, and IgG4 are shown in grey. Polymorphisms exist at a variants that mediate ADCC in the presence of effector cells 15 number of positions (Kim et al., 2001, J. Mol. Evol. 54:1-9), less effectively. and thus slight differences between the presented sequences It is a further object of the present invention to provide Fc and sequences in the prior art may exist. The possible begin variants that have improved function and/or solution proper nings of the Fc region are labeled, defined hereinas either EU ties as compared to the aglycosylated form of the parent Fc position 226 or 230. polypeptide. Improved functionality herein includes but is FIG. 4. Residues at which amino acid modifications were not limited to binding affinity to an Fc ligand. Improved made in the Fc variants of the present invention, mapped onto solution properties herein includes but is not limited to sta the Fc/FcyRIIIb complex structure 1IIS. Fe is shown as a gray bility and solubility. In an one embodiment, said Fc variants ribbon diagram, and FcyRIIIb is shown as a black ribbon. bind to an FcyR with an affinity that is within about 0.5-fold Experimental library residues are shown in black, the N297 of the glycosylated form of the parent Fc polypeptide. In an 25 carbohydrate is shown in grey. alternate embodiment, said aglycosylated Fc variants bind to FIG. 5. Expression of Fc variant and wild type (WT) pro an FcyR with an affinity that is comparable to the glycosylated teins of alemtuzumab in 293T cells. Plasmids containing parent Fc polypeptide. In an alternate embodiment, said Fc alemtuzumab heavy chain genes (WT or variants) were co variants bind to an FcyR with an affinity that is greater than the transfected with plasmid containing the alemtuzumab light glycosylated form of the parent Fc polypeptide. 30 chain gene. Media were harvested 5 days after transfection. The present invention also provides methods for engineer For each transfected sample, 10 ul medium was loaded on a ing optimized Fc variants. It is a further object of the present SDS-PAGE gel for Western analysis. The probe for Western invention to provide experimental production and Screening was peroxidase-conjugated goat-anti human IgG (Jackson methods for obtaining optimized Fc variants. Immuno-Research, catalog #109-035-088). WT: wild type The present invention provides isolated nucleic acids 35 alemtuzumab: 1-10: alemtuzumab variants. Hand L indicate encoding the Fc variants described herein. The present inven antibody heavy chain and light chain, respectively. tion provides vectors comprising said nucleic acids, option FIG. 6. Purification of alemtuzumab using protein A chro ally, operably linked to control sequences. The present inven matography. WT alemtuzumab proteins was expressed in tion provides host cells containing the vectors, and methods 293T cells and the media was harvested 5 days after transfec for producing and optionally recovering the Fc variants. 40 tion. The media were diluted 1:1 with PBS and purified with The present invention provides novel Fc polypeptides, protein A (Pierce, Catalog #20334). O: original sample before including antibodies, Fc fusions, isolated Fc, and Fc frag purification; FT: flow through; E: elution; C: concentrated ments, that comprise the Fc variants disclosed herein. Said final sample. The left picture shows a Simple Blue-stained novel Fc polypeptides may find use in a therapeutic product. SDS-PAGE gel, and the right shows a western blot labeled The present invention provides compositions comprising 45 using peroxidase-conjugated goat-anti human IgG. Fc polypeptides that comprise the Fc variants described FIG.7. Production of deglycosylated antibodies. Wildtype herein, and a physiologically or pharmaceutically acceptable and variants of alemtuzumab were expressed in 293T cells carrier or diluent. and purified with protein A chromatography. Antibodies were The present invention contemplates therapeutic and diag incubated with peptide-N-glycosidase (PNGase F) at 37° C. nostic uses for Fc polypeptides that comprise the Fc variants 50 for 24h. For each antibody, a mock treated sample (-PNGase disclosed herein. F) was done in parallel. WT: wild-type alemtuzumab; #15, #16, #17, #18, #22: alemtuzumab variants F241E/F243R/ BRIEF DESCRIPTION OF THE DRAWINGS V262E/V264R, F241E/F243Q/V262T/V264E, F241R/ F243Q/V262T/V264R, F241E/F243Y/V262T/V264R, and FIG.1. Antibody structure and function. Shown is a model 55 I332E respectively. The faster migration of the PNGase F of a full length human IgG1 antibody, modeled using a treated versus the mock treated Samples represents the deg humanized Fab structure from pdb accession code 1 CE1 lycosylated heavy chains. (James et al., 1999, J Mol Biol 289:293-301) and a human FIG.8. Alemtuzumab expressed from 293T cells binds its IgG1 Fc structure from pdb accession code 1D/V2 (DeLano antigen. The antigenic CD52 peptide, fused to GST, was et al., 2000, Science 287: 1279-1283). The flexible hinge that 60 expressed in E. coli BL21 (DE3) under IPTG induction. Both links the Fab and Fc regions is not shown. IgG1 is a uninduced and induced samples were run on a SDS-PAGE homodimer of heterodimers, made up of two light chains and gel, and transferred to PVDF membrane. For western analy two heavy chains. The Ig domains that comprise the antibody sis, either alemtuzumab from Sotec (a-CD52, Sotec) (final are labeled, and include V, and C for the light chain, and V. concentration 2.5 ng/ul) or media of transfected 293T cells Cgammal (CY1), Cgamma2 (CY2), and Cgamma3 (Cy3) for 65 (Campath, Xencor) (final alemtuzumab concentration the heavy chain. The Fc region is labeled. Binding sites for approximately 0.1-0.2 ngful) were used as primary antibody, relevant proteins are labeled, including the antigen binding and peroxidase-conjugated goat-antihuman IgG was used as US 8,188,231 B2 15 16 secondary antibody. M: pre-stained marker, U: un-induced curves represent the fits of the data to a one site competition sample for GST-CD52: I: induced sample for GST-CD52. model. PBS was used as a negative control. FIG.9. Expression and purification of extracellular region FIG. 18. AlphaScreen assay measuring binding to human of human V158 FcyRIIIa. Tagged FcyRIIIa was transfected in V158 FcyRIIa by select Fc variants in the context of ritux 293T cells, and media containing secreted FcyRIIIa were 5 imab. The binding data were normalized to the upper and harvested 3 days later and purified using affinity chromatog lower baselines for each particular antibody, and the curves raphy. 1: media; 2: flow through; 3: wash;4-8: serial elutions. represent the fits of the data to a one site competition model. Both simple blue-stained SDS-PAGE gel and western result PBS was used as a negative control. are shown. For the western blot, membrane was probed with FIG. 19. AlphaScreen assay measuring binding to human 10 V158 FcyRIIIa by select Fc variants in the context of cetux anti-GST antibody. imab. The binding data were normalized to the upper and FIG. 10. Binding to human V158 FcyRIIIa by selectalem lower baselines for each particular antibody, and the curves tuzumab Fc variants from the experimental library as deter represent the fits of the data to a one site competition model. mined by the AlphaScreenTM assay, described in Example 2. PBS was used as a negative control. In the presence of competitor antibody (Fc variant or WT 15 FIGS. 20a-20b. AlphaScreen assay showing binding of alemtuzumab) a characteristic inhibition curve is observed as select alemtuzumab Fc variants to the V158 (FIG. 20a) and a decrease in luminescence signal. Phosphate buffer saline F158 (FIG. 20b) allotypes of human FcyRIIIa. The binding (PBS) alone was used as the negative control. The binding data were normalized to the upper and lower baselines for data were normalized to the maximum and minimum lumi each particular antibody, and the curves represent the fits of nescence signal for each particular curve, provided by the the data to a one site competition model. PBS was used as a baselines at low and high antibody concentrations respec negative control. tively. The curves represent the fits of the data to a one site FIGS. 21a-21d. FIGS. 21a and 21b show the correlation competition model using nonlinear regression. These fits pro between SPR Kd's and AlphaScreen IC50s from binding of vide IC50s for each antibody, illustrated for WT and S239D selectalemtuzumab Fc variants to V158 FcyRIIIa (FIG. 21a) by the dotted lines. 25 and F158 FcyRIIIa (FIG. 21b). FIGS. 21c and 21d show the FIGS.11a and 11b. AlphaScreen assay showing binding of correlation between SPR and AlphaScreen fold-improve select alemtuzumab (FIG. 11a) and trastuzumab (FIG. 11b) ments over WT for binding of selectalemtuzumab Fc variants Fc variants to human Val158 FcyRIIIa. The binding data were to V158 FcyRIIIa (FIG. 21c) and F158 FcyRIIIa (FIG. 21d). normalized to the upper and lower baselines for each particu Binding data are presented in Table 3. The lines through the lar antibody, and the curves represent the fits of the data to a 30 data represent the linear fits of the data, and the r values one site competition model. PBS was used as a negative indicate the significance of these fits. control. FIGS. 22a and 22b. AlphaScreen assay showing binding of FIG. 12. AlphaScreen assay showing binding of select selectalemtuzumab Fc variants to human V158 FcyRIIIa. The alemtuzumab Fc variants to human FcyRIIb. The binding data binding data were normalized to the upper and lower base were normalized to the upper and lower baselines for each 35 lines for each particular antibody, and the curves represent the particular antibody, and the curves represent the fits of the fits of the data to a one site competition model. PBS was used data to a one site competition model. PBS was used as a as a negative control. negative control. FIGS. 23a-23b. Cell-based ADCC assays of select Fc vari FIG. 13. AlphaScreen assay showing binding of select ants in the context of alemtuzumab. ADCC was measured alemtuzumab Fc variants to human R131 FcyRIIa. The bind 40 using the DELFIACR EuTDA-based cytotoxicity assay (Per ing data were normalized to the upper and lower baselines for kinElmer, Mass.), as described in Example 3, using DoHH-2 each particular antibody, and the curves represent the fits of lymphoma target cells and 50-fold excess human PBMCs. the data to a one site competition model. FIG. 23a is a bar graph showing the raw fluorescence data for FIG. 14. AlphaScreen assay measuring binding of select the indicated alemtuzumab antibodies at 10 ng/ml. The alemtuzumab Fc variants to human FcRn, as described in 45 PBMC bar indicates basal levels of cytotoxicity in the Example 2. The binding data were normalized to the upper absence of antibody. FIG. 23b shows the dose-dependence of and lower baselines for each particular antibody, and the ADCC on antibody concentration for the indicated alemtu curves represent the fits of the data to a one site competition Zumab antibodies, normalized to the minimum and maximum model. PBS was used as a negative control. fluorescence signal for each particular curve, provided by the FIG. 15. AlphaScreen assay measuring binding of select 50 baselines at low and high antibody concentrations respec alemtuzumab Fc variants to bacterial protein A, as described tively. The curves represent the fits of the data to a sigmoidal in Example 2. The binding data were normalized to the upper dose-response model using nonlinear regression. and lower baselines for each particular antibody, and the FIGS. 24a-24c. Cell-based ADCC assays of select Fc vari curves represent the fits of the data to a one site competition ants in the context of trastuzumab. ADCC was measured model. PBS was used as a negative control. 55 using the DELFIACR EuTDA-based cytotoxicity assay, as FIGS. 16a-16b. AlphaScreen assay comparing binding of described in Example 3, using BT474 and Sk-Br-3 breast select alemtuzumab Fc variants to human V158 FcyRIIa carcinoma target cells and 50-fold excess human PBMCs. (FIG. 16a) and human FcyRIIb (FIG. 16b). The binding data FIG. 24a is a bar graph showing the raw fluorescence data for were normalized to the upper and lower baselines for each the indicated trastuzumab antibodies at 1 ng/ml. The PBMC particular antibody, and the curves represent the fits of the 60 bar indicates basal levels of cytotoxicity in the absence of data to a one site competition model. PBS was used as a antibody. FIGS. 24b and 24c show the dose-dependence of negative control. ADCC on antibody concentration for the indicated trastu FIGS. 17a-17b. AlphaScreen assay measuring binding to Zumab antibodies, normalized to the minimum and maximum human V158 FcyRIIIa (FIGS. 17a and 17b) and human fluorescence signal for each particular curve, provided by the FcyRIIb (FIG. 17c) by select Fc variants in the context of 65 baselines at low and high antibody concentrations respec trastuzumab. The binding data were normalized to the upper tively. The curves represent the fits of the data to a sigmoidal and lower baselines for each particular antibody, and the dose-response model using nonlinear regression. US 8,188,231 B2 17 18 FIGS.25a-25c. Cell-based ADCC assays of select Fc vari noma target cells, and the level of cytotoxicity was measured ants in the context of rituximab. ADCC was measured using using the LDH Cytotoxicity Detection Kit, according to the the DELFIACR EuTDA-based cytotoxicity assay, as described manufacturer's protocol (Roche Diagnostics GmbH, Pen in Example 3, using WIL2-S lymphoma target cells and Zberg, Germany). The graph shows the dose-dependence of 50-fold excess human PBMCs. FIG.25a is a bar graph show ADCC on antibody concentration for the indicated trastu ing the raw fluorescence data for the indicated rituximab Zumab antibodies, normalized to the minimum and maximum antibodies at 1 ng/ml. The PBMC bar indicates basal levels of fluorescence signal for each particular curve, provided by the cytotoxicity in the absence of antibody. FIGS. 25b and 25c baselines at low and high antibody concentrations respec show the dose-dependence of ADCC on antibody concentra tively. The curves represent the fits of the data to a sigmoidal tion for the indicated rituximab antibodies, normalized to the 10 dose-response model using nonlinear regression. minimum and maximum fluorescence signal for each particu FIG. 31. Cell-based ADCP assay of select variants. The lar curve, provided by the baselines at low and high antibody ADCP assay was carried out as described in Example 7, using concentrations respectively. The curves represent the fits of a co-labeling strategy coupled with flow cytometry. Differen the data to a sigmoidal dose-response model using nonlinear tiated macrophages were used as effector cells, and Sk-Br-3 regression. 15 cells were used as target cells. Percent phagocytosis repre FIGS. 26a-26b. Cell-based ADCC assay of select trastu sents the number of co-labeled cells (macrophage+Sk-Br-3) Zumab (FIG. 26a) and rituximab (FIG. 26b) Fc variants show over the total number of Sk-Br-3 in the population (phagocy ing enhancements in potency and efficacy. Both assays used tosed--non-phagocytosed). homozygous F158/F158 FcyRIIIa PBMCs as effector cells at FIGS. 32a-32c. Capacity of select Fc variants to mediate a 25-fold excess to target cells, which were Sk-Br-3 for the binding and activation of complement. FIG. 32a shows an trastuzumab assay and WIL2-S for the rituximab assay. Data AlphaScreen assay measuring binding of selectalemtuzumab were normalized according to the absolute minimal lysis for Fc variants to C1q. The binding data were normalized to the the assay, provided by the fluorescence signal of target cells in upper and lower baselines for each particular antibody, and the presence of PBMCs alone (no antibody), and the absolute the curves represent the fits of the data to a one site competi maximal lysis for the assay, provided by the fluorescence 25 tion model. FIGS. 32b and 31c show a cell-based assay mea signal of target cells in the presence of Triton X1000, as Suring capacity of select rituximab Fc variants to mediate described in Example 3. CDC. CDC assays were performed using Alamar Blue to FIG. 27. AlphaScreen assay showing binding of select monitor lysis of Fc variant and WT rituximab-opsonized alemtuzumab Fc variants to human V158 FcyRIIIa. The bind WIL2-S lymphoma cells by human serum complement ing data were normalized to the upper and lower baselines for 30 (Quidel, San Diego, Calif.). The dose-dependence on anti each particular antibody, and the curves represent the fits of body concentration of complement-mediated lysis is shown the data to a one site competition model. PBS was used as a for the indicated rituximab antibodies, normalized to the negative control. minimum and maximum fluorescence signal for each particu FIG. 28. ADCC. Cell-based ADCC assays of select Fc lar curve, provided by the baselines at low and high antibody variant trastuzumab antibodies as compared to WT trastu 35 concentrations respectively. The curves represent the fits of Zumab. Purified human peripheral blood monocytes (PB the data to a sigmoidal dose-response model using nonlinear MCs) were used as effector cells, and Sk-Br-3 breast carci regression. noma cells were used as target cells. Lysis was monitored by FIGS. 33a-33c. Enhanced B cell depletion by Fc variants in measuring LDH activity using the Cytotoxicity Detection Kit macaques, as described in Example 9. FIG. 33a shows the (LDH, Roche Diagnostic Corporation, Indianapolis, Ind.). 40 percent B cells remaining in Macaca Fascicularis monkeys Samples were run in triplicate to provide errorestimates (n=3, during treatment with anti-CD20 WT and S239D/I332E rit +/-S.D.). The figure shows the dose dependence of ADCC at uximab antibodies, measured using markers CD20+ and various antibody concentrations, normalized to the minimum CD40+. FIG.33b shows the percent natural killer (NK) cells and maximum levels of lysis for the assay. The curves repre remaining in the monkeys during treatment, measured using sent the fits of the data to a sigmoidal dose-response model 45 markers CD3-/CD16+ and CD3-f(D8+. FIG.33c shows the using nonlinear regression. dose response of CD20+ B cell levels to treatment with FIGS. 29a-29b. Cell-based ADCC assay of select trastu S239D/I332E rituximab. Data are presented as the average of Zumab Fc variants against different cell lines expressing vary 3 monkeys/sample. ing levels of the Her2/neu target antigen. ADCC assays were FIGS. 34a, 34b, and 34c. AlphScreen assay measuring run as described in Example 5, with various cell lines express 50 binding of select alemtuzumab (FIG. 34a) and trastuzumab ing amplified to low levels of Her2/neu receptor, including (FIGS. 34b and 34c) Fc variants to mouse FcyRIII, as Sk-Br-3 (1x106 copies), SkOV3 (-1x105), OVCAR3 (-1x described in Example 10. The binding data were normalized 104), and MCF-7 (-3x103 copies). FIG.29a provides a west to the upper and lower baselines for each particular antibody, ern blot showing the Her2 expression level for each cell line: and the curves represent the fits of the data to a one site equivalent amounts of cell lysate were loaded on an SDS 55 competition model. PBS was used as a negative control. PAGE gel, and Her2 was detected using trastuzumab. Human FIG. 35. Cell-based ADCC assays of select Fc variants in PBMCs allotyped as homozygous F158/F158 FcyRIIIa were the context of trastuzumab using mouse PBMCs as effector used at 25-fold excess to target cells. The bar graph in FIG. cells. ADCC was measured using the DELFIACR EuTDA 29b provides ADCC data for WT and Fc variant against the based cytotoxicity assay using Sk-Br-3 breast carcinoma tar indicated cell lines, normalized to the minimum and maxi 60 get cells and 8-fold excess mouse PBMCs. The bar graph mum fluorescence signal provided by minimal lysis (PBMCs shows the raw fluorescence data for the indicated trastuzumab alone) and maximal lysis (Triton X1000). antibodies at 10 ng/ml. The PBMC bar indicates basal levels FIG. 30. Cell-based ADCC assays of select Fc variants in of cytotoxicity in the absence of antibody, and TX indicates the context of trastuzumab using natural killer (NK) cells as complete cell lysis in the presence of Triton X1000. effector cells and measuring LDH release to monitor cell 65 FIG. 36. AlphaScreen assay measuring binding to human lysis. NK cells, allotyped as heterozygous F158/F 158 V158 FcyRIIIa by select trastuzumab Fc variants expressed in FcyRIIIa, were at an 4-fold excess to Sk-Br-3 breast carci 293T and CHO cells, as described in Example 11. The bind US 8,188,231 B2 19 20 ing data were normalized to the upper and lower baselines for wherein nonspecific cytotoxic cells that express FcyRs rec each particular antibody, and the curves represent the fits of ognize bound antibody on a target cell and Subsequently the data to a one site competition model. PBS was used as a cause lysis of the target cell. negative control. By ADCP or antibody dependent cell-mediated phago FIGS. 37a-37b. Synergy of Fc variants and engineered 5 cytosis as used herein is meant the cell-mediated reaction glycoforms. FIG. 37a presents an AlphaScreen assay show wherein nonspecific cytotoxic cells that express FcyRs rec ing V158 FcyRIIIa binding by WT and Fc variant (V209, ognize bound antibody on a target cell and Subsequently S239/I332E/A330L) trastuzumab expressed in 293T, CHO, cause phagocytosis of the target cell. and Lec-13 CHO cells. The data were normalized to the upper By “amino acid modification' herein is meant an amino and lower baselines for each antibody, and the curves repre 10 acid Substitution, insertion, and/or deletion in a polypeptide sent the fits of the data to a one site competition model. PBS sequence. The preferred amino acid modification herein is a was used as a negative control. FIG.37b presents a cell-based substitution. By “amino acid substitution' or “substitution” ADCC assay showing the ability of 239T. CHO, and Lec-13 herein is meant the replacement of an amino acid at a particu CHO expressed WT and V209 trastuzumab to mediate lar position in a parent polypeptide sequence with another ADCC. ADCC was measured using the DELFIAR EuTDA 15 amino acid. For example, the substitution I332E refers to a based cytotoxicity assay as described previously, with Sk variant polypeptide, in this case an Fc variant, in which the Br-3 breast carcinoma target cells. The data show the dose isoleucine at position 332 is replaced with a glutamic acid. In dependence of ADCC on antibody concentration for the some embodiments, the WT identity need not be defined. For indicated trastuzumab antibodies, normalized to the mini example, the substitution 332E refers to a variant polypeptide mum and maximum fluorescence signal for each particular in which position 332 is mutated to glutamic acid. curve, provided by the baselines at low and high antibody By “antibody' herein is meant a protein consisting of one concentrations respectively. The curves represent the fits of or more polypeptides substantially encoded by all or part of the data to a sigmoidal dose-response model using nonlinear the recognized immunoglobulin genes. The recognized regression. immunoglobulin genes, for example in humans, include the FIG.38. AlphaScreen assay showing binding of aglycosy 25 kappa (K), lambda (X), and heavy chain genetic loci, which lated alemtuzumab Fc variants to human V158 FcyRIIIa. The together comprise the myriad variable region genes, and the binding data were normalized to the upper and lower base constant region genes mu (LL), delta (Ö), gamma (Y), sigma (O), lines for each particular antibody, and the curves represent the and alpha (C.) which encode the IgM, Ig|D, IgG, IgE, and IgA fits of the data to a one site competition model. PBS was used isotypes respectively. Antibody herein is meant to include full as a negative control. 30 length antibodies and antibody fragments, and may refer to a FIG. 39. AlphaScreen assay comparing human V158 natural antibody from any organism, an engineered antibody, FcyRIIIa binding by select alemtuzumab Fc variants in gly or an antibody generated recombinantly for experimental, cosylated (solid symbols, Solid lines) and deglycosylated therapeutic, or other purposes as further defined below. The (open symbols, dotted lines). The binding data were normal term “antibody' includes antibody fragments, as are known ized to the upper and lower baselines for each particular 35 in the art, such as Fab, Fab'. F(ab'). Fv, scFv, or otherantigen antibody, and the curves represent the fits of the data to a one binding Subsequences of antibodies, either produced by the site competition model. modification of whole antibodies or those synthesized de FIGS. 40a-40c. Sequences showing improved anti-CD20 novo using recombinant DNA technologies. Particularly pre antibodies. The light and heavy chain sequences of rituximab ferred are full length antibodies that comprise Fc variants as are presented in FIG. 4.0a and FIG. 40b respectively, and are 40 described herein. The term “antibody’ comprises mono taken from translated Sequence 3 of U.S. Pat. No. 5,736,137. clonal and polyclonal antibodies. Antibodies can be antago Relevant positions in FIG. 40b are bolded, including S239, nists, agonists, neutralizing, inhibitory, or stimulatory. The V240, V264-I, H268, E272, K274, N297, S298, K326, A330, antibodies of the present invention may be nonhuman, chi and I332. FIG. 40c shows the improved anti-CD20 antibody meric, humanized, or fully human, as described below in heavy chain sequences, with variable positions designated in 45 more detail. bold as X1,X2, X3, X4, X5, X6, X7, X8, X9, Z1, and Z2. The Specifically included within the definition of “antibody” table below the sequence provides possible substitutions for are aglycosylated antibodies. By "aglycosylated antibody’ as these positions. The improved anti-CD20 antibody sequences used herein is meant an antibody that lacks carbohydrate comprise at least one non-WT amino acid selected from the attached at position 297 of the Fc region, wherein numbering group of possible substitutions for X1, X2, X3, X4, X5, X6. 50 is according to the EU system as in Kabat. The aglycosylated X7, X8, and X9. These improved anti-CD20 antibody antibody may be a deglycosylated antibody, that is an anti sequences may also comprise a substitution Z1 and/or Z2. body for which the Fc carbohydrate has been removed, for These positions are numbered according to the EU index as in example chemically or enzymatically. Alternatively, the agly Kabat, and thus do not correspond to the sequential order in cosylated antibody may be a nonglycosylated or unglycosy the sequence. 55 lated antibody, that is an antibody that was expressed without FIG. 41 depicts the set of Fc variants that were constructed Fc carbohydrate, for example by mutation of one or residues and experimentally tested. that encode the glycosylation pattern or by expression in an FIG. 42 depicts SEQID NO:8; the particular Xaa residues organism that does not attach carbohydrates to proteins, for are as shown in Table 10. example bacteria. 60 Specifically included within the definition of “antibody” DETAILED DESCRIPTION OF THE INVENTION are full-length antibodies that contain an Fc variant portion. By “full length antibody' herein is meant the structure that In order that the invention may be more completely under constitutes the natural biological form of an antibody, includ stood, several definitions are set forth below. Such definitions ing variable and constant regions. For example, in most mam are meant to encompass grammatical equivalents. 65 mals, including humans and mice, the full length antibody of By ADCC or “antibody dependent cell-mediated cyto the IgG class is a tetramer and consists of two identical pairs toxicity’ as used herein is meant the cell-mediated reaction of two immunoglobulin chains, each pair having one light and US 8,188,231 B2 21 22 one heavy chain, each light chain comprising immunoglobu described below. By “Fe polypeptide' as used herein is meant lin domains V, and C, and each heavy chain comprising a polypeptide that comprises all or part of an Fc region. Fc immunoglobulin domains V, Cyl (Cl), Cy2 (C2), and polypeptides include antibodies, Fc fusions, isolated Fcs, and Cy3 (C3). In some mammals, for example in camels and Fc fragments. llamas, IgG antibodies may consist of only two heavy chains, 5 By “Fe fusion' as used herein is meant a protein wherein each heavy chain comprising a variable domain attached to one or more polypeptides or Small molecules is operably the Fc region. By “IgG” as used herein is meant a polypeptide linked to an Fc region or a derivative thereof. Fc fusion is belonging to the class of antibodies that are substantially herein meant to be synonymous with the terms “immunoad encoded by a recognized immunoglobulin gamma gene. In hesin”, “Ig fusion”, “Ig chimera', and “receptor globulin' humans this class comprises IgG1, IgG2, IgG3, and IgG4. In 10 (sometimes with dashes) as used in the prior art (Chamow et mice this class comprises IgG1, IgG2a, IgG2b, IgG3. al., 1996, Trends Biotechnol 14:52-60; Ashkenazi et al., 1997, By “amino acid and “amino acid identity” as used herein Curr Opin Immunol 9:195-200. incorporated by reference). is meant one of the 20 naturally occurring amino acids or any An Fc fusion combines the Fc region of an immunoglobulin non-natural analogues that may be present at a specific, with a fusion partner, which in general can be any protein or defined position. By “protein herein is meant at least two 15 small molecule. The role of the non-Fc part of an Fc fusion, covalently attached amino acids, which includes proteins, i.e. the fusion partner, may be to mediate target binding, and polypeptides, oligopeptides and peptides. The protein may be thus it is functionally analogous to the variable regions of an made up of naturally occurring amino acids and peptide antibody. bonds, or synthetic peptidomimetic structures, i.e. “analogs. By “Fe gamma receptor or “FeyRas used herein is meant such as peptoids (see Simon et al., 1992, Proc Natl Acad Sci any member of the family of proteins that bind the IgG anti USA 89 (20):9367, incorporated by reference) particularly body Fc region and are substantially encoded by the FcyR when LC peptides are to be administered to a patient. Thus genes. In humans this family includes but is not limited to “amino acid', or “peptide residue', as used herein means both FcyRI (CD64), including isoforms FcyRIa, FcyRIb, and Fcy naturally occurring and synthetic amino acids. For example, RIc; FcyRII (CD32), including isoforms FcyRIIa (including homophenylalanine, citrulline and noreleucine are consid 25 allotypes H131 and R131), FcyRIIb (including FcyRIIb-1 and ered amino acids for the purposes of the invention. “Amino FcyRIIb-2), and FcyRIIc; and FcyRIII (CD16), including iso acid also includes imino acid residues such as proline and forms FcyRIIIa (including allotypes V158 and F158) and hydroxyproline. The side chain may be in either the (R) or the FcyRIIIb (including allotypes FcyRIIIb-NA1 and FcyRIIIb (S) configuration. In the preferred embodiment, the amino NA2), as well as any undiscovered human FcyRs or FcyR acids are in the (S) or L-configuration. If non-naturally occur 30 isoforms or allotypes. An FcyR may be from any organism, ring side chains are used, non-amino acid Substituents may be including but not limited to humans, mice, rats, rabbits, and used, for example to prevent or retard in vivo degradation. monkeys. Mouse FcyRs include but are not limited to FcyRI By “effector function” as used herein is meant a biochemi (CD64), FcyRII (CD32), FcyRIII (CD16), and FcyRIII-2 cal event that results from the interaction of an antibody Fc (CD16-2), as well as any undiscovered mouse FcyRs or FcyR region with an Fc receptor or ligand. Effector functions 35 isoforms or allotypes. include but are not limited to ADCC, ADCP, and CDC. By By “Feligand’ or “effector ligand as used herein is meant “effector cell as used herein is meant a cell of the immune a molecule, preferably a polypeptide, from any organism that system that expresses one or more Fc receptors and mediates binds to the Fc region of an antibody to forman Fc/Fc ligand one or more effector functions. Effector cells include but are complex. Binding of an Fc ligand to Fc preferably elicits or not limited to monocytes, macrophages, neutrophils, den 40 more effector functions. Fc ligands includebut are not limited dritic cells, eosinophils, mast cells, platelets, B cells, large to Fc receptors, FcyRs, FcCRs, FceRS, FcRn, C1q, C3, man granular lymphocytes, Langerhans cells, natural killer (NK) nan binding lectin, mannose receptor, staphylococcal protein cells, and YYT cells, and may be from any organism including A, Streptococcal protein G, and viral Fcy R. Fc ligands also but not limited to humans, mice, rats, rabbits, and monkeys. include Fc receptor homologs (FcRH), which are a family of By “library' herein is meant a set of Fc variants in any form, 45 Fc receptors that are homologous to the FcyRs (Davis et al., including but not limited to a list of nucleic acid oramino acid 2002, Immunological Reviews 190: 123-136, incorporated by sequences, a list of nucleic acid oramino acid substitutions at reference). Fc ligands may include undiscovered molecules variable positions, a physical library comprising nucleic acids that bind Fc. that encode the library sequences, or a physical library com By "IgG” as used herein is meant a polypeptide belonging prising the Fc variant proteins, either in purified or unpurified 50 to the class of antibodies that are substantially encoded by a form. recognized immunoglobulin gamma gene. In humans this By “Fe' or “Fe region', as used herein is meant the class comprises IgG1, IgG2, IgG3, and IgG4. In mice this polypeptide comprising the constant region of an antibody class comprises IgG1, IgG2a, IgG2b, IgG3. By “immunoglo excluding the first constant region immunoglobulin domain. bulin (Ig) herein is meant a protein consisting of one or more Thus Fc refers to the last two constant region immunoglobu 55 polypeptides Substantially encoded by immunoglobulin lin domains of IgA, Ig|D, and IgG, and the last three constant genes. Immunoglobulins include but are not limited to anti region immunoglobulin domains of IgE and IgM, and the bodies. Immunoglobulins may have a number of structural flexible hinge N-terminal to these domains. For IgA and IgM, forms, including but not limited to full length antibodies, Fc may include the J chain. For IgG, as illustrated in FIG. 1, antibody fragments, and individual immunoglobulin Fc comprises immunoglobulin domains Cgamma2 and 60 domains. By “immunoglobulin (Ig) domain herein is meant Cgamma3 (Cy2 and Cy3) and the hinge between Cgammal a region of an immunoglobulin that exists as a distinct struc (CY1) and Cgamma2 (CY2). Although the boundaries of the tural entity as ascertained by one skilled in the art of protein Fc region may vary, the human IgG heavy chain Fc region is structure. Ig domains typically have a characteristic -sand usually defined to comprise residues C226 or P230 to its wich folding topology. The known Ig domains in the IgG carboxyl-terminus, wherein the numbering is according to 65 class of antibodies are V, Cyl, Cy2, Cy3, V, and C. the EU index as in Kabat. Fc may refer to this region in By “parent polypeptide' or “precursor polypeptide' (in isolation, or this region in the context of an Fc polypeptide, as cluding Fc parent or precursors) as used herein is meant a US 8,188,231 B2 23 24 polypeptide that is Subsequently modified to generate a vari S239D/A33OL/I332E is the same Fc variant as S239DfI332E? ant. Said parent polypeptide may be a naturally occurring A330L, and so on. For all positions discussed in the present polypeptide, or a variant or engineered version of a naturally invention, numbering is according to the EU index or EU occurring polypeptide. Parent polypeptide may refer to the numbering scheme (Kabat et al., 1991, Sequences of Proteins polypeptide itself, compositions that comprise the parent of Immunological Interest, 5th Ed., United States Public polypeptide, or the amino acid sequence that encodes it. Health Service, National Institutes of Health, Bethesda, Accordingly, by "parent Fc polypeptide' as used herein is incorporated by reference). The EU index or EU index as in meanta Fc polypeptide that is modified to generate a variant, Kabat refers to the numbering of the EU antibody (Edelman and by “parent antibody” as used herein is meant an antibody et al., 1969, Proc Natl AcadSci USA 63:78-85, incorporated that is modified to generate a variant antibody. 10 by reference). As outlined above, certain positions of the Fc molecule can The present invention is directed to optimized Fc variants be altered. By “position' as used herein is meant a location in useful in a variety of contexts. As outlined above, current the sequence of a protein. Positions may be numbered antibody therapies suffer from a variety of problems. The sequentially, or according to an established format, for present invention provides a promising means for enhancing example the EU index as in Kabat. For example, position 297 15 the anti-tumor potency of antibodies is via enhancement of is a position in the human antibody IgG1. Corresponding their ability to mediate cytotoxic effector functions such as positions are determined as outlined above, generally through ADCC, ADCP, and CDC. The present invention shows that alignment with other parent sequences. antibodies with an Fc region optimized for binding to certain By “residue' as used herein is meant a position in a protein FcyRs may better mediate effector functions and thereby and its associated amino acid identity. For example, Aspar destroy cancer cells more effectively in patients. The balance agine 297 (also referred to as Asn297, also referred to as between activating and inhibiting receptors is an important N297) is a residue in the human antibody IgG1. consideration, and optimal effector function may result from By “target antigen” as used herein is meant the molecule an Fc with enhanced affinity for activation receptors, for that is bound specifically by the variable region of a given example FcyRI, FcyRIIa/c, and FcyRIIIa, yet reduced affinity antibody. A target antigen may be a protein, carbohydrate, 25 for the inhibitory receptor FcyRIIb. Furthermore, because lipid, or other chemical compound. FcyRS can mediate antigen uptake and processing by antigen By “target cell as used herein is meant a cell that expresses presenting cells, enhanced Fc/FcyRaffinity may also improve a target antigen. the capacity of antibody therapeutics to elicit an adaptive By "variable region' as used herein is meant the region of immune response. For example, several mutations disclosed an immunoglobulin that comprises one or more Ig domains 30 in this application, including S298A, E333A, and K334A, Substantially encoded by any of the VK, VW, and/or V genes show enhanced binding to the activating receptor FcyRIIIa that make up the kappa, lambda, and heavy chain immuno and reduced binding to the inhibitory receptor FcyRIIb. These globulin genetic loci respectively. mutations maybe combined to obtain double and triple muta By "variant polypeptide' as used herein is meant a tion variants that show additive improvements in binding. A polypeptide sequence that differs from that of a parent 35 particular variant is a S298A/E333A/K334A triple mutant polypeptide sequence by virtue of at least one amino acid with approximately a 1.7-fold increase in binding to F 158 modification. The parent polypeptide may be a naturally FcyRIIIa, a 5-fold decrease in binding to FcyRIIb, and a occurring or wild-type (WT) polypeptide, or may be a modi 2.1-fold enhancement in ADCC. fied version of a WT polypeptide. Variant polypeptide may Although there is a need for greater effector function, for refer to the polypeptide itself a composition comprising the 40 some antibody therapeutics reduced or eliminated effector polypeptide, or the amino sequence that encodes it. Prefer function may be desired. This is often the case for therapeutic ably, the variant polypeptide has at least one amino acid antibodies whose mechanism of action involves blocking or modification compared to the parent polypeptide, e.g. from antagonism but not killing of the cells bearing target antigen. about one to about ten amino acid modifications, and prefer In these cases depletion of target cells is undesirable and can ably from about one to about five amino acid modifications 45 be considered a side effect. For example, the ability of anti compared to the parent. The variant polypeptide sequence CD4 antibodies to block CD4 receptors on T cells makes herein will preferably possess at least about 80% homology them effective anti-inflammatories, yet their ability to recruit with a parent polypeptide sequence, and most preferably at FcyR receptors also directs immune attack against the target least about 90% homology, more preferably at least about cells, resulting in T cell depletion (Reddy et al., 2000, J 95% homology. Accordingly, by “Fe variant as used herein 50 Immunol 164: 1925-1933, incorporated by reference). Effec is meant an Fc sequence that differs from that of a parent Fc tor function can also be a problem for radiolabeled antibodies, sequence by virtue of at least one amino acid modification. An referred to as radioconjugates, and antibodies conjugated to Fc variant may only encompass an Fc region, or may exist in toxins, referred to as immunotoxins. These drugs can be used the context of an antibody, Fc fusion, isolated Fc, Fc frag to destroy cancer cells, but the recruitment of immune cells ment, or other polypeptide that is Substantially encoded by 55 via Fc interaction with FcyRs brings healthy immune cells in Fc. Fc variant may refer to the Fc polypeptide itself, compo proximity to the deadly payload (radiation or toxin), resulting sitions comprising the Fc variant polypeptide, or the amino in depletion of normal lymphoid tissue along with targeted acid sequence that encodes it. cancer cells (Hutchins et al., 1995, Proc Natl AcadSci USA The Fc variants of the present invention are defined accord 92: 11980-11984: White et al., 2001, Annu Rey Med 52:125 ing to the amino acid modifications that compose them. Thus, 60 145, incorporated by reference). This problem can potentially for example, I332E is an Fc variant with the substitution be circumvented by using IgG isotypes that poorly recruit I332E relative to the parent Fc polypeptide. Likewise, complement or effector cells, for example IgG2 and IgG4. An S239D/A330L/I332E (also referred to as 239D/330L/332E) alternate solution is to develop Fc variants that reduce or defines an Fc variant with the substitutions S239D, A330L, ablate binding (Alegreet al., 1994, Transplantation 57: 1537 and I332E (239D, 330L, and 332E) relative to the parent Fc 65 1543; Hutchins et al., 1995, Proc Natl Acad Sci USA 92: polypeptide. It is noted that the order in which substitutions 11980-1 1984; Armour et al., 1999, Eurj Immunol 29:2613 are provided is arbitrary, that is to say that, for example, 2624; Reddy et al., 2000, J Immunol 164:1925-1933; Xu et US 8,188,231 B2 25 26 al., 2000, Cell Immunol 200:16-26: Shields et al., 2001, J Biol any of the antibody classes, including but not limited to Chem 276:6591-6604) (U.S. Pat. No. 6,194.551; U.S. Pat. sequences belonging to the IgG (including human Subclasses No. 5,885,573; PCT WO99/58572), all incorporated by ref IgG1, IgG2, IgG3, or IgG4), IgA (including human Sub erence. A critical consideration for the reduction or elimina classes IgA1 and IgA2), Ig|D, IgE, IgG, or IgM classes of tion of effector function is that other important antibody antibodies. Most preferably the parent Fc polypeptides of the properties not be perturbed. Fc variants should be engineered present invention comprise sequences belonging to the that not only ablate binding to FcyRs and/or C1q, but also human IgG class of antibodies. For example, the parent Fc maintain antibody stability, Solubility, and structural integ polypeptide may be a parent antibody, for example a human rity, as well as ability to interact with other important Fc IgG1 antibody, a human IgA antibody, or a mouse IgG2a or ligands such as FcRn and proteins A and G. 10 IgG2b antibody. Said parent antibody may be nonhuman, In addition, the invention utilizes engineered glycoforms chimeric, humanized, or fully human as described in detail that can enhance Fc/FcyR affinity and effector function. An below. The parent Fc polypeptide may be modified or engi aglycosylated Fc with favorable solution properties and the neered in Some way, for example a parent antibody may be capacity to mediate effector functions would be significantly affinity matured, or may possess engineered glycoforms, all enabling for the alternate production methods described 15 as described more fully below. Alternatively, the parent Fc above. By overcoming the structural and functional short polypeptide may be an Fc fusion, for example an Fc fusion comings of aglycosylated Fc, antibodies can be produced in wherein the fusion partner targets a cell Surface receptor. bacteria and transgenic plants and animals with reduced risk Alternatively, the parent Fc polypeptide may be an isolated Fc of immunogenicity, and with effector function for clinical region, comprising little or no other polypeptide sequence applications in which cytotoxicity is desired such as cancer. outside the Fc region. The parent Fc polypeptide may be a The present invention describes the utilization of protein naturally existing Fc region, or may be an existing engineered engineering methods to develop stable, Soluble Fc variants variant of an Fc polypeptide. What is important is that the with effector function. Currently, such Fc variants do not exist parent Fc polypeptide comprise an Fc region, which can then in the art. be mutated to generate an Fc variant. Fc Variants of the Present Invention 25 The Fc variants of the present invention may be an anti The Fc variants of the present invention may find use in a body, referred to herein as an “antibody of the present inven variety of Fc polypeptides. An Fc polypeptide that comprises tion'. Antibodies of the present invention may comprise an Fc variant of the present invention is herein referred to as immunoglobulin sequences that are substantially encoded by an “Fe polypeptide of the present invention’. Fc polypeptides immunoglobulin genes belonging to any of the antibody of the present invention include polypeptides that comprise 30 classes, including but not limited to IgG (including human the Fc variants of the present invention in the context of a Subclasses IgG1, IgG2, IgG3, or IgG4), IgA (including larger polypeptide, such as an antibody or Fc fusion. That is, human subclasses IgA1 and IgA2), Ig|D, IgE, IgG, and IgM Fc polypeptides of the present invention include antibodies classes of antibodies. Most preferably the antibodies of the and Fc fusions that comprise Fc variants of the present inven present invention comprise sequences belonging to the tion. By “antibody of the Present invention” as used herein is 35 human IgG class of antibodies. Antibodies of the present meant an antibody that comprises an Fc variant of the present invention may be nonhuman, chimeric, humanized, or fully invention. By “Fc fusion of the Present invention' as used human. As will be appreciated by one skilled in the art, these herein refers to an Fc fusion that comprises an Fc variant of different types of antibodies reflect the degree of “human the present invention. Fc polypeptides of the present inven ness' or potential level of immunogenicity in a human. For a tion also include polypeptides that comprise little or no addi 40 description of these concepts, see Clark et al., 2000 and tional polypeptide sequence other than the Fc region, referred references cited therein (Clark, 2000, Immunol Today 21:397 to as an isolated Fc. By "isolated Fc of the present invention 402, incorporated by reference). Chimeric antibodies com used herein is meant an Fc polypeptide that comprises an Fc prise the variable region of a nonhuman antibody, for example variant of the present invention, and comprises little or no V and V, domains of mouse or rat origin, operably linked to additional polypeptide sequence other than the Fc region. Fc 45 the constant region of a human antibody (see for example polypeptides of the present invention also include fragments U.S. Pat. No. 4,816,567, incorporated by reference). Said of the Fc region. By “Fe fragment of the Present invention” as nonhuman variable region may be derived from any organism used herein is meant an Fc fragment that comprises an Fc as described above, preferably mammals and most preferably variant of the present invention. As described below, any of rodents or primates. In one embodiment, the antibody of the the aforementioned Fc polypeptides of the present invention 50 present invention comprises monkey variable domains, for may be fused to one or more fusion partners or conjugate example as described in Newman et al., 1992, Biotechnology partners to provide desired functional properties. 10:1455-1460, U.S. Pat. No. 5,658,570, and U.S. Pat. No. Fc variants may be constructed in a parent Fc polypeptide 5,750,105, incorporated by reference. In a preferred embodi irrespective of its context. That is to say that, the sole criteria ment, the variable region is derived from a nonhuman source, for a parent Fc polypeptide is that it comprise an Ec region. 55 but its immunogenicity has been reduced using protein engi The parent Fc polypeptides described herein may be derived neering. In a preferred embodiment, the antibodies of the from a wide range of Sources, and may be substantially present invention are humanized (Tsurushita & Vasquez, encoded by one or more Fc genes from any organism, includ 2004, Humanization of Monoclonal Antibodies, Molecular ing but not limited to humans, rodents including but not Biology of B Cells, 533-545, Elsevier Science (USA), incor limited to mice and rats, lagomorpha Such as rabbits and 60 porated by reference). By “humanized' antibody as used hares, camelidae Such as camels, llamas, and dromedaries, herein is meant an antibody comprising a human framework and non-human primates, including but not limited to Pros region (FR) and one or more complementarity determining imians, Platyrrhini (New World monkeys), Cercopithecoidea regions (CDR's) from a non-human (usually mouse or rat) (Old World monkeys), and Hominoidea include the Gibbons, antibody. The non-human antibody providing the CDR's is Lesser and Great Apes, with humans most preferred. The 65 called the “donor” and the human immunoglobulin providing parent Fc polypeptides of the present invention may be Sub the framework is called the “acceptor. Humanization relies stantially encoded by immunoglobulin genes belonging to principally on the grafting of donor CDRS onto acceptor US 8,188,231 B2 27 28 (human) V, and V frameworks (Winter U.S. Pat. No. 5.225, tors, purinergic receptors, adrenergic receptors, histaminer 539, incorporated by reference). This strategy is referred to as gic receptors, opiod receptors, chemokine receptors, “CDR grafting”. “Backmutation of selected acceptor frame glutamate receptors, nicotinic receptors, the 5HT (serotonin) work residues to the corresponding donor residues is often receptor, and estrogen receptors. A fusion partner may be a required to regain affinity that is lost in the initial grafted Small-molecule mimetic of a protein that targets a therapeu construct (U.S. Pat. No. 5,530,101; U.S. Pat. No. 5,585,089; tically useful target. Specific examples of particular drugs U.S. Pat. No. 5,693,761; U.S. Pat. No. 5,693,762; U.S. Pat. that may serve as Fc fusion partners can be found in L. S. No. 6,180,370; U.S. Pat. No. 5,859,205; U.S. Pat. No. 5,821, Goodman et al., Eds., Goodman and Gilman's The Pharma 337; U.S. Pat. No. 6,054,297: U.S. Pat. No. 6,407,213, incor cological Basis of Therapeutics (McGraw-Hill, New York, porated by reference). A large number of other methods for 10 ed. 9, 1996, incorporated by reference). Fusion partners humanization are known in the art (Tsurushita & Vasquez, include not only small molecules and proteins that bind 2004, Humanization of Monoclonal Antibodies, Molecular known targets for existing drugs, but orphan receptors that do Biology of B Cells, 533-545, Elsevier Science (USA), incor not yet exist as drug targets. The completion of the genome porated by reference), and any of such methods may find use and proteome projects are proving to be a driving force in in the present invention for modifying Fc variants for reduced 15 drug discovery, and these projects have yielded a trove of immunogenicity. The humanized antibody optimally also orphan receptors. There is enormous potential to validate will comprise at least a portion of an immunoglobulin con these new molecules as drug targets, and develop protein and stant region, typically that of a human immunoglobulin, and Small molecule therapeutics that target them. Such protein thus will typically comprise a human Fc region. In a most and Small molecule therapeutics are contemplated as Fc preferred embodiment, the immunogenicity of an Fc variant fusion partners that employ the Fc variants of the present of the present invention is reduced using a method described invention. Fc fusions of the invention may comprise immu in U.S. Ser. No. 11/004,590, filed Dec. 3, 2004, entitled noglobulin sequences that are Substantially encoded by “Methods of Generating Variant Proteins with Increased Host immunoglobulin genes belonging to any of the antibody String Content and Compositions Thereof.” incorporated by classes, including but not limited to IgG (including human reference. In an alternate embodiment, the antibodies of the 25 Subclasses IgG1, IgG2, IgG3, or IgG4), IgA (including present invention may be fully human, that is the sequences of human Subclasses IgA1 and IgA2), Ig|D, IgE. IgG, and IgM the antibodies are completely or Substantially human. A num classes of antibodies. Most preferably the Fc fusions of the ber of methods are known in the art for generating fully present invention comprise sequences belonging to the human antibodies, including the use of transgenic mice human IgG class of antibodies. A variety of linkers, defined (Bruggemann et al., 1997, Curr Opin Biotechnol 8:455-458, 30 and described below, may be used to covalently link Fc to a incorporated by reference) or human antibody libraries fusion partner to generate an Fc fusion. coupled with selection methods (Griffiths et al., 1998, Curr The Fc variants of the present invention may find use in an Opin Biotechnol 9:102-108, incorporated by reference). isolated Fc. that is an Fc polypeptide that comprises little or The Fc variants of the present invention may be an Fc no additional polypeptide sequence other than the Fc region fusion, referred to herein as an “Fc fusion of the present 35 and that comprises an Fc variant of the present invention. invention’. Fc fusions of the present invention comprise an Fc Isolated Fc of the present invention are meant as molecules polypeptide operably linked to one or more fusion partners. wherein the desired function of the molecule, for example the The role of the fusion partner typically, but not always, is to desired therapeutic function, resides solely in the Fc region. mediate binding of the Fc fusion to a target antigen. (Chamow Thus the therapeutic target of an isolated Fc of the present et al., 1996, Trends Biotechnol 14:52-60; Ashkenazi et al., 40 invention is likely to involve one or more Fc ligands. An 1997, Curr Opin Immunol 9:195-200, incorporated by refer isolated Fc that comprises the Fc variant may require no ence). For example, the approved drug alefacept (marketed as additional covalent polypeptide sequence to achieve its AMEVIVE(R) is an immunosuppressive Fc fusion that con desired outcome. In a preferred embodiment, said isolated Fc sists of the extracellular CD2-binding portion of the human comprises from 90-100% of the Fc region, with little or no leukocyte function antigen-3 (LFA-3) linked to the Fc region 45 “extra sequence. Thus, for example, an isolated Fc of the of human IgG1. The approved drug etanercept (marketed as present invention may comprise residues C226 or P230 to the ENBREL(R) is an Fc fusion comprising the extracellular carboxyl-terminus of human IgG1, wherein the numbering is ligand-binding portion of human tumor necrosis factor recep according to the EU index as in Kabat. In one embodiment, tor (TNFR) linked to human IgG1 Fc. Virtually any protein, the isolated Fc of the present invention may contain no extra polypeptide, peptide, or Small molecule may be linked to Fc 50 sequence outside the Fc region. However it is also contem to generate an Fc fusion. Fusion partners include but are not plated that isolated Fc's may not also comprise additional limited to receptors and extracellular receptor domains, adhe polypeptide sequences. For example, an isolated Fc may, in sion molecules, ligands, enzymes, cytokines, chemokines, or addition to comprising an Fc variant Fc region, comprise Some other protein or protein domain. The fusion partner may additional polypeptide sequence tags that enable expression, also play a role as a chemoattractant. Undiscovered ligands or 55 purification, and the like. receptors may serve as fusion partners for the Fc variants of The Fc variants of the present invention may find use in a the present invention. Small molecules may serve as fusion fragment of the Fc region, that is an Fc polypeptide that partners, and may include any therapeutic agent that directs comprises an Fc fragment that comprises an Fc variant of the the Fc fusion to a therapeutic target. Such targets may be any present invention. Clearly a requirement of an Fc fragment of molecule, preferrably an extracellular receptor, that is impli 60 the present invention is that it contains the position(s) at cated in disease. Two families of surface receptors that are which the amino acid modifications of the Fc variant are targets of a number of approved small molecule drugs are made. An Fc fragment of the present invention may comprise G-Protein Coupled Receptors (GPCRs), and ion channels, from 1-90% of the Fc region, with 10-90% being preferred, including K+, Na+, Ca+ channels. Nearly 70% of all drugs and 30-90% being most preferred. Thus for example, an Fc currently marketed worldwide target GPCRs. Thus the Fc 65 fragment of the present invention may comprise an Fc variant variants of the present invention may be fused to a small IgG1 Cy2 domain, an Fc variant IgG1 Cy2 domain and hinge molecule that targets, for example, one or more GABA recep region, an Fc variant IgG1 Cy3 domain, and so forth. In one US 8,188,231 B2 29 30 embodiment, an Fc fragment of the present invention addi protein atoms of the proteins. Regardless of how equivalentor tionally comprises a fusion partner, effectively making it an corresponding residues are determined, and regardless of the Fc fragment fusion. As with isolated Fcs, Fc fragments may or identity of the parent Fc variant in which the Fc variants are may not contain extra polypeptide sequence. made, what is meant to be conveyed is that the Fc variants Fc variants of the present invention may be substantially 5 discovered by the present invention may be engineered into encoded by genes from any organism, preferably mammals, any second parent Fc variant that has significant sequence or including but not limited to humans, rodents including but not structural homology with said Fc variant. Thus for example, if limited to mice and rats, lagomorpha including but not limited a variant antibody is generated wherein the parent antibody is to rabbits and hares, camelidae including but not limited to human IgG1, by using the methods described above or other camels, llamas, and dromedaries, and non-human primates, 10 methods for determining equivalent residues, said variant including but not limited to Prosimians, Platyrrhini (New antibody may be engineered in a human IgG2 parent anti World monkeys), Cercopithecoidea (Old World monkeys), body, a human IgA parent antibody, a mouse IgG2a or IgG2b and Hominoidea including the Gibbons and Lesser and Great parent antibody, and the like. Again, as described above, the Apes. In a most preferred embodiment, the Fc variants of the context of the parent Fc variant does not affect the ability to present invention are substantially human. The Fc variants of 15 transfer the Fc variants of the present invention to other parent the present invention may be substantially encoded by immu Fc variants. For example, the variant antibodies that are engi noglobulin genes belonging to any of the antibody classes. In neered in a human IgG1 antibody that targets one epitope may a most preferred embodiment, the Fc variants of the present be transferred into a human IgG2 antibody that targets a invention comprise sequences belonging to the IgG class of different epitope, into an Fc fusion that comprises a human antibodies, including human Subclasses IgG1, IgG2, IgG3, 20 IgG1 Fc region that targets yet a different epitope, and so and IgG4. In an alternate embodiment, the Fc variants of the forth. present invention comprise sequences belonging to the IgA The Fc variants of the present invention may find use in a (including human Subclasses IgA1 and IgA2), Ig), IgE. IgG, wide range of products. In one embodiment the Fc variant of or IgM classes of antibodies. The Fc variants of the present the invention is a therapeutic, a diagnostic, or a research invention may comprise more than one protein chain. That is, 25 reagent, preferably atherapeutic. Alternatively, the Fc variant the present invention may find use in an Fc variant that is a of the present invention may be used for agricultural or indus monomer oran oligomer, including a homo- or hetero-oligo trial uses. An antibody of the present invention may find use C. in an antibody composition that is monoclonal or polyclonal. In the most preferred embodiment, the Fc polypeptides of The Fc variants of the present invention may be agonists, the invention are based on human IgG sequences, and thus 30 antagonists, neutralizing, inhibitory, or stimulatory. In a pre human IgG sequences are used as the “base' sequences ferred embodiment, the Fc variants of the present invention against which other sequences are compared, including but are used to kill target cells that bear the target antigen, for not limited to sequences from other organisms, for example example cancer cells. In an alternate embodiment, the Fc rodent and primate sequences, as well as sequences from variants of the present invention are used to block, antago other immunoglobulin classes such as IgA, IgE, IgGD, 35 nize, or agonize the target antigen. In an alternately preferred IgGM, and the like. It is contemplated that, although the Fc embodiment, the Fc variants of the present invention are used variants of the present invention are engineered in the context to block, antagonize, or agonize the target antigen and kill the of one parent Fc variant, the variants may be engineered in or target cells that bear the target antigen. “transferred to the context of another, second parent Fc Targets variant. This is done by determining the “equivalent” or “cor- 40 Virtually any antigen may be targeted by the Fc variants of responding residues and substitutions between the first and the present invention, including but not limited to proteins, second Fc variants, typically based on sequence or structural Subunits, domains, motifs, and/or epitopes belonging to the homology between the sequences of the two Fc variants. In following list of targets: 17-IA, 4-1 BB, 4Dc, 6-keto-PGF1a, order to establish homology, the amino acid sequence of a 8-iso-PGF2a, 8-oxo-dG, A1 Adenosine Receptor, A33, ACE, first Fc variant outlined herein is directly compared to the 45 ACE-2, Activin, Activin A, Activin AB, Activin B, Activin C, sequence of a second Fc variant. After aligning the sequences, Activin RIA, Activin RIA ALK-2, Activin RIBALK-4, using one or more of the homology alignment programs Activin RIIA, Activin RIIB, ADAM, ADAM10, ADAM12, known in the art (for example using conserved residues as ADAM15, ADAM17/TACE, ADAM8, ADAM9, ADAMTS, between species), allowing for necessary insertions and dele ADAMTS4, ADAMTS5, Addressins, aFGF, ALCAM, ALK, tions in order to maintain alignment (i.e., avoiding the elimi- 50 ALK-1, ALK-7, alpha-1-antitrypsin, alpha-V/beta-1 antago nation of conserved residues through arbitrary deletion and nist, ANG, Ang, APAF-1, APE, APJ, APP. APRIL, AR, ARC, insertion), the residues equivalent to particular amino acids in ART Artemin, anti-Id, ASPARTIC, Atrial natriuretic factor, the primary sequence of the first Fc variant are defined. Align av/b3 integrin, Axl b2M, B7-1, B7-2, B7-H, B-lymphocyte ment of conserved residues preferably should conserve 100% Stimulator (BlyS), BACE, BACE-1, Bad, BAFF, BAFF-R, of such residues. However, alignment of greater than 75% or 55 Bag-1, BAK. Bax, BCA-1, BCAM, Bcl, BCMA, BDNF, as little as 50% of conserved residues is also adequate to b-ECGF, bFGF, BID, Bik, BIM, BLC, BL-CAM, BLK, BMP define equivalent residues. Equivalent residues may also be BMP-2 BMP-2a, BMP-3 Osteogenin, BMP-4. BMP-2b, defined by determining structural homology between a first BMP-5, BMP-6 Vgr-1, BMP-7 (OP-1), BMP-8 (BMP-8a, and second Fc variant that is at the level of tertiary structure OP-2), BMPR, BMPR-IA (ALK-3), BMPR-IB (ALK-6), for Fc variants whose structures have been determined. In this 60 BRK-2, RPK-1, BMPR-II (BRK-3), BMPs, b-NGF, BOK, case, equivalent residues are defined as those for which the Bombesin, Bone-derived neurotrophic factor, BPDE, BPDE atomic coordinates of two or more of the main chain atoms of DNA, BTC, complement factor 3 (C3), C3a, C4, C5, C5a, a particular amino acid residue of the parent or precursor (N C10, CA125, CAD-8, Calcitonin, cAMP. carcinoembryonic on N, CA on CA, C on C and O on O) are within 0.13 nm and antigen (CEA), carcinoma-associated antigen, Cathepsin A, preferably 0.1 nm after alignment. Alignment is achieved 65 Cathepsin B, Cathepsin C/DPPI, Cathepsin D. Cathepsin E, after the best model has been oriented and positioned to give Cathepsin H. Cathepsin L. Cathepsin O. Cathepsin S, Cathe the maximum overlap of atomic coordinates of non-hydrogen psin V. Cathepsin X/ZIP CBL, CCI, CCK2, CCL CCL1,

US 8,188,231 B2 33 34 CD153), TNFSF9 (4-1BB Ligand CD137 Ligand), TP-1, In some cases such as the CD30 antigen, this clustering with t-PA, Tpo, TRAIL, TRAIL R, TRAIL-R1, TRAIL-R2, free antibody may be insufficient to cause apoptosis in vitro. TRANCE, transferring receptor, TRF, Trk, TROP-2, TSG, For in vitro assays Sufficient clustering can be mediated by TSLP. tumor-associated antigen CA 125, tumor-associated crosslinking the antibody or by immobilizing it at high den antigen expressing Lewis Y related carbohydrate, TWEAK, sity to a surface such as the well of a microtiter plate. How TXB2, Ung, uPAR, uPAR-1, Urokinase, VCAM, VCAM-1, ever, in vivo this effect may be mediated by binding of the VECAD, VE-Cadherin, VE-cadherin-2, VEFGR-1 (filt-1), antibody to the Fc ligands, for example FcyRs expressed on a VEGF, VEGFR, VEGFR-3 (fit-4), VEGI, VIM, Viral anti nearby cell. Antibody Fc variants that bind more tightly to Fc gens, VLA, VLA-1, VLA-4, VNR integrin, von Willebrands ligands may thus more effectively cluster the signaling target factor, WIF-1, WNT1, WNT2, WNT2B/13, WNT3, WNT3A, 10 and lead to enhanced induction of apoptosis. Such a mecha WNT4, WNT5A, WNTSB, WNT6, WNT7A, WNT7B, nism could be tested experimentally by adding antibody with WNT8A, WNT8B, WNT9A, WNT.9A, WNT9B, WNT10A, and without enhanced Fc ligand binding to cells expressing WNT10B, WNT11, WNT16, XCL1, XCL2, XCR1, XCR1, the desired target that signals, and/or adding an Fc receptor XEDAR, XIAP, XPD, and receptors for hormones and and a corresponding antibody that will cluster the Fc receptor. growth factors. 15 Alternative means for clustering Fc receptor include immo One skilled in the art will appreciate that the aforemen bilization on beads, and over-expression in a non-effector cell tioned list of targets refers not only to specific proteins and line. After allowing apoptosis to occur, measurement of the biomolecules, but the biochemical pathway or pathways that relative apoptosis of target expressing cells would enable a comprise them. For example, reference to CTLA-4 as a target quantitiative determination of the effect. antigen implies that the ligands and receptors that make up the Antibodies that cause cell death through their interaction T cell co-stimulatory pathway, including CTLA-4, B7-1, with targets may have an additional benefit. The signals B7-2, CD28, and any other undiscovered ligands or receptors released by Such dying cells attract macrophages and other that bind these proteins, are also targets. Thus target as used cells of the immune system. These cells can then takeup the herein refers not only to a specific biomolecule, but the set of dead or dying cells in an antibody mediated manner. This has proteins that interact with said target and the members of the 25 been shown to result in cross-presentation of antigen and the biochemical pathway to which said target belongs. One potential for a host immune response against the target cells. skilled in the art will further appreciate that any of the afore Such auto-antibodies in response to antibody therapy have mentioned target antigens, the ligands or receptors that bind been reported for the antigen targets Her2 and CD20. For this them, or other members of their corresponding biochemical reason it may be advantageous to have Fc variants with pathway, may be operably linked to the Fc variants of the 30 altered receptor specificities to specifically stimulate cross present invention in order to generate an Fc fusion. Thus for presentation and an immune response rather than the undes example, an Fc fusion that targets EGFR could be constructed ired effect of tolerance induction. by operably linking an Fc variant to EGF, TGF-B, or any other Other therapeutic antibodies exert their effects by inhibit ligand, discovered or undiscovered, that binds EGFR. ing interaction between a receptor and its cognate ligand, Accordingly, an Fc variant of the present invention could be 35 ultimately blocking signaling of the receptor. Suchantibodies operably linked to EGFR in order to generate an Fc fusion that are used to treat many disease states. In this case it may be binds EGF, TGF-B, or any other ligand, discovered or undis advantageous to utilize antibodies that do not recruit any host covered, that binds EGFR. Thus virtually any polypeptide, immune functions. A secondary effect of Such an antibody whether a ligand, receptor, or Some other protein or protein may be actually inducing signalling itself through receptor domain, including but not limited to the aforementioned tar 40 clustering. In this case the desired therapeutic effect of block gets and the proteins that compose their corresponding bio ing signaling would be abrogated by antibody mediated sig chemical pathways, may be operably linked to the Fc variants naling. As discussed above, this clustering may be enhanced of the present invention to develop an Fc fusion. by antibody interaction with cells containing an Fc receptor. Choosing the right target antigen for antibody therapy is a In this case, use of an Fc variant that binds less tightly or not complex process and encompasses many variables. For anti 45 at all to the Fc receptor would be preferable. Such an antibody cancer treatment it is desirable to have a target whose expres would not mediate signaling, and its mechanism of action sion is restricted to the cancerous cells. Some targets that have would thereby be restricted to blockage of receptor/ligand proven especially amenable to antibody therapy are those interactions. Signaling receptors for which this would be with signaling functions. Other therapeutic antibodies exert most appropriate would likely be monomeric receptors which their effects by blocking signaling of the receptor by inhibit 50 can only be dimerized but not substantially clustered by a ing the binding between a receptor and its cognate ligand. primary antibody. Mulitimeric receptors may be significantly Another mechanism of action of therapeutic antibodies is to clustered by the primary antibody and may not require addi cause receptor down regulation. Although many therapeuti tional clustering by Fc receptor binding. cally effective antibodies work in part by signaling through Another potential mechanism of action of therapeutic anti their target antigen, this is not always the case. For example, 55 bodies is receptor downregulation. Such may be the case, for Some target classes such as cell Surface glycoforms do not example, with the insulin-like growth factor receptor. Cell generate any biological signal. However, altered glycoforms growth depends on continued signaling through the receptor, are often associated with disease states Such as cancer. whereas in its absence cells cease to grow. One effect of Another significant target type are those that internalize either antibodies directed against this receptor is to downregulate its as a normalfunction or in response to antibody binding. In the 60 expression and thereby ablate signaling. Cell recovery from case of targets that are soluble rather than cell surface bound cytotoxic therapy requires stimulation of this receptor. Down the recruitment of effector functions would not result in any regulation of this receptor prevents these cells from recovery cell death. and renders the cytotoxic therapy substantially more effec Some targets that have proven especially amenable to anti tive. For antibodies for which this is the primary mechanism body therapy are those with signalling functions. For 65 of action, decreased Fc receptor binding may prevent the example, antibody cross-linking of the Her2/neu antigen may sequestration of antibody by nontarget binding to Fc recep generate an apoptotic signal that results in cancer cell death. tOrS. US 8,188,231 B2 35 36 Although many therapeutically effective antibodies work candidates’. Thus in a preferred embodiment, the Fc in part by signaling through their target antigen, this is not polypeptides of the present invention may find use in a range always the case. For example, some target classes such as cell of clinical products and candidates. For example, a number of Surface glycoforms do not generate any biological signal. antibodies that target CD20 may benefit from the Fc polypep However, altered glycoforms are often associated with dis tides of the present invention. For example the Fc polypep ease states Such as cancer. In other cases, interaction of anti tides of the present invention may find use in an antibody that bodies with different epitopes of the same target antigen may is substantially similar to rituximab (Rituxan.R., IDEC/Ge confer different signaling effects. In Such cases where this is nentech/Roche) (see for example U.S. Pat. No. 5,736,137), a little or no elicited signaling by binding of antibody or Fc chimeric anti-CD20 antibody approved to treat Non fusion to target antigen, Fc polypeptides of the present inven 10 Hodgkin’s lymphoma; HuMax-CD20, an anti-CD20 cur tion may find utility in providing novel mechanisms of effi rently being developed by Genmab, an anti-CD20 antibody cacy for otherwise non-efficacious molecules. described in U.S. Pat. No. 5,500,362, AME-133 (Applied One approach that has been taken in generating therapeutic Molecular Evolution), ha20 (Immunomedics, Inc.), Huma antibodies to such nonsignaling targets is to couple the anti LYM (Intracel), and PRO70769 (PCT/US2003/04.0426, body to a cytotoxic agent Such as a radio-isotope, toxin, oran 15 entitled “Immunoglobulin Variants and Uses Thereof). A enzyme that will process a Substrate to produce a cytotoxic number of antibodies that target members of the family of agent in the vicinity of the tumor. As an alternative to a epidermal growth factor receptors, including EGFR (ErbB cytotoxic moiety, Fc variants of the present invention may 1), Her2/neu (ErbB-2), Her3 (ErbB-3), Her4 (ErbB-4), may provide increased recruitment of immune functions that are benefit from the Fc polypeptides of the present invention. For inherently less toxic to the host while still effective at destroy example the Fc polypeptides of the present invention may find ing target cancer cells. Such Fc variants may, for example, be use in an antibody that is Substantially similar to trastuzumab more efficient at recruiting NK cells or at activating phago (Herceptin R, Genentech) (see for example U.S. Pat. No. cytosis or initiating CDC. Alternatively, if a cytotoxic agentis 5,677,171), a humanized anti-Her2/neu antibody approved to utilized, it may be advantageous to use an Fc variant that treat breast cancer; pertuzumab (rhuMab-2C4, OmnitargTM), provides reduced or altered Fc ligand binding. This may 25 currently being developed by Genentech; an anti-Her2 anti reduce or ablate the cytotoxic effects of the agent on immune body described in U.S. Pat. No. 4,753,894; cetuximab (Er cells that express Fc receptors, thereby reducing toxicity to bitux(R), Imclone) (U.S. Pat. No. 4,943,533; PCT WO the patient. Furthermore, reduction of Fc ligand binding may 96/40210), a chimeric anti-EGFR antibody in clinical trials help to minimize the generation of an immune response to the for a variety of cancers; ABX-EGF (U.S. Pat. No. 6,235,883), toxic agent or enzyme. As mentioned above, cell death may 30 currently being developed by Abgenix-Immunex-Amgen, result in recruitment of host immune cells; antibody mediated HuMax-EGFr (U.S. Ser. No. 10/172,317), currently being cross-presentation in such a case may be increased with developed by Genmab: 425, EMD55900, EMD62000, and immune response rather than immune tolerance if in addition EMD72000 (Merck KGaA) (U.S. Pat. No. 5,558,864; Mur to a cytotoxic moiety the therapeutic antibody has increased thy et al. 1987, Arch Biochem Biophy's. 252 (2):549-60; Fc receptor binding affinity or altered receptor specificity. 35 Rodecket al., 1987, J Cell Biochem. 35 (4):315-20; Kettle Another significant target type are those targets that inter borough et al., 1991, Protein Eng. 4 (7):773-83); ICR62 (In nalize, eitheras a normal part of their biological function or in stitute of Cancer Research) (PCT WO95/20045; Modjtahedi response to antibody binding. For Such targets, many efforts et al., 1993, J. Cell Biophys. 1993, 22 (1-3):129-46; Modjta have been made to couple cytotoxic agents such as RNase, hedi et al., 1993, BrJ Cancer: 1993, 67 (2):247-53: Modjta ricin and calicheamicin, which can only exert their effect after 40 hedietal, 1996, BrJCancer, 73 (2):228-35; Modjtahedietal, internalization. For Such reagents, Fc ligand binding may 2003, Int J Cancer, 105 (2):273-80); TheraCIM hR3 (YM reduce efficacy due to nonproductive sequestration of the Biosciences, Canada and Centro de Immunologia Molecular, therapeutic by Fc ligands. In this case it may be advantageous Cuba (U.S. Pat. No. 5,891,996; U.S. Pat. No. 6,506,883: to utilize Fc variants that provide decreased Fc ligandaffinity. Mateo et al., 1997, Immunotechnology, 3 (1):71-81); mab Conversely, antibody pre-association with Fc ligands prior to 45 806 (Ludwig Institue for Cancer Research, Memorial Sloan their binding to target antigen presented on cells may serve to Kettering) (Jungbluth et al. 2003, Proc Natl AcadSci USA. inhibit internalization of the target. In this case, increased Fc 100 (2): 639-44); KSB-102 (KS Biomedix); MR1-1 (IVAX, ligand affinity may serve to improve pre-association and National Cancer Institute) (PCT WO 0162931A2); and thereby recruitment of effector cells and the host immune SC100 (Scancell) (PCTWO 01/88138). In another preferred response. 50 embodiment, the Fc polypeptides of the present invention In the case of targets that are soluble rather than cell surface may find use in alemtuzumab (Campath(R), Millenium), a bound, recruitment of effector functions would not result humanized monoclonal antibody currently approved for directly in cell death. However, there may be utility in stimu treatment of B-cell chronic lymphocytic leukemia. The Fc lating the generation of hostantibodies to the target. For some polypeptides of the present invention may find use in a variety disease states, successful treatment may require administra 55 of antibodies or Fc fusions that are substantially similar to tion of the therapeutic antibody for extremely long periods of other clinical products and candidates, including but not lim time. Such therapy may be prohibitively costly or cumber ited to muromonab-CD3 (Orthoclone OKT3(R), an anti-CD3 some. In these cases, stimulation of the host immune response antibody developed by Ortho Biotech/Johnson & Johnson, and the generation of antibodies may result in improved effi ibritumomab tiuxetan (Zevalin R), an anti-CD20 antibody cacy of the therapeutic. This may be applicable as an adjuvant 60 developed by IDEC/Schering AG, gemtuzumab ozogamicin to vaccine therapy. Antibody Fc variants that mediate such an (Mylotargr), an anti-CD33 (p67 protein) antibody devel effect may have increased affinity for Fc ligands or altered Fc oped by Celltech/Wyeth, alefacept (AmeviveR), an anti ligand specificity. LFA-3 Fc fusion developed by Biogen), abciximab (Re A number of antibodies and Fc fusions that are approved oPro(R), developed by Centocor/Lilly, basiliximab for use, in clinical trials, or in development may benefit from 65 (Simulect(R), developed by Novartis, palivizumab (Syn the Fc variants of the present invention. These antibodies and agis(R), developed by MedImmune, infliximab (Remicade(R). Fc fusions are herein referred to as “clinical products and an anti-TNFalpha antibody developed by Centocor, adali US 8,188,231 B2 37 38 mumab (HumiraR), an anti-TNFalpha antibody developed anti-Her2 antibody being developed by Medarex and by Abbott, HumicadeTM, an anti-TNFalpha antibody devel Immuno-Designed Molecules, HuMaxTM-CD4, an anti-CD4 oped by Celltech, etanercept (Enbrel(R), an anti-TNFalpha Fc antibody being developed by Medarex and Genmab, HuMax fusion developed by Immunex/Amgen, ABX-CBL, an anti IL15, an anti-IL 15 antibody being developed by Medarex and CD147 antibody being developed by Abgenix, ABX-IL8, an Genmab, CNTO 148, an anti-TNFC. antibody being devel anti-IL8 antibody being developed by Abgenix, ABX-MA1, oped by Medarex and Centocor/J&J, CNTO 1275, an anti an anti-MUC18 antibody being developed by Abgenix, Pem cytokine antibody being developed by Centocor/J&J, tumomab (R1549, 'Y-muHMFG1), an anti-MUC1 In devel MOR101 and MOR102, anti-intercellular adhesion mol opment by Antisoma, Therex (R1550), an anti-MUC1 anti ecule-1 (ICAM-1) (CD54) antibodies being developed by body being developed by Antisoma, AngioMab (AS1405), 10 being developed by Antisoma, HuBC-1, being developed by MorphoSys, MOR201, an anti-fibroblast growth factor Antisoma. Thioplatin (AS1407) being developed by Anti receptor 3 (FGFR-3) antibody being developed by Morpho Soma, Antegren(R) (natalizumab), an anti-alpha-4-beta-1 Sys, Nuvion(R) (visilizumab), an anti-CD3 antibody being (VLA-4) and alpha-4-beta-7 antibody being developed by developed by Protein Design Labs, Hu ZAFTM, an anti Biogen, VLA-1 mAb, an anti-VLA-1 integrin antibody being 15 gamma interferon antibody being developed by Protein developed by Biogen, LTBR mAb, an anti-lymphotoxin beta Design Labs, Anti-C.5 B1 Integrin, being developed by Protein receptor (LTBR) antibody being developed by Biogen, CAT Design Labs, anti-IL-12, being developed by Protein Design 152, an anti-TGF-B2 antibody being developed by Cam Labs, ING-1, an anti-Ep-CAM antibody being developed by bridge Antibody Technology, J695, an anti-IL-12 antibody Xoma, and MLNO1, an anti-Beta2 integrin antibody being being developed by Cambridge Antibody Technology and developed by Xoma, all references incorporated by reference. Abbott, CAT-192, an anti-TGFB1 antibody being developed Application of the Fc polypeptides to the aforementioned by Cambridge Antibody Technology and Genzyme, CAT antibody and Fc fusion clinical products and candidates is not 213, an anti-Eotaxin1 antibody being developed by Cam meant to be constrained to their precise composition. The Fc bridge Antibody Technology, LymphoStat-BTM an anti-Blys polypeptides of the present invention may be incorporated antibody being developed by Cambridge Antibody Technol 25 into the aforementioned clinical candidates and products, or ogy and Human Genome Sciences Inc., TRAIL-R1mAb, an into antibodies and Fc fusions that are substantially similar to anti-TRAIL-R1 antibody being developed by Cambridge them. The Fc polypeptides of the present invention may be Antibody Technology and Human Genome Sciences, Inc., incorporated into versions of the aforementioned clinical can AvastinTM (bevacizumab, rhuMAb-VEGF), an anti-VEGF didates and products that are humanized, affinity matured, antibody being developed by Genentech, an anti-HER recep 30 engineered, or modified in Some other way. Furthermore, the tor family antibody being developed by Genentech, Anti entire polypeptide of the aforementioned clinical products Tissue Factor (ATF), an anti-Tissue Factor antibody being and candidates need not be used to construct a new antibody developed by Genentech, XolairTM (Omalizumab), an anti or Fc fusion that incorporates the Fc polypeptides of the IgE antibody being developed by Genentech, RaptivaTM present invention; for example only the variable region of a (Efalizumab), an anti-CD11a antibody being developed by 35 clinical product or candidate antibody, a Substantially similar Genentech and Xoma, MLN-02 Antibody (formerly LDP variable region, or a humanized, affinity matured, engineered, 02), being developed by Genentech and Millenium Pharma or modified version of the variable region may be used. In ceuticals, HuMax CD4, an anti-CD4 antibody being devel another embodiment, the Fc polypeptides of the present oped by Genmab, HuMax-IL15, an anti-IL 15 antibody being invention may find use in an antibody or Fc fusion that binds developed by Genmab and Amgen, HuMax-Inflam, being 40 to the same epitope, antigen, ligand, or receptor as one of the developed by Genmab and Medarex, HuMax-Cancer, an anti aforementioned clinical products and candidates. Heparanase I antibody being developed by Genmab and In one embodiment, the Fc polypeptides of the present Medarex and Oxford GcoSciences, HuMax-Lymphoma, invention are used for the treatment of autoimmune, inflam being developed by Genmab and Amgen, HuMax-TAC, matory, or transplant indications. Target antigens and clinical being developed by Genmab, IDEC-131, and anti-CD40L 45 products and candidates that are relevant for Such diseases antibody being developed by IDEC Pharmaceuticals, IDEC include but are not limited to anti-C4f37 integrin antibodies 151 (Clenoliximab), an anti-CD4 antibody being developed such as LDP-02, anti-beta2 integrin antibodies such as LDP by IDEC Pharmaceuticals, IDEC-114, an anti-CD80 anti 01, anti-complement (C5) antibodies such as 5G1.1, anti body being developed by IDEC Pharmaceuticals, IDEC-152, CD2 antibodies such as BTI-322, MEDI-507, anti-CD3 anti an anti-CD23 being developed by IDEC Pharmaceuticals, 50 bodies such as OKT3, SMART anti-CD3, anti-CD4 anti-macrophage migration factor (MIF) antibodies being antibodies such as IDEC-151, MDX-CD4, OKT4A, anti developed by IDEC Pharmaceuticals, BEC2, an anti-idio CD11a antibodies, anti-CD14 antibodies such as IC 14, anti typic antibody being developed by Imclone, IMC-1C11, an CD18 antibodies, anti-CD23 antibodies such as IDEC 152, anti-KDR antibody being developed by Imclone, DC101, an anti-CD25 antibodies such as Zenapax, anti-CD40L antibod anti-flk-1 antibody being developed by Imclone, anti-VE cad 55 ies such as 5c8, Antova, IDEC-131, anti-CD64 antibodies herin antibodies being developed by Imclone, CEA-CidetM such as MDX-33, anti-CD80 antibodies such as IDEC-114, (labetuZumab), an anti-carcinoembryonic antigen (CEA) anti-CD 147 antibodies such as ABX-CBL, anti-E-selectin antibody being developed by Immunomedics, Lympho antibodies such as CDP850, anti-gpIb/IIIa antibodies such Cide.M (Epratuzumab), an anti-CD22 antibody being devel as ReoProf Abcixima, anti-ICAM-3 antibodies such as ICM3, oped by Immunomedics, AFP-Cide, being developed by 60 anti-ICE antibodies such as VX-740, anti-FcR1 antibodies Immunomedics, MyelomaCide, being developed by Immu such as MDX-33, anti-IgE antibodies such as rhuMab-E25, nomedics, LkoCide, being developed by Immunomedics, anti-IL-4 antibodies such as SB-240683, anti-IL-5 antibodies ProstaCide, being developed by Immunomedics, MDX-010, such as SB-240563, SCH55700, anti-IL-8 antibodies such as an anti-CTLA4 antibody being developed by Medarex, ABX-IL8, anti-interferon gamma antibodies, anti-TNF MDX-060, an anti-CD30 antibody being developed by 65 (TNF, TNFC, TNFa. TNF-alpha) antibodies such as CDP571, Medarex, MDX-070 being developed by Medarex, MDX CDP870, D2E7, Infliximab, MAK-195F, and anti-VLA-4 018 being developed by Medarex, OsidemTM (IDM-1), and antibodies such as Antegren. US 8,188,231 B2 39 40 Fc variants of the present invention may be utilized in TNF bodies such as Crucells anti-EpCAM and Edrecolomab, and inhibitor molecules to provide enhanced properties. It has anti-CD52 antibodies such as Campath(R) (alemtuzumab). been shown that the effector function associated with In one embodiment, the the Fc polypeptides of the present FcyRIIIa may negatively impact the effectiveness of certain invention are directed againstantigens expressed in the hema TNF inhibitor molecules used in the treatment of rheumatoid 5 tological lineage. Target antigens and clinical products and arthritis or psoriatic arthritis patients that have a high-affinity candidates that are relevant for Such application may include polymorphism (158 F:V discussed herein elsewhere) and but are not limited to: anti-CD33 antibodies such as Smart vice-versa (Z. Tutuncu et al., 2004, “FcRPolymorphisms and M195, anti-CD40L antibodies such as AntovaTM, IDEC-131, Treatment Outcomes in Patients with Inflammatory Arthritis anti-CD44 antibodies such as Blvatuzumab, anti-CD52 anti Treated with TNF Blocking Agents, oral presentation on 10 bodies such as Campath(R) (alemtuzumab), anti-CD80 anti bodies such as IDEC-114, anti-CTLA-4 antibodies such as Oct. 18, 2004 at the 2004 ACR Meeting, San Antonio, Tex.; MDX-101, anti-CD20 antibodies such as Bexocar, Rit abstract published in Arthritis & Rheumatism, September uxan.R., Zevalin R, and PRO70769, anti-CD22 antibodies 2004, incorporated by reference). In general for autoimmune such as Lymphocide"M, anti-CD23 antibodies such as IDEC conditions such as rheumatoid arthritis or psoriatic arthritis, 15 152, anti-CD25 antibodies such as Zenapax(R) (daclizumab), combining a TNF inhibitor with an Fc variant that provides and anti-MHC (HLA-DR) antibodies such as apolizumab. reduced binding to one or more FcyRs as compared to the In one embodiment, the Fc polypeptides of the present parent enhances the effectiveness of therapy. Ideally, reduced invention are directed against antigens expressed in Solid or even ablated binding to one or more FcyRs, for example tumors. Target antigens and clinical products and candidates FcyRIIIa, with a TNF inhibitor molecule would produce the that are relevant for Such application may include but are not best results. limited to: anti-EpCAM antibodies such as Crucells anti Useful TNF inhibitor molecules include any molecule that EpCAM and Edrecolomab, anti-CEA antibodies such as can inhibits the action of TNF-alpha in a mammal. Suitable tumab and Pentacea, anti-EGFR antibodies such as ABX examples include the Fc fusion Enbrel(R) (etanercept) and the EGF, Cetuximab, IMC-C225, Merck Mab 425, anti-Muc1 antibodies Humira R. (adalimumab) and Remicade(R) (inflix 25 antibodies such as BravaRex, TriAb, anti-Her2 antibodies imab). Monoclonal antibodies (such as Remicade and such as HerceptinR, MDX-210, anti-GD-2 ganglioside anti Humira) engineered using the Fc variants of the present bodies such as 3F8 and TriGem, anti-GD-3 ganglioside anti invention to reduce Fc binding, may translate to better effi bodies such as mitumomab, anti-PSMA antibodies such as cacy. Effector function of Humira, Remicade, and Enbrel was MDX-070, anti-CA125 antibodies such as oregovomab, anti not considered in the development of these drugs, let alone 30 TAG-72 antibodies such as MDX-220, and anti-MUC-1 anti modulation of effector function. By using an Fc variant of the bodies such as cantuzumab. present invention that reduces binding to one or more FcyRs In a preferred embodiment, the target of the Fc variants of in the context of an antibody or Fc fusion that acts on autoim the present invention is itself one or more Fc ligands. Fe mune conditions, efficacy may be enhanced as compared to polypeptides of the invention can be utilized to modulate the the currently commercialized products. Useful TNF inhibitor 35 activity of the immune system, and in Some cases to mimic the molecules preferably include Dominant Negative TNF mol effects of IVIg therapy in a more controlled, specific, and ecules (as defined in U.S. Ser. No. 09/798,789, filed Mar. 2, efficient manner. IVIg is effectively a high dose of immuno 2000; Ser. No. 09/981,289, filed Oct. 15, 2001; Ser. No. globulins delivered intravenously. In general. IVIg has been 10/262,630, filed Sep. 30, 2002; and Ser. No. 10/963,994, used to downregulate autoimmune conditions. It has been filed Oct. 12, 2004, all incorporated by reference). The Domi 40 hypothesized that the therapeutic mechanism of action of nant Negative TNF molecules (DN-TNF) have no intrinsic IVIg involves ligation of Fc receptors at high frequency (J. effector activity, and act to “save' transmembrane TNF Bayry et al., 2003, Transfusion Clinique et Biologique 10: (tmTNF) (i.e., if the killing of cells that contain timTNF has a 165-169: Binstadt et al., 2003, J. Allergy Clin. Immunol, negative effect on disease outcome for rheumatoid or psori 697-704). Indeed animal models of Ithrombocytopenia pur atic arthritis). A DN-TNF molecule associated with an Fc 45 pura (ITP) show that the isolated Fc are the active portion of variant that reduces or ablates FcyR binding to the receptor is IVIg (Samuelsson et al., 2001, Pediatric Research 50 (5), preferred. 551). For use in therapy, immunoglobulins are harvested from In one embodiment, the Fc polypeptides of the present thousands of donors, with all of the concomitant problems invention function therapeutically, in whole or in part, associated with non-recombinant biotherapeutics collected through ADCC activity. Target antigens and clinical products 50 from humans. An Fc variant of the present invention should and candidates that are relevant for Such application may serve all of the roles of IVIg while being manufactured as a include but are not limited to: anti-CD20 antibodies such as recombinant protein rather than harvested from donors. Bexocar, Rituxan R), Zevalin R, and PRO70769, anti-CD33 The immunomodulatory effects of IVIg may be dependent antibodies such as Smart M195, anti-CD22 antibodies such as on productive interaction with one or more Fc ligands, includ Lymphocide"M, anti-CD30 antibodies such as AC-10 and 55 ing but not limited to FcyRs, complement proteins, and FcRn. SGN-30, anti-EGFR antibodies such as ABX-EGF, Cetux In some embodiments, Fc variants of the invention with imab, IMC-C225, Merck Mab 425, anti-EpCAM antibodies enhanced affinity for FcyRIIb can be used to promote anti such as Crucells anti-EpCAM, anti-HER2 antibodies such as inflammatory activity (Samuelsson et al., 2001, Science 291: Herceptin and MDX-210, and anti-CEA antibodies such as 484-486) and or to reduce autoimmunity (Hogarth, 2002, cantumab and Pentacea. 60 Current Opinion in Immunology, 14:798-802). In other In one embodiment, the Fc polypeptides of the present embodiments, Fc polypeptides of the invention with invention function therapeutically, in whole or in part, enhanced affinity for one or more FcyRs can be utilized by through CDC activity. Target antigens and clinical products themselves or in combination with additional modifications and candidates that are relevant for Such application may to reduce autoimmunity (Hogarth, 2002, Current Opinion in include but are not limited to: anti-CEA antibodies such as 65 Immunology, 14:798-802). In alternative embodiments, Fc cantumab and Pentacea, anti-CD20 antibodies such as Bexo variants of the invention with enhanced affinity for FcyRIIIa car, Rituxan R, Zevalin R, and PRO70769, anti-EpCAM anti but reduced capacity for intracellular signaling can be used to US 8,188,231 B2 41 42 reduce immune system activation by competitively interfer relative to a parent Fc polypeptide, wherein said amino acid ing with FcyRIIIa binding. The context of the Fc variant modification(s) provide one or more optimized properties. An drammatically impacts the desired specificity. For example, Fc variant of the present invention differs in amino acid Fc variants that provide enhanced binding to one or more sequence from its parent Fc polypeptide by virtue of at least activating FcyRs may provide optimal immunomodulatory one amino acid modification. Thus Fc variants of the present effects in the context of an antibody, Fc fusion, isolated Fc, or invention have at least one amino acid modification compared Fc fragment by acting as an FcyRantagonist (van Mirre et al., to the parent. Alternatively, the Fc variants of the present 2004, J. Immunol. 173:332-339). However, fusion or conju invention may have more than one amino acid modification as gation of two or more Fc variants may provide different compared to the parent, for example from about one to fifty effects, and for Such an Fc polypeptide it may be optimal to 10 amino acid modifications, preferrably from about one to ten utilize Fc variants that provide enhanced affinity for an inhibi amino acid modifications, and most preferably from about tory receptor. one to about five amino acid modifications compared to the The Fc variants of the present invention may be used as parent. Thus the sequences of the Fc variants and those of the immunomodulatory therapeutics. Binding to or blocking Fc parent Fc polypeptide are Substantially homologous. For receptors on immune system cells may be used to influence 15 example, the variant Fc variant sequences herein will possess immune response in immunological conditions including but about 80% homology with the parent Fc variant sequence, not limited to idiopathic thrombocytopenia purpura (ITP) and preferably at least about 90% homology, and most preferably rheumatoid arthritis (RA) among others. By use of the affinity at least about 95% homology. enhanced Fc variants of the present invention, the dosages The Fc variants of the present invention may be optimized required in typical IVIg applications may be reduced while for a variety of properties. An Fc variant that is engineered or obtaining a substantially similar therapeutic effect. The Fc predicted to display one or more optimized properties is variants may provide enhanced binding to an FcyR, including herein referred to as an “optimized Fc variant'. Properties but not limited to FyRIIa, FcyRIIb, FcyRIIIa, FcyRIIIb, and/or that may be optimized include but are not limited to enhanced FcyRI. In particular, binding enhancements to FcyRIIb would or reduced affinity for an FcyR. In a preferred embodiment, increase expression or inhibitory activity, as needed, of that 25 the Fc variants of the present invention are optimized to receptor and improve efficacy. Alternatively, blocking bind possess enhanced affinity for a human activating FcyR, pref ing to activation receptors such as FcyRIIIb or FcyRI may erably FcyRI, FcyRIIa, FcyRIIc, FcyRIIIa, and FcyRIIb, most improve efficacy. In addition, modulated affinity of the Fc preferably FcyRIIIa. In an alternately preferred embodiment, variants for FcRn and/or also complement may also provide the Fc variants are optimized to possess reduced affinity for benefits. 30 the human inhibitory receptor FcyRIIb. These preferred In one embodiment, Fc variants that provide enhanced embodiments are anticipated to provide Fc polypeptides with binding to the inhibitory receptor FcyRIIb provide an enhanced therapeutic properties in humans, for example enhancement to the IVIg therapeutic approach. In particular, enhanced effector function and greater anti-cancer potency. the Fc variants of the present invention that bind with greater In an alternate embodiment, the Fc variants of the present affinity to the FcyRIIb receptor than parent Fc polypeptide 35 invention are optimized to have reduced or ablated affinity for may be used. Such Fc variants would thus function as FcyRIIb a human FcyR, including but not limited to FcyRI, FcyRIIa, agonists, and would be expected to enhance the beneficial FcyRIIb, FcyRIIc, FcyRIIa, and FcyRIIb. These embodiments effects of IVIgas an autoimmune disease therapeutic and also are anticipated to provide Fc polypeptides with enhanced as a modulator of B-cell proliferation. In addition, such therapeutic properties in humans, for example reduced effec FcyRIIb-enhanced Fc variants may also be further modified 40 tor function and reduced toxicity. In other embodiments, Fc to have the same or limited binding to other receptors. In variants of the present invention provide enhanced affinity for additional embodiments, the Fc variants with enhanced one or more FcyRs, yet reduced affinity for one or more other FcyRIIb affinity may be combined with mutations that reduce FcyRs. For example, an Fc variant of the present invention or ablate to other receptors, thereby potentially further mini may have enhanced binding to FcyRIIIa, yet reduced binding mizing side effects during therapeutic use. 45 to FcyRIIb. Alternately, an Fc variant of the present invention Such immunomodulatory applications of the Fc variants of may have enhanced binding to FcyRIIa and FcyRI, yet the present invention may also be utilized in the treatment of reduced binding to FcyRIIb. In yet another embodiment, an oncological indications, especially those for which antibody Fc variant of the present invention may have enhanced affinity therapy involves antibody-dependant cytotoxic mechanisms. for FcyRIIb, yet reduced affinity to one or more activating For example, an Fc variant that enhances affinity to FcyRIIb 50 FcyRs. may be used to antagonize this inhibitory receptor, for Preferred embodiments comprise optimization of Fc bind example by binding to the Fc/FcyRIIb binding site but failing ing to a human FcyR, however in alternate embodiments the to trigger, or reducing cell signaling, potentially enhancing Fc variants of the present invention possess enhanced or the effect of antibody-based anti-cancer therapy. Such Fc reduced affinity for FcyRs from nonhuman organisms, variants, functioning as FcyRIIb antagonists, may either 55 including but not limited to rodents and non-human primates. block the inhibitory properties of FcyRIIb, or induce its Fc variants that are optimized for binding to a nonhuman inhibitory function as in the case of IVIg. An FcyRIIb antago FcyR may find use in experimentation. For example, mouse nist may be used as co-therapy in combination with any other models are available for a variety of diseases that enable therapeutic, including but not limited to antibodies, acting on testing of properties Such as efficacy, toxicity, and pharmaco the basis of ADCC related cytotoxicity. FcyRIIb antagonistic 60 kinetics for a given drug candidate. As is known in the art, Fc variants of this type are preferably isolated Fc or Fc frag cancer cells can be grafted or injected into mice to mimic a ments, although in alternate embodiments antibodies and Fc human cancer, a process referred to as Xenografting. Testing fusions may be used. of Fc variants that comprise Fc variants that are optimized for Optimized Properties one or more mouse FcyRs, may provide valuable information The present invention provides Fc variants that are opti 65 with regard to the efficacy of the protein, its mechanism of mized for a number of therapeutically relevant properties. An action, and the like. The Fc variants of the present invention Fc variant comprises one or more amino acid modifications may also be optimized for enhanced functionality and/or US 8,188,231 B2 43 44 Solution properties in aglycosylated form. In a preferred ficity of a given Fc variant will provide different properties embodiment, the aglycosylated Fc variants of the present depending on whether it composes an antibody, Fc fusion, or invention bind an Fc ligand with greater affinity than the an Fc variants with a coupled fusion or conjugate partner. For aglycosylated form of the parent Fc variant. Said Fc ligands example, toxin, radionucleotide, or other conjugates may be include but are not limited to FcyRs, C1q, FcRn, and proteins less toxic to normal cells if the Fc variant that comprises them A and G, and may be from any source including but not has reduced or ablated binding to one or more Fc ligands. As limited to human, mouse, rat, rabbit, or monkey, preferably another example, in order to inhibit inflammation or auto human. In an alternately preferred embodiment, the Fc vari immune disease, it may be preferable to utilize an Fc variant ants are optimized to be more stable and/or more soluble than with enhanced affinity for activating FcyRs, such as to bind the aglycosylated form of the parent Fc variant. 10 these FcyRs and prevent their activation. Conversely, an Fc Fc variants of the invention may comprise modifications variant that comprises two or more Fc regions with enhanced that modulate interaction with Fc ligands other than FcyRs, FcyRIIb affinity may co-engage this receptor on the surface of including but not limited to complement proteins, FcRn, and immune cells, thereby inhibiting proliferation of these cells. Fc receptor homologs (FcRHs). FcRHs include but are not Whereas in Some cases an Fc variants may engage its target limited to FcRH1, FcRH2, FcRH3, FcRH4, FcRH5, and 15 antigen on one cell type yet engage FcyRs on separate cells FcRH6 (Davis et al., 2002, Immunol. Reviews 190: 123-136). from the target antigen, in other cases it may be advantageous Preferably, the Fc ligand specificity of the Fc variant of the to engage FcyRs on the Surface of the same cells as the target present invention will determine its therapeutic utility. The antigen. For example, if an antibody targets an antigen on a utility of a given Fc variant for therapeutic purposes will cell that also expresses one or more FcyRs, it may be benefi depend on the epitope or form of the Target antigen and the cial to utilize an Fc variant that enhances or reduces binding disease or indication being treated. For some targets and to the FcyRs on the surface of that cell. This may be the case, indications, enhanced Fcy R-mediated effector functions may for example when the Fc variant is being used as an anti be preferable. This may be particularly favorable for anti cancer agent, and co-engagement of target antigen and FcyR cancer Fc variants. Thus Fc variants may be used that com on the Surface of the same cell promote signaling events prise Fc variants that provide enhanced affinity for activating 25 within the cell that result in growth inhibition, apoptosis, or FcyRs and/or reduced affinity for inhibitory FcyRs. For some other anti-proliferative effect. Alternatively, antigen and targets and indications, it may be further beneficial to utilize FcyR co-engagement on the same cell may be advantageous Fc variants that provide differential selectivity for different when the Fc variant is being used to modulate the immune activating FcyRs; for example, in some cases enhanced bind system in some way, wherein co-engagement of target anti ing to FcyRIIa and FcyRIIIa may be desired, but not FcyRI, 30 gen and FcyR provides some proliferative or anti-prolifera whereas in other cases, enhanced binding only to FcyRIIa tive effect. Likewise, Fc variants that comprise two or more may be preferred. For certain targets and indications, it may Fc regions may benefit from Fc variants that modulate FcyR be preferable to utilize Fc variants that enhance both FcyR selectivity or specificity to co-engage FcyRs on the Surface of mediated and complement-mediated effector functions, the same cell. whereas for other cases it may be advantageous to utilize Fc 35 The Fc ligand specificity of the Fc variants of the present variants that enhance either FcyR-mediated or complement invention can be modulated to create different effector func mediated effector functions. For some Targets or cancer indi tion profiles that may be Suited for particular target antigens, cations, it may be advantageous to reduce or ablate one or indications, or patient populations. Table 1 describes several more effector functions, for example by knocking out binding preferred embodiments of receptor binding profiles that to C1q, one or more FcyRs, FcRn, or one or more other Fc 40 include improvements to, reductions to or no effect to the ligands. For other targets and indications, it may be preferable binding to various receptors, where such changes may be to utilize Fc variants that provide enhanced binding to the beneficial in certain contexts. The receptor binding profiles in inhibitory FcyRIIb, yet WT level, reduced, or ablated binding the table could be varied by degree of increase or decrease to to activating FcyRs. This may be particularly useful, for the specified receptors. Additionally, the binding changes example, when the goal of an Fc variant is to inhibit inflam 45 specified could be in the context of additional binding mation or auto-immune disease, or modulate the immune changes to other receptors such as C1q or FcRn, for example system in Some way. by combining with ablation of binding to C1q to shut off Clearly an important parameter that determines the most complementactivation, or by combining with enhanced bind beneficial selectivity of a given Fc variant to treat a given ing to C1q to increase complement activation. Other embodi disease is the context of the Fc variant, that is what type of Fc 50 ments with other receptor binding profiles are possible, the variant is being used. Thus the Fc ligand selectivity or speci listed receptor binding profiles are exemplary. TABLE 1. Affinity Affinity Enhancement Reduction Cell Activity Therapeutic Activity FcyRI only Enhanced dendritic cell activity Enhanced cell-based and uptake, and Subsequent immune response against presentation of antigens target Enhanced monocyte and macrophage response to antibody FcyRIIIa Enhanced ADCC and Increased target cell lysis phagocytosis of broad range of cell types FcyRIIIa FcyRIIb Enhanced ADCC and Increased target cell lysis phagocytosis of broad range of cell types US 8,188,231 B2 45 46 TABLE 1-continued Affinity Affinity Enhancement Reduction Cell Activity Therapeutic Activity FcyRIIb Reduced activity of all FcyR Enhancement of target cell FcyRIIc bearing cell types except NK lysis selective for NK cell cells accessible target cells Possible activation of NK cells via FcyRIIc receptor signaling FcyRIIb Possible NK cell specific Enhanced target cell lysis FcyRIIIa activation and enhancement of selective for NK cell NK cell mediated ADCC accessible target cells FcyRIIIb Neutrophil mediated Enhanced target cell phagocytosis enhancement destruction for neutrophil accessible cells FcCR Neutrophil mediated Enhanced target cell phagocytosis enhancement destruction for neutrophil accessible cells FcyRI FcyRIIb Enhanced dendritic cell activity Enhanced cell-based FcyRIIa and uptake, and Subsequent immune response against FcyRIIIa presentation of antigens to T arget cells Enhanced monocyte and macrophage response to antibody FcyRIIb FcyRI Reduced activity of monocytes, FcyRIIa macrophages, neutrophils, NK, Eliminated or reduced cell FcyRIIIa dendritic and other gamma mediated cytotoxicity receptor bearing cells against target bearing cells

The presence of different polymorphic forms of FcyRs ligands, for example FcRHS. Additional modifications may provides yet another parameter that impacts the therapeutic provide altered or optimized affinity and/or specificity to the utility of the Fc variants of the present invention. Whereas the 30 Fc ligands. Additional modifications may provide altered or specificity and selectivity of a given Fc variant for the differ optimized effector functions, including but not limited to ent classes of FcyRs significantly affects the capacity of an Fc ADCC, ADCP, CDC, and/or serum half-life. Such combina variant to target a given antigen for treatment of a given tion may provide additive, synergistic, or novel properties. In disease, the specificity or selectivity of an Fc variant for one embodiment, the Fc variants of the present invention may different polymorphic forms of these receptors may in part 35 be combined with known Fc variants (Duncan et al., 1988, determine which research or pre-clinical experiments may be Nature 332:563-564; Lund et al., 1991, JImmunol 147:2657 appropriate for testing, and ultimately which patient popula 2662; Lundet al., 1992, Mol Immunol 29:53-59; Alegreet al., tions may or may not respond to treatment. Thus the speci 1994, Transplantation 57:1537-1543; Hutchins et al., 1995, ficity or selectivity of Fc variants of the present invention to Proc Natl Acad Sci USA 92:11980-1 1984; Jefferis et al., Fc ligand polymorphisms, including but not limited to FcyR, 40 1995, Immunol Lett 44:111-117; Lund et al., 1995, FasebJ C1q, FcRn, and FcRH polymorphisms, may be used to guide 9:115-119; Jefferis et al., 1996, Immunol Lett 54:101-104; the selection of valid research and pre-clinical experiments, Lund et al., 1996, J Immunol 157:4963-4969; Armour et al., clinical trial design, patient selection, dosing dependence, 1999, Eurj Immunol 29:2613-2624: Idusogie et al., 2000, J and/or other aspects concerning clinical trials. Immunol 164:41 78-4184: Reddy et al., 2000, J Immunol 164: Additional Modifications 45 1925-1933; Xu et al., 2000, Cell Immunol 200:16-26: Iduso In addition to comprising an Fc variant of the present gie et al., 2001, J Immunol. 166:2571-2575; Shields et al., invention, the Fc polypeptides of the present invention may 2001, J Biol Chem 276:6591-6604; Jefferis et al., 2002, comprise one or more additional modifications. Said modifi Immunol Lett 82:57-65; Presta et al., 2002, Biochem Soc cations may be amino acid modifications, or may modifica Trans 30:487-490; Hinton et al., 2004, J Biol Chem 279: tions that are not amino acid modifications such as modifica 50 6213-6216) (U.S. Pat. No. 5,624,821; U.S. Pat. No. 5,885, tions that are made enzymatically or chemically. 573; U.S. Pat. No. 6,194,551; PCTWO 00/42072; PCT WO Combinations of additional amino acid modifications and 99/58572; U.S. 2004/0002587 A1), U.S. Pat. No. 6,737,056, modifications that are not amino acid modifications are con PCT US2004/000643, U.S. Ser. No. 10/370,749, and PCT/ templated. Such additional modification(s) likely provide US2004/005112), all incorporated by reference. For Some improvement in the Fc polypeptide, for example an 55 example, as described in U.S. Pat. No. 6,737,056, PCT enhancement in its stability, solubility, function, or clinical US2004/000643, U.S. Ser. No. 10/370,749, and PCT/ use. The present invention contemplates a variety of improve US2004/0051 12, the substitutions S298A, S298D, K326E, ments that made be made by coupling the Fc variants of the K326D, E333A, K334A, and P396L provide optimized FcyR present invention with additional modifications. binding and/or enhanced ADCC. Furthermore, as disclosed in The Fc variants of the present invention may be combined 60 Idusogie et al., 2001, J. Immunology 166:2571-2572, incor with other amino acid modifications in the Fc region that porated by reference, substitutions K326W, K326Y. and provide altered or optimized interaction with one or more Fc E333S provide enhanced binding to the complement protein ligands, including but not limited to FcyRs, C1q or other C1q and enhanced CDC. Finally, as described in Hinton et al., complement proteins, FcRn, FcRhomologues (FcRHs), and/ 2004, J. Biol. Chem. 279 (8): 6213-6216, incorporated by or as yet undiscovered Fc ligands. It is noted that Fc polypep 65 reference, substitutions T250O, T250E, M428L, and M428F tides of the present invention may themselves have as yet provide enhanced binding to FcRn and improved pharmaco unknown useful interaction properties with one or more Fc kinetics. US 8,188,231 B2 47 48 Because the binding sites for FcyRs, C1q, and FcRn reside Additional modifications may comprise amino acid modi in the Fc region, the differences between the IgGs in the Fc fications wherein residues in an Fc polypeptide are modified region are likely to contribute to differences in FcyR- and to the corresponding residue in a homologous Fc polypeptide. C1q-mediated effector functions. It is also possible that the Effector functions such as ADCC, ADCP, CDC, and serum modifications can be made in other non-Fc regions of an Fc 5 half-life differ significantly between the different classes of variant, including for example the Fab and hinge regions of an antibodies, including for example human IgG1, IgG2, IgG3. antibody, or the Fc fusion partner of an Fc fusion. For IgG4, IgA1, IgA2, Ig), IgE, IgG, and IgM (references— example, as disclosed in U.S. Ser. No. 11/090,981, hereby Michaelsen et al., 1992, Molecular Immunology 29 (3): 319 incorporated by reference, the Fab and hinge regions of an 326). Human IgG1 is the most commonly used antibody for antibody may impact effector functions such as antibody 10 dependent cell-mediated cytotoxicity (ADCC), antibody therapeutic purposes, and engineering studies wherein vari dependent cell-mediated phagocytosis (ADCP), and comple ants have been constructed that show enhanced effector func ment dependent cytotoxicity (CDC). Thus modifications out tion have been carried out predominantly in this context. As side the Fc region of an Fc variant of the present invention are described above, it is possible to determine corresponding or contemplated. For example, antibodies of the present inven 15 equivalent residues in Fc polypeptides that have significant tion may comprise one or more amino acid modifications in sequence or structural homology with each other. By the same the V, C, V, C1, and/or hinge regions of an antibody. token, it is possible to use Such methods to engineer addi Other modifications may provide additional or novel bind tional amino acid modifications in an Fc polypeptide to pro ing determinants into an Fc variant, for example additional or vide additional optimized properties, for example as novel Fc receptor binding sites, for example as described in described in U.S. Ser. No. 60/621,387, filed Oct. 21, 2004. In U.S. Ser. No. 60/531,752, filed Dec. 22, 2003, entitled “Fc one embodiment, amino acid modifications can be made that variants with novel Fc receptor binding sites”. In one embodi replace one or more residues in an Fc polypeptide of the ment, an Fc variant of one antibody isotype may be engi present invention with one or more residues in another neered such that it binds to an Fc receptor of a different homologous Fc polypeptide. In an alternate embodiment, isotype. This may be particularly applicable when the Fc 25 hybrid Fc polypeptides are constructed. Such that one or more binding sites for the respective Fc receptors do not signifi regions of an Fc polypeptide of the present invention are cantly overlap. For example, the structural determinants of replace with the corresponding regions of a homolous Fc IgA binding to FcyRI may be engineered into an IgG Fc polypeptide. For example, some studies have explored IgG1. variant. IgG2, IgG3, and IgG4 variants in order to investigate the The Fc variants of the present invention may comprise 30 determinants of the effector function differences between modifications that modulate the in vivo pharmacokinetic them. See for example Canfield & Morrison, 1991, J Exp Med properties of an Fc variant. These include, but are not limited 173: 1483-1491; Chappel et al., 1991, Proc Natl Acad Sci to, modifications that enhance affinity for the neonatal Fc USA 88 (20):9036-9040; Chappel et al., 1993, J Biol Chem receptor FcRn (U.S. Ser. No. 10/020,354; 268: 25124-25131; Tao, Canfield, and Morrison, 1991, J Exp WO2001 US0048432; EP2001000997063: U.S. Pat. No. 35 Med 173: 1025-1028; Tao et al., 1993, J Exp Med 178: 661 6,277.375: U.S. Ser. No. 09/933,497; WO1997US0003321; 667: Redpath et al., 1998, Human Immunology, 59,720-727. U.S. Pat. No. 6,737,056; WO2000US0000973; Shields et al., In one embodiment, the Fc variants of the present invention 2001, J Biol Chem 276 (9): 6591-6604; Zhou et al., 2003, J comprise one or more engineered glycoforms. By “engi Mol. Biol. 332: 901-913). These further include modifica neered glycoform as used herein is meant a carbohydrate tions that modify FcRn affinity in a pH-specific manner. In 40 composition that is covalently attached to an Fc variant, some embodiments, where enhanced in vivo half-life is wherein said carbohydrate composition differs chemically desired, modifications that specifically enhance FcRn affinity from that of a parent Fc variant. Engineered glycoforms may at lower pH (5.5-6) relative to higher pH (7-8) are preferred be useful for a variety of purposes, including but not limited (Hinton et al., 2004, J Biol Chem 279 (8): 6213-6216: to enhancing or reducing effector function. Engineered gly Dall’Acqua et al., 2002J Immuno 169: 5171-5180: Ghetie et 45 coforms may be generated by a variety of methods known in al., 1997, Nat Biotechnol 15 (7): 637-640; the art (Umafia et al., 1999, Nat Biotechnol 17:176-180; WO2003US0033037; WO2004US001 1213). For example, Davies et al., 2001, Biotechnol Bioeng 74:288-294; Shields et as described in Hinton et al., 2004, “Engineered Human IgG al., 2002, J Biol Chem 277:26733-26740; Shinkawa et al., Antibodies with Longer Serum Half-lives in Primates' J Biol 2003, JBiol Chem 278:3466-3473); (U.S. Pat. No. 6,602,684; Chem 279 (8): 6213-6216, substitutions T250O, T250E, 50 U.S. Ser. No. 10/277,370; U.S. Ser. No. 10/113,929; PCTWO M428L, and M428F provide enhanced binding to FcRn and 00/61739A1; PCT WO 01/29246A1; PCTWO 02/31140A1; improved pharmacokinetics. Additionally preferred modifi PCT WO 02/30954A1); (PotelligentTM technology Biowa, cations are those that maintain the wild-type Fc's improved Inc., Princeton, N.J.; GlycoMAbTM glycosylation engineer binding at lower pH relative to the higher pH. In alternative ing technology (GLYCART biotechnology AG, Zürich, Swit embodiments, where rapid in Vivo clearance is desired, modi 55 Zerland). Many of these techniques are based on controlling fications that reduce affinity for FcRn are preferred. (U.S. Pat. the level of fucosylated and/or bisecting oligosaccharides that No. 6,165,745; WO1993US0003895; EP 1993000910800; are covalently attached to the Fc region, for example by WO1997US0021437; Medesan et al., 1997, J Immunol 158 expressing an Fc variant in various organisms or cell lines, (5): 2211-2217: Ghetie & Ward, 2000, Annu Rev Immunol engineered or otherwise (for example Lec-13 CHO cells or rat 18: 739-766; Martin et al. 2001, Molecular Cell 7: 867-877; 60 hybridoma YB2/0 cells), by regulating enzymes involved in Kim et al. 1999, Eur J. Immunol 29: 2819-2825). Preferred the glycosylation pathway (for example FUT8 C.1.6-fuco variants that enhance FcRn are described in U.S. Ser. No. Syltranserase and/or B1-4-N-acetylglucosaminyltransferase 60/627,763, filed Nov. 12, 2004; 60/642,886, filed Jan. 11, III GnTIII), or by modifying carbohydrate(s) after the Fc 2005; 60/649,508, filed Feb. 2, 2005; 60/662,468, filed Mar. variant has been expressed. Engineered glycoform typically 15, 2005, and 60/669,311 filed Apr. 6, 2005, entitled “Fc 65 refers to the different carbohydrate or oligosaccharide; thus Variants with Altered Binding to FcRn, all hereby incorpo an Fc variant, for example an antibody or Fc fusion, may rated by reference. comprise an engineered glycoform. Alternatively, engineered