CA2103059C - Method for making humanized antibodies - Google Patents

Abstract

Variant immunoglobulins, particularly humanized antibody polypeptides are provided, along with methods for their preparation and use. Consensus immunoglobulin sequences and structural models are also provided.

Description

W~ 92/22653 ~ ~ Q ~ ~ ~ ~ PCI'/US92/U5126 METHOD FOIL h9Ai~ING-HUMANIZED ANTIBODIES.
i Field of the Invention This invention relates to methods for the preparation and use of variant antibodies and 1o finds application particularly in the fields of immunology and cancer diagnosis and therapy.
t3~kg_r~~ of the Inven~i~n Naturally occurring antibodies (immunogfobulins) comprise two heavy chains linked Z5 'together by disulfide bonds and two light chains, one light chain being linked to each of the heavy chains by disulfide bonds. Each heavy chain has at one end a variable domain (V,~) followed by a number of constant domains. Each light chain has a variabie domain 4~J~) at one end and a constant domain at its other end; the constant domain of the light chain is aligned with the first constant domain of the heavy chain, and the light chain variable domain is 2o aligned with the variable domain of the heavy chain. Particular amino acid residues are believed to form an interface between the light end heavy chain variable domains, see e.g.
Chothia et al., J lNol. Biol. 186:651-868 (1985); Novotny and Naber, 'roc.
Matt. Aca~f Sci.
USA 8:4592-4596 (1985).
The constant domains are not involved directly in binding the antibody to an antigen, 25 but are involved in various effector functions, such as participation of the antibody in antibody-dependent ~ellufar cytotoxicity. The variabl~ domains of each pair of tight and heavy chains are involved directly pn binding the antibody to the antigen. The domains of natural light and heavy chains have the sarcie general structure, and each domain comprises four framework (FR) regions, whose sequences are somewhat conserved, connected by three hyper-variable 30 or complementarity determining regions tCDRs) (see tCabat, E. A. et al., See~s~errces of Pr~teens of Imenunoldgecal lnte~est, National institutes of He~9th, Bethesda, ~!I~, (198?)). The four framework regions largely adopt a ~-sheet conformation and the CCRs form loops connecting, and in some cases forming part of, the ~-sheet structure. The CDRs in each chain are held ire close proximity by the framework regions and, with the CDRs from the other chain, contribute z ~c-rlu~~~eom26 W~ 92!22653 to the formation of the antigen binding site.
Widespread use has been made of monoclonal antibodies, particularly those derived from radents including mice, however they are frequently antigenic in human clinical use. For example, a major limitation in the clinical use of rodent monoclonal antibodies is an anti-globulin response during therapy (Miller, R. A. et al., Blood 62:988-995 (1983); Schroff, R. W. et al., Cancer Res. 45:879-885 41985)).
The art has attempted to overcome this problem by constructing "chimeric"
antibodies in which an animal antigen-binding variable domain is coupled to a human constant domain (Cabilly et al., U.a. patent No. 4,816,56?; IVlorri~on, S. L. et al., Proc.
Natl. .~lcad. Sci. UBA
81:6851-6855 41984); Boutianne, G. L. et al., Nature 312:643-646 (1984);
Neuberger, M. S.
et al., Nature 314:268-270 (1985)). The term "chimeric" antibody is used herein to describe a polypeptide comprising at least the antigen binding portion of an antibody molecule linked to at least part of another protein (typically an immunogiobuiin constant domain).
The isotype of the human constant domain may be selected to tailor the chimeric °~ antibody for participation in antibody-dependent cellular cytotaxicity (ADCC) and complement-dependent cytotoxicity (see e.g. Bruggemann, l~If. et al., J. Exp.
IVled.
186:1351-1361 (1987); Riechmann, L. et al., Nature 332:323-327 (1988); Love et al., IVlethods in Enzymology 178:515-527 (1989): Bindon et al., J. Exp. MecJ.
168:127-142 1988).
2o In the typical embodiment, such chimeric antibodies contain about one third rodent 4or other non-human species) sequence and thus are capable of eliciting a significant anti-globulin response in humans. For example, in the case of the murine anti-CD3 antibody, ~KT3, much of the resulting anti-globulin response is directed against the variable region rather than the constant region tJaffers, G. J. et al., transplantation 41:572-578 11986)).
In a further eff~art to resolve the antigen binding functions of antibodies and to minimize the us~ of h~terologous sequences in human antibodies, Winter and colleagues (Jones, P. T.
et al., Nature 321:5'22-525 (1986): Riechmann, L. et al., Nature 332:323-327 11988);
Verhoeyen, iVi. etal., Science 239:1534-1536 (1988)) have substituted rodent CDRs or CDR
sequences for the corresponding segments of a human antibody. As used herein, the term 30 ''humanized" antibody is an embodiment of chimeric antibodies wherein substantiaNy less than an intact human variable domain has been substituted by the corresponding sequence from a non-human species. In practice, hdmanized antibodies are typically human antibodies in which some CDR residues and possib6y some FR residues are substituted by residues from analogous sites in rodent antibodies.

,..,..r. '',, . . . ;'~, . . : .~ . ~', W092/2265~ 3 ~ ~ ~'~ ~~ ~ ~ PCT/US92/OS126 The therapeutic promise of this approach is supported by the clinical efficacy of a hs..~nanized antibody specific for the CAMPATH-1 antigen with two non-Hodgkin lymphoma patients, one of whom had previously developed an anti-globulin response to the parental rat antibody tRiechmann, L. et al., Nature 332:323-327 (1988); Hale, G. et ai., Lancet 1:1394-1399 (1988)). A murine antibody to the interfeukin 2 receptor has also recently been.
humanized (Queen, C. et ai., Prac. Nath Acad. Sci. USA 8f:10029-10033 (1989)) as a potential immunosuppressive reagent. Additional references related to humanization of antibodies include Co et al., Prac. Nat/. Acad. Sci. USA 88:2869-2873 11991 );
Gorman et al., Prac. Nat/. Acad. Sci. USA 88:4181-4185 (1991 ); Daugherty et al., G1/uc%ic Acids Research 19(9):2471-2476 t 1991 ); Brown et al., Proc. Nat/. Acad Sci. USA 88:2663-2667 t 1991 );
,lunghans et al., Cancer Research 50:1495-1502 11990).
In some cases; substituting CDRs from rodent antibodies for the human CDRs in human frameworks is sufficient to transfer high antigen binding affinity (Jones, P.
T, et al., Nature 321:522-525 t 1986): Verhoeyen, M. etal.; Science 239:1534-153611988)), whereas in other -cases it has been necessary to additionally replace one tRiechmann, L. et al"
Nature 332:323-327 t1988D? or several (Queen, C. et ai., Proc. Natl. Acad, Sci. USA
86:10029-10033 (1989)) framework region tFR1 residues. See also Co et al., supra.
For a given antibody a small number of FR residues are anticipated to be important for antigen binding. Firstly for example, certain antibodies have been shown to contain a few FR
residues which directly contact antigen in crystal structures of antibody-antigen complexes te:g:; reviewed in Davies, D. R: et al., Ann. Rev. Biochem. 59:439-473 11990)). Secondly, a number of FR residues have been proposed by Chothia, Lesk and colleagues lChothia, C. &
Lesk, A. M.; J. Mol. Bi~l. 196:901-917 11987); Chothia, C. et al.~, Nature 342:877-883 ti 989); Tramontano, A. et al.; J. Mol. Biol. 215:175-182 (1990)) as critically affecting the cpnformation of particular tDRs and thus their contribution to antigen binding. See also Margolies ef al.; Proc. lVatl. ~8cad. Sci. USA 72:2180-2184 11975).
It is also known that; in a few instances, an antibody variable domain (either V" or V~) may contain glycosylation sites; and that this glycosylation may improve or abolish antigen binding, Pluclcthun, Biotechnology 9:545-51 (1991 ); Spiegelberg etal., Biochemistry 9:4217-4223 (1970); Wal)ic ef al., J: Exp. Med. 168:1099-1109 11988); Sox etal., Prac. Nat/. Acad.
Sci: USA 66:975-982' (19701; Margni et al:, Ann. Rev. lmmunol. 6:535-554 (1988).
Ordinarily, however, glycosyl-. ion has: no influence on the antigen-binding properties of an antibody. Pluckthun, supra, t : a91 ).
The three-dimensional structure of irnmunoglobulin chains has been studied, and crystal Pi.'T/US92/05126 ~f structures for intact immunoglobulins, for a variety of immunoglobulin fragments, and for antibody-antigen complexes have been published (see e.g., Saul et al., Journal of Biological Chemistry 25:585-97 (1978); Sheriff etal., Prac. Nat/. Acad Sci. USA 84:8075-79 11987);
Segal et al,, Proc. Nat/. Acad. Sci. USA 71:4298-4302 11974); Epp et al., Biochemistry 14(22):4943-4952 (1975); Marquart et al., J. Mol. Biol. 141:369-391 11980);
Furey et al., J. Mol. Biol. 167:661-692 (1983); Snaw and Amzel, Protein: Structure, Function, and Genetics 1:267-279, Alan R. Liss, Inc. pubs. 11986): Chothia and Lesk, J. Mal.
Biol. 196:901-917 (1987); Chothia et al., Nature 342:877-883 (1989); Chothia et al., Science 233:755-58 (1986); Huber et al., Nature 264:415-420 11976); Bruccaleri et al., Nature 335:564-568 11988) and Nature 336:266 (1988): Sherman etal., JournalofBiological Chemistry 263:4064-4074(1988); Amzel and Pa)jak, Ann. Rev. Biochem. 48:961-67 (1979); Silverton etal., Proc.
Nat/. Acad Sci. USA ?4:5140-5144 11977); and Gregory et al., Molecular Immunology 24:821-829 (1987). It is known that the function of an antibody is dependent on its three dimensional structure, and that amino acid substitutions can change the three-dimensional ''structure of an antibody, Snow and Amzel, supra. It has previously been shown that the antigen binding affinity of a humanized antibody can be increased by mutagenesis based upon molecular modelling (Riechmann; L. etal., Nature 332:323-327 (1988): Queen, C.
etal., Proc.
Natl. Acad. Sci. USA 86:10029-10033 (1989)).
Humanizing an antibody with retention of high affinity for antigen and other desired biological activities is at present difficult to achieve using currently available procedures.
Methods are needed for ratibrralizing the selection of sites for substitution in preparing such antib~dies and thereby increasing, the efficiency of antibody humanization.
The proto-onco~ene HER,2 (human epidermal growth factor receptor 2) encodes a protein tyrosine kinase tp'185HER2'that is related to and somewhat homologous to the human epidermal growth factor receptor lsee CoussenS, L. et al., Science 230:1132-1139 (1985);
Yamamoto; T. et al., Nature 319:230-234 (1986); King, C. R, et al., Science 229:974-976 1985)). HER2 is also known in the field as c-erbB-2, and sometimes by the name of the rat homolog, neu. AmpiifiGation and/or overexpression of HER2 is associated with multiple human malignancies and appears to be integrally involved in progression of 25-30°~ of human breast ' 30 and ovarian cancers tSP'amon, D. J. et al.; Science 235:177-182 (1987), Siamon, D. J. et al., Science 244:707-712 E 1989)): Furthermore, the extent of amplification is inversely correlated with the observed median patient survival time (Slamon, supra, Science 1989).
The murine monoclonal antibody known as muMAb4D5 tFendly, 13. M. et al., Cancer Res. 50:1550-1558 t1 X90)), directed against the extracellular domain, (ECD) of p185HER2, WO 92/22653 . ~ .i ~ ~ ~ ~ ~ PCT/US92/~512G
specifically inhibits the growth of tumor cell lines overexpressing p185RER2 in monolayer culture or in soft agar (Hudziak, R. M. et al., Molec. Cell. Biol. 9:1165-1172 (1989D; Lupu, R.
et al., Science 243:1552-1555 (1990)). MuMAb4D5 also has the potential of enhancing tumor cell sensitivity to tumor necrosis factor, an important effector molecule in 5 macrophage-mediated tumor ce!! cytotaxicity (Hudziak, supra, 1989; Shepard, H. M. and t_ewis, G. D. J. Clinical Immunology 8:333-395 (1988)). Thus muMAb4D5 has potential far . clinical intervention in and imaging of carcinomas in which p185H~R2 is averexpressed. The muMAb4D5 and its uses are described in PGT application WO 89/06692 published 27 July 1989. This murine antibody was deposited with the ATCC and designated ATCC CRL
10463.
However, this antibody may be immunogenic in humans.
tt is therefore an object of this invention to provide methods for the preparation of antibodies which are less antigenic in humans than non-human antibodies but have desired antigen binding and other characteristics and activities.
It is a further object of this irwention to provide methods for the efficient humanization ~of antibodies; i.e. selecting non-human amino acid residues for importation into a human antibody background sequence in such a fashion as to retain or improve the affinity of the non human donor antibody for a given antigen.
It is another object of this invention to provide humanized antibodies capable of binding pIB~HER2 2o C?ther objects; features, and characteristics of the present invention will became apparent upon consideration of the following description and the appended claims.
~ummaw of the invention The objects of this invention are accomplished by a method for making a humanized antibody comprising amino acid sequence of an import, non-human antibody and a human antibody, c~mprising the steps of:
a. obtaining the amino acid sequences of at least a portion of an import antibody variable darnain and of a consensus variable domain;
b. identifying Complementarily Determining Region (CDR) amino acid sequences in the import and the human variable domain sequences;
c. substituting an impart CDFi amino-acid sequence far the corresponding human CDR amino acid sequence;
d. aligning the ameno acid sequences of a Framework Region (FR) of the import ".': -'. .'. , .n'. . , "' ; .;~ ,:.. . :.. ~', ... ;;
.. ... ,.,, .,-.,;... ..... ,:.. > "...; .. ,.,..... ~ , ...' :.!.~: . , ,.;
~...,.. . . ' : .. ~ ~ . .. . a W~ 92/22653 ; ', . 6 PGT/US921~5126 antibody and the corresponding FR of the consensus antibody;
e. identifying impart antibody FR residues in the aligned FR sequences that are non-homologous to the corresponding consensus antibody residues;
f. determining if the non-homologous import amino acid residue is reasonably expected to have at least one of the following effects:
1. non-covalantly binds antigen directly, 2. interacts with a CDR; or 3. participates in the VL - VH interface: and g, for any non-homologous import antibody amino acid residue which is reasonably expected to have at least one of these effects, substituting that residue for the corresponding amino acid residue in the consensus antibody I~R sequence.
f3ptionally, the method of this invention comprises the additional steps of determining if any non-homologous residues identified in step te) are exposed on the surface of the domain or buried within it, and if the residue is exposed but has none of the effects identified in step r 15 rtf), retaining the consensus residue.
Additionally; in certain embodiments the method of this invention comprises the feature wherein the corresponding consensus antibody residues identified in step (e) above are salect~d from the group consisting of 4L, 35L, 36L, 38L, 43L, 44L, 46L, 58L, 62L, 63L, 64L, 65L, 66L, 67L, 68L, 69L, COL; 71 L, ?3L, 85L; 8?L, 98L, 2H, 4H, 24H, 36H, 37H, 39H, 43H, 20 45H, 49H, 58H, 60H; 67H, 68H; 69H, ?~DH, ?3H, ?4H, 75H, 76H, ?8H, 91 H, 92H, 93H, and t03H tutilizing the numbering system set forth in Rabat, E. A. et al., SeQuences of Proteins of trrrmunologica! Inter~~t (National Institutes of Health, Bethesda, MD, 1987)).
in certain embodiments; the method of this invention comprises the additional steps of searching either or both of the import, non-human end the consensus variable domain 25 sequences for glycosylation sites, deternnining if the glycosytation is reasonably expected to be irnportarvt for the desired antigen binding and biological activity of the antibody (i.e., determinincd if the glycosyiation site binds to antigen or changes a side chain of an amino acid residue that binds to antigen, or if the glycosytation bnhanees or weakens antigen binding, or is important for maintaining antibody affiriityD: If the import sequence bears the glycosylation 3p site, it is preferred to Substitute that site for the corresponding residues in the consensus human if the glycosyiation site is re~sonat~ly expeoted to be important. if only the consensus sequence, end hot the import; bears the gtycosyfati~n site, it is preferred to eliminate that glycosyiation site or substitute therefor'ths corresponding amino acid residues from the import s~q~ience:

,.' .. :s: ;..,.,. . ... ., ::; ,;~' .. . 'v . . '.': :v..-; .,::'.' ,;: ~: ,:
v.
W~ X2/22653 ~ ~, o ~ ~ PLT/US92/05126 Another embodiment of this invention comprises aligning import antibody and the consensus antibody FR sequences, identifying import antibody FR residues which are non-homologous with the aligned consensus FR sequence, and far each such non-homologous import antibody FR residue, determining if the corresponding consensus antibody residue represents a residue which is highly conserved across all species at that site, and if it is so conserved, preparing a humanized antibody which comprises the consensus antibody amino acid residue at that site.
Certain alternate embodiments of the methods of this invention comprise obtaining the amino acid sequence of at feast a portion of an import, non-human antibody variable domain 1o having a CDR and a FR, obtaining the amino acid sequence of at least a portion of a consensus antibody variable domain having a CDR and a FR, substituting the non-human CDR
for the human CDR in the consensus antibody variable domain, and then substituting an amino acid residue for the consensus amino acid residue at at least one of the following sites:
a, tin the FR of the variable domain of the tight chain) 4L, 35L, 36L, 38L, 43L, '~ ~ 44L, 58L, 46L, 62L, 63L, 64L, 65L, 66L, 67L, 68L, 69L, ?OL, 71 L, 73L, 85L, 87L, 98L, or b. tin the FR,of the variable domain of the heavy chain) 2H, 4H, 24H, 36H, 37H, 39H; 43H, 45H; 49H, 58H, 60H, 67H, 68H, 69H, 70H, ?3H, 74H, 75H, 76H, ?8H, 91 H, 92H, 93H, and 103H.
In preferred embodiments, the non-CDR residue substituted at the consensus FR
site is the residue found at the corresponding location of the non-human antibody.
Dptionally, this just-recited embodiment comprises the additional steps of following the method steps appearing at the beginning of this summary and determining whether a particular amino acid residue can reasonably be expected to have undesirable effects.
This invention also relates to a humanized antibody comprising the CDR
sequence of an irnp~rt, non-human antibody and the FR sequence of a human antibody, wherein an amino acid residue within the human FR sequence located at any one of the sites 4L, 35L, 36L, 38L, 43L, 44L, 46L, 58L, 62L, 63L64L, 65L; 66L, 67L, 68L, 69L, 70L, 71 L, 73L, 85L, 87L, 98L, 2H, 4H, 24H, 36H, 37H, 39H, 43H; 45H, 49H, 58H, 60H, 67H, 68H, 69H, 70H, 73H, 74H, 75H, 76H, 78H, 91 H92H, 93H, and 103H has been substituted by another residue.
In preferred embodiments, the residuq substituted at the human FR site is the residue found at the corresponding location of the non-human antibody from which the non-human CDR was obtaified. In other embodiments, no human FR residue other than those set forth in this group has been substituted.

1 and SEQ. 1D NO. 3, respectively). FIGURE 1 B shows the comparison between the VH
domain amino acid residues of the muMAb4d5, huMAb4D5, and a consensus sequence (Fig_ 1 B, SEQ. ID N0. 6, SEO. 1D N0. 2 and SEQ. ID NO. 4, respeciively)_ Both Figs 1 A and 1 B
use the generally accepted numbering scheme from Kabat, E. A., et al , Sequences of Proteins of Immunologicallnterest (National Institutes of Health, Bethesda; MD
(1987)). !n both Fig_ 1 A and Fig. i B, tfie CDR residues determined according to a standard sequence definition (as in Kabat, E. A. et al., Sequences of Proteins oflmmunologicallnterest (National Institutes of Health, Bethesda, MD, 1987)) are indicated by the first underlining beneath the sequences, and the CnR residues determined according to a structur<~I
definition (as in Choihia, C. & Lesk, A. M., J. MoL Biol. 19fi:901-917 (1987)) are indicated by the second, lower underlines. The mismatches between genes are shown by the vertical lines.
FIGURE 2 shows a scheme for humanization of muMAb4D5 Vt. and Vti by gene conversion mutagenesis.
FIGURE 3 shows the inhibition of SK-BR-3 proliferation by MAb4D5 variants.
Relative i 5 cell proliferation was determined as described (Hudziak, R. N1. et al., Molec. Cell. Biol.
9:1165-1172 (1989)) and data (average of triplicate determinations) are presented as a percentage of results with untreated cultures for muMAb4D5 (I), huMAb4D5-8 (n) and huMAb4D5-1 (I).
FIGURE 4 shows a stereo view of a-carbon tracing for a model of huMAb4D5-8 V~
and VH _ The CDR residues (Kabat, E. A. et al., Sequences of Proteins of ImmunologicaJ Interest (National Institutes of Health, Bethesda, MD, 1987)) are shown in bold and side chains of VH
residues A?1, T73, A78. S93, Y102 and V~ residues Y55 plus R66 (see Table 3) are shown.
FIGURE 5 shows an amino acid sequence comparison of V~ (top panel) and VH
(tower pane!) domains of the murine anti-CD3 monoclonal Ab UCHT1 (muxCD3, Shalaby et al., J.
Exp. Med. 175, 217-225 (1992) with a humanized variant of this;antibody (ImxC:1).sv I Also shown are consensus sequences (most commonly occurring residue or pair of residues) of the most abundant human subgroups, namely V~ K i and VH III upon which the humanized sequences are based (Kabat, E. A. et aL, Sequences of Proteins of ImmunolopicaJ Interest, 5'" edition, National Institutes of Health, Bethesda, MD, USA (1991)). The light chain sequences--muxCD3, )~tlxCD3 v 1 and huKl--correspond to SEQ.ID.NOs 16, 17, and i 8, respectively. 'The heavy chain sequences muxCD3, huxCD3v1 and hutd correspond to SEQ.
ID. NOs 19, 20, and 21 respectively. Residues which differ between muxCD3 and huxCD3v1 are identified by an asterisk ('), whereas those which differ between humanized and consensus sequences are identified by a sharp sign (ll). A bullet (°) denotes that a residue at this position has been found to contact antigen in one or more crystallographic structures of antibodylantigen complexes (Kabat et al., 1991; Mian, i_ S. et al , J. Mol_ Biol.
217, 133-151 (1991)). The location of CDR residues according to a sequence definition SUBSTITUTE SHEET

2/22653 ~ ~ ~ ~ ~ 'a ~ pCT/US92/0512b 'WU 9 amino acid residues of the muMAb4d5, huMAb4D5, and a consensus sequence (Fig.
1 B, SEQ
ID NO. 6, SEO.. ID NO. 2 and SECT. ID NO. 4, respectively). Both Figs 1 A and 1 B use the generally accepted numbering scheme from Kabat, E. A., et ai., Sequences of Proteins of lmmunological Interest (National Institutes of Health, Bethesda, MD (1987)).
In both Fig. 1A
and Fig. 1 B, the CDR residues determined according to a standard sequence definition tas in Kabat, E. A. et al" Sequences of Proteins of lmmunolagical Interest (National Institutes of Health, Bethesda, MD, 1987)) are indicated by the first underlining beneath tha sequences, and the CDR residues determined according to a structural definition (as in Chothia, C. & Lesk, A.
M., J, Mol. Biol. 996:901-917 (198?I) are indicated by the second, Lower underlines. The to mismatches between genes are shown by the vertical lines.
FIGURE 2 shows a scheme for humanization of muMAb4D5 VE and VH by gene conversion mutagenesis.
FIGURE 3 shows the inhibition of SK-BR-3 proliferation by MAb4D5 variants.
Relative cell proliferation was determined as described tHudziak, R. M. et al., Mo%c.
Cell. Biol.
'9 1165-1172 ( 1989)) and data (average of triplicate determinations) are presented as a percentage of results with untreated cultures for muMAb4D5 (I), huMAb4D5-8 (n) and huMAb4D5-1 (I).
FIGURE 4 shows ~ stereo view of a-carbon tracing for a model of huMAb4D5-8 V~
and VH . The CDR residues (Kabat, E. A: et aL, Sequences of Proteins of Immunological Interest (NationaE Institutes of Health, Bethesda, MD, 1987)) are shown in bold and side chains of VH
residues A71, T 73, A78, S93: Y 102 and VL residues Y55 plus R66 (s~e Table 3) are shown.
FIGURE 5 shows an amino acid Sequence comparison of V~ (top panel) and VH
(lower panel) domains of the murine anti-CD3 monoclonal Ab UCHT1 (muxCD3, Shalaby et al., J.
Exp. Med. 175, 217-225 ( 1992) with a humanized variant of this antibody (huxCD3v9). Also shown are consensus sequences (most commonly occurring residue or pair of residues) of the most abundant human subgroups, namely VL K 1 and VH II) upon which the humanized sequences are based (Kabat, E. A. et al.; Sequences of Proteins of Immuno%gical Interest, 5'"
edition; National institutes of Health; Bethesda, MD, USA (1991 )). The light chain sequences--muxCD3, huxCD3v9 and huKl--correspond to SEQ.ID.NOs 16, 17, and 18, respectively. The hea~ey chain sequences--muxCD3; huxCD3v9 and huKt--correspond to SEa.ID.NOs 19, 20, and 21; respectively: Residues which differ between muxCD3 and huxCD3v9 are identified by an asterisk ('" ), whereas those which differ between humanized and consensus sequences are identified by a sharp sign (#). A b~rllet (a) denotes that a residue at this position has been found to contact antigen in- one or more crystallographic structures of antibodyiantigen °nf, ,~'.~;r~~rt~:~~.a~y 7,~' a.'i%ji4rr~..~,~i&5il~su :~~'~d~~t~r ~
n,s~!~~r~t~.:al~a;. . <.u',~i~":x:..~,.~~r.~i<. ..":rk;sso.,:<:.a ; .tea., .:~..

17V0 92/22653 , . , complexes iKabat et al., 1991; Mian, t. S. et al., J. Mol. Biol. 217, 133-151 (1991 )). The location of CDR residues according to a sequence definition (Kabat et al., 1991 ) and a structural definition tChothia and t-esk, supra 1987) are shown by a line and carats t") beneath the sequences, respectively.
FIGURE 6A compares marine and humanized amino acid sequences for the heavy chain of an anti-CD18 antibody. H52H4-160 tSEa. ID. NO. 22) is the marine sequence, and pH52-8.0 tSECI. ID. N0. 23) is the humanized heavy chain sequence, pH52-8.0 residue 143S is the final amino acid in the variable heavy chain domain VH, and residue 144A is the first amino acid in the constant heavy chain domain CH1.
FIGURE 6s compares marine and humanized amino acid sequences for the light chain of an anti-CD18 antibody. H52L6-158 tSEQ. ID. NO. 24> is the marine sequence, and pH52-9.0 tSEQ. ID. NO. 25) is the Humanized light chain sequence. pH52-9.0 residue 128T is the final amino acid in the light chain variable domain VL, and residue 129V is the first amino acid in the light chain constant domain CL, ' Defiled Descrio~on of the Invention pefinition~
In general, the following words or phrases have the indicated definitions when used in the description, examples, and claims:
The mutine monoclonal antibody known as muMAb4D5 tFendly, B. M. et al., Cancer Res. 50:1550-1558 119901) is directed against the extracellular domain tECD) of p185HER2, The muMAb4D5 and its uses are described in PCT application W0 89/06692 published 27 July 1989. This marine antibody was deposited with the ATCC and designated ATCC CRt-10463.
Inthis description and claims, the terms muMAb4D5~ chMAb4D5 and huMAb4D5 represent marine, chimerized and humanized versions pf the mor~oclonai antibody 4D5, respectively.
A humanized antibody- for the purposes herein is an immunoglobulin amino acid sequence variant ar fragment thereof which is capable, of binding to a predetermined antigen and which comprises a FR region having substantially the amino acid sequence of a human immunoglobulin and a 'CDR having substantially the amino acid sequence of a non-human imenunogtobulirr.
'Generally, a humanized antibody has one ~r more amino acid residues introduced into it from a source which is non-human. These non-human amino acid residues are referred to hereirt as "impbrt" residues, which are typically taken from an "import°' antibody domain, W~ 92/22653 ~ ~ ~ ~ ~ ~ ~ P~/vS92/05126 particularly a variable domain. An import residue, sequence, or antibody has a desired affinity and/or specificity, or other oesirable antibody biological activity as discussed herein.
In general, the humanized antibody will comprise substantially all of at least one, and typically two, variable domains (Fob, Fab', F(ab')z, Fabc, Fv) in which all or substantially all of the CDR regions correspond to these of a non-human immunoglobulin and all ar substantially all of the FR regions are those of a human immunoglobulin consensus sequence.
The humanized antibody optimally also will comprise at Least a portion of an immunoglobulin constant region (Fc), typically that of a human immunoglobulin. Ordinarily, the antibody will contain both the light chain as well as at least the variable domain of a heavy chain. The i0 antibody also may include the CH1, hinge, CH2, CH3, and CH4 regions of the heavy chain.
The humanized antibody will be selected from any class of tmmunoglobufins, including IgM, IgG, IgD, IgA and Ig~, and any isotype, including IgG 1, IgG2, IgG3 and IgG~4. Usually the constant domain is a complement fixing constant domain where it is desired that the humanized antibody exhibit cytotoxic activity, and the class is typically IgG,. Where such ''cytotoxic activity is not desirable, the constant domain may be of the IgGz class. The humanized antibody may comprise sequences from mare than one class or isotype, and selecting particular constant domains to optimize desired effector functions is within the ordinary skill in the art.
The FR and CDR regions of the humanized antibody need not correspond precisely to the parent~t sequences, e.g., the import CDR or the consensus FR may be mutagenized by substitution, insertion or deletion of at least one residue so that the CDR or FR residue at that site does not correspond to either the consensus or the import antibody. Such mutations, however, wilt not be extensive. Usually, at least 75°~ of the humanized antibody,residues will correspond to those of the pareritat FR and CDR sequences, more often 90%, and mast preferably greater than 95,%.
In general, humanized antibodies prepared by the method of this invention are produced by a procoss of analysis of the parental sequences and various conceptual humanized products using three dimensional models of the parental and humanized sequences. Three dimensional immunoglabulin models are commonly available and are familiar to those skilled in the art.
Computer programs are available which illustrate and display probable three dimensional conformational structures of selected candidate immunoglobulin sequences.
Inspection of these displays permits analysis of the likely role of the residues in the functioning of the candidate immunoglobulin sequence, i.e., the analysis of residues that influence the ability of the candidate immunogtobulin to bind its antigen.

PC1'/US92/05126 Iz Residues that influence antigen binding are defined to be residues that are substantially responsible for the antigen affinity or antigen specificity of a candidate immunoglobulin, in a positive or a negative sense. The invention is directed to the selection and combination of FR
residues from the consensus and import sequence so that the desired immunoglobutin characteristic is achieved. Such desired characteristics include increases in affinity and greater specificity for the target antigen, although it is conceivable that in some circumstances the opposite effects might be desired. In general, the CDR residues are directly and most substantially involved in influencing antigen binding talthough not all CDR
residues are so involved and therefore need not be substituted into the consensus sequence).
However, FR
to residues also have a significant effect and can exert their influence in at least three ways:
They may noncovalently directly bind to antigen, they may interact with CDR
residues and they may affect the interface between the heavy and light chains.
A residue that nancovalently directly binds to antigen is one that, by three dimensional analysis, is reasonably expected to noncovalently directly bind to antigen.
Typically, it is a necessary to impute the position of antigen from the spatial location of neighboring CDRs and the dimensions and structure of the target antigen. In general, only those humanized antibody residues that are capable of forming salt bridges, hydrogen bands, or hydrophobic interactions are likely to be involved in non-covalent antigen binding, however residues which have atoms which are separated from antigen spatially by 3.2 Angstroms or less may also non-covalently 2p interact with antigen. Such residues typically are the relatively larger amino acids having the side chains with the greatest bulk; such as tyrosine, arginine, and lysine.
Antigen-binding FR
residues also typically will have side chains that are oriented into an envelope surrounding the solvent oriented face of a CDR which extends about 7 Angstroms into the solvent from the CDR domain and about 7 Angstroms on either side of the CDR domain, again as visualized by three dimensional modeling.
A residue that interacts with a CDR generally is a residue that either affects the conformation of the CDR polypeptide backbon~ ar forms a noncovalent bond with a CDR
residue side chain: Conformation-affecting residues ordinarily are those that change the spatial position of any CDR backbone atom IN; Ca, C, 0, C,B') by more than about 0.2 Angstroms.
' 30 ' Backbone atoms of CDR sequences are displaced far example by residues that interrupt or modify organized structures such as beta sheets, helices or loops. Residues that can exert a profound affect on the conformation of neighboring sequences include proline and giycine, both of which are capable of introducing bends into the backbone. Other residues that can displace backbone atoms are those that are capable of participating in salt bridges and hydrogen bonds.

WO ~Z/2~653 _ I~ ~ ~ ~ ~ ~ J ~~ . ~ ~ f'~/US92/05126 A residue that interacts with a CDR side chain is one that is reasonably expected to form a noncovalent bond with a CDR side chain, generally either a salt bridge or hydrogen bond. Such residues are identified by three dimensional positioning of their side chains. A salt or ion bridge could be expected to form between two side chains positioned within about 2.5 -3.2 Angstroms of one another that bear opposite charges, for example a lysinyl and a.
giutamyl pairing. A hydrogen bond could be expected to form between the side chains of residue pairs such as Beryl or threonyl with aspartyl or glutamyl for other hydrogen accepting residues). Such pairings are well known in the protein chemistry art and will be apparent to the artisan upon three dimensional modeling of the candidate immunoglobulin.
1o lmmunoglobulin residues that affect the interface between heavy and light chain variable regions t"the V~ - V" interface") are those that affect the proximity or orientation of the fiwo chains with respect to ane another. Certain residues involved in interchain interactions are already known and include V~ residues 34, 36, 38, 44, 46, 87, 89, 91, 96, and 98 and V" residues 35, 37, 39, 45, 47, 91, 93; 95, 100, and 103 (utilizing the nomenclature set forth '' n Kabat et al.; Sequences of Proteins of lmmunological Interest (National Institutes of Health, Bethesda, MD, 1987)1. Additional residues are newly identified by the inventors herein, and include 43t., 85L, 43H and 60H. While these residues are indicated for IgG
only, they are applicable across species: In the practice of this invention, import antibody residues that are reasonably expected to be involved in interchain interactions are selected for substitution into 20' the consensus sequence: It is believed that heretofore no humanized antibody has been prepared-with an intrachain-a>fecting residue selected from an import antibody sequence.
Since it is not entirely possible to predict in advance what the exact impact of a given substitution will be it may be necessary to make the substitution and assay the candidate antibody for the desired characteristic. These steps, however, are per se routine and welt within the ordinary skill of the art.
CDR and FR residues are deternnined according to a standard sequence definition tKabat et al. , Sequences of Proteins of lmmcen~lagical Interest, National institutes of Health, Bethesda MD (1987), and a structural definition tas in Chothia and Lesk, J. Mol. Bi~l.
196:901-917 (1987). Where these two methods result' in slightly different identifications of a CDR, the 3o structural definition is preferred, but the residues identified by the sepuence definition method are considered important FR residues for determination of which framework residues to import into a consensus ~equence.
Throughout this description; reference is made to the numbering scheme from Kabat, E. A., et el., Sequences of Proteins of lmmunological Interest (National Institutes of Health, WO 92/22653 ~ ~ ~ ~ ~ ~ ~ ~' I P(.'T/US92/OS126 ''i Bethesda, MD 11987) and (1991 ). In these compendiums, Kabat lists many amino acid sequences for antibodies for each subclass, and lists the most commonly occurring amino acid for each residue position in that subclass. Kabat uses a mefihod for assigning a residue number to each amino acid in a Listed sequence, and this method for assigning residue numbers has become standard in the field. The Kabat numbering scheme is followed in this description.
For purposes of this invention, to assign residue numbers to a candidate antibody amino acid sequence which is not included in the Kabat compendium, one follows the following steps. Generally, the candidate sequence is aligned with any immunoglobulin sequence or any consensus sequence in Kabat. Alignment may' be done by hand, or by computer using commonly accepted computer programs; an example of such a program is the Align 2 program discussed in this description. Alignment may be facilitated by using some amino acid residues which are common to mast Fab sequences. For example, the light and heavy chains each typically have two cysteines which have the same residue numbers; in V~ domain the two cysteines are typically at residue numbers 23 and 38, and in the VH domain the two cysteine residues are typically numbered 22 and 92. Framework residues generally, but not always, have approximately the same number of residues, however the CDRs will vary in size. For example, in the case of a CDR from a candidate sequence which is longer than the CDR in the sequence in Kabat to which it is aligned, typically suffixes are added to the residue number to indicate the insertion of additional residues (see, e.g. residues 'i OOabcde in Fig. 5). For candidate sequences which, for example, align with a Kabat sequence for residues 34 and 3f but have no residue betv~reen ther~n to align with residue 35, the number 35 is simply not assigned to a residue.
Thus, in humanization of an import variable sequence, where one cuts out an entire human or consensus CDR and r~places it with an import CDR sequence, (a) the exact number z5 of residues may be swapped, leaving the numbering the same, Ib) fewer import amine acid residues may be introduced than are cut, in w~Oich ease there will be a gap in the residue numbers, or (c) a larger number of'amino acid residues may be introduced then were cut, in which case the numbering will involve the use of suffixes such as 100abcde.
The terms "consensus sequence°' and "consensus antibody" as used herein refers to ' 30 ' an amine acid sequence which comprises the most frequently occurring amino acid residues at each location in ail immunoglobulins of any particular subclass or subunit structure. The consensus sequence may be based on immunogiobulins of a particular species or of many species. A "consensus" sequence, structure, or antibody is understood to encompass a consensus human sequence as described in certain embodiments of this invention, and to refer P~'1US92/0512G
'NVO 92/22653 ' 15 to an amino acid sequence which comprises the most frequently occurring amino acid residues at each location in all human immunoglobulins of any particular subclass or subunit structure.
This invention provides consensus human structures and consensus structures which consider other species in addition to human.
The subunit structures of the five immunogiobulin classes in humans are as follows:
IgG Y Y1, Y2, Y~, Y4 x or ~l (Y~2) (Ya~a) IgA a a1, a2 x or A (a~rz)" , (~2JI2)~

IgM ~ none ~ x or e1 (~.~zx~)s , G~.~z~z)s IgD a none x or /I (c3ax2) , (d~Jlz) IgE a none x or ~l (e~re2) , (E~12) (" may equal 1, 2, or 3) Pn preferred embodiments of an IgGY1 human c~nserysus sequence, the consensus variable domain sequences are derived from the most abundant subclasses in the sequence compilation of Kabat etal., Sequencesnf~roteins oflmmuncl~gicalfnterest, National Institutes of Health, Bethesda MD (1987), namely V~ x Subgroup I and Vy group Ill. In such preferred embodiments, the V~ consensus domain has the amino acid sequence:

SGSGTDFTLTISSLQPEDFATYYCQQYNSLPYTFGQGTKVEIKRT (SEQ. ID NO. S);
the VH consensus domain has the amino acid sequence:
EV~1LVESGGGLVQPGGSLRLSCAA~GFTFSf~YAMSWVRQAPGKGLEWVAVISENGGYTRYAD
SVKGRFTISADTSKNTAYLQMNSLRAEDTAVYYCSRWGGDGFYAMDVWGQGTLVTVSS (SEQ.
ID NC. 4).
These sequences include consensus CDRs as well as consensus FR residues (see for eacample irk Fig. 1 ).
While not wishing to be limited to any particular theories, it may be that these preferred embodiments are less likely to be imrnunogenic ire an individuaa than less abundant subclasses.
however, in other embodiments, the consensus sequence is derived from other subclasses of human immunogiobulin variable domains: In yet other embodiments, the consensus sequence is derived from human constant domains.
Identity or homology with respect to a specified amino acid sequence of this invention is defined herein as the percentage of amino acid residues in a candidate sequence that are ' PCTlUS92/U5126 WO 92/22653 - , . ' , identical with the specified residues, after aligning the sequences and introducing gaps, if necessary, to achieve the maximum percent homology, and not considering any conservative substitutions as part of the sequence identity. None of N-terminal, C-terminal or internal extensions, deletions, or insertions into the specified sequence shall be construed as affecting homology. All sequence alignments called for in this invention are such maximal homology alignments. While such alignments may be done by hand using conventional methods, a suitable computer program is the "Align 2°' program far which protection is being sought from the U.S. Register of Copyrights (Align 2, by Genentech, Inc., application filed 9 December 1991 ).
"Non-homologous" import antibody residues are those residues which are not identical to the amino acid residue at the analogous or corresponding location in a consensus sequence, after the import and consensds sequences are aligned.
The term °'computer representation" refers to information which is in a form that can be manipulated by a computer. The act of staring a computer representation refers to the act ~ of placing the information in a form suitable for manipulation by a computer.
This invention is also directed to novel polypeptides, and in certain aspects, isolated novel humanized anti-p185HER2 antibodies are provided. These novel anti-p185HER2 antibodies are sometimes collectively referred to herein as huMAb4D5, and also sometimes as the light or heavy chain variable domains of huMAb4D5, and are defined herein to be any 2o polypeptide sequence which possesses a biological property of a polypeptide comprising the following polypeptide sequence:
DIO.MTt~SPSSLSASVGDRVTITCRASaDVNTAVAWYQQKPGKAPKLLIYSASFLESGVP
SRFSGSRSGTDFTLTISSLQPEDFATYYCQQHYTTPPTFGO,GTKVEIKRT (SEQ. ID NO. 1, which is the light chain variable domain of huMAb4.D5); or EVQLVESGGGLV~PGGSLRLSCAASGFNIKDTYiHWVRQAPGKGLEWVARIYPTNGYTR
YADSVKGRFTfSADTSKNTAYLQMNSLRAEDTAVYYCSRWGGDGFYAMDVWGQGTLV
TVSS (SECT. ID NO. 2, which is the heavy chain variable domain of huMAb4D5?.
30 "Biological property", as relates for example to anti-p185RER2, for the purposes herein means an in a~ivo effector or antigen-binding function or activity that is directly or indirectly performed by huMAb4D5 (whether in its native or denatured conformation).
Effector functions include p185HER2 binding, any hormonal or hormonal antagonist activity, any mitogenic or agonist or antagonist activity, any cytotoxic activity. An antigenic function means possession WO 92/22653 ~ ~ '~ ~ I . , PCT~US92/05126 I~
of an epitope or antigenic site that is capable of cross-reacting with antibodies raised against t.::polypeptide sequence of huMAb4D5.
Biologically active huMAb4D5 is defined herein as a polypeptide that shares an effectar function of huMAb4D5. A principal known effector function of huMAb4D5 is its ability to bind to p185HERZ_ Thus, the biologically active and antigenically active huMAb4D5 palypeptides that are the subject of certain embodiments of this invention include the sequence of the entire translated nucleotide sequence of huMAb4D5; mature huMAb4D5; fragments thereof having a consecutive sequence of at Least 5, 10, 15, 20, 25, 30 or 40 amino acid residues comprising sequences from muMAb4D5 plus residues from the human FR of huMAb4D5; amino acid sequence variants of huMAb4D5 wherein an amino acid residue has been inserted N- or C-terminat ta, ar within, huMAb4D5 ar its fragment as defined above; amino acid sequence variants of huMAb4D5 or its fragment as defined above wherein an amino acid residue of huMAb4D5 or its fragment as defined above has been substituted by another residue, including ~~predetermined mutations by, e.g., site-directed or PCR mutagenesis;
derivatives of huMAb4D5 or its fragments as defined above wherein huMAb4D5 or its fragments have been covalent modified, by substitution, chemical, enzymatic, or other appropriate means, with a moiety other than a naturally occurring amino acid: and gtycosylation variants of huMAb4D5 tinsertion s~f a gtycosytation site or dehtion of any glycosylation site by deletion, insertion or substitution of suitable residues). Such fragments and variants exclude any potypeptide heretofore identified, includir~g~muMAb4D5 or any known polypeptide fragment, which are anticipatory order 35 U.S.C.102 as well as polypeptides obvious thereaver under 35 U.S.C.
103.
An "isolated" polypeptide means polypeptide which has been identified and separated and/or recovered from a component of its natural environment. Contaminant components of its natural environment are materials which would interfere with diagnostic or therapeutic uses for the polypeptide, and may include enzymes, hormones, and other proteinaceous or nonproteinaceous salutes. In preferred embodiments, far example, a polypeptide product comprising huMAb4D5 wilt be purified from a cell culture or other synthetic environment t1 to greater than 95% by weight of protein as determined by the Lowry method, and most preferably mare than 99°~ by weight, i2) to a degree sufficient to obtain at least 15 residues of N-terminal ar internal amino acid sectuence by use of a gas- ar liquid-phase sequenator tsuch as a commercially availabt~, Applied Biasystems sequenator Model 470, 477, or 4731, or t3) to homogeneity by SDS-PAGE under reducing or nonreducing conditions using Coomassie blue or, preferably, silver stain. isolated huMAb4D5 includes huMAb4D5 in,~tu within recombinant P~GT/US9~/OS1Z5 WO 92/2653 ' -t$
Gcll$ since at least or~~ component of the huMAb4D5 natural envirorlrr~ent will not be present.
Drdinarity, however, Isolated huMAb4I~5 will be prepared by at least one purification step.
in accordance with this invention. huMAb4D5 nucteiC acid is I~NA or DNA
containing greater than ten bases that encodes a biologically or antigenically active huMAb4D5, is complementarlr to nucleic acid sequence encoding such huMAb4D5. or hybridizCS
tb nuCIeiC
acid sequence encoding such huMAb~t05 and remains sfabty bound to it under stringent conditions, and comprises nuClt~ic acid from s muMAb4D5 CDR and a human FR
region.
Preferably. the huMAb4a5 riucleic~f~cid ~sncodes a palypeptida sharing ax least 75°~6 sequenaa idantixy, more preferably at least BO°/s, still more praf8tably at l6ast 8596, oven more i0 preferably at 90%, and most prCfarabty 959b, with the huMAb4D5 amino acid sos~uance.
Preferably, a nucleic acid molecule that hybridizes to the huMAb4D5 nucleic said contains at . least 20, more proferably.. Vin, and most preferably 9D bases. Such hybridizing or comrilementary nucleic acid, however, is further defined as being novel under 35 U-S.C. 10~
and unobvious under 35 U:S.~. t 03 over any prior art nuCleiC acid.
is ~' ~ Stringent conditions era those that f 1 i employ law ionic svength and high temperature for washing, for exempts, O.Oi S M NaC110.00t 5 M sodium citratel0/1 %
NaDodSO; at 80° G:
t2) employ during hybridixatlon a ,denaturing agent such as formamide, for example, &096 lvol/vol) .formamide with 0:9 96~ bavina serum albuminl0ll °.6 ~colh~l9 °~ polyvinylpyrrolidona150 mM sodium~phosphate. buffer at pH 6.,6 with 75o rnM NaGI, T5 mM sodium citrate at 42° G:
D' v w or: (3) employ 5096 form>~mida, 6.x';SSC 10.76 M, NeCI, 0.0'Y5 M sodium Gitratef, 50 mM
sodiury phosphate lphl f.$i. O-t °.~ sQdlum pyrophosphate, 5 x D$nhardt's solution, sonicaxed .. $atmowsperm DNA I50.g1m11, 0.1:% SDS. and 1096 dextran sulfate at 4~ C, with washes at - . ~,~ C in 0.2 x SSC arid 0.~1.% S~S..
w . ~ Tha~term .'cant~ol sequerzc~s" refers to UNA soquertces necessary for the expression ~f:':an~operabty linkqdycoding. sequence yin a.parti.uuler host organism. The Contr4l Setluencss ' , that era suitable for-~~prokaryotes, for example; include a promoter.
aptionaliy an operattrr $equence,~ a ribosome binding site; and possibly, other as yet poorly understood sequences.
Eukaryotic cells. are knows to utiliaa promoters, polyadenylation signals and enhancers.
Nucleic acid is "oparably linked." when it is placed into a functional relationship with ~' so ancrthar nucleic acrd sequence: °FCir example, 17NA for a presaqu~nce or secretory leader is Qperpbly,lihked to DNA for a polypeptide if it is expressed as $ grpprptein that participates in the. secretion' of the ~ potypeptide; a promoter ~or enhanaor is aperably linked to a Coding sequence if It affects the transcription of the seguence; ar a ribosome binding site is aperably linked ~to a ,coding ~ sequersce ii it is positioned. so as to facilitate translation. Generally, *-trademark _WO 92/22653 ~ ~ ~ ~ ~ ~ ~ PCT%US92/05126 1 '~
"operably linked" means that the DNA sequences being linked are contiguous and, in the case of a secretory leader, contiguous and in reading phase. However enhancers do not have to be contiguous. Linking is accomplished by ligation at convenient restriction sites. If such sites do not exist, then synthetic oligonucleotide adaptors or linkers are used in accord with conventional practice.
An "exogenous" element is defined herein to mean nucleic acid sequence that is foreign to the cell, or homologous to the cell but in a position within the host cell nucleic acid in which the element is ordinarily not found.
As used herein, the expressions "cell,"~ "cell line," and "cell culture" are used 1o interchangeably and all such designations include progeny. Thus, the words "transformants"
and "transformed cells" include the primary subject cell and cultures derived therefrom without regard for the number of transfers. It is also understood that alt progeny may not be precisely identical in DNA content, due to deliberate or inadvertent mutations. Mutant progeny that have the same function or biological activity as screened for in the originally transformed cell '"'are included. Where distinct designations are intended, it will be clear from the context.
"Oiigonucleotides" are short-length, single- or double-stranded polydeoxynucleotides that are chemically synthesized by known methods (such as phosphotriester, phosphate, or phosphoramidite chemistry, using solid phase techniques such as described in EP 266,432 published 4 May 1988, o~ via deoxynucleoside H-phosphonate intermediates as described by Froehler et a6.; Nu_cl__ Acids Res., 14: 5399-5407 ( 19861). They are then purified on polyacrylamide gels:
The technique of "polymerase bhain reaction," or "PCR," as used herein generally refers to a procedure wherein minute amounts of a specific piece of nucleic acid, RNA
and/or DNA, are amplified as described in U.S. Pat. No. 4;683,195 issued 28 July 1987.
Generally, sequence information from the ends of the region of interest or beyond needs to be available, such that oligonucleotide primers can be designed; these primers will be identical or similar in sequence to opposite strands of the template to be amplified. The 5' terminal nucleotides of the two primers may coincide with the ends of the amplified material. PCR can be used to amplify specific RNA sequences; specific DNA sequences from total genomic DNA, and cDNA
3o transcribed from total cellular RNA, bacteriophage or plasmid sequences, etc. See generally Mullis et aL, Cold Sarinca Harbor- Symn. Quant. Biol., ,~: 263 (1987); Erlich, ed., P R
Technoioav: tStockton f~ress, NY; 1989). As used herein, PCR is considered to be one, but not the only, example of a nucleic acid polyrryerase reaction method for amplifying a nucleic acid 'test sample, comprising he use of ~ known nucleic acid (DNA or RNA) as a primer and ~:~U~~J~
WO 92/22$53 ~ PGT/US92/0~126 ,, zo utilizes a nucleic acid polymerise to amplify or generate a specific piece of nucleic acid or to amplify or generate a specific piece of nucleic acid which is complementary to a particular nucleic acid.
Su~able Metho~i~ for Practicing ~h~ Invention Some aspects of this invention include obtaining an import, non-human antibody variable domain, producing a desired humanized antibody sequence and for humanizing an antibody gene sequence are described below. !~ particularly preferred method of changing a to gene sequence, such as gene conversion from a non-human or cons~nsus sequence into a humanized nucleic acid sequence, is the cassette mutagenesis procedure described in ~xampl8 1. Additionally, methods are given for obtaining and producing antibodies generally, which apply equally to native non-human antibodies as well as to humanized antibodies.
Generally, the antibodies and antibody variable domains of t6rois invention are i5 'conventionally prepared in recombinant cell culture, as described in more detail below.
Recombinant synthesis is preferred for reasons of safety and economy, but it is known to prepare peptides by chemical synthesis and to purify them from natural sources; such preparations are included within the definition of antibodies herein.
2o Molecular Modeling An integral step in our approach to antibody humanization is construction of computer graphics models of the import and humanized antibodies. These models are used to determine if the six complementarity-determining regions (CDRs? can be successfully transplanted from the import framework to a human one and to determine which framework residues from the 25 import antibody: if any, need-to be incorporated into the humanized antibody in order to maintain CDR conformation. In addition, analysis of the sequences of the import and humanized antibodies and reference to the- models can help to discern which framework residues are unusual and hereby might be involved in antigen binding or maintenance of proper antibody structure.
30 All of the humanized antibody models of this invention are based ~an a single three-dimensional'computer graphics structure hereafter referred to as the consensus structure. This consensus structure is ~ key distinction from the qpproach of previous workers in the field, who ty~icalty begin by selecting a human antibody structure which has an amino acid sequence which is similar to tl~e sequence of their import antibody.

WO 92/22653 . ~ ~ ~ ~ ~ ~ PC"f/US92/05126 z~
The consensus structure of one embodiment of this invention was built in five steps as described below.
Step 1: Seven Fab X-ray crystal structures from the Brookhaven Protein Data Bank were used (entries ZFB4, 2RHE, 3FAB, and 1 REI which are human structures, and 2MCP, 1 FBJ, and 2HFL which are murine structures). For each structure, protein mainchain geometry and hydrogen bonding patterns were used to assign each residue to one of three secondary structure types: alpha-helix, beta-strand or other (i.e. non-helix and non-strand). The immunoglobutin residues used in superpositioning and those included in the consensus structure are shown in Table 1.

aWWO 9WZ26s3 PCT/'U~'~2/OS126 ~Z
Table I
Immunoglobulin Residues in Superpositioningand Those Used Included in the Consensus Structure ~l,t~ domain Iga 2FB4 2RIiE 2MCP 3FAB iFBJ 2I~'~. 11tE1 Consensush 60-56 62-58 57-72 53-65 60=65 50-55 61-66 59-77 RMSe 0.40 0.60 0.53 0.54 0.48 0.50 VB domain Iga 2hB4 2MCP 3FAB 1FBJ 2I~r. Consensusb ~~- 3-8 58-71 70-73 57-?0 58-71 58-71 66-?1 78-84 80-86 77-8~ 78-84 78-84 75-82 82_99 94-101 91-98 92-99 92-99 88-94 RMSe 0.43 0.85 0.62 0.91 RMSd 0.91 0.73 0.77 0.92 a Four-letter code Bank file.
for Protein Data b Residue numbers the crystalstructures ire from the Protein Data for taken Bank files. Residuenumbers f~r the consensus structure are according to Kabat et al.

c Root-mean-s9uare ation for (N,Ca;C) superimposed 2FB4.
devi in ~ atoms on d Root~aean-square atioa~ for (Id,Ca;C) superira~poscd2IIFL.
devi in ;~ atones on W(~ 92/22653 '~ ~ 0 ~ ~ j ~ PCT/US92/05126 z~
Step 2: Having identified the alpha-helices and beta-strands in each of the seven structures, the structures were superimposed on one another using the INSIGHT
computer program (Biosym Technologies, San Diego, CA) as follows: The 2FB4 structure was arbitrarily chosen as the template (or reference? structure. The ~FB4 was held fixed in space and the S other six structures rotted and translated in space so that their common secondary structural elements ti.e. alpha-helices and beta-strands) were oriented such that these common elements ware as close in position to one another as possible. (This superpositioning was performed using accepted mathematical formulae rather than actually physically moving the structures by hand.) t0 Step 3: With the seven structures thus superimposed, for each residue in the template (ZFB41 Fab one calpulates the distance from the template alpha-carbon atom (Ca) to the analogous Ca atom in each of th~ other six superimposed structures. This results in a table of Ca Ca distances for each residue pdsition in the sepuence. Such a table is necessary in order to determine which residue positions will be included in the consensus model. Generally, 15 '~if all Ca-Ca distances for a given residue position were ~ 1.0~, that position was included in the consensus structure. If for a given position only one Fab crystal structure was > 1.0~, the position was included but the outlying crystal structure was not included in the next step (for this position anlyl. In gene~al~ the seven ~B-strands were included in the consensus structure while some of the loops connecting the ~B-.strands, e.g.
complementarity-determining 20 regions (CDRs?, were not included in view of Ca divergence.
Step 4: For each residue which was included in the consensus structure after step 3, the average of the coordinates for individual mainchain N, Ca, C, O and C~
atoms were calculated. D~ae to the averaging procedure,- as well as variation in bond Isngth, band angle and dihedral angle among the crystal structures, this "average" structure contained some bond 25 lengths and angles which deviated from standard geometry. For purposes of this invention, "standard geometry" is understood to include geometries commonly accepted as typical, such as the compilation of bond lengths and angles from small molecule structures in Weiner, S.J.
et, al.; J. Artier. Chem. Soc., 106: ?65-784'(1984).
Step 5: In order to correct these deviations; the final step was to subject the 30 "average" structure t'o 50 cyc6e~ of energy minimization (DISCOVER program, Biosym Technologies) using the AMBER fllVeiner, S.J. er: al., J. Amer. Chem. S~c., 106: 765-784 (1984.x? parameter set with only the Ca coordinates fixed (i:e. all other atoms are allowed to mtDVe) (energy minimization is described belowl. This allowed any deviant bond lengths and angles to assume a standard (chemically acceptable) geometry. See Table It.

PCT/US92/OSl2G

2 'i Table II
Average Bond Lengths and Angles for "Average" (Before) and Energy-Minimized Consensus (After SO Cycles) Structures VLK V1,K Vg VH Standard before after before after Geometry (~) (A> (~) (~) (~) N-Ca 1.459(0.012)1.451(0.004)1.451(0.023)1.452(0.004)1.449 Ca-C 1.515(0.012)1.523(0.005)1.507(0.033)1.542(0.005)1.522 ~C 1.208(0.082)1.229(0.003)1.160(0.1??)1.231(0.003)1.229 C-N 1.288(0.049)1.33?(0.002)1.282(0.065)1.335(0.004)1.335 Ca-C~ 1.508(0.026)1.530(0.002)1.499(0.039)1.530(0.002)1.526 b C-N-Ca 123.5(4.2) 123.8(1.1) 125.3(4.6) 124.0(1.1) 121.9 N-Ca-C 110.0(4.0) 109.5(1.9) 110.3(2.8) 109.5(1.6) 110.1 Ca-C-N 116.6(4.0) 116.6(1.2) 11?.6(5.2) 116.6(0.8) 116.6 C~C 123.1 (4.1 123.4(0.6) 122.2(4.9) 123.3(0.4) 122.9 N ) N-Ca-C~110.3(2.1) 109.8(0.?) 110.6(2.5) 109.8(0.6) 109.5 C~-Ca-C111.4(2.4) 111.1(0.?) 111.2(2.2) .111.1(0.6)111.1 ~ialues in parentheses are standard deviations. Note that while some bond length and angle averages did not change appreciably after energy-minimization, the corresponding standard deviations arc reduced due to deviant geometries assuming standard values after energy-minimization. standard geometry values are from the AMBER forcefield as implennented in I)ISCOVEFt (Biosym Technologies).

~r:s::. -_-_"._,_ __,__ _..
.WO 92/22653 ~ ~ ~ ~ ~ ~ ~ P~/US9zio~1z6 zs The consensus structure might conceivably be dependent upon which crystal structure was chosen as the template an which the others were superimposed. As a zest, the entire procedure was repeated using the crystal structure with the worst superposition versus 2Ft34, i.e. the 2HFL Fab structure, as the new template Ireference). The two consensus structures corlnp8re favorably (root-mean-squared deviation of x.11 ~ for alt N, Ca and C
atoms).
Note that the consensus structure only includes mainchain IN, Co. C. a. CB
atoms) coordinates far only those residues which are part of a conformation common to all seven X
' ray cryøtal structures. Fvr iha Fob structures, these include the common ~
strands Iwhich comprisB two ~-shoat$1 and a few non-~DFi lobes which connect these -strands.
The >.o consensus structure does not include GDRs yr sidachains, both of which vary in their conformation. among the seven structures. Also, note that the consensus structure includes only the VL and VH domains.
This aansansus structure is u$ed as the archetype. It is not particular to any species.
aryd has only the basic sh~p~ without side chains. starting with this consensus structure the is ~-rinodal of any import, human, or humanized Fab can be constructed as follows. lJsing the amino said seQuence of tha..particular antibody VL and VH domains of interest, a computer graphics prv~ram iau~b as INSIGHT*Bio$ym Technologies) is used to add sidechains and CDRs to the congarysus strirctura. When a 'sidechatn is added, its conformation is chosen on the basis of knbwn Fab erystal structures lsee the Background acotian fw publications of such . Zo . ~ crystal struaturesl and rotamer libraries (Ponder. J.W. & Richards, F. M.. J. Mwi. Biol. 1g3:
77~-791 (t sa$7I1. The modet also. is constructed so that the atoms of the sidechain are positioned s4. as to not collid,a with other atoms in the s=ob. , CDRS are coifed to fthe model Inow having the backbone plus 5ida chains) as follows.
The- size Ii.e. number-.of amino acids? of each import GDR is compared to canonical CDR
as structuras.~tabuiated by Chothia et al., Nature, 342:877-883 ('I 9B9)1 and which ware derived from. Fab cnistals. Each CDR svquanca is also reviewed for the prssancs or absence of certain specific arinino acid rasiduas v~ihich era identified by Chathia as structurally important: e.g. light chain residues 29 (CCiR1 i and' 95' (GDR3), dnd heavy chain residues 26. 27.
29 IGDR1 f and 55 ICDR2?. For light chain .CpR2, and heavy chain CDR3, only the size of the GDR i5 3o compered to~the Chvtliia. canonical structure. if the size and sequence li_e. inctusiar< of the specific, structurally impartant~.residuas as denoted by Chothia et al.l of the import CdR agrees in size and ha$ the same structurally. important residues asr those of a canonical CPR, than the mainchsiwconfonnation of the irnpoft CDR in the model is taken to be the same as that of the canonical CDR- This means that the import seduance is assigned the structural configuration *-trademark _ ,, . ~v,.-,. -.;~:
,,, r ': ~: r..~.. ~:.; x, , . /,'~.~ZS:":' 7, ~~\ e.h.
~.~'1'~.":.' _1i :': .~,..,.. ., ... .
f..:rwl:~ ~~ , . .. ... ..
S ~.~' L. f '~,~iy,,' ' y..,ffS ~.~:,,~1,~; ~~~.
~ . ,S.~'n':~:~:~lf:n 7 .... ... 'f~.
5...: ..
wa gxiz~~~ ~ ~ ~ ~ Pcr~us9zro~rzra of the canonical ~(?R, which is then incorporated in the evolving model.
Wowevar, if no matcriing canonical CDR can be assigned far the import CDR, than one of two options can bs exercised. First, using a program such as INSIGHT
iRiosym Technologie$), the 8rookhavan protein Data Bank can bs searched far loops with a similar size to that of the import CDR and these loops can be evaluated as possible conformations for the .
import CDR in thn model. Minimally. such loops must sxhibtt a conformation in which na loop atom overlaps with other protein atoms. Second, one can use available Programs which calculate possible laop~ conformations. assuming a given loop size, using methods such as tiescribsd by ~r~coleri et at , Natwe 335: b6a-ggg t'1988).
Wh~n all CDRS and sidechains have been added to the aonsansus structure to give the final rrvodal timport. human or, humanixad), the model is preferably subjected to energy minimization using programs which era available commerCi811Y (a-g~ Ci~CaVpR;
BioBym Tpchnolog)as)'. This technique .uses complex mathematical formulae to refine the model by performing such tasks as ohecking.that ail atoms ors within appropriate distances from cue ~, 15 'another and checking that band lengths and angles are within chemically acceptable limits.
nnodels of ~a humanized, import, or human antibody sequence era used in the practice of thi3 invent;on to understand the impact' of selected amino acid re$idues of tire activity of the seqwanca being modet~d. For exarhple, such a model can show residues which may be important in antigen binding. br for ma)ntaining the conformation of the $ntit~ody, as discussed _ in mare detail below. Modeling can also be used to explore the potential impact of Gharlgin$
. sny amino~acid residue in the antibody sequence.
M ini tn the praatJce~ af,. tills : invention, the first stop in humanizing an irnPort antibody is deriving a aonsensirs amine acid. sequence into which to incorporate the import sequences.
HBO ~ ~Qdel. is.'gdnera~tad far.these sequences using the methods described above. in certain ~~b~imqnts of this lnvenxian~ the consensus human sequences are derived from the mast $bundant~ ~~~)assas- iri the' seqWenca ccmp)lation of Kabat et al. IKabat. !!.
A. et al..
Sd4uences of Proteins of lri~rmuriological interest National Institutes of Health: Etethesda. MD, 3p ~ 1987)1,, namely V4 K subgroup 1 arid VM group III, and have th$
setluances indicated in the ~
. . definitions above.. , . . Whileahese steps may. ba. takbn in different order, typically a structure for the Candidate h~ariized antibod~r is~'crpated by transfsrring~ the at least one GOR from the non-human.
import sequence into thd consensus human structure, after the entire corresponding human . *--trademark i .t~' ~ ...' ,. ~_,..~... .. . .;~.,' ' ...:.:.: , ':_~'_;... ... -.;....~
.~..~:.~ . ...:. , :~;' . ;-,,~.. .. ~:r:. ' . .,:, . ~, ~',,.;..' .::... ,.' ,:..; ,...,...:~ ..: w~...~.~. '....,,. .. ; .''. .
.:..:.~v . ~.~.~':.... ~,:. ~:.~ . '. ... ,.. ., , ....
WO 92/22653 ~ ~ ~ 1 Q j ~ ~ ~ PC'T/US921U512~b CDR has been removed. The humanized antibody may contain human replacements of the non-human import residues at positions within CDRs as defined by sequence variability (Kabat, E. A. et al., Seguences of Proteins of Immunoiagica! Interest (National Institutes of Health, Bethesda, MD, 1987)) or as defined by structural variability (Chothia, C. & t-esk, A. M., J. Mol.
Biol. 196:901-917 (1987)). For example, huMAb4D5 contains human replacements of the muMAb4D5 residues at three positions within CDRs as defined by sequence variability (Kabat, E. A. et al., Seguences of Proteins of lmmunoiogical Interest tNational Institutes of Health, Bethesda, MD, 1987)) but not as defined by structural variability (Chothia, C.
& Lesk, A. M., J. Mol. Biol. 196:901-917 (1987)): V~-CDR1 K24R, V~-CDR2 R54L and V~-CDR2 T56S.
Differences between the non-human import and the human consensus framework residues are individually investigated to determine their possible influence on CDR conformation and/or binding to antigen. Investigation of such possible influences is desirably performed through modeling, by examination of the characteristics of the amino acids at particular locations, or determined experimentally through evaluating the effects of substitution or t5 '"mutagenesis of particular amino acids.
In certain preferred embodiments of this invention, a humanised antibody is made comprising amino acid sequence of an import, non-human antibody and a human antibody, utilizing the steps of:
a. obtaining the amino acid sequences of at least a portion of an import antibody variable domain and of a consensus human variable domain;
b. identifying Complementarily Determining Region (CDR) amino acid sequences in the import and the human variable domain sequences;
c. substituting an import CDR amino acid sequence for the corresponding human CDR amino acid sequence;
d. aligning the amino acid sequences of a Framework Region (FR) of the import antibody and the corresponding FR of the consensus antibody;
e. identifying import antibody FR residues in the aligned FR sequences that are non-homologous to the corresponding consensus antibody residues;
f. determining if the non-homologous import amino acid residue is reasonably 3o expected to have at least one of the following effects:
1. r9on-covalently binds antigen directly, 2. interacts with a CDR; or 3. participates in the V~ - VH interface; and g. far any'non-homologous import antibody amino said residue which is reasonably ~~.03~~j WO 92l226S3 ~CT/US92/05126 2$
expected to have at feast one of these effects, substituting that residue for the corresponding amino acid residue in the consensus antibody FR sequence.
aptionally, one determines if any non-homologous residues identified in step te) are exposed on the surface of the domain or buried within it, and if the residue is exposed but has none of the effects identified in step tf), one may retain the consensus residue.
Additionally, in certain embodiments the corresponding consensus antibody residues identified in step te) above are selected from the group consisting of 4L, 35L, 36L, 38L, 43L, 44L, 46L, 58L, 62L, 63L, 64L, 65L, 66L, 6?L, 68L, 69L, ?4L, ?1 L, ?3L, 85L, 8?L, 98L, 2H, 4H, 24H, 36H, 3?H, 39H, 43H, 45H, 49H, 58H, 6tJH, 6?H, 68H, 69H, ?OH, ?3H, ?4H, ?5H, io ?6H, 78H, 91 H, 92H, 93H, and 1 ~D3H (utilizing the numbering system set forth in Rabat, E.
A. et al., SeQuences of Proteins of lmmunologica! Interest (National Institutes of Health, 8ethesda, nIID, 19871).
In preferred embodiments, the method of this invention comprises the additional steps of searching either or both of the import, non-human bnd the consensus variable domain ~~sequences for glycosylatian sites, determining if the glycosylation is reasonably expected to be important for the desired antigen binding and biological activity of the antibody ti.e., determining if the glycosylation site binds to antigen or changes a side chain of an amino acid residue that binds to antigen, or if the glycosyiation enhances or weakens antigen binding, or is important for maintaining antibody affinity). If the import sequence bears the glycosylation site, it is preferred to substitute that site for the corresponding residues in the consensus human sequence if the glycosylation site is reasonably expected to be important. if only the consensus sequence; and not the import, bears the glycosylation site, it is preferred to eliminate that glycosylation site or substitute therefor the corresponding amino acid residues from the import sequence.
Another preferred embodiment of the methods of this invention comprises aligning import antibody and the consensus antibody FR sequences, identifying import antibody FR
residues which are non-homologous with the aligned consensus FR sequence, and for each such non-homologous import antibody FR residue, determining if the corresponding consensus antibody residue represents a residue which is highly conserved across all species at that site, 3o and if it is so conserved, preparing a humanized antibody which comprises the consensus antibody amino acid residue at that site.
tn certain alternate embodiments, one need not utilize the modeling and evaluation steps described above, and may instead proceed with the steps of obtaining the amino acid sequence of ~t least a portion of an import, non-human antibody variable domain having a GDR and a FR, WC? X2/22653 ~ ~ ~ PCT/US92/~5126 obtaining the amino acid sequence of at least a portion of a consensus human antibody variable domain having a CDR and a FR, substituting the non-human CDR for the human CDR
in the consensus human antibody variable domain, and then substituting an amino acid residue for the consensus amino acid residue at at least one of the following sites:
a. tin the FR of the variable domain of the light chain) 4L, 35L, 36L, 38L, 43L, 44L, 58L, 46L, 62L, 63L, 64L, 65L, 66L, 67L, 68L, 69L, 70L, 71 L, 73L, 85L, 87L, 98Lr or b. !in the FR of the variable domain of the heavy chain) 2H, 4H, 24H, 36H, 37H, 39H, 43H, 45H, 49H, 58H, 60Hr 67H, 68H, 69H, 70H, 73H, 74Hr 75H, 76H, l0 78H, 91 H, 92H, 93H, and 103H.
Preferably, the non-C~R residue substituted at the consensus FR site is the residue found at the corresppnding location of the non-human antibody» If desired, one may utilize the other method steps described above for determining whether a particular amino acid residue can reasonably be expected to have undesirable effects, and remedying those effects.
' f if after making a humanized antibody according to the steps above and testing its activity one is not satisfied with . the humanized antibody, one preferably reexamines the potential effects of the amino acids at the specific locations recited above.
Additionally, it is desirable to reinvestigate any buried residues which are reasonably expected to affect the ~h -V" interface but may not directly affect CDR conformation. It is also desirable to reevaluate the humanized antibody utilizing the steps of the methods claimed herein.
In certain embodiments of this invention, amino acid residues in the consensus human sequence are substituted for by other amino acid residues. In preferred embodiments, residues from a particular non-human import sequence are su~st~tute~i, however there are circumstances where it is desired to evaluate the effects of other amino acids. For example, if after making a humanized antibody according to the steps above and testing its activity one is hot satisfied with the humanized antibody, one may compare the sequences of other classes or subgroups of human antibodies, or classes or subgroups of antibodies from the particular non-human species, and determine which other amino acid side chains and amino acid residues are found at particular locations and substituting such other residues.
Antibodies Certain aspects of this invention are directed to natural antibodies and to monoclonal antibodies, as illustrated in the Examples below and by antibody hybridomas deposited with the ATCC tae described below). Thus, the references throughout this description to the use ~~Q ~~ ~~ , V~IG 92/22653 PCT/US92/0512b of monoclonal antibodies are intended to include the use of natural or native antibodies as well as humanized and chimeric antibodies. As used herein, the term "'antibody"
includes the antibody variable domain and other separable antibody domains unless specifically excluded.
In accordance with certain aspects of this invention, antibodies to be humanized timport 5 antibodies) are isolated from continuous hybrid cell tines formed by the fusion of antigen-primed immune lymphocytes with myeloma cells.
in certain embodiments, the antibodies of this invention are obtained by rautine screening. Polyclonat antibodies to an antigen generally are raised in animals by multiple subcutaneous (sc) or intraperitoneal (ip) injections'of the antigen and an adjuvant. It may be 1o useful to conjugate the antigen or a fragment containing the target amino acid sequence to a protein that is immunogenib in the species to be immunized, e.g., keyhole limpet hemocyanin, serum albumin, bovine thyroglobulin, or soybean trypsin inhibitor using a bifunctional or derivatizing agent, for example, maleimidobenzoyl sultosuccinimide ester (conjugation through cysteine residues), N-hydroxysuccinimide tthrough lysine'residuesl, glutaraldehyde, succinic ~r iS anhydride, SOC12, or R'N = C = NR; where R and R' are different alkyl groups.
The raute and schedule of the host animal or cultured antibody-producing cells therefrom are generally in keeping with established and conventional techniques for antibody stimulation and production. While rnice are frequently employed as the test madel° it is contemplated that any mammalian subjECt including human subjects or antibody-producing 20 cells obtained therefrom can be manipulated according to the processes of this invention to serve as the basis for production of mammalian, including human, hybrid cell lines.
Animals are typically immunized against the immunogenic conjugates or derivatives by combining 1 mg or 1 ,ug of conjugate (for rabbits or mice, respectively) with 3 volumes of Freund°s complete adjuvant and injecting the solution intradermally at multiple sites. One 25 month later the animals are boosted with 1 /5 to 1 /10 the original amount of conjugate in Freund's complete adjuvant for other suitable adjuvant) by subcutaneous injection at multiple sites. 7 to 14 days later animals are-bled and the serum is assayed for antigen titer. Animals erg boosted until the titer plateaus. Preferably; the animal is boasted with the conjugate of the same antigen, but conjugated to a different protein and/or through a different cross-linking 30 agent. Conjugates also can be made in recombinant cell culture as protein fusions. Also, aggregating agents such as alum are used to enhance the immune response.
After immunization; monoclonal antibodies are prepared by recovering immune lymphoid cells--typically spleen cells or lymphocytes from lymph node tissue--from immunized animals and immortalizing the ceps in con~rentional fashion, e.g. by fusion with myeloma cells or by f .,r ~. . ~ .~.. , ~.~~ ~, ~ .. ... . . ,.,..,. ~~ ' .. ., . . ,." y WO 92/22653 ~ ~ ~ '~ ~ '~ '~ PCT/US92/U5126 3!
Epstein-Barr tEB)-virus transformation and screening for clones expressing the desired antibody.
The hybridoma technique described originally by ICohler and Milstein, Eur. J.
Irrrmunol. 6:51 1 ( 1976) has been widely applied to produce hybrid cell lines that secrete high levels of monoclonal antibodies against many specific antigens.
It is possible to fuse cells of one species with another. However, it is preferable that .
the source of the immunized antibody producing cells and the myeloma be from the same species.
The hybrid call lines can be maintained in culture in rritrp in cell culture media. The cell lines of this invention can be selected and/or maintained in a composition comprising the to continuous cell line in hypoxanthin~-aminapterin thymidine tMAT) medium. In fact, once the hybridoma cell line is established, it can be maintained on a variety of nutritionally adequate media. Moreover, the hybrid cell lines can be stored and preserved in any number of conventional ways. including freezing and storage under liquid nitrogen.
Frozen calf lines can be revived and cuttured indefinitely with resumed synthesis and secretion of monoclonal antibody. The secreted antibody is recovered from tissue culture supernatant by conventional methods such as precipitation, ton exchange chromatography, affinity chromatography, or the like. The antibodies described herein are also recovered from hybridoma celP
cultures by conventional methods for purification of IgG or IgM as the case may bs that heretofore have been used to purify these irnmunoglobulins from pooled plasma, e.g. ethanol or polyethylene 2o glycol precipitation procedures. ~'he purified antibodies are sterile filtered, and optionally are conjugated to a detectable marker such as an enzyme or spin label for use in diagnostic assays of the antigen in test samples.
While routinely rodent monoclonal antibodies are used as the source of the import antibody, the invention is not limited to any species. Additionally, techniques developed for the production of chirneric antibodies ~Morrison etal., Froc. Nall. cad. Sci., 81:6851 t1984);
Neuberger ef al., Nature 312:6~4 (1984); Takeda at al., Nature 314:452 t1985)) by splicing the genes from a mouse antipody molecule of appropriate antigen specificity together with ga3nes from a human antibody molecule o~f appropriate biological activity tsuch as ability to activate human complement and mediate ADCC) can be used; such antibodies are within the scope of this invention.
Techniques for creating recombinant DNA versions of the antigen-binding regions of antibody molecules (known ae Fab fragments) which bypass the generation of monoclonal antibt~dies are encornpassed within the practice of this invention. One extracts antibody-specific messenger RW 0. molecules from immune system cells taken from an immunized animal, .. :,.'... ;..; .... . ,w.., ' :, .. ,.. .'' ..:.: , :,; ;.: ,. .,'; ; ;;.:, .,..;; ,, . . :... , ,, . ...
WO 92/22653 , . PC1'/US92/05126 .
3z transcribes these into complementary DNA (cDNA), and clones the ,DNA into a bacterial expressions system. One example of such a technique suitable for the practice of this invention was developed by researchers at Scripps/Stratagene, and incorporates a proprietary bacteriophage lambda vector system which contains a leader sequence that causes the expressed Fab protein to migrate to the periplasmic space (between the bacterial cell membrane and the cell watt) or to be secreted. One can rapidly generate and screen great numbers of functional FAb fragments for those which bind the antigen. Such FAb fragments with specificity for the antigen are specifically encompassed within the term "antibody" as it is defined, discussed, and claimed herein. ' Aming Acid Seauence Variants Amino acid sequence variants of the antibodies and polypeptides of this invention (referred to in herein as the target polypeptidel are prepared by introducing appropriate nucleotide changes into the DNA encoding thb target polypeptide, or by in vitro synthesis of the desired target polypeptide. Such variants include, for example, humanized variants of non-human antibodies, as well as deletions from, or insertions or substitutions of, residues within particular amino acid sequences. Any combination of deletion, insertion, and substitution can be made to arrive at the final construct, provided that the final construct possesses the desired characteristics. The amino acid changes also may alter post-translational processes of the target po[ypeptide; such as changing he number ar position of glycosylation sites, altering any membrane anchoring characteristics, and/or altering the intro-cellular location of the target polypeptid~ by inserting, deleting, ;or otherwise affecting any leader sequence of the native target polypeptide.
in designing, amino acid sequence variants of target polypeptides, the location of the mutation site and th~ nature of ' the mutation will depend an the target polypeptide Gharacteristicts) to' be modified. The sites for mutation can be modified individually or in series, e.g.; by t 1 ) substituting first v~ith conservative amine acid choices and then with more radical selections depending upon the results achieved, d2) deleting the target residue, or f3) inserting residues of the same'or a different class adjacent to the located site, ar combinations of options 1-3. tn certain embodiments, these choices are guided by the methods far creating humanized sequences set forth above.
A useful method for identification of certain residues or regions of the target polype~rtide that are preferred locafians far mutagenesis i~ called "alanine scanning mutagenesis" as described by Cunningham and Wells tS~,ience, 244: 1081-1085 (19891).

WU 92!22653 3 3~ ~ I~ ~ ~ j ~ PCT/t1S92/05126 Here, a residue or group of target residues are identified (e.g., charged residues such as arg, asp, his, lys, and glu) and replaced by a neutral or negatively charged amino acid (most preferably ~tanine or polyalanine) to affect the interaction of the amino acids with the surrounding aqueous environment in or outside the cell. Those domains demonstrating functional sensetivity to the substitutions then are refined by introducing further or other variants at or for the sites of substitution. Thus, while the site for introducing an amino acid sequence variation is predetermined, the nature of the mutation per se need not be predetermined. For example, to optimize the performance of a mutation at a given site, ala scanning or random mutagenesis may be conducted at the target codon ar region and the expressed target polypeptide variants are screened for the optimal combination of desired activity.
There are two principal variables in the construction of amino acid sequence variants:
the location of the mutation site and the nature of the mutation. In general, the location and nature of the mutation chosen will depend upon the target polypeptide characteristic to be -modified.
Amino acid sequence deletions of antibodies are generally not preferred, as maintaining the generally configuration of an antibody is believed to be necessary for its activity. Any deletions will be selected so as to preserve the structure of the target antibody.
Amino acid sequence insertions include amino- and/or carboxyl-terminal fusions ranging zo in length from one residue to polypeptides containing a hundred or more residues, as well as intrasequence insertions of singte or multiple amino acid residues.
Intrasequence insertions li.e., insertions within the target polypeptide sequence) may range generally from about 1 to 't 0 residues, more preferably 1 to 5, mast preferably 1 to 3, Examples of terminal insertions include the target polypeptide with an N-terminal methionyl residue, an artifact of the direct expression of target potypeptide in bacterial recombinant cell culture, and fusion of a heterologous N-terminal signal sequence to the N-terminus of the target polypeptide molecule to facilitate the secretion of the mature target polypeptide from recombinant host cells. Such signal sequences generally will; be otitair~ed from, and thus homologous to, the intended host cell species: Suitable sequences include STIt or Ipp for ~ call, alpha factor for yeast, and viral ssgnals such 'as herpes, gD for mammalian cells.
Other insertional variants of 'the target polypeptide include the fusion to the N- or C-tecminus of the target polypeptide of immunogenic polypeptides, e.g., bacterial polypeptides such as beta~lactamase ar an enzyme encoded by the E: coli trp locus, or yeast protein, and C-terminal fusions with proteins having a tong half-fife such as immunogiobulin constant ~~ 92/22653 ~ ~ ~ '~ ~ ') ~ ~ PCT/US92/05126 3 ~9 regions Ior other immunoglobulin regions), albumin, or ferritin, as described in WO 89/02922 published 6 April 1989.
Another group of variants are amino acid substitution variants. These variants have at least one amino acid residue in the target polypeptide molecule removed and a different residue inserted in its place. The sites of greatest interest for substitutional mutagenesis include sites identified as the active sites) of the target polypeptide, and sites where the amino acids found in the target poiypeptide from various species are substantially different in terms of side-chain bulk, charge, and/or hydrophobicity. Other sites for substitution are described infra, considering the effect of the substitution of ' the antigen binding, affinity and other characteristics of a particular target antibody.
Other sites of interest are those in which particular residues of the target polypeptides obtained from various species ara identical. These positions may be important for the biological activity of the target polypeptide. These sites, especially those falling within a sequence of at least three other identicalty conserved sites, are substituted in s relatively conservative manner. If such substitutions result in a change in biological activity, then other changes are introduced and the products screened until the desired effect is obtained.
Substantial modifications ire furictian or immunolagical identity of the target polypeptide are accomplished by selecting substitutions that differ significantly in their effect on maintaining (a) the structure of the palypeptide backbone in the area of the substitution, for example, as a sheet or helical conformation, (b) the charge or hydrophobicity of the molecule at the target site, or tcl the bulk of th~ side chain. Naturally occurring residues are divided into groups based on common side chain properties:
t1 ) hydrophobic: norfeucine, met, ala> val, leu; ile; , t2) neutral hydrophilic: cys, ser, thr;
t3) acidic: asp, glu;
t4) basic: a5n, gln, his:' lys, arg;
(5) residues that influence chain orientation: gly, pro; and (6) aromatic: trp; tyr, phe.
Non-conservative substitutions will entail exchanging a member of one of these classes ' 30 for another. Such substituted residues may be introduced into regions of the target polypeptide that are homologous with other antibodies of the same class or subclass, or, more preferably, into the non-homologous regions of the molecule.
any cysteine residues not involved' in maintaining the proper conformation of target polypeptide also may be substituted, generally with serine, to improve the oxidative stability ' ,;.. ...
WO 92/22653 . ~ ~ ~ J ~ ~ ~ PCg'/US92/U5126 3s of the molecule and prevent aberrant crasslinking.
DNA encoding amino acid sequence variants of the target polypeptide is prepared by a variety of methods known in the art. These methods include, but are not limited to, isolation from a natural source tin the case of naturally occurring amino acid sequence variants) or preparation by oligonucleatide-mediated (or site-directed) mutagenesis, PCR
mutagenesis, and cassette mutagenesis of an earlier prepared variant or a non-variant version of the target polypeptide. A particularly preferred method of gene conversion mutagenesis is described below in Example 1. These techniques may utilized target potypeptide nucleic acid tDNA or RNA), or nucleic acid complementary to the target polypeptide nucleic acid.
to Otigonucleotide-mediated mutagenesis is a preferred method for preparing substitution, deletion, and insertion variants of target polypeptide DNA. This technique is well known in the art as described by Adetman et al., _D_N~A, ~: 183 (1983). Briefly, the target polypeptide DNA
is altered by hybridizing an oligonucleotide encoding the desired mutation to a DNA template, where the template is the single-stranded form of a plasmid or bacteriophage containing the unaltered or native DNA sequence of the target potypeptide. After hybridization, a DNA
potymerase is used to synthesize an entire second complementary strand of the template that will thus incorporate the otigonucleotide primer, and wilt code for the selected alteration in the target polypeptide DNA:
Generally, oligonucleotides of at least 25 nucleotides in length are used. An optimal oligonucteotide wilt have 12 to 15 nucleotides that are camptetely complementary to the template on either side of the nucleotidets) coding for the mutation. This ensures that the otigonucteotide will hybridize properly to the single-stranded DNA template molecule. The oligonucleotides are readily synthesized using techniques known in the art such as that described by Crew et al. tProg: Na~rl. Acad. Sci. USA, ~: 6765 (19781).
Single-stranded DNA template may also be generated by denaturing doubts-stranded plasmid for other) DNA usirfg standard techniques.
For alteration of the native DNA sequence tto generate amino acid sequence variants, fer example), the oligonucteotide is hybridized to the single-stranded template under suitable hybridization conditions: A DNA polymerizing enzyme, usually the Ktenow fragment of DNA
polymerase t, is then added to synthesize the complementary strand of the template using the otigonucleotide as a primer for synthesis. A heteroduptex molecule is thus formed such that one stand of DNA encodes he mutated form of the target polypeptide, and the other strand tthe original template) encodes the native, unaltered sequence of the target potypeptide. This heteroduplex molecule is then transformed into a suitable host cell, usually a prokaryote such :,';~: .; ,'.'.,,.,., ': .~.... . . '..:,~... ..
WO 92/22653 ~ ~ ~ '~ ~ '~ ~ 1'CT/1JS92/05126 3~
as E, coil ,lM101. After the cells are grown, they are plated onto agarase plates and screened using the oliganucleotide primer radiolabeled with 32-phosphate to identify the bacterial colonies that contain the mutated DNA. The mutated region is then removed and placed in an appropriate vector for protein production, generally an expression vector of the type typically employed for transformation of an appropriate host.
The method described immediately above may be modified such that a homoduplex molecule is created wherein both strands of the ptasmid contain the mutation(s). The modifications are as follows: The single-stranded oligonucleotide is annealed to the single-stranded template as described above. A mixture of three deoxyribonucleotides, deoxyriboadenosins tdATP); deoxyriboguanosine tdGTP), and deoxyribothymidine tdTTP), is combined with a modified thin-deoxyribocytosine called dCTP-taS) twhich can be obtained from Amersham Corporation). This mixture is added to the template-oligonucieotide complex.
Uport addition of DNA polymerase to this mixture, a strand of DNA identical to the template except for the mutated bases is generated. ln~ addition, this new strand Ot UNA wul contain dCTP-taS) instead of dCTP, which serves to protect it from restriction endonuclease digestion.
After the template strand of the double-stranded heteroduplex is nicked with an appropriate restriction enzyme; the template strand can be digested with VIII
nuclease or another appropriate nuclease past the region that contains the sitets) to be mutagenized. The reaction is then stopped to leave a molecule that is only partially single-stranded. A complete double-stranded DNA homoduplex is then formed using DNA polymerase in the presence of all four deoxyribonucleotide triphosphates, ATP, and DNA tigase. This homoduplex molecule can then be transformed into a suitable host cell such as E. coil JM10'6, as described above.
DNA encoding target polypeptide variants with more than one amino acid to be substituted may be generated i~ one of several ways. If the amino acids are located close tog~ther in the polypep~tide chain, they may be mutated simultaneously using one o6igonuctdotide that podes for alt of the desired amino acid substitutions.
If, however, the amino acids are located some distance from each other (separated by more than about ten amino acids), it is more difficult to 'generate a single oligonucleotide that encodes all of the desired changes. Instead, one of two alternative methods may be employed.
tn the first method, a separate oligonucleotide is generated for each amino acid to be substituted. The oligonucleotides are then annealed to the single-stranded template DNA
simultaneously, and the second strand of DNA that is synthesized from the template will encode alt of the desired amino acid substitutions.

WU 92/22653 ~ ~ ~ ~ ~ J ~ PCTlUS92lOSf26 3 '~
The alternative method involves two or more rounds of mutagenesis to produce the desired mutant. The first round is as described for the single mutants: wild-type DNA is used for the template, an oligonucleotide encoding the first desired amino acid substitutions) is annealed to this template, and the heteroduplex DNA molecule is then generated. The second round of mutagenesis utilizes the mutated DNA produced in the first round of mutagenesis as the template. Thus, this template already contains one or more mutations. The oligonucleotide encoding the additional desired amino acid substitutions) is then annealed to this template, and the resulting strand of DNA now encodes mutations from both the first and second rounds of mutagenesis. This resultant DNA can be used as a template in a third round Zo of mutagenesis, and so on.
PCR mutagenesis is also suitable for making amino acid variants of target polypeptide.
While the following discussion refers to DNA, it is understaad that the technique also finds application with RNA. The PCR technique generally refers to the following procedure (see Erlich, supra, the chapter by R. Higuchi, p. 61-70): When small amounts of template DNA are ~5 Bused as starting material in a PCR, primers that differ slightly in sequence from the corresponding region in a template DNA can be used to generate relatively large quantities of a specific DNA fragment that differs from the template sequence only at the positions whore the primers differ from the template. Far introduction of a mutation into a plasmid DNA, one of the primers is designed to overlap the position of the mutation and to contain the mutation;
20 the sequence of th~ other primer must be identical to a stretch of sequence of the apposite strand of the ptasmid, but this sequence can be Located anywhere along the plasmid DNA. It is preferred, however, that the sequence of the second primer is located within 2~0 nucleotides from that of the first, such that in the end the entire amplified region of DNA
bounded by the primers can be easily sequenced. PCR amplification using a primer pair like 25 the one just described results in a population of DNA fragments that differ at the position of the mutation specified by the primer, and possibly at other positions, as template copying is somewhat error-prone.
If the ratio of template to product material is extremely low, the vast majority of product DNA fragments incorporate the desired mutation(s). This product material is used to 30 replace the corresponding region in the plasmid that served as PCR template using standard DNA technology. Mutations at separate positions can be introduced simultaneously by either using a mutant second primer, or performing a second PCR with different mutant primers and ligating the two resulting PCR fragments simultaneously to the vector fragment in a three (or rreore?-part ligation.

,v, n ' ~:. , ,.,~~.", ~.,...., ..~ ,~,~,~'~~~v~. ;~... ; .., ,.,~:.' ~.:w,, .
. .~...;.,:,. ;.-i,~~..~~ ~~.,;:
PCTlUS92l05126 WO 92!22653 In a specific example of PCR mutagenesis, template plasmid DNA (1 ,ug) is lineariaed by digestion with a restriction endonuclease that has a unique recognition site in the plasmid DNA outside of the region to be amplified. Of this material, 100 ng is added to a PCR mixture containing PCR buffer, which contains the four deoxynucleotide tri-phosphates and is included in the GeneAmp'~ kits (obtained from Perkin-Elmer Cetus, Norwalk, CT and Emeryville, CA), and 25 pmole of each aligonuclsotide primer, to a final volume of 50 Nl. The reaction mixture is overlayed with 35 NI mineral oil. The reaction is denatured for 5 minutes at 100~C, placed briefly an ice, and then 1 NI Thermus aquaticus (Taql DNA polymerise t5 unitslNl, purchased from Perkin-Elmer Cetus, Norwalk, CT and Emeryvilie, CA) is added below the mineral oil layer.
to The reaction mixture is then inserted into a DNA Thermal Cycler (purchased from Perkin-Eimer Cetus) programmed as fol!~ws: 2 min. at 55~C, then 30 sec. at 72~C, then 19 cycles of the following: 30 sec. at 94~C, 30 sec. at 55~C, and 30 sec. at 72~C.
At the end of the program, the reaction vial is removed from the thermal cycler and the aqueous phase transferred to a new vial, extracted with phenol/chlorafarm t50:50:vo1), and 'ethanol precipitated, and the DNA is recovered by standard procedures. This material is subsequently subjected to the appropriate treatments for insertion into a vector.
Another method for preparing variants, cassette mutagenesis, is based on the technique described by Wells etal. tGen_e,~4: 315 (19851). The starting material is the plasmid for other vector? comprising the target polypeptide DNA to be mutated. The cadants) in the target polypeptide DNA to be mutated are identified. There must be a unique restriction endonuclease site an each side of the identified mutation sitets). If no such restriction sites exist, they may be generated using the above-described oliganucleatide-mediated mutagenesis method to introduce them at appropri to locations in the target polypeptide DNA. After the restriction sites have been introduced into the plasmid, the piasmid is cut at these sites to finearize it. A double-stranded oligonucleatide encoding the sequence of the DNA between the restriction sites but containing tho desired mutatiants) is synthesized using standard procedures. The two strands are synthesized separately and then hybridized together using standard techniques. This double-stranded oligonucleotide is referred to as the cassette. This cassette is designed to have3' and 5' ends that are compatible with the ends of the linearized plasmid, such that it can be ditectiy ligated to the plasmid. This plasmid now contains the mutoted target polypeptide DNA sequence.
Insertion of DNA into a Cloning 'Vehicle The cDNA or gertomic DNA encoding the target palypeptide is inserted into a replicable ,.
.., . . "e~. ...
. . . " . ,. . ,,,. ~ , .'. , . .:. . . .,. . ,. ,. ,..
:,.::. ,.~' '.'.~~ ~~..;~, . ',. .. .. ~...; , ~ 7, .. r..., .' :.,, .,..,.
.., . . . .. . . . . ~ ~
~_~.~ ~~~J~
'WO 92!22653 PCT/US92105126 vector for further cloning tamplification of the DNA) or for expression. Many vectors are available, and selection of the appropriate vector will depend on 1 ) whether it is to be used for DNA amplification or for DNA expression, 2) the size of the DNA to be inserted into the vector, and 3) the host cell to be transformed with the vector. Each vector contains various components depending on its function (amplification of DNA or expression of DNAI and the bast cell for which it is compatible. The vector components generally include, but are not limited to, one or mare of the following: a signal sequence, an origin of replication, one or more marker genes, an enhancer element, a promoter, and a transcription termination sequence.
(a) ,signal Se uenc~ Component In general, the signal sequence may be a component of the vector, or it may be a part of the target polypeptide DNA that is inserted into the vector.
The target poiypeptides of this invention may be expressed not only directly, but also i5 gas a fusion with a heteraiagous polypeptide, preferably a signal sequence or ether polypeptide having a specific cleavage site at the N~terminus of the mature protein or polypeptide. In general, the signal sequence may be a component of the vector, or it may be a part of the target polypeptide DNA that is inserted into the vector. Included within the scope of this invention are target paiypeptides with any native signal sequence deleted and replaced with a heterotogaus signs! sequence. The heterologaus signal sequence selected should be one that is recognized and processed (i.e. cleaved by a signal peptidase) by the host cell. For prokaryotic host cells that do not reGOgnize and process the native target paiypeptide signet sequence, the signal sequence is substituted by a prokaryotic signal sequence Selected, for example, from the group of the alkaline phosphatase, penicillinase, Ipp, or heat-stable 'enterotoxin If leaders, For yeast seG~etion the native target polypeptide signal sequence may be substituted by the yeast invprtase, dlpha,factor, or acid phasphatase leaders. in mammalian cell expression the native signal sequence is satisfactory, although ether mammalian signal sequences may be suitable.
(b) ~?rigin of Replication Comaonent Both expressiart and cloning vectors contain s nucleic acid sequence that enables the vector to replicate in one or more 'selected host cells. Generally, in cloning vectors this sequence is one that enables the vector to replicate independently of the host chromosomal ,,;, ;., . , . :, ,,:
~ ~;.,. . .. . ,. . ,~, , ~; .. .'' , . , ...,' , ..y( .y' ,. !.. ..~.~. ..., ':. ' ..,' ~ ,. . .. . . y" . , ~..
PC'f'JUS92/OSI z6 wo 9zJZZ~s3 2 ~ ~ ~ ~ '~ ~
~0 DNA, and includes origins of replication or autonomously replicating sequences. Such sequences are well known for a variety of bacteria, yeast, and viruses. The origin of replication from the plasmid pBR322 is suitable for most Gram-negative bacteria, the 2N
plasmid origin is suitable for yeast, and various viral origins tSV40, polyoma, adenovirus, VSV
or BPV) are useful for cloning vectors'in' mammalian cells. Generally, the origin of replication component is not needed for mammalian expression vectors (the SV40 origin may typically be used only because it contains the early promoters.
Most expression vectors are "shuttle" vectors, i.e. they are capable of replication in at least one class of organisms but can be transfected into another organism for expression. For io example, a vector is cloned in E. coli and then the same vector is transfected into yeast or mammalian cells for expression even though it is not capable of replicating independently of the host cell chromosome.
DNA may also be amplified by insertion' into the host genome. This is readily accomplished using Bacillus species as hosts, for example, by including in the vector a DNA
sequence that is complementary to a sequence found in Bacillus genomic DNA.
Transfection of Bacillus with this vector results in homologous recombination with the genome and insertion of the target poiypeptide DNA. However, the recovery of genomic DNA encoding the target polypeptide i~ more complex than that of an exogenously replicated vector because restriction enzyme digestion is required to excise the target poiypeptide DNA.
tc) ~~ c.t~~ene Component Expression and cloning vectors should contain a selection gene, also termed a selectable marker. This gene encodes a protein necessary for the survival or growth of transformed host cells grown 'in a selective culture medium. Host cells not transformed with the vector containing the selection gene will not survive in the culture medium. Typical selection genes encode proteins that ta) confer resistance to antibiotics or other toxins, e.g. ampiciltin, neomycin, methotrexate; or tetracycline; tb) complement auxotrophic deficiencies, or tc) supply critical nutrients not available frorro complex media, e.g. the gene encoding D-alanine racemase for Bacilli.
One example of a selection scheme utilizes a drug to arrest growth of a host cell.
Those calls that are successfully-transformed with a heterologous gene express a protein conferring drug resistance and thus survive the selection regimen. Examples of such dominant selection use the drugs neomycin (Southern et al., ~. Motec. Appl. Genet., 1:
S27 (19821), Frtt/ ,~ 9Q ';. S V .III ;Yr''~,2~W~P .ir.'P(j~.a=.,..
~nt,"','.~."
~~a~~JJ
''YO 92/2265 P~1'/US921o5126 '-I I
mycophenolic acid (Mulligan et al., cien~_e., 209: 1422 [ 19801) or hygromycin (Sugden et al. , Mol. Cell. Biol., 5_: 410-413 (19851). The three examples given above employ bacterial genes under eukaryotic control to convey resistance to the appropriate drug 6418 or neomycin (geneticin), xgpt (mycophenolic acid), or hygromycin, respectively.
Another example of suitable selectable markers for mammalian cells are those that enable the identification of cells competent to take up the target polypeptide nucleic acid, such as dihydrofolate reductase (DHFR) or thymidine kinase. The mammalian cell transformants are placed under selection pressure which only the transformants are uniquely adapted to survive by virtue of having taken up the marker. Selection pressure is imposed by culturing the transformants under conditions in which the concentration of selection agent in the medium is successively changed; thereby leading to amplification of both the selection gene and the DNA that encodes the target polypeptide. Amplification is the process by which genes in greater demand for the production of a protein critical for growth are reiterated in tandem ~vyithin the chromosomes of successive generations of recombinant cells.
Increased quantities 15 of the target polypeptide are synthesized from the amplified DNA.
For example, cells transformed with the DHFR selection gene are first identified by cultdring all of the transformants in a culture medium that contains methotrexate (Mtx), a competitive antagonist of Di~FR. An appropriate host cell when wild-type DHFR
is employed is the Chinese hamster ovary (CHO) cell line deficient in DHFR activity, prepared and 20 propagated as described by Urlaub and Chasin, Proc. Natl. Acad. Sci. USA, 77: 4216 ~1980~.
Tne transfocrned cells are then exposed to increased levels of methotrexate.
This leads to the synthesis of multiple copies of the DHFR gene, and, concomitantly, multiple copies of other DNA comprising thd expression vectors, such as the DNA encoding the target polypeptide.
This amplification technique can be used with any otherwise suitable host, e.g., ATCC No.
CCL61 CHO-K1, notwithstanding the presence of endogenous DHFR if, for example, a mutant DHFR gene that is highly resistant to Mtx is employed (EP 117,060?.
Alternatively, host cells (particularly wild-type hosts that contain endogenous DHFR) transformed or co-transformed with DNA sequendes encoding the target polypeptide, wild-type DHFR protein, and another selectable marker such as arninoglycoside 3' phosphotransferase (APH) can be selected by cell growth in medium containing a selection agent for the selectable marker such as an aminoglycosidic antibibtic, e:g., kanamycin, neomycin, or 6418. See U.S. Pat.
No.

4,965;199.
A suitable s~iection gene for use in yeast is the trill gene present in the yeast plasmid YRp7 (Stinchcc~mb et al:; Na ure, 282: 39 (19791; Kingsman et al., Gene, 7:
141 [1979D; or '.. r.;.; , ~
~r,..~ ~.5',. b . .1.~~. ' r ~1.7: . ~~i:9,.. , r . .. . ' .. . '~ . . s... .

~I Z
Tschemper et al., Gene, 10: 157 [1980]?. 'The trp1 gene provides a selection marker for a mutant strain of yeast lacking the ability to grow in tryptophan, for example, ATCC No. 44076 or PEP4-1 (Jones, Genetics, 85: 12 [19771). The presence of the trp1 lesion in the yeast host cell genome then provides an effective environment for detecting transformation by growth in the absence of tryptophan. Similarly, Leu2-deficient yeast strains tATCC
20,622 or 38,626) are complemented by known plasmids bearing the Leu2 gene.
td) Promoter ramoonent Io Expression and cloning vectors usually contain a promoter that is recognized by the host organism and is operably linked to the target polypeptide nucleic acid.
Promoters are untranslated sequences located upstream t5') to the start codon of a structural gene (generally within about 100 to 1000 bpl that control the transcription and translation of a particular nucleic acid sequence, such as that encoding the target polypeptide, to which they are ~operably linked. Such promoters typically fall into two classes, inducible and constitutive.
Inducible promoters ere promoters that initiate increased levels of transcription from DNA
under their control in response to some change in culture conditions, e.g. the presence or absence of a nutrient or a change in temperature. At this time a large number of promoters recognized by a variety of potential host cells are well known. These promoters are operably 2Q linked to DNA encoding the target polypeptide by removing the promoter from the source ANA
by restriction enzyme digestion and inserting the isolated promoter sequence into the vector.
Both the native target polypeptide promoter sequence and many heterologous promoters may be used to direct amplification andlor expression of the target polypeptide DNA. However, heterologous promoters are preferred, as they generally permit greater transcription and higher yields of expressed target polypeptide as compared to the native target polypeptide promoter.
Promoters suitable for use with prokaryotic haste include the ~B-lactamase and lactose promoter systems tChang et al., _Naturg, 275: 615 [1978]; and Goeddel et al., Nature, 2 1:
544 [ 19791), alkaline phosphatase, a tryptophan ttrp) promoter system (Goeddel, Nu,~leic Aids Res., 8: 4057 119801 and ER 36,761 and hybrid promoters such as the tac promoter tdeBoer ' 3o et al., Proc. Natl. Acad: Sci. USA; ,$Q: 21-25 [19831). However, other known bacterial pr~moters are suitable. Their nucleotide sequences have been published, thereby enabling a skilled worker operably: to ligate them to DNA encoding the target polypeptide tSiebenlist et al.Ceii; 20: 269 (1980)) using linkers or adaptors to supply any required restriction sites.
Promoters lot use in bacterial systems also generally will contain a Shine-Dalgarno tS.D.) d1'O 92122653 . ~ ~ ~ ~ ~~ ~ ~ PC.'T/US92105126 ~I 3 sequence operably linked to the DNA encoding the target polypeptide.
Suitable promoting sequences for use with yeast hosts include the promoters for 3-phasphoglycerate kinase (Hitzeman et al., J. Biol. Chem., 2~~'5: 2073 [1980)) or other glyco)ytic -enzymes (Hess et al., J Adv. Enzyme Reg~., 7: 149 (19681; and Holland, Biochemistrv,17: 4900119781), such asenolase, glyceraldehyde-3-phosphate dehydrogenase, hexokinase, pyruvate decarbaxylase, phosphafructakinase, glucose-6-phosphate isomerase, 3-phosphoglycerate mutase, pyruvate kinase, triosephosphate isamerase, phasphoglucase isomerase, and glucakinase.
Other yeast promoters, which are inducib)e' promoters having the additional advantage of transcription controlled by growth conditions, are the promoter regions for alcohol dehydrogenase 2, isocytochrome C, acid phosphatase, degradative enzymes associated with nitrogen metabolism, metallathianein, glyceraldehyde-3-phosphate dehydrogenase, and enzymes responsible for maltose and gafactose utilization. Suitable vectors and promoters for use in yeast expression are further described in Hitzeman et al., EP 73,857A.
Yeast enhancers also are advantageously used with yeast promoters.
Promoter sequences are known for eukaryotes. Virtually ail eukaryotic genes have an AT-rich region located approximately 25 to 30 bases upstream from the site where transcription is initiated. Another sequence found 70 to 80 bases upstream from the start of transcription of many genes is a CXCAAT region where X may be any nucleotide.
At the 3' end of mast eukaryatic genes is an AATAAA sequence that may be the signal for addition of the poly A tail to the 3' end of the coding seauence. All of these sequences are suitably inserted into mammalian expression vectors.
Target polypeptide transcription from vectors in mammalian host cells is controlled by promoters obtained from the genomes of viruses such as polyoma virus, fowlpox virus (Ul~
z5 2,211,504 published S July 1989), adenovirus (such as Adenovirus 2), bovine papit)oma virus, avian sarcoma virus, cytamegatovirus, a retrovirus, hepatitis-B virus and mast preferably Simian Virus 40 (SV40), from heterologous mammalian promoters, e.g. the actin promoter ar an immunogl~bulin promoter, from heat-shock promoters, and from the promoter normally associated with the target palypeptide sequence, provided such promoters are compatible with the host cell systems.
The early and late promoters of the SV40 virus are conveniently obtained as an restriction fragment that also contains the SV40 viral origin of replication.
Fiers et al" Na ure, 27:113 11978); Mulligan and Berg, fence, 2~9: 1422-1427 (1980): Pavlakis et al., Proc.
N~,~1 Acad Sci USA, 78: 7398-7402 (1981 ). The immediate early promoter of the human .'.~. , ~ . ~~. . .::. ~,....,. u,... ...w, n:' ..."' .,:....,.,. . ~['~.",.
~.' ,... '..:.. ' . .. . . ' ... ... .~ . ..
~ r. ,. . ~.~. ~~ ~ '. . .' '.i ne , -'.:.. . , : ~ , f',.'., :. ... . .. ...
~..'.~ ..'.. :. ' : ? , n,;,. ..;~ n . ~ :'. .' . ' .. . '. : . .. . .
sd.~~Jil:~~~
W~ 92/22653 PCT/US92/05126 cytomegalovirus is conveniently obtained as a Hindlll E restriction fragment.
Greenaway et al. , Gene, 18: 355-360 (198Z). A system for expressing DNA in mairnmalian hosts using the bovine papilloma virus as a vector is disclosed in U.S. 4,419,446. A
modification of this system is described in U.S. 4,601,978. See also Gray et al,, Nature, ~: 503-508 (1982) on expressing cDNA encoding immune interferon in monkey cetts; , Reyes et al., N ure, 297: ~
598-601 (1982) on expression of human ~-interferon cDNA in mouse cells under the control of a thymidine kinase promoter from herpes simplex virus, Canaani and Berg, Pr2q,. Natl. Acad.
~;i. USA, 79: 5166-5170 (1982) on expression of the human interferon,81 gene in cultured mouse and rabbit cells, and Gorman et al., Plc. N~tl. Acad. S~;i. USA, ,~Q:
6777-6781 ( 1982) 1o on expression of bacterial CAT sequences in CV-1 monkey kidney cells, chicken embryo fibroblasts, Chinese hamster ovary cells, HeLa cells, and mouse NIH-3T3 cells using the ffious sarcoma virus long terminal repeat as a promoter. ~
(e) Enhan~er Element Camnonent ' Transcription of DNA encoding the target polypeptide of this invention by higher eukaryotes is often increased by inserting an enhancer sequence into the vector. Enhancers are cis-acting elements of DNA, usually about from 10-300 bp, that act on a promoter to increase its transcription. Enhancers are relatively orientation and position independent having 2p been found 5' tLaimins et al., Per ~. Natt. Acad. Sri- USA, 7$: 993 ( 19811) and 3' (Lusky et a/..; Mol: Cell Bio., ~: 1108 t1983)t to the transcription unit, within an intron (Banerji et al., Cell, ,~: 729 f 19t~31) as well as within the coding sequence itself (Osborne et al., M I. ell Bio., 4: 1233 [1984)). Many enhancer sequences are now known from mammalian genes (globin, etastase; albumin, o-fetoprotein and insulin). Typically, however, one will use an enhancer frt~m s eukaryotic cell virus: Examples include the SV40 enhancer on the late side of the replication origin (bp '100-270), the cytpmegalovirus early promoter enhancer, the polyoma enhancer on the late side of the replication origin, and adenovirus enhancers. See also Yaniv, N_ature_, 297:' 17-18 11982) on enhancing elements for activation of eukaryotic promoters. The enharycer may be spliced into the vector at a position 5' or 3' to the target potypeptide DNA, but is preferably located at a site 5' from the promoter.
(f) Transcription 'termination ComQOnent Expression vectors used in eukaryotic host cells (yeast, fungi, insect, plant, animal, WO X2/22653 ~ ~ ~ ~ ~ a ~ PCTlUS92/05126 human, or nucleated cells from other multicellular organisms) will also contain sequences necessary for the termination of transcription and for stabilizing the mRNA.
Such sequences are commonly available from the 5' and, occasionally 3' untranslated regions of eukaryotic or viral DNAs or cDNAs. These regions contain nucleotide segments transcribed as polyadenylated fragments in the untranslated portion of the mRNA encoding the target polypepride. The 3' untranslated regions also include transcription termination sites.
Construction of suitable vectors containing one or more of the above listed components the desired coding and control sequences employs standard ligation techniques.
Isolated plasmids or DNA fragments are cleaved, tailored, and religated in the form desired to generate the plasmids required.
For analysis to confirm correct sequences in plasmids constructed, the ligation mixtures are used to transform E, coli K12 strain 294 tATCC 31,446) and successful transformants selected by ampicillin or tetracycline resistance where appropriate. Plasmids from the transformants are prepared; analyzed by restriction endonuclease digestion, and/or sequenced °lay the method of Messing e# a/., Nucleic Acids Res., ~,: 309 11981 ) or by the method of Maxam et al., Methods in Enzvmoloqv_, ,~5_: 499 11980).
Particularly useful in the practice of this invention are expression vectors that provide for the transient expression in mammalian cells of DNA encoding the target polypeptide. in general; trarvsient expression involves the use of an expression vector that is able to replicate efficiently in a host cell; such that the host cell accumulates many copies of the expression vector and, in turn, synthesizes high levels of a desired polypeptide encoded by the expression vector. Transient expression systems, comprising a suitable expression vector and a host cell, allow for the convenient positive identification of polypeptides encoded by cloned DNAs, as well as for the; rapid screening of such .polypeptides for desired biological or physiological properties. Thus: transient expression systems are particularly useful in the invention for purposes of identifying analogs and variants of the target polypeptide that have target polypsptide-like activity:
~ther methods, vectors; and host cells suitable for adaptation to the synthesis of the target poiypeptide in recombinant vertebrate cell culture are described in Gething etal., N re, ' 3o ' 2."~,~: 620-625 (19811; Mantel et al.-, N re, 2 1: 40-46 [19791;
Levinson et al.; EP 117,060;
and EP 117,058. A particularly useful plasmid for mammalian cell culture expression of the target poiypeptide is pRKS tEP-pub. no. 307,247) or pSVi6B.
Seie~tsrt~ and Transformation of Host Cells v ~I ~P
Suitable host cells for cloning or expressing the vectors herein are the prokaryote, yeast, or higher eukaryote cells described above. Suitable prokaryotes include eubacteria, such as Gram-negative or Gram-positive organisms, for example, E. toll, Bacilli such as B. subtilis, Pseudomonas species such as P. aeruginosa, Salmonella typhimurium, or Serratia marcescans.
One preferred E. toll cloning host is E. toll 294 (ATCC 31,446), although other strains such as E, cofi B, ,E. toll X1776 (ATCC 31,537), and E. toll W3110 (ATCC 27,325) are suitable.
These examples are illustrative rather than limiting. Preferably the host cell should secrete minimal amounts of proteoPytic enzymes. ,A(ternatively, in vitro methods of cloning, e.g. PCR
or other nucleic acid polymerase reactions, are suitable.
1o In addition to prokaryotes, eukaryotic microbes such as filamsntous fungi or yeast are suitable hosts for target polypeptids-encoding vectors. Saccharomyces cerevisiae, or common baker's yeast, is the most commonly used among lower eukaryotic host microorganisms.
However, a number of other genera, species, and strains are commonly auaiiable and useful herein, such as Schizosaccharomyces pombe (Beach and Nurse, Na ur , ,2~Q: 140 (1981 ); EP
1S P139,383 published May 2, 19851, Kluyveromyces hosts (U.S. 4,943,529) such as, e.g., K.
lactis (Louvencourt et al., ~. B~~teriol., 737 (1983)), K. arragilis, K.
bulgaricus, K.
thern~otolerans, and K. marxianus, yarrowia tEP 402,2261, Pichia pastoris IEP
183,070;
Sreekrishna et al., J. Basic Microbiol., 28: 265-278 (1988)1, Candida, Trichoderma reesia fEP
244,2341, Neurospora crassa (Case et al., Proc,, Natl. Asad. Sci. USA, ~: 5259-52(3 20 (1979)1, and filamentous fungi such as, e.g, Neurospora, Penicillium, Tolypocladium [WO
91100357 published 10 January 19911, and Aspergillus hosts such as A. nidulans (l3allance etal., Biochem. Bi~phvs. Res. Commun., 112_: 284-289 (1983); Tilburn et al., en , ~6: 205-221 (1983); Yelton et al., Proc. Natl. e4cad. Sci. USA, ~1: 1470-1474 (1984)) and A. niger (Kelly and Hynes, EMBO J., 4: 475-479 (1985)x.
25 Suitable host cells for the expression of glycosylated target polypeptide are derived from multicellular organisms. Such host calls are capable of complex processing and glycosylation activities. Pn principle; any higher eukaryotic cell culture is workable, whether from vertebrate or invertebrate culture. Examples of invertebrate cells include plant and insect cells.
Numerous baculoviral strains and variants and corresponding permissive insect host cells from 30 hosts such as Spodoptera frugiperda (caterpillar), Aedes aegypti (mosquito), Aedes albopictus (mosquito), Drosophila melanogaster (fruitfly), and Bombyx mori host cells have been identified. See, e.g., Luckow et al., B~,Ttechnologv. _6: 47-55 (1988); Miller et al., in ~n Ericlineerine. Setlow, J.K. et al., eds., Vol. 8 (Plenum Publishing, 1986), pp. 277-279; and Maeda et al., Nature, 315: 592-594 (1985). A variety of such viral strains are publicly J .~, ~ . . , .
'~ , f ~ ~f~r . ~... .~. .... ,... . ." . ... . . . , . ,. . ..
WtJ 92!22653 ~ ~ ~ ~ ~ ~ f'CT/US92105126 available, e.g, the L-1 variant of Autographa californica NPV and the Bm-5 strain of Bombyx mori NPV, and such viruses may be used as the virus herein according to the present invention, particularly for transfection of Spodoptera frugiperda cells. Plant cell cutrtres of cotton, corn, potato, soybean, petunia, tomato, and tobacco can be utilized as hosts.
Typically, plant cells are transfected by incubatian with certain strains of the bacterium Agrobacterium tumefaciens, which has been previously manipulated to contain the target polypeptide DNA. During incubation of the plant cell culture with A, tumefaciens, the DNA
encoding target polypeptide is transferred to the plant cell host such that it is transfected, and will, under appropriate conditions, express the target polypeptide DNA. In addition, regulatory and signal sequences compatible with plant cells are available, such as the nopaline synthase promoter and polyadenylation signal sequences. Depicker et aO , ~. Mol. Aopl.
Gen., 1_: 561 ! 1982). In addition, DNA segments isolated from the upstream region of the T-DNA 780 gene are capable of activating or increasing transcription levels of plant-expressible genes in recombinant DNA-containing plant tissue: See EP 321,196 published 21 June 1989.
'"~ However, interest has been greatest in vertebrate cells, and propagation of vertebrate cells in culture (tissue cultural has become a routine procedure in recent years (Tissue Culture, Academic Press, Kruse and Patterson, editors 11973)1. Examples of useful mammalian host cell tines are monkey kidney CV 1 line transformed by SV40 (COS-7, ATCC CRL
1651 ); human embryonic kidney line (293 or 293 cells subcloned for growth in suspension culture, Graham etal., J~Gen Virol., ~: 59 (19771): baby hamster kidney cells (BHK, ATCC CCL
10); Chinese hamster ovary cells/-D!-:FR tCHO, Urtaub and Chasin, PrQc. Natl. Acad. ci.
USA, 77: 4216 (19801): mouse settoli cells tTM4, Mather, Biol: Reorod., 23: 243-251 (19801);
monkey kidney cells tCV1 ATCC CCL 70); African green monkey kidney cells (VERO-76, ATCC CRL-1587);
human cervical carcinoma cells tHELA, ATCC CCL 2): canine kidney cells (MDCK, ATCC CCL
34); buffalo rat liver cells (BRL 3A, ATCC CRL 1442); human lung cells tW138, ATCC CCL
75); human liver cells tHep G2, HB 80651; mouse mammary tumor (MMT 060562, ATCC
CCL51); TRI cells (blather et ~1., Annals N:Y. Acad. Sci., ,: 44-68 (19821):
MRC 5 cells;
FS4 cells; and a human hepatoma cell line tHep G2). Preferred host cells are human embryonic kidney 293 and Chinese hamster ovary cells.
Host cells are transfected and preferably transformed with the above-described expression or cloning vectors of this 'invention and cultured in conventional nutrient media modified as appropriate for inducing promoters, selecting transformants, or amplifying the genes encoding the desiied sequences.
Transfection refers to the taking up of an expression vector by a host cell whether or Wo ~zm.~~ ~ 3 U j ~ . »crms9vosizs not any COdin9 sequences are in fact eXpr~ssed. Numerous methods of transfeetion are known to the ordinarily skilled artisan, far example, CaPO, and electroporarion.
Successful transfeCtion is generally recognized whorl any indication of the oporati4n of this vector occurs within the host Cell.
Transformation means introducing DIVA into an organism so that the L7NA is replicable, either as an extrachromosbmai element or by chromosomal integrant. Depending on the host cell used, transformation iS done using Standard techniques appropriate to such oeus. The Calcium treatment employing calcium chloride, as described ire section 9.8? of Sambrook et aL, supra, is generally used for prokaryotes or other cells that contain substanGat colt-wall io bar~l8rs. tnfaction with Agrobacteritrm tVmefaciens is used for transformation of certain plant coils, as described by Shaw et al., Gene. 2~' : 315 (19831 and WO 89/Q5859 published 2S
June 1989. For mammalian cells without such cell walls, the calcium phosphate precipitation method described in sections 16.33--t 8,37 of Samhrook et arl, supra, is prefarrgd. General aspects of mammalian cell host system transforrnations have been described bY
Axa1 in U.S. .
15 '4,399.216 issued 16 August '1983. Transformations into yeast are typically carried out according to the method of Van Solingen et at-, .,,1,~t3act., .1~: 946 11977) and Hsiaa et al., Prod. Natl_ A dad. Sci...CUSAL.~: 3829 ~~ 9791. tfawaver, other methods for introducing DNA
into cells such as by nuclear injection, alflctfOp4fatiOn. or protoplast fusion may also be used.
., 20 Culturing the Hpst Cells Prokaryotic tails used.ta produce th9 target polypeptide of thisvnventtan are cu~tured in.suitable media as described gerraraliy in Sambrodk et al., ~utrra.
The mammalian host tails used to produce the target polypeptide of this invention 25 ~ may be'cultured in a variety of media. Corrzme~cially available media such as Ham's F10.
(Sigma). Minimal Essential Medium tIMEnn), Sigma), f;PMI-16~i0 tSigma), and Dulbecco's Modified iragle'$ Medium t(DMEM3, Sigma) are suitable for culturing the host celis_ in addition, any of the media,describ~d-in Harri arid Walface. Meth. Erix.; ~$: 44 (1979), Barnes and Sato, Anal;~BiOCrlem.. ~: 255 11980),; U.S. 4;7fi7;704; 4,857,866; 4.927.762: or 4.580.655:
3o Wp gpl03430; WO 87/~Da195; may be used as culture media for the host cells.
Any of these media may be supplemented as necessary with hormones andlor other growth factors (such as insulin, tr2~nsferring or epidermal growth flGtor), salts (such as sodium ohlaride, calcium, magnesium, and phosphate), buffers (such as HOPES), nucleosides (such a5 adenosine and thymidine), antibiotics (such as GentamycinT"' drug), trace elements ...
I~YO 92l22S53 ~ ~ ~ ~ ~ ~ ~ PCflfJS92/4S126 (defined as inorganic compounds usually present at final concentrations in the micromolar range), and glucose or an equivalent energy source. Any other necessary supplements may also be included at appropriate concentrations that would be known to those skilled in the art.
The culture conditions, such as temperature, pFi, and the like, are those previously used with the host cell selected for expression, and will be apparent to the ordinarily skilled artisan.
The host cells referred to in this disclosure encompass cells in in vitro culture as well as cells that are within a host animal.
It is further envisioned that the target polypeptides of this invention may be produced by homologous recombination, or with recombinant production methods utilizing control io elements introduced into cells already containing DNA encoding the target poiypeptide currently in use in the field. For example, a powerful promoter/enhancer element, a suppressor, or an exogenous transcription modulatory element is inserted in the genome of the intended host cell in proximity and orientation sufficient to influence the transcription of DNA
encoding the desired target potypeptide. The control element does not encode the target ~~olypeptide of this invention, but the DNA is present in the host cell genome. One next screens for cells making the target polypeptide of this invention, or increased or decreased levels of expression, as desired:
DetPCting Gene Amnlifi,~tion/Ex_pression Gene amplification and/or expression may be measured in a sample directly; for exampld, by conventional Southern blotting, northern blotting to quantitate the transcription of mRNA-tThomas; Prod; Natl. Acad. Sci. USA, 77: 5201-5205 (1980j), dot blotting tDNA
analysis), or in situ hybridization; using an appropriately labeled probe, based on the sequences provided herein, Various labels may be employed, most commonly radioisotopes, particularly 32P~ however, other techniques may also be employed, such as using biotin-modified nucleotides'for introduction into a polynucleotide. The biotin then serves as the site for binding to avidin or antibodies, which may b~ labeled with a wide variety of labels, such as radionuctides, fluo~escers, enzymes, or the tike. Alternatively, antibodies may be employed that can recognize specific duplexes; including DNA duplexes, RNA duplexes, and DNA-RNA
hybrid duplexes or DNA-protein duplexes. The antibodies in turn may be labeled and the assay may be carried out where the duplex is bound to a surface, so that upon the formation of dd~iex an the surface, the presence of antibody bound to the duplex can be detected.
Gene expression, atternatively, may be measured by immunological methods, such as r~.t i~ ~ ~ ~C't/US91~45126 wo 92izz~~3 so immunohistoehgmicat staining of tissue sectibns and asset' of cell CUItUr9 Or ksadY fluids, to quantitate directly th~ expression of gene product. With immunohistochemical staining techniques, a cell sample is prepared, tYPically by dehydration and fixation, followed try reaction with labeled antlb4dies specific far the gene product coupled, where the labels are u$uafly visually detectable, such as enzymatic labels. fluorescorrt labels.
luminescent labals~
and the like. A particularly sensitive staining technique Suitable for use in the present invention is described by Hsu et a~ , ~~.' . W 7~4-X38 11980).
Antibodies useful for immunohistochemicat staining andlor assay of sample fluids may ba either monoclonal or palyclonal, and may be prepared in any mammal.
~o~weniantly, the to antibodies may be prepared ageinst,a native target~patypeptide yr against a synthetic peptide based on the 4~NA sequences, provided herein as described further in Section 4 below-D,urifir.~tinr'~ a ThB~noIyoeDtidB
. The target palypaptid~.pref~arably is racovsred fr~n the culture medium as a secreted '~olypeptide, although it also may be recovered from host call lysatss when directly expressed without a secretory signal.
WfYen the tgrgmt pQIYPePtide is expressed in a recombinant cell other than one of human .origin, the target polypepti~e is completely frees of proteins or polypeptides of human origin.
However, it i$ necessary tc purify the target polypeptide from recombinant cell proteins or Z.p. . ~ polypaptides to obtain preparations that ar,e substantially homogeneous ass to the target pol~rpeptide. As.a first step; ths-cultu~a rnedium or lysate is centrifuged to remove particulate celi~dabris. The membfene ahd soluble protein fraotians are then separated.
The target . paiypeptide maY then be purified frorriahe soluble protein fraction and from the membrane fraction of the ~ cultur~. iysate, depending on whether the target polypeptida is membrane b. The ialt~awing: , proceduies ~ are exemplary of suitable purification procedures:
fractionation.on immuriQaffiiiity ~or iori-exchange columns; ethanol precipitation; reverse phase ~. Hpl:C: , ahromatograptiy: ' an silica, ~ or on a ration exchange resin such as D>"AE;
chrom$tafocusing: 5DS-i'~GE;. ~mvinanium sulfate precipitation: gel filtration using. for example. Sephadex G~75; and.protein A 6aphsrase*columns to remove contaminants such as ;- 30 ~ IQG. ; .
~, ~ :Target palypeptide wariar<ts iii:which residues have been deleted, inserted or substituted are recovered in the.sa~np. ;fashion, taking account of any substantial changes in properties . ac~asiprved by the variation_ Few example, preparation,af a target polypaptide fusion with another,ptatein or polYpeDtide, e.g,a,bacterial or vita! antigen, facilitates purification; an *_trademark ' '' , ... '~' ~ . ' .
VlrO 92f226S3 ~ ~ ~ ~ PCT/gJS92fOS126 immur~oaffinity column containing antibody to the antigen (or containing antigen, where the target polypeptide is an antibody) can be used to adsorb the fusion.
Immunoaffinity columns such as a rabbit polyclonat anti-target polypeptide column can be employed to absorb the target polypeptide variant by binding it to at least one remaining immune epitope. A protease S inhibitor such as phenyl methyl sutfonyl fluoride (~MSF) also may be useful to inhibit proteolytic degradation during purification, and antibiotics may be included to prevent the growth of adventitious contaminants. One skilled in the art wile appreciate that purification methods suitable for native target polypeptide may require modification to account for changes in the character of the target polypeptide or its variants upon expression in recombinant cell culture.
Covalent ~llodifica~ions of Tara~e~"F'otvDeptides Covalent modifications of target polypeptides are included within the scope of this invention. One type of covalent modification included within the scope of this invention is a target potypeptide fragment. Target polypeptide fragments having up to about 4O amino acid residues may be conveniently prepared by chemical synthesis, or by enzymatic or chemical cleavage of the full-length target polypeptide or variant target polypeptide.
Other types of covalent modifications of the target polypeptide or fragments thereof are introduced into the molecule by reacting specific amino acid residues of the target polypeptide or fragments thereof with an organic derivatizing agent that is capable of reacting with selected side chains or the N- or C-terminal residues.
Cysteinyl residues most commonly are reacted with a-haloacetates (and corresponding amines), such as chBe~roacetic acid or chloroacetamide, to give carboxymethyl or carboxyamidomethyl derivatives. Gysteinyl residues also are derivatized by reactitm with bromotrifluoroacetone, a-bromo,8-(a-imidozoyl)propionic acid, chloroacetyl phosphate, N-alkylmaleimides, 3-vitro-Z-pyridyl disulfide, methyl 2-pyridyl disutfid~, p-chloromercuribenzoate, ~-chtoromercuri-4-nitrophenol, or chloro-?-nitrobenzo-2-oxa-1,3-diazole.
Histidyt residues are derivatized by reaction with diethylpyrocarbonate at pH

5.5-?.0 because this agent is relatively specific for the histidyl side chain. pare-bromophenacyl 3,0 bromide also is useful; the reaction is preferably performed in 0.1 M
sodium cacodytate at pH
fiØ
bysinyl and amino terminal residues are reacted with succinic or other carboxylic acid anhydrides. C~erivatization with these agents has the effect of reversing the charge of the lysinyl residues. Other suitable reagents for derivatizing Q-amino-containing residues include . ... . .., ~'~~~~ . , :~ .'~~ . o:~:' : ;. r ' , . ~ . ' ~. _ : _~ . , ~.,,.
.. ,,~._:.. . ~.;:. . .
W~ 92/22653 ~ ~ ~ ~ ~ ,'~ ~ )PC°A'/US92/05r26 ,..~ .
Z
imidoesters such as methyl picolinimidate; pyridoxal phosphate; pyridoxal;
chloroborohydride;
trinitrobenzenesulfonic acid; O-methylisourea; 2,4-pentanedione; and transaminase-catalyzed reaction with glyoxylate.
Arginyl residues are modified by reaction with one or several conventional reagents, 5 among them phenylglyoxal, 2,3-butanedione, 1,2-cyclohexanedione, and ninhydrin.
Derivatization of arginine residues requiros that the reactian be performed in alkaline conditions because of the high pKa of the guanidine functional group. Furthermore, these reagents may react with the groups of lysine as well as the arginine epsilon-amino group.
Th~ specific modification of tyrosyl residues may be made, with particular interest in introducing spectral labels into tyrosyl residues by reaction with aromatic diazonium compounds or tetranitromethane. Most commonly, N-acetylimidizole and tetranitromethane are used to form O-acetyl tyrosyl species and 3-nitro derivatives, respectively. Tyrosyl residues are iodinated using'z51 or'3'I to prepare labeled proteins for use in radioimmunoassay, the chloramine T method described above being suitable.
Carboxyl side groups taspartyl or glutamyl) are selectively modified by reaction with carbodiimides tR'-N = C = N-R'), where R and R' are different alkyl groups, sash as 1-cyclohexyt-3-(2-morpholinyl-4-ethyl) carbadiimide or 1-ethyl-3-(4-azonia-4,4-dimethylpentyl) carbodiimide. Furthermore, aspartyl and glutamyl residues are converted to asparaginyl and glutarninyl residues by reaction with ammonium ions., Derivatization with bifunctional agents is useful for crosslinking target palypeptide to a water-insoluble support matrix ar surface for use in the method for purifying anti-target paiyp~ptids antibodies, and vice versa. Commonly used crosslinking agents include, e.g., 1,1-bisidiazoacetyl)-2-phenylethane, glutaraldehyd~, N-hydroxysuccinimide esters, for example, esters with ~-azidosalicylic acid; homobifunctional imidaesters, including disuccinimidyl esters such as 3,3'-dithicrbistsuccinimidylpropionate), and bifunctional maleimides such as bis-N-rnaleimido-1,8-octane. Derivatizing agents such as methyl-3-tIp-azidophenyl)dithiolpropioimi-date yield photoactivatable intermediates that are capable of farming crosslinks in the presence of light. Alternatively, reactive water-insoluble matrices such as cyanogen bromide-activated carbohydrates and the reactive substrates described in U.S. 3,969,28?;
3,691,016;
4,195,128; 4,247,642; 4,229.53?; and 4,330,440 are employed far protein immobilization.
Glutaminyl and asp~araginyl residues are frequent6y deamidated to the carresponding glutamyl and aspartyl residues, respectively. Alternatively, these residues are deamidated under mildly acidic conditions. Either form of these residues falls within the scope of this inventian.

.,''VO92/22653 ~ ~ ~ ~ P~;f/f.JS92/05i26 Other modifications include hydroxylation of proline and lysine, phosphorylation of hydroxyl groups of Beryl or threonyl residues, methylation of the a-amino groups of lysine, arginine, and histidine side chains tT.E. Creighton, Pr ins: Structure and Molecular Pranerj~i, es, W.H. Freeman & Co., San Francisco, pp. 79-86 f 1983)), acetylation of th~ N-terminal amine, and amidation of any C-terminal carboxyl group.
Another type of covalent modification of the target polypeptide included within the scope of this invention comprises altering the native glycosylation pattern of the polypeptide.
By altering is meant deleting one or more carbohydrate moieties found in the native target polypeptide, andlor adding one or more glycosylation sites that are not present in the native target pofypeptide.
Clycosylation of polypeptides is typically either N-linked or O-linked. N-linked refers to the attachment of the carbohydrate moiety to the side chain of an asparagine residue. The tri-peptide sequences asparagine-X-serine and asparagine-X-threonine, where X
is any amino acid except proline, are the recognition sequences for enzymatic attachment of the carbohydrate moiety to th~ asparagir~e side chain. Thus, the presence of either of these tri-peptide sequences in a polypeptide creates a potential glycosylation site. O-finked glycosylation refers to the attadhment of one of the sugars N-acetylgalactosamine, galactose, or xytose; to a hydroxyamino acid, most commonly serine or threonine, although hydroxyproline or 5-hydroxylysine may also' be used.
Addition of g[ycosylation sites to the target polypeptide is conveniently accomplished by altering the amino acid sequence such hat it contains one or more of the above-described tri-peptide sequences (for N-linked glycosylation sites). The alteration may also be made by the addition of; or substitution by, one or more serine or threonine residues to the native target s''' polypeptide sequence tfor O-linked glycosylation sites?. Fat ease, the target polypeptide amino acid sequencb is prgferabl~ altered throughchanges at the DNA level, particularly by mutating the DNA encodieg the target polypeptide at p~esalected bases such that codons are generated than will translate into the desired amino acids. The DNA mutationtsl may be made using methods described above under the heading of "Amino Acid Sequenc8 Variants of Target 'Potypeptide".
Another means of ' increasing the number of carbohydrate moieties on the target polypeptide is by chemical or enzymatic coupling of glycosides to the polypeptide. These procedures are advantageous in that they do not require production of the polypeptide in a host cell that has glycosyiation capabilities for N- and O- linked glycosylation.
Depending on the coupling mode used, the sugartsf may be attached to ta? arginine and histidine, tb? free ~i'T..,..,.',~ , ~ ,' a :.~... .. ~ .-,. ~ ,:.': .. :..,. .::, ~ . . ~. ~.
.'.~~. ~. ,.:. ;'; _, . ,~. ..' . .,'.,,.'~~. , wo ~z>zz~s3 2 ~. ~ ~ ~ ~ ~ FCT/US9zl05126 carboxyl groups, tc) free sulfhydryl groups such as those of cysteine, td) free hydroxyl groups such as those of serine, threonine, or hydroxyproline, te) aromatic residues such as those of phenytalanine, tyrosine, or tryptophan, or (f) the amide group of glutamine.
These methods are described in WO 87/05330 published 11 September 1987, and in Aplin and Wriston (CF1C_ Crit Rev Biochem., pp. 259-306 (1981]).
Removal of carbohydrate moieties present on the native target polypeptide may be accomplished chemically or enzymatically. Chemical deglycosylation requires exposure of the polypeptide to the compound trifluoromethanesulfonic acid, or an equivalent compound. This treatment results in the cleavage of most or' ail sugars except the linking sugar (N-acetylglucosamine or N-acetylgalactosamine), while leaving the polypeptide intact. Chemical deglycosytation is described by Hakimuddin et al. (Arch Biochem. Bionhys., 259:52 (198?]) and by Edge et al. (Anal. Biachem., 11 :131 ( 19811). Enzymatic cleavage of carbohydrate moieties on polypeptides can be achieved by the use of a variety of endo- and exo-giycosidases as described by Thatakura et al. (fVleth. Enzvmol., 18:350 (1987]).
Glycosylation at potential ~lycosbtation sites may be prevented by the use of the compound tunicamycin as describ~d by Duskin et al. (J. Bial. them.. 2_x:3105 ( 19821).
Tunicamycin blocks the formation of protein-N-glycoside linkages.
Another type of covalent modification of the target polypeptide comprises linking the target poiypeptide to various nonproteinaceous polymers, e.g. polyethylene glycol, _ polypropylene glycol or polyoxyalkylenes, in the manner set forth in U.S.
4,640,835;
4:496,689; 4,301;144; 4;670,417; 4,791,192 ar 4,179,337.
The target polypeptide also may, be entrapped in microcapsules prepared, for example, by coacervation techniques or by interfacial polymerization Ifor example, hydroxymethylcellulase o~ getatir~-microcapsules and poly-(methytmethacylateD
microcapsufes, ~S ' respectively), in colloidal drug delivery systems tfor example, tiposomes, albumin microspheres, microemutsionsnano-particles and nanocapsules), or in macroemulsions. Such techniques are disclosed in Reminaton's Pharmaceutical Sciences, 16th edition, Osol, A., Ed., (1980).
Target polypeptide preparations are also useful in generating antibodies, for screening #or binding partners, as standards in assays for the target polypeptide te.g.
by labeling the target polypeptide for use as a standard in a radioimmunoassay, enzyme-linked immunoassay, or radioreceptar assay), in affinity purification techniques, and in competitive-type receptor binding' assays when labeled with radioiodine, enzymes, fluorophores, spin labels, and the like.
Since it is often difficult to predict in advance the characteristics of a variant target potypeptide, it will be appreciated that some screening of the recovered variant will be needed ~~~~,a : .:F: ;. :' .. .: . .. : . ~ w.r: . ..:.; . ",; , ... ..,, ::; . : ;;
: .
~~t..
,.,~ WO 92/22653 ~ ~ ~ ~ ~ ~ PL'I'/US92/OS126 S S' to select the optimal variant. For example, a change in the immunologicai character of the target polypeptide molecule, such as affinity for a given antigen or antibody, is measured by a competitive-type immunoassay. The variant is assayed for changes in the suppression or enhancement of its activity by comparison to the activity observed for the target polypeptide in the same assay. Other patential modifications of protein or palypeptide properties such as redox or thermal stability, hydrophobicity, susceptibility to proteolytic degradation, stability in recombinant calf culture or in plasma, or the tendency to aggregate with carriers or into multimers are assayed by methods well known in the art.
la Diac~na_~ i~,~nd Relatee~~l~,g_s of the Anti odies The antibodies of this invention are useful in diagnostic assays for antigen expression in specific cells or tissues. The antibodies are detectably labeled andJor are immobilized on an insoluble matrix.
The antibodies of this invention find further use far the affinity purification of the - 15 antigen from recombinant cell cult~rre or natural sources. Suitable diagnostic assays for the antigen and its antibodies depend on the particular antigen or antibody.
Generally, such assays include competitive and sandwich assays, and steric inhibition assays.
Competitive and sandwich methods employ a phase-separation step as an integral part of the method white steric inhibition assays are conducted in a single reaction mixture.
Fundamentally, the same 20 procedures are used for the assay of the antigen and for substances that bind the antigen, although certain methods wilt be fav~red depending upon the molecular weight of the substance being assayed: ' Therefore, the substance to be tested is referred to herein as an ar~alyte, irrespective of its status otherwise as an antigen or antibody, and proteins,$hat bind to the analyte, are denominated binding partners, whether they be antibodies, cell surface 25 ' receptors, or antigens.
Analyxical methods far he antigen or its antibodies all use one or more of the following reagents: labeled anaiyte analogue; immobilized ana)yte analogue, labeled binding partner, immobilized bindiryg partner and steric conjugates. The labeled reagents also are known as "tracers."
The label used _tand this is also useful to label antigen nucleic acid for use as a probe) is any detectable functionality 'that dogs not interfere with the binding of ana)yte and its binding partner. Numerous labels are known for use in immunoassay, examples including moieties that gay be detected directly, such as fluorochrome, chemiluminescent, and radioactive labels, as well as moieties, such as enzymes, that must be reacted or derivatized r?,~r k , ., ,.. ,. . ',.
PCT/ US92/0512b W~ 92/22653 5~
to be detected. Examples of such labels include the radioisotopes 3ZP, '4C, '25i, 3H, and '3'I, fluorophores such as rare earth chelates or fluorescein and its derivatives, rhodamine and its derivatives, dansyl, umbelliferone, luceriferases, e.g., firefly luciferase and bacterbal luciferase tU.S. Pat. No. 4,737,456);.luciferin, 2,3-dihydrophthalazinedianes, horseradish peroxidase 5' tHRP), alkaline phasphatase, ~-galactosidase, glucoamylase, lysozyme, saccharide oxidases, e.g., glucose axidase, galactose oxidase, and glucose-6-phosphate dehydrogenase, heterocyclic oxidases such as uricase and xanthine oxidase, coupled with an enzyme that employs hydrogen peroxide to oxidize a dye precursor such as HRP, lactoperoxidase, or micraperoxidase, biotin/avidin, spin labels, bacterrophage labels, stable free radicals, and the like.
Conventional methods are available to bind these labels covalently to proteins or polypeptides. For instance, coupling agents such as dialdehydes, carbadiimides, dimaleimides, bis-imidates, bis-diazotized benzidine, and the like may be used to tag the antibodies with the above-described fluorescent, chemiiuminescent, and enzyme labels. See, for example, U.S.
Pat. Nos. 3,940,475 tfluarimetry) and 3,645,090 (enzymes); Hunter et al., Nature, 144: 945 (1962); L?avid etal., Biach~mistry, 1~: 1014-1021 (1974); Pain etal., ~
Immunol. Methods, 40: 219-230 (1981); and Nygren, J. Histochem. and Cvtochem., ~0: 407-412 (1982).
Preferred labels herein are enzymes such as horseradish peroxidase and alkaline phosphatase.
The cpnjugation of such label; including the enzymes, to the antibody is a standard manipulative procedure far one of ordinary skill in immunoassay techniques.
See, for example, O'Sultivan et a/., "Methods for the Preparation of Enzyme-antibody Conjugates for Use in Enzyme Imrnunoassay." in Methods in Enzvmolaav, ed. J.J. Langone and H. Van Vunakis, Vol.
73 (Academic Press, New: Yprk; New York, 1981 ), pp. 147-166. Such bonding methods are suitable for use with the antibodies and polypeptides of this invention.
Immobilization of reagents is required for certain assay methods.
Immobilization entails separating the binding partner from any anatyte that remains Eras in solution.
This conventionally is accomplished by either insoiubilizing the binding partner or analyte analogue before the assay procedure, as by adsorption to a water-insoluble matrix or surface (Bennich etal.., U.S. 3,710,760), by covalent coupling (for example, using glutaraldehyde cross-finking), or by insolubilizing the partner or analogue afterward, e.g., by immunoprecipitation.
Other assay methods, known as competitive or sandwich assays, are well established and widely used in thd commercial diagnostics industry.
Competitive assays rely on the ability of a tracer analogue to compete with the test sample anaiyte for a limited number of binding sites on a common binding partner. The binding ry~,:~,;~.:v" ,. . ., ,.., wo ~2i226~~ ~ ~. ~ ~ ~ J ~~ ~orius~2i~5~z6 S~
partner generally is insolubilized before or after the competition and then the tracer and analyte bound Lo the binding partner are separated from the unbound tracer and analyte. This separation is accomplished by decanting (where the binding partner was preinsoiubilized? or by centrifuging (where the binding partner was precipitated after the competitive reactiony.
The amount of test sample analyte is inversely proportional to the amount of bound tracer as measured by the amount of marker substance. Dose-response curves with known amounts of analyte are prepared and compared with the test results to quantitatively determine the amount of analyte present in the test sample. These assays are called ELISA
systems when enzymes are used as the detectable markers. ' to Another species of competitive assay, called a "homogeneous" assay, doss not require a phase separation. Here, a conjugate of an enzyme with the analyte is prepared and used such that when anti-analyte binds to the analyte the presence of the anti-analyte modifies the enzyme activity. In this case, the antigen or its immunologically active fragments are conjugated with a bifunctional organic bridge to an enzyme such as peroxidase.
Conjugates are selected for use with antibody so that binding of the antibody inhibits or potentiates the enzyme activity of the label. This method per se is widely practiced under the name of EMIT.
Steric conjugates are used in steric hindrance methods for homogeneous assay.
These conjugates are synthesized by covalently linking a low-molecular-weight hapten to a small analyte so that antibody to hapten substantially is unable to bind the conjugate at the same 2o time as anti-analyte. Under this assay procedure the analyte present in the test sample wilt bind anti-analyte, thereby allowing anti-hapten to bind the conjugate, resulting in a change in the character of the conjugate hapten, e.g., a change in fluorescence when the hapten is a fluorophore.
., Sandwich assays particularly are useful for the determination of antigen or antibodies.
In sequential sandwich assays an immobilized binding partner is used to adsorb test sample analyte, the tdst sample is removed as by washing, the bound analyte is used to adsorb labeled binding partner, and bound nnaterial is then separated from residual tracer.
The amount of bound tracer is directly proportional to test sample analyte. In "simultaneous'° sandwich assays the test sample is not separated before adding the labeled binding partner. A sequential sandwich assay using an anti-antigen monoclonal antibody as one antibody and a polyclonal anti-antigen antibody ass he other is,useful in testing samples for particular antigen activity.
The foregoing aye merely exemplary diagnostic assays for the import and humanized antibodies of this inventibn. other methods now or hereafter developed for the determination of these analytes are included within the scope hereof, including the bioassays described -_._ ~ ~'.~ ~ 1 S.f, Y.mrr v.,-r.r WO 9Z/~2653 S$
above.
~mmunotoxins This invention is also directed to immunochamica) derivatives of the antibodies of this invention such as immunatoxins fconjug3tes at the antibody and a cytotoxic moiety).
A_rrtfbodies which carry the apprapriato effector functit5ns, such as with-their constant domains, are also used to induce lysis;thirough the natural complement process, and to interact with antibody dependent cytotoxic cells normally present-For example, purified, sterile filtered antibodies are optionally conjugated to a 1a cytotoxin such as ricin for use in AIDS therapy. 'The methods of this invention, for example, era 'suitable for obtaining humanized antibodies for use as immunotoxins for use in AIDS therapy, The cytotoxic moiety. of the immunatoxin may be a aYtotoxic drug or an enzymaticaNy active toxin of bacterial, fungal, plant or animal origin, or an enzymatically active fragment Qt such a toxin. i:nzymatically active toxins and fragments thereof used are diphtheria A chain, nonbinding active fragments of diprithoria~ .toxin. exotoxin A drain (from Psetrrlamo~as aemginosa), ~ricin A chain, abrin A chain, modaccin A chain, alpha-s9rcin, Aleurites fordii proteins,.dianthin fxoteins;.~hyiclacaamericana proteins iF'APi; PAPI1, and PAF-S), momordica ~.: zo charantia inhibitor, curcirr. . crotin, sapaonaria afficinalis inhibitor, gelonin, mitogellin, restrictocin, phor<omycin. enomycin and the tricothecenes. In another embodiment, the antibod'ses ire conjugated to small molecule anticancer drugs such as cis-platin or 5FU.
. Conjugates of the monoclonal antibody and'such cYtotoxic moieties are made using a variety of bifunotional protein coupling agents. Examples of such reagents are SPDP, IT , bifunctional derivatives af. imidoesters : such . as dimethyl ~ adipimidate HCI, active esters such as . ~ disuccinimidyl ,suberate, aldehydes.such as ~tutaraidehyde. his-azido compounds such as bis fp-azidobenzoyll hekanediamina, bis-diaz4riium-derivatives such as his- tp-diazonium4enzoyll w ~ethylenediamine, .diisocyanates such, as tolylene 2,6-diisocyanate and bis-active fluorine compounds such ~as l,~-difluoro 2;4-dinitrobenzer~e. The (YSing portion of s toxin may be joined to the Fob fragment of the antib6dies.
Imrrmnoioxins can be cnada gin, a variety of ways; as discussed heroin.
Gommoniy . .known crosslinking reager~ta can.be used~to yield stable conjugates.
Advantageously. rE'~onocional antibodies specifically binding the domain of the antigen which is exposed tiry the infected cell 'surface, are con9ugated to ricin A
chain. Most . VVO 92/22653 2 ~ ~ ~ 0 ~ ~ PCT/US92/OS1Z6 advantageously the ricin A chain is deglycosylated and produced through recombinant means.
An advantageous method of making the ricin immunotoxin is described in Vitetta et al., Science 238:1D98 11987).
When used to kill infected human cells in vitro for diagnostic purposes, the conjugates will typically be added to the cell culture medium at a concentration of at least about 10 nM.
The formulation and mode of administration for in vitro use are not critical.
Aqueous formulations that are compatible with the culture or perfusion medium will normally be used.
Cytotoxicity may be read by conventional techniques.
Cytotoxic radiopharmaceuticals for treating infected cells may be made by conjugating radioactive isotopes te.g. l, Y, Pr) to the antibodies. Advantageously alpha particle-emitting isotopes are used. The term 'cytotoxic moiety" as used herein is intended to include such isotopes.
in a preferred embodiment, ricin A chain is deglycosylated or produced without oligosaccharides, to decrease its clearance by irrelevant clearance mechanisms (e.g., the liver).
in another embodiment. whale ricin tA chain plus B chain) is conjugated to antibody if the gatactose binding property of B-chain can be blocked t"blocked ricin").
In a further embodiment toxin-conjugates are made with Fab or Ftab')2 fragments.
Because of their relatively small size these fragments can better penetrate tissue to reach infected cells. -In another embodiment, fusogenic liposomes are filled with a cytotoxic drug and the liposom~s are coated with antibodies specifically binding the particular antigen.
Antibody Denenden~ Cellular Cvtotoxicitv ".w/
Certain aspects of this invention involve antibodies which are ta) directed against a particular antigen acrd tb) belong to a subclass or i~otype that is capable of mediating the lysis of cells to which the antibody molecule 'binds. IVlore specifically, these antibodies should belong to a subclass or isotype than upon c~mplexing with cell surface proteins, activates serum complement and/or mediates antibody dependent cellular cytotoxicity tADCC) by activating effector cells such as natural kilter cells or macrophages.
90 Bioto~ical activity of antibodies is known to be determined, to a large extent, by the constant domains or Fd region of the antibody molecule tUananue and t3enacerraf, Textbook of Jmmuno%gy; 2nd i=dition, Williams & Wilkinsp. 218 11984)). This includes their ability to activate complement and to mediate antibody-dependent cellular cytotoxicity tADCC) as effected by leukocytes. Antibodies of different classes and subclasses differ in this respect, wo ~ziza~s3 ~ ~. ~ ~~ ~ ~~ ~ ~crius9a/asiab coo as do antibodies from the same subclass but different species; according to the present invention, antibodies of, those classes having the desired biological activity are prepared.
Preparation of these antibodies involves the selection of antibody constant domains are their incorporation in the humanized antibody by known technique. For example, mouse immunoglobulins of the IgG3 and IgG2a class are capable of activating serum complement upon binding to the target cells which express the cognate antigen, and therefore humanized antibodies which incorporate IgG~ and IgG2a effector functions are desirable for certain therapeutic applications.
In general, mouse antibodies of the IgG2a and IgG3 subclass and occasionally IgG 1 can to mediate ADCC, and antibodies of the IgG3, IgG2a, and IgM subclasses bind and activate serum complement. Complement activation generally requires the binding of at least two IgG
molecules in close proximity on the target cell. However, the binding of only one IgM molecule activates serum complement.
The ability of any particular antibody to mediate lysis of the target cell by complement activation and/or ADCC can be assayed. The cells of interest are grown and labeled in vitro;
the antibody is added to the cell culture in combination with either serum complement or immune cells which may be activated by the antigen antibody complexes.
Cytolysis of the target cells is detected by the release of label from the lysed cells. 6n fact, antibodies can be screened using the patient's own serum as a source of complement and/or immune cells. The antibody that is capable of activating complement or mediating ADCC in the in vitro test can then be used therapeutically in that particular patient.
This invention specifically encompasses consensus Fc antibody domains prepared and used according to the teachings of this invention.
-a Thera,~eutic and~ther Uses of the Antibodies When used in ~ivo for therapy, tha antibodies of the subject invention are administered to the patient in therapeutically effective amounts (i.e. amounts that have desired therapeutic effect?. They will normally be administered parenterally. The dose and dosage regimen will depend upon the degree of the infecti~n, the characteristics of the particular antibody or 3o immunotoxin used, e.g., its therapeutic index, the patient, and the patient's history.
Advantageously the antibody or immunotoxin is administered continuously over a period of 1 ~2 weeks, intravenously to treat cells in the vasculature and subcutaneously and intraperitoneally to treat regional lymph nodes. Optionally, the administration is made during the course of adjunct therapy such as combined cycles of radiation, chemotherapeutic treatment, or . W4 92/22653 ~ ~ ~ ~ ~ ~ v PC'd'lUS92>05126 (0 4 administration of tumor necrosis factor, interferon or other cytoprotective or immunomodulatory agent.
For parentera! administration the antibodies will be formulated in a unit dosage injectable form (solution, suspension, emulsion) in association with a pharmaceutically acceptable parenteral vehicle. Such vehicles are inherently nontoxic, and non-therapeutic. Examples of such vehicles are water, saline, Ringer's solution, dextrose solution, and 5~~
human serum albumin. Nonaqueous vehicles such as fixed oils and ethyl oleate can aBso be used. Liposomes shay be used as carriers. The vehicle may contain minor amounts of additives such as substances that enhance isotonicity and chemical stability, e.g., buffers and preservatives.
to The antibodies will typically be farmulated in such vehicles at concentrations of about 1 mgiml to 10 mgiml.
Use of lgAl1 antibodies may be preferred for certain applications, however !gG
molecules by being smaller may be more abte than IgM molecules to localize to certain types of infected cells.
There is evidence that complement activation in vivo~leads to a variety of biological effects, including the induction of an inflammatory response and the activation of macrophages tUananue and l3enecerraf, Textbook of lmmunalogy, 2nd Edition, Williams &
Wilkins, p. 218 (1984)). The increased vasodilation accompanying inflammation may increase the ability of various agents to localize in infected cells. Therefore, antigen-antibody combinations of the . 20 type specified by this invention can b~ used therapeutically in many ways. Additionally, purified antigens tHakomori, Ann. Rev, lmmunol. 2:103 (1984)) or anti-idiotypic antibodies (Nepom et aP., Proc. Nat/. Acad. Sci. 81:2864 (1985); Koprowski et al., Proc.
lUatl. Acad. Sci.
81:216 ( 1984)) relating to such antigens could be used to induce an active immune response in human patients. Such a response includes the formation of antibodies capable of activating human complement and mediating ADCC and by such mechanisms cause infected cell destruction.
Optionally, the antibodies of this invention are useful in passively immunizing patients, as exemplified by the administration of humanized anti-HlV antibodies.
The antibody compositions used in therapy are formulated and dosages established in 3o a fashion consistent with good medics! practice taking into account the disorder to be treated, the condition of the individual patient, the site of delivery of the composition, the method of administration and other factors known to practitioners. The antibody compositions are prepared for administration according to the description of preparation of polypeptides for administration, infra.

W~ 92/22653 ~ ~ ~ ~ ~ j ~ PCTlUS92/0512t~
(0 2 Deposit of Materials As described above, cultures of the muMAb4D5 have been deposited with the American Type Culture Collection, 12301 Parklawn Drive, Rockville, MD, USA
(ATCC).
This deposit was made under the provisions of the Budapest Treaty on the International Recognition of the Deposit of Microorganisms for the Purpose of Patent Procedure and the Regulations thereunder (Budapest.Tr~aty). This assures maintenance of viable cultures for 30 years from the date of the deposit. The organisms will be made available by ATCC under the terms of the Budapest Treaty, and subject to an agreement between Genentech, Inc. and ATCC, which assures permanent and unrestricted availability of the progeny of the cultures l0 to the public upon issuance of the pertinent U.S. patent or upon laying open to the public of any U.S. or foreign patent application, whichever comes first, and assures availability of the progeny to one determined by the U.S. Commissioner of Patents and Trademarks to be entitled thereto according to 38 USC ~ 122 and the Commissioner's rules pursuant thereto (including 37 CFR ~ 1.12 with particular reference to 886 OG 638).
In respect of those designations in which a European patent is sought, a sample of the deposited microorganism will be made available until the publication of the mention of the grant of the European patent or until the date on which the application has been refused or withdrawn or is deemed to be withdrawn, only by the issue of such a sample to an expert nominated by the person requesting the sample. (Rule 28(4) EPCD
The assignee of the present application has agreed that if the cultures on deposit should die or be lost or destroyed when cultivated under suitable conditions, they will be promptly replaced on notification with a viable specimen of the same culture.
Availability of the deposited strain is not to be construed as a license to practice the invention in contravention ,~ s of the rights granted under the authority of any government in accordance with its patent Isws.
The foregoing written spepification is considered to be sufficient to enable one skilled in the art to practice the invention. The present invention is not to be limited in scope by the constructs deposited, since the deposited embodiments are intended to illustrate only certain aspects of the invention and any constructs that are functionally equivalent are within the ,, scope of this invention. The deposit of material herein does not constitute an admission that the written description herein contained is inadequate to enable the practice of any aspect of the invention, including the best mode (hereof, nor is it to be construed as limiting the scope of the Maims to the specific illustrations that they represent. Indeed, various modifications of the invention in addition to those shown and described herein will become apparent to those ~.,:..;, ,. ..; ~ ..-.. ...,;..:.;: .,.:.: . ;;. :... .....:".,. .. ._ .. .
~~.~~f~ )~
..,W~ 92/22653 4 ~ PC~'/USl2/05126 skilled in the art from the foregoing description and fall within the scope of the appended claims.
It is understood that the application of the teachings of the present invention to a specific problem or situation will be within the capabilities of one having ordinary skill in the art in light of the teachings contained herein. Examples of the products of the present invention and representative processes for their isolation, use, and manufacture appear below, but should not be construed to limit the invention.
EXAMPLES
EXAMPLE 1. HUMANIZATION OF muMAb4D5 Here we report the chimerization of muMAb4D5 tchMAb4D5> and the rapid and simultaneous humanization of heavy tVM) and light tV~) chain variable region genes using a novel "gene conversion mutagenesis" strategy. Eight humanized variants thuMAb4D5) were constructed to probe the importance of several FR residues identified by our molecular modeling or previously proposed to be critical to the conformation of particular CDRs tees Chothia, C. & Lesk, A. M., J. Mol. Biol. '196:901-917 (19871; Chothia, C. et al., Nature 342:877-883 11989); Tramontano, A. et al., J Mol. 6~ial. 215:175-182 11990)).
Efficient 2p transient expression of humanized variants in non-myeloma cells allowed us to rapidly inv~stigate the relationship between binding affinity for p185HEgz ECD.and anti-proliferative activity against p185HER2 ~verexpressing carcinoma cells.
flliATERIALS and tUIETHOD9 ,;
Cloning of ilariobte l~egi~ro Genes: The muMAb4D5 VH and VL genes were isolated by potymerase chain reaction (PCR) arnptific~tion of mRNA from the corresponding hybridoma iFer~dly, ~. M. et al., Cancor Res. 5~:1550-1558 11990)) as described by Orlandi et al.
tOrtandi; R. et ~P.; Proc: '!Nato Acid. Sci. LISA 86:3833-3837 11989)). Amino terminal sequencing of muMA64D5 V~ and V~ was used to design the sense strand PCR
primers, whereas the anti-sense PCR primers were based upon consensus sequences of murine framework residues dOrtandi, R. et al., Proc. lUarl. Acad. Sci. USA 86:3833-3837 11989);
Kabat, E. A.' et al., Sequences'of ProPeins of Ira~murrologica! Interest tPlational Institutes of Health, 8ethesda; MD; 1987)) incorporating restriction sites for directional cloning shown by underlining ahd listed after the sequences: V L sense, 5'-TCCGATATCCAGCTGACCCAG~'CTCCA-3' tSEQ. 1D NO. 7), EcoRV; VL anti-sense, 5'-Wo 9zn~s53 ~ ~, ~ ~ ~ ~ ~ P~G'1'/US92/0512b GT'iTGATCTCCAGCTT~~t~iSCDCCGAA-S' (SEO. IQ NO. 8). Asp77 B; vH sense, 5'-AGGTSMARSe.TSAGTGWGGw3' ISEa~ iQ NO- 9f~ Pstl and YH anti-sense, 5'-TGAGGAGAC~~.GTGG~'CCCTTGGCCCCAG-3' (SECZ. ID NO. 10), BstEll; where H =
A or C or T. 8 - C or G. D = A or G or T, M ~ A ar C. R = A or G and W = A or T. The PCR products were cloned iota pli'C11,9 tVielra..f~ & Messing. J., Methods En~ymol. '153:3-1.1 , (19871) and five cibnes for each variakFla domain sequenced bY the dideoxy method (5anger.
F. ei .al., Prac. Natl. Aced. Saj. LISA 74:6463-5487 ! ~ 977)).
Molecular Modelling. Modats for muMAb4D5 VH and VL domains were c~structed separately from oonsensws coordinates based upon seven Fob structures from the Brookhavan 1o protein dad bank (entries 1 F~~ 2RHt~, 2MCP. 31=AB, 1 FBJ, 2HFL and 1 REI).
The pab fragment KCAL lMarquart, M. et al., J. Mal. Bial. '141:3$9-391 (1980)) was first Chosen as a template for V~ and VH domains and additional structures were then superimposed upon this structure using their main chain atom coordinates 11NSIGH~'iprogram. Biosym Techwlogies).
The distance from the template Catv the analvgaus Ca in each of thg superirnposad structures 15. was calculated for each residue position. If alt for nearly all? Ccr-Ca distances far a given residue were 5 1 R, then that position .was included in the aansensus structure. In most cases the ~-sheet framework.rssidues satisfied tflese criteria whereas the CDR loops did not. For each of these selected 'rasidu'as the average coordinates fvr individual N, Ca. C, O and CB
atoms were calGUl~ted and then corrected far resultant deviations from non-standard bond ' geormetry >;y 50 cycles of energy minimization using the DISCOVER"program (Biosym Technologies) .with the yAM,B>'Rk:fotcefield (Weiner, s. J. et al., J. Amer.
Chem. 5oc.
106:766-?84 (1984)) and Cccoordir~ates fixed. The side chains of highly conserved residues, ' , such as the disul8de-bridseii b~lstdine reaidues~ ware then incorporated into the .resultant consensus structure. . Next the saqu~ncBS of muMAb4D6 V~ and VH ware incorporated 5sxartin~ with the CDR fes'idues and using the tabulations of CDR
conformations from Chotnia ex a!: (Chothia:.C~.'et al,, IVatLre' 342:8?7-883 (1989)1 as a guide. Side-chain conformations - were chosen on.the~ basis.of I=ab crystal structures, rotemer libraries (Ponder, ,!. W. ~ Richards.
F. nn,; .1.; Mah viol 193:?75-.791 (19$7)i and pac.kir~g considerations. Since VR-CDR3 could not be assigned a definite #ackbone; conformation from these criteria, two modals were created '~o ~ frpm a search of similar,sized leaps using the INSIGHT program. A third rnadal was derived using packing and solvent expusurd considerations. Each modal was then subjected to 5000 cybles of energy minimization: , In humanizing mulVIAb4D~.''aanssnsUS human sequences were first derived from the mdst~abundant sutxlasses in. he sequence caniliilation of Kabat et al. (Kabat.
E. A. et al_, *-trademarks W4 92122653 ~ ~ ~ J ~ J ~ PC1'/LJ592/0512G
!~ S
SeQuences of Proteins of Immunological Interest (National Institutes of Health, Bethesda, MD, 1987)), namely V~ x subgroup I and VH group III, and a molecular model generated for these sequences using the methods described above. A structure for huMAb4D5 was created by transferring the CDRs from the muMAb4D5 model into the consensus human structure. All huMAb4D5 variants contain human replacements of muMAb4D5 residues at three positions within CDRs as defined by sequence variability (Kabat, E. A. et al., Sequences of Proteins of lmmuno%gicallnterest (National Institutes of Health, Bethesda, MD, 1987)) but not as defined by structural variability (Chothia, C. & Lesk, A. M., J. Mol. Biol. 196:901-917 (1987)):
VL-CDR1 K24R, VL-CDR2 R54L and VL-CDR2 TSfS. Differences between muMAb4D5 and l0 the human consensus framework residues (Fig. 1 ) were individually modeled to investigate their possible influence on CDR conformation and/or binding to the p185HER2 ECD.
Construction of Chimeric Genes. Genes encoding chMAb4D5 light and heavy chains were separately assembled in previously described phagemid vectors containing the human cytomegalovirus enhancer and promoter, a 5' intron and SV40 polyadenylation signal (Gorman, 1~ C. M, et al., DNA & Prot. Engin. Tech. 2:3-10 (1990)). Briefly, gene segments encoding muMAb4D5 VL (Fig. 1 A) and REI human K1 light chain CL (Palm, W. & Hilschmann, N., Z.
Physiol. Chem. 356:167-191 (1975)~ were precisely joined as were genes for muMAb4D5 VH
(Fig. i B) and human y1 constant region (Capon, D. J. et al., Nature 337:525-531 (1989)) by simple subcloning (Boyle, A.; in Current Protocols in Molecular Biology, Chapter 3 (F. A.
20 Ausubel et al., eds., Greene Publishing & Wiley-Interscience, New York, 1990)) and site-directed mutagenesis (Carter; P., in Mutagenesis: A Practical Approach, Chapter 1 (IRL
Press, Oxford, UK 1991 )). The y1 isotype was chosen as it has been found to be the preferred human isotype for supporting ADCC and complement dependent cytotoxicity using matched sets of chimeric EBruggemanrv, M. et al.; J. Exp. Med. 166:1351-1361 (1987)) or humanized 25 antibodies (Riechmann; L. eral.; Nature 332:32-327 (1988)). The PCR-generated VL and VH
fragments (Fig. 1 ) were subsequently mutagenized so that they faithfully represent the soquence of muMAb4D5 determined at the protein level: VH Q1 E, VL V104L and tvariants are denoted by the -amino acid residue and number followed by the replacement amino acid). The human yl constant regions are identical to those reported by Ellison et al.
30 (Ellison, J. W. et al., Nucleic Acids Res. 13:4071-4079 (1982)) except for the mutations E359D and M361 L (Eu numbering, as in Kabat, E. A. et al., Sequences of Proteins of Immunologicallnterest (National Institutes of Health, Bethesda, MD, 1987)) which we installed to convert the antibody from the naturally rare A allotype vo the much more common non-A
allotypd (Tramont~no, A. et al., J. Mol: Biol. 215:175-182 (1990)). This was an attempt to Vd~ 92!22653 ~ ~ U ~ ~ J ~ PCIf/US92/OS126 reduce the risk of anti-allotype antibodies interfering with therapy.
Construction of Humanized Genes. Genes encoding chMAb4D5 light chain and heavy chain Fd fragment (VH and CH1 domains) were subcloned together into pUC119 (Vieira, J. &
Messing, J., Methods Enzymoi. ''53:3-11 (1987)) to create pAK1 and simultaneously S humanized in a single step rtFig. 2). Briefly, sets of 6 contiguous oligonucleotides were designed to humanize VM and V~ (Fig. 1 ). These oligonucleatides are 28 to 83 nucleotides in length, contain zero to 19 mismatches to the marine antibody template and are constrained to have 8 or 9 perfectly matched residues at each end to promote efficient annealing and iigation of adjacent oligonucleotides. The sets of VH and VL hurnanization oligonucleotides (5 pmol each) wars phosphorylated with either ATP or y-32P-ATP (Carter, P.
Methods Enzymol.
154:382-403 (1987)) and separately annealed with 3.7 pmol of pAK1 template in 40,u1 10 mM Tris-HCl (pH 8.0) and 10 mM MgCl2 by cooling from 100 ~C to room temperature over 30 min. The annealed oligonucleatides were joined by incubation with T4 DNA
ligase ( 12 units; New England Biolabs) in the prqsence of 2 NI 5 mM ATP and 2 /el 0.1 M
DTT for 10 min at 14 ~C. After electrophoresis on a 6°!° acrylamide sequencing get the assembled oligonucleotides were looted by autoradiography and recovered by electroelution. The assembled oligonucleotides ( -- 0.3 pmol each) were simultaneously annealed to 0.15 pmol single-stranded deoxyuridine-containing pAK1 prepared according to Kunkel et al. (Kunkel, T.
A. et aL, Methods Enzymol. 154:367-382 (1987)) in 10 pl 40 mM Tris-HCI (pH
7.5) and ~16 24 mM MgCl2 as above. He'teroduplex DNA was constructed by extending the primers with T7 DNA polymerase and transformed into E, coil 8MH 71-18 mutt as previously described (Carter; P., in Mutagenesis: A Practice! Approach, Chapter 1 (IRL Press, Oxford, UK 1991 )).
The resultant phag~iwid DNA pool was enriched first for huVL by restriction purification using 4_~
Xhol and then for huV~ key restriction selection using Stul as described in Carter, P., in Mutagene5is: A Practical Approach, Chapter 1 tIRL Press, Oxford, UK 1991 );
and in Wells, J. A. et al , Phil: Traps: R. Soc. Lond A 317:415-423 41986). Resultant clones containing both huVL and huVH genes were identified by nucleotide sequencing tSanger, F.
et al., Proc.
IVatl, aAcad. Sci. US~I 74:5463-5467 (19771) and designated pAK2. Additional humanize variants were generated by site-directed mutagenesis (Carter, P., in Mutagenesis: A Practice!
Approach, Chapter 1 (IRL Press, Oxford, UK 1991 )). The muMAb4D5 VL and VH
gene sbgme~nts in the transient expression vectors described above were then precisely replaced with their humanized versions.
Exlaression and Purification of IUlAb4D5 Variants. Appropriate MAb4D5 light and heavy chain cDNA expression vectors were co-transfected into an adenovirus transformed human .,.WO 92/22b53 ~ ~ ~ ~ ~ ~ ~ PGTIU~a9214S~~~
embryonic kidney cell line, 293 (Graham, F. L, et al.. J. Gen. Viro~ 36:68-72 f't 977)) using a high efficiency procedure IGarman, G. M. er al., DNA & Prot. Engin. Tech. 2:3-117 11990);
Gorman, C., in DNA Clanln9. val II, pp 143-'1 g0 (~. M. Glover, ed., IRL
Press, Oxford. UK
19$5?). Media were harvested daily for up to 5 days and the cells re-fed with serum free g madla. An'~bddigs were recovered from the media and affinity purified on protein a sepharose*
CL-4B IPharmacia) as described by the manufacturer. The eluted antibody was buffer-exchanged into phosphate-buffered saline by G25 gel filtration, concentrated by ultrafiltration (Cerrtriprep~i0 or ~er'atriao~i 00. Amiconi, sterile-filtered (Miliex~"sV, Millipore) and stored at 4 ~C. The concentration of antibody was determined by using bath total immunoglobuiin and antigen binding .ELISAs. The staruiard used was huMAb4D5-5, whose concentration had been rJetgrmined by amino acid composition analysis.
Cal! ProtifW"at~n Assay. The off8ct of MAb4D5 variants upon proliferation of the human mammary adenocarcinoma coil line, SK-BR-3, was investigated ss previousty.described ~IFendly, B. M. et al.: Cancer Res. 50:1554-'t 55$ I1 ~g0)f using saturating MAb4~5 rs concentrat;Qns.
AftINty Messuraments. The antigen binding affinity of MAb41~5 variants was determined using a sebrotad.form of the p185HER~ ECD prepared as described in Fendiy. ~.
. M..et al., J. 8iol. Rasp; INod.' 9:449~155.t't9941. Briefly, antibody and plg5HeR2 ECD were . incubated ~in so(ut(on until eqbilibriurh was found to be reached. The concentration of free antibody was th~dn determined'. by ,ELfSA .using immobilized p1$5HE~2 ECD and used to calouiata affinity tKd) according to Friguet et al. (Friguet. 8, et al. .t.
Immunal. Methods 77:305-319 I19135fi: ~ , .
.. , ,; ~. . ;
. . . ~ itF,SULTS
.a:5 , . 1 .. t~hrrhanixation tif'iinuMAb4iD5: Tha muMAb4a5 V~ arid vH gene segments ware first aloned~ by_' PC'R ~ and Sequoi~cBdi tFig. , .11. Ttie variable genes were xhan simultaneously . ,h~e~ized~by.:gene~avnvarsian_mutagenesis,usirip preassemblad ofigonucleotides tFig. xl. A
311 ~mer,: oiigoycleotide;~ontaining ~9 rniarria~chas to the template directed 24 simultaneous amino acid changes, required to humanize muMAb4D5 VL. Humanization of muMAb4D5 VH
requirad.32 amino acid .~char~ga~..~fuph were installed with a 361~mer containing 59 ~.rnismatches to the muN~Ab4D5 tbmplate: Two. out of 8 clones sequenced precisely encode huMAb4D~-5, although:~one of these clones cvittained a aingte nucleotide imporfectian. The 6.other clones were ess~ntialfy humamied but contained a small number of grrQrs_ ~ 3 nucleotide changes and ~ f~ single. nucleotides uelation per kilobe$e.
Additibnai humanized '*-trademarks ' WO 92!22653 ~ ~ ~ ~ ~ ''~ '~ PCTltJS92lg5126 .
!r 8 variants (Table 31 were constructed by site-directed mutagenesis of huMAb4D5-5.
Expression levels of huMAb4D5 variants were in the range of 7 to 15 pglml as judged by ELISA using immobilized p185HER2 ECD. Successive harvests of five 10 cm plates allowed 200 Ng to 500 mg of each variant to be produced in a week. Antibodies affinity purified on protein A gave a single band on a rCoomassie blue stained SDS polyacrylamide gel of mobility consistent with the expected Mr of -150 kDa. Electrophoresis under reducing conditions gave 2 bands consistent with the expected Mr of free heavy (48 kDa) and light (23 kDa) chains Inot shown). Amino terminal sequence analysis ( 10-cycles) gave the mixed sequence expected (see Fig. 1 ) from an equimolar combination of light and heavy chains (not shown).
huMAb4D5 Variants. In general, the FR residues were chosen from consensus human sequences (Kabat, E. A. ex al., Sequences of Proteins of lmmuno!pgica!
Interest (National Institutes of Health, Bethesda, MD, 1987)) and CDR residues Pram muMAb4D5.
Additional variants were constructed by replacing selected human residues in huMAb4D5-1 with their muMAb4D5 counterparts. These are VH residues ? 1, 73, 78, 93 plus 102 and VE
residues 55 plus 66 identified by our molecular modeling. VH residue 71 has previously been proposed by others (Tramontano, A. et al., J. Mol. Biol. 215:175-182 (1990)) to be critical to the conformation of VH-CnR2: Amino acid sequence differences between huMAb4D5 variant molecules aye shown in Table 3, together with their p185HER2 ECD binding affinity and 2tt maximal anti-proliferative activities against SK-BR-3 cells. Very similar Kd values were obtained for binding of MAb4D5 variants to either SK-BR-3 cells or to p185HER2 ECD (Table 3). However, Kd estimates derived from binding of MAb4D5 variants to p185HER2 ECD were more reproducible with smelter standard errors and consumed much smaller quantities of antibody than binding measurements with whole cells.
The most potent humanized variant designed by molecular modeling, huMAb4D5-8, contains S FR residues from muMAb4D5. This antibody binds the p185HER2 ECD 3-fold mare tightly than does muMAb4D5 itself (Table 3) and has comparable anti-proliferative activity with SK-BR-3 cells (Fig. 3).' In contrast; huMAb4D5-1 is the most humanized but least potent muMAb4D5 va'iant, created by simply installing the muMAb4D5 CDRs into the consensus human sequences. huMAb4D5-l binds the p185RER2 ECD 80-fold less tightly than does the murine antibody and has no detectable anti-proiiferative activity at the highest antibody concentration investigated (16 Ng/m)).
The anti-pPOliferative activity of huMAb4D5 variants against p185HER2 overexpressing SK-i3R-3 cells is not simply co~retated with their binding affinity for the p185HER2 ECD. For -~WO 92/22653 PCf/US92/~D5126 example, installation of three marine residues into the VH domain of huMAb4D5-2 (D73T, L78A and A93S) to create huMAb4D5-3 does not change the antigen binding affinity but does confer significant anti-proliferative activity (Table 3).
The importance of VH residue 71 (Tramontano, A. et el., J. M~!. Biol. 215:175-S (1990)) is supported b~> the observed 5-fold increase in affinity far p185HER2 ECD on .
replacement of R71 in huMAb4D5-1 with the corresponding marine residue, alanine (huMAb4D5-2). In contrast, replacing VH L78 in huMAb4D5-4 with the marine residue, alanine (huMAb4D5-5), does not significantly change the affinity for the p185HER2 ECD or change anti-proliferative activity, suggesting that residue 78 is not of critical functional significance to huMAb4D5 and its ability to interact properly with the extracellular domain of p185HERZ.
VL residue 66 is usually a glycine in human and marine K chain sequences (Kabat, E.
A. et al., Sequences of Proteins of lmmunological Inferest (National Institutes of Health, Bethesda, MD, 1987)) but an arginine occupies this position in the muMAb4D5 k light chain.
The side chain of residue 66 is likely to affect the conformation of VL-CDR1 and V~-CDR2 and the hairpin turn at 68-69 (Fig. 4). Consistent with the importance of this residue, the mutation VL G66R (huMAb4D5-3 --~ huMAb4D5-5) increases the affinity for the p185HER2 ECD by 4-fold with a concomitant increase in anti-proliferative activity.
From maPecular modeling it appears that the tyrosyl side chain of muMAb4D5 VL
residue 55 may either stabilize he conformation of VH-CDR3 or provide an interaction at the VL-V~ interface. The latter function may be dependent upon the presence of VH
Y102. In the context of huMAb4D5-5 the mutations VL E55Y (huMAb4D5-6) and VH V102Y
(huMAb4D5-7) individualty increase the affinity for p185HER2 ECD by 5-fold and 2-fold respectively,, whereas together thuMAb4D5-81 they increase the affinity by 11-fold. This is consistent with either proposed role of VL Y55-and VH Y102.
Secondary Immune Funcdorr of huMAb4D5-8: MuMAb4D5 inhibits the growth of human beast tumor cells which overexpress pl B~HER2 tHudziak, R. M. et el., Malec. Cell.
Biol. 9:1165-1172 (19891). The antibody, however, does not offer the possibility of direct termor cytotoxic effects: This possibility does arise in huMAb4D5-8 as a result of its high affinity (Kd _ 0.1 NM) and its human IgGl subtype. Table 4 compares the ADCC, mediated by huMAb4D5-8 with muMAb4D5 on a narrnal lung epithelial cell line, WI-38, which expresses a low level of p185HER2'and on SK-~R-3, which expresses a high level of p185HER2, The results demonstrate that: (1 D huMAb4D5 has a greatly enhanced ability to carry out ADCC as compared with its marine patent; and (2) that this activity may be selective for cell types WVtD 92!22653 ~ ~ ~ ~ '~ ~ PCg'lUS92l05126 ~0 which overexpress p185H~R2 DlSCUSStON
MuMAb4D5 is potentially useful for human therapy since it is cytostatic towards S human breast and ovarian tumor ~tir°tes overexpressing the /-?ERA-encoded p185RER2 receptor-like tyrosine kinase. Since both breast and ovarian carcinomas are chronic diseases it is anticipated that the optimal MAb4D5 variant molecule for therapy will have low immunogenicity and will be cytotoxic rather than solely cytostatic in effect.
Humanization of muMAb4D5 should accomplish these goals. We have identified 5 different huMAb4D5 variants which bind tightly to p185~ER2 ECD (Kd 5 1 nM) and which have significant anti-proliferative activity (Table 3). Furthermore huMAb4D5-8 but not muMAb4D5 mediates ADCC against human tumor cell lines overexpressing p185HER2 in the presence of human effector cells (Table 4) as anticipated for a human y1 isotype (Bruggemann, M.
et al., ,!. Exp.
Med. 166:1351-1361 (1987); Riechmann, ~.. et al., Nature 332:323-327 (1988)):
Rapid humanization of huMAb4D5 was facilitated by the gene conversion mutagenesis strategy developed here using long preassembled oligonucleotides. This method requires less than half the amount of synthetic DNA as does total gene synthesis and does not require convenient restriction sites in the target DNA. Our method appears to be simpler and more reliable than a variant protocol recently reported (Rostapshov, V. M. e? al., FEBS Gett.
249:379-382 (1989)). Transientexpression of huMAb4D5 in human embryonic kidney calls permitted the isolation of a few hundred micrograms of huMAb4D5 variants for rapid characterization by growth inhibition and antigen binding affinity assays.
Furthermore, different Combinations of light and heavy chain were readily tested by co-transfec~tion of corresponding cDNA expression erectors.
Z5 The crucial role of molecular modeling in the humanization of muMAb4D5 is illustrated by the designad variant huMAb4D5-8 which binds the p185HER2 ECt3 250-fold more tightly than the simple CDR I~op swap variant; huMAb4D5-1. It has previously been shown that the antigen binding affinity of a humanized antibody can be increased by mutagenesis based upon molecular modelling (Riedhmann, L: et~l., Na?ure 332:323-327 (1988); C~ueen, C. etal., Proc.
Natl. ~lcaa! Sci. USA 86:10029-10033 ( 1989)). Here we have extended this earlier work by others with a designed humanized antibody which binds its antigen 3-fold more tightly than the parent rodent antibody. While this result is gratifying, assessment of the success of the molecular modeling must await the outcome of X-ray structure determination.
From analysis of huMAb4D5 variants (Table 3) it is apparent that their anti-proliferative activity is not a .~JV~ 92/2265 ~ ~ ~ ~ ~ 7 ~ PCl"/US92JO5126 simple function of their binding affinity for p185HER2 ECD. I=or example the huMAb4D5-8 variant binds p185HER2 8-fold more tightly than muMAb4D5 but the humanized variant is slightly less potent in blocking the proliferation of SIB-BR-3 cells.
Additional huMAb4D5 variants are currently being constructed in an attempt to identity residues triggering the S anti-proliferative activity and in an attempt to enhance this activity.
In addition to retaining tight receptor binding and the ability to inhibit cell growth, the huMAb4D5-8 also confers a secondary immune function (ADCC). This allows for direct cytotoxic activity of the humanized molecule in the presence of human effector cells. The apparent selectivity of the cytotoxic activity for cell types which overexpress p185HER2 allows to for the evolution of a straightforward clinic approach to those human cancers characterized by overexpression of the HER2 protooncogene.
.... _ .. ~.: .. .,-. . ..

ewr~ ~zizz~s3 ~c~ius9zrossz~
~z Table 3. p185HEtt2 FCD binding affinity and anti-protiferative activities of MAb4D5 variants VH Residue° VL Residue°
MAb4D5 71 73 78 93 102 55 66 Rdt Relative cell Variant FR3 FR3 F"R3 FR3 CDR3 CDR2 FR3 nM
proliferationt i0 huMAb4D5-1 R D L A V E G 25 102 huMAb4D5-2 Ala D L A V E G 4.7 101 huMA,b4D5-3 Ala Thr Ala Ser V E G 4.4 66 huMAb4D5-4 Ala Thr L Ser V E Arg 0.82 56 huMAb4D5-S Ala Thr Ala Ser V E Arg 1.1 48 ~5 huMAb4D5-6 Ala Thr Ala Ser V Tyr Arg 0.22 51 huMAb4D5-7 Ala Thr Ala Ser Tyr E Arg 0.62 53 huMAb4D5-8 Ala Thr Ala Ser Tyr Tyr Arg 0.10 54 muMAb4D5 Ala Thr A1a Ser Tyr Tyr Arg 0.30 37 hluman and murine ~e~idues are shown in one letter and three letter amino acid code rd~pectiv~ly.
f Kd values for th~ ~185H~R2 EC~ vvere determined using the method of Friguet et aJ. t~3) and the standard error of each estimate is ~ ~ 1 ~°~.
t Proliferation of SK-8R-3 cells incubated for 98 hr with PVlAb4D5 ~rariants shown as a 25 percentage of the untreated control as described (Hudziak, R. M. et al., IVlulec. dell, Biol.
J:1 i 85-1172 X1989)). Data represent the maximal anti-proliferative effect for each ~rariant (see Fig: ~A) calculated as the mean of triplicate determinations at a A~Ab4D5 concentration of 8 yg/m!. Data are a!I taken from the same experiment with an estimated standard error of CA 02103059 2003-05-20 _..

coupling reaction by S100-HR (Pharmacial size exclusion chromatagraphy (2_5 cm x 100 cm) in the presence of PBS. The BsF/ab')2 samples were passed through a 0.2 mm filter flash frozen in liquid nitrogen and stored at -70' C.
Flow cytometric analysis of Flab' lsbindinp to Jurkat cells The Jurkat human acute T cell leukemia cell line was purchased from the American Type Cutture Collection (Rockville, MO) (ATCC TIB 152) and grown as recommended by the ATCC. Aliquots of 108 Juricat cells were incubated with appropriate concentrations of BsF(ab')z (anti-p185"~ / anti-CD3 variant) or control mono-specific anti-p185"E''~ F(ab')z in PBS plus 0.196 (w/v) bovine serum albumin and 10 mM sodium azide for 45 min at 4 ' C.
The cells were washed and then incubated with fluorescein-conjugated goat anti-human F(ab')z (Organon Teknika, West Chester, PA) for 45 min at 4 -C. Cells were washed and ' analyzed on a FACScari (Becton Dickinson and Co., Mountain View, CA). Cells (8 x 10') were acquired by list mode and gated by forward Light scatter versus side light scatter excluding dead cells and debris.
RESULTS
Design of humanized anti-CD3 variants The most potent humanized anti-CD3 variant previously identified, v1, differs from the marine parent antibody, UCHT1 at 19 out of 107 amino acid residues within V~
and at 37 out of 122 positions within V" (Shalaby et al.,supra) 1992). Here we recruited back additional marine residues into anti-CD3 v1 in an attempt to improve the binding affinity for CD3. The strategy chosen was a compromise between minimizing both the number of additional marine residues recruited and the number of anti-CD3 variants to be analyzed. We focused our attentions on a few CDR residues which were originally kept as human sequences in our minimalistic humanization regime. Thus human residues in V" CDR2 of anti-C03 v1 were replaced en bloc with their marine counterparts to give anti-CD3 v9:
'f57S:AGON:DGIQ:S62K:VG3F:GG_SD (SEQ ID NO: ~?0) (Fig. 5~ Similarly, the human residue E55 in V~ CDR2 of anti-CD3 vl was replaced with histidine from the marine anti-C:D3 antibody to generate anti-CD3 v 11. In addition, Vtc framework region (FR) residues 75 and 7G in anti-CD3 v1 were also replaced with their marine counterparts to create anti-CD3 v8: K75S:N7GS. Vtt residues75 and 7G are located in a loop close to Vti CDRI and CDR2 and therefore might influence antigen binding.
Additional variants created by combining mutations at these three sites are described helovv.
Preparation of BsFlab71 fra8ments Soluble and functional anti-p185"E'u and anti-CD3 Fab' fragments were recovered directly from corresponding E. colt fermentation pastes with the single hinge cysteine predominantly in the free thiol form (75-100 96 Fab'-SH) by affinity purification on Streptococcal protein G at pH 5 in the presence of EDTA (Carter et aL, 1992b, supra).
Thioether-Linked BsF(ab')~ fragments were then constructed by directed coupling using o-PDM
Si~i35~i 1 i ~J i:~ Sri~T

V'a'~ 92/226x3 ~ ~ Q ~ ~ 7 ~ P~'/1.JS92/05126 Table 4. Selectivity of antibody dependent tumor cell cytotoxicity mediated by huiVlAb4D5-8 WI-3$° SK-BR-3 E~fector:Target ratf,o'~ m~b4D5 huMAb4D5-8 m~a~3A'b4D5 huMAb4D5-8 A.t 25:1 <1.0 9.3 7.5 40.6 12.5:1 <1.0 11.1 4.7 36.8 6.25:1 <1.0 8.9 0.9 35.2 g0 3.13:1 <1.0 8.5 4.6 19.6 B. 25:1 <1.0 3.1 6.1 33.4 12.5:1 <1.0 1.7 5.5 26.2 6.25:1 1.3 2.2 2.0 21.0 3.13:1 <1.0 0.8 2.4 13.4 Sensitivity to ADCC of two human cell lines (WI-38, normal lung epithelium;
and SK-8R-3, human breast tumor cell line) are compared. WI-38 expresses a low level of p"185~ERZ 10.6 pg p~r pg cell protein) and SK-8Fi-3 expresses a high Devel of p185~ER2 !64 pg p185HER2 per pg cell protdin), as determined by ELISA lFendly et at., J. Biol. Reap. Mod.
9:449-455 (1.90)).
t ADCC assays were' carried out as described in l3rieggemann et al., J. Exp.
Med.
186:1851-1361 l1987D. Effector to target ratios were of 1!.-2 activated human peripheral blood lymphocytes to either WI-38 fibroblasts or SK-8R-8 tumor cells in 96-well microtiter plates for 4 hours at 3T QC. Values given represent percent specific cetl lysis as determined by '1Cr release. i"stimated standard error in these quadruplicate determinations was s t 1096.
t Monoclonal antibody concentrations used were 0.1 yg/ml !AD and 0.1 pg/ml lBi.

WU 92l22b53 '~ ~ ~ ~ ~ ~ ~ PGTlUS92/OSt2b ~5 EXAMPLE 2. Schematirt Method for Humanizinct an Antibody Seguence This example illustrates one stepwise elaboration of the methods for creating a humanised sequence described above. It will be understood that not all of these steps are essential to the claimed invention, and that steps may be taken in different order.
9 . ascertain a consensus human variable domain amino acid sequence and prepare from it a consensus structural model.
2. prepare model of import (the non-human domain to be humanized) variable domain sequences and note structural differences with respect ~o consensus human model.
3. identify CDR sequences in human and in import, both by using Kabat . (supra, 1 ~87> and crystal structure criteria. If there is any difference in CDR identity from the different criteria, use of crystal structure definition of the CDR, but retain the Kabat residues as important framework residues to impart.
4. substitute import CDR sequences for human CDR sequences to obtain initial "humanised" sequence.
5. compare import c~bn~CDR variabl~ domain sequence to the humanized sequence qnd note divergenGes.
6. Proceed through the following analysis for each amino acid residue where the import diverges from the humanized.
28 aIf the humanised residue represents a residue which is generally highly conserved across all species, use the residue in the humanized sequence. If the residue is not conserved across atl species; proceed with the analysis described in 6b.
b. If the residue is not generally conserved across all species, ask if the residue is generally conserved in humans.
i. If the residue is generally conserved in humans but the import ~esid~ae differs, examine the structural models of the i,~port and human sequences and determine if the import eesidue v~ould be likely to affect the binding or biological =,. ... ,,.., , ,:,: ;: ~ . -::- ;: , . , .., .,_ : , .... , . ... ,. ;. .. :......:
. , . >,. ,., .:: . .. , ..... ... ... , .

'6rVCD 92f22653 ~'CT/US92i05126 activity of the CDRs by considering 1 ) could it bind antigen directBy and 2) could it affect the conformation of the CDR.
If the conclusion is that an affect on the CDRs is likely, substitute the import residue. If the conclusion is that a CDR affect is unlikely, leave the humanized residue unchanged.
ii. If the residue is also not generally conserved in humans, examine the structural models of the import and human sequences and determine if the import residue would be likely to affect the binding or biological activity of the CDRs be considering 11 could it bind antigen directly and 2) could it affect the conformation of the CDR. If the conclusion is that an affect on the CDRs is likely, substitute the import residue. If the conclusion is that a CDR affect is unlikely, proceed to the next step.
a) examine the structural models of the import and human sequences and determine if the residue is exposed on the surface of the domain or is buried within. If th~ residue is exposed, use the residue in the humanized sequence. If the residua is buried, proceed to the next step.
(i) Examine the structural models of the impart and human sequences and determine if the residue is likely to affect the ~J~ - V~, interface. Residues involved with the ~interfac~ include; 34L, 36L, 38L, 43L: 33L: 36L, 85l., 87L, 89l., 91 L, 86L, 88L, X51°!, 3~'t~. 39FI, 43H, 45H, 4711, 60i~, 91 H; 93H~ 95H, 100H, and 103N. If no effect is likely, use the residue in the humanized sequence. If some affect is likely, substitute the import residue.

7. search the impot't sequence; the consensus sequence and the humanized sequence for gtycosylation sites outside the CDRs, and . determine if this gly~osylation site is likely to have any affect on . .".;' a%.. ..t,.~~~t ..'.y., .,...': . .:::. ....a....:. . .,:.:.
..,:.ar~. .''.'.: . ; ~:'. , ''...~.., , :',',:-.;.;: ~.~.:...~..;.,. ,,....
,...:.,.,.. , -:"'.y . :;.".~,. . ....' PCI'l1J892l05126 ~. ~VCD 92/22653 antigen binding andlor biological activity. If no effect is likely, use the human sequence at that site; if some affect is likely, eliminate the glycosylation site or use the import sequence at that site.

8. After completing the above analysis, determine the planned humanized sequence and prepare and test a sample. If the sample does not bind well to the target antigen, examine the particular residues listed below, regardless of the puestion of residue identity between the import and humanized residues.
a. Examine particular peripheral (note-CDR) variable domain residues that may, due to their position, possibly interact directly with a macromolecular antigen, including the following residues (where the " indicates residues which have been found to interact with antigen based on crystal structures):
i. Variable light domain: 36, 46, ~9°, 63-70 ii. Variable heavy domain: 2, 47°, 68, 70, 73-7.,.6.
b. Examine particular variable domain residues which could interact uvith, or otherwise affect, the conformation of variable domain CDRs, including the following (not including CDR residues themselves; since it is assumod that, because the CDRs interact with one mother, a~,y resis~ue in one CDR could potentially affect the c~nformation of another CDR residue) (L= LIGHT, H = HEAVY, residues appearing in b~Id are indicated to be structurally important according the Chothia et al., Mature .. ' 34~;1~77 (1989); and residues appearing in italic were altered ~5 during humanizati~n by Queen et al. (PDLD, Proc. Mall. Acad. Sci.
USA gg:10029 (1989) and Proc. Matl. Acad. 8ci, tJSA 88:2869 (1991 ):);
i. Variable tight domain:
aD CDR-1 (residues 24L-3~L): 2L, 4L, 66L-69L, 71 L
b) GDR-2 (residues 50L-56L): 35L, 46L, 47L, 48L, 49L, 58L, 62L. 64L-661:, 71 L, 73L
cl CDR-3 (residues 89L-97L): 2L, ~L, 36L, 98L, 37H, 45H, 47H, 58H, 6QH
ii. Variable heavy domain:

~.~,.,,,::..:, ., ,., :...
°~(~ 92/22b5~ ~ ~ ~ ~ ~ ~~ ~ PCI'/US92/05126 ,~,,>..;
a) CDR-1 (residues 26H-35H): 2H, 4H, 2~H, 36H, 71 H, 73H, 76H, 78H, g2H, ;94H
b) CDR-2 (residues 50H-55H): 49H, 69H, 69H, 7't H, 73H, 78H
c1 CDR-3 (residues 95H-102H): examine atl residues as possibl~~ interaction partners with this loop, because this loop varies in sire and conformation much more than the other CDRs.

9. If after step 8 the humanized variable~domain still is lacking in desired binding, repeat step 8. In addition, re-investigate any' buried residues which might affect the 'J~ -,/H interface (but ~rvhich would not directly affect CDR conformation). Additionally, evaluate the accessibility of non-CDR residues to solvent.

WO 92/22653 ~ 1 0 ~ ~ ~ ~ PCf'/US92/05126 EXAMPLE ~. En4ineering a Humanized Bist~ecific Ftab')2 Fragment This example demonstrates the construction of a humanized bispecific antibe:'y (BsF4ab')av1 by separate E, coii expression of each Fab' arm followed by directed chemical coupling in vitro. BsFlab')2 v1 ianti-CD3 anti-p185"ER2) was demonstrated to retarget the cytotoxic activity of human CD3+ CTL in vitro against the human breast tumor cell line, SK-BR-3, which overexpresses the p185"Epa product of the protooncogene HER2. This example demonstrates the minimalistic humanization strategy of installing as few murine residues as possible into a human antibody in order to recruit antigen-binding affinity and biological properties comparable to that of the murine parent antibody: This strategy proved very successful for the anti-p185"ERaarm of BsF(ab')2v1. tn contrast BsF(ab')2 v1 binds to T cells via its anti-CD3 arm much less efficiently than does the chimeric BsF(ab')Z
which contains the variable domains of the murine parent anti-CD3 antibody.
Here w~ have constructed additional BsF(ab)2 fragments containing variant anti-CD3 arms with selected murine residues restored in an attempt to improve antibody binding to T cells: One such variant, Bs Ftab')Z v9, was created by replacing six residues in the second hypervariable loop of the anti-CD3 heavy 2t? chain variable domain of BsFfab')a v1 with their counterparts from the murine parent anti-CD3 antibody. BsF4ab')Z v9 binds to T cel)s (Jurkat) much more efficiently than does BsF(ab')~ v1 and almost as efficiently as the chimeric BsFtab')2. This improvement in the efficiency of T cell binding of the humanized BsFtab'I~ is an important step in its development as a potential therapeutic agent for the treatment of p185"'ER~-overexpressing cancers.

Bispecefic antibodies tBsAbs) with specificities for tumor-associated antigens and surface markers on immune effector cells have proved effective for retargeting effector ;cells to kill tumor targets bath in vitro and in vivo treviewed by Fanger; M. W. et aJ., tmrrrunol. Today 10: 92-99 11989);

Fanger; M: W. et' al., lmmunol. Today 12: 51-54 (1991 ); and Nelson, H., Cancer Gells 3: 163-1 ?2 41991 )).' t3s~(ab') z fragments have often been used in preference to intact t3sAbs in retargeted cellular cytotoxicity to avoid the risk of killing innocent bystander cells binding to the Fc region of the antibody. An additional advantage of BsFtab')a over intact BsAbs is that they r~:~!:~'° ~ .:,,' . ;: ' : ,.~-r:, . . ;~ ' ~'~:':' ,::, ;;;. . , ..:v.. ,;.~:. .
W4 92/22653 ~ ~ ~ ~ ~ ~ ~ PCT/U592/05126 ,....,.
$~
are generally much simpler to prepare free of contaminating monospecific molecules (reviewed by Songsivilai, S. and Lachmann, P. J., Clin. Exp.
Immunol. 79: 315-321 (1990) and Nolan, O. and O°Kennedy, R., Biochim.
Biophys. Acta 1040: 1-11 (1990)).
BsF(ab')Z fragments are traditionally constructed by directed chemical , coupling of Fab' fragrnents obtained by'limited proteolysis plus mild reduction of the parent rodent monoclonal Ab (Brennan, M. et al , Science 229, 81-83 (1985) and Glennie, M. J, etal.-',°J. lmmunol, 139: 2367-23?5 (1987)).
One such BsF(ab')a fragment (anti-glioma associated antigen / anti-CD3) was found to have clinical efficaoy in glioma patients (Nitta, T. etal., Lancet 335:
36$-371 (1990) and another BsF(ab')2 (anti-indium chelate ! anti-carcinoembryonic antigen) allowed clinical imaging of colorectal carcinoma (Stickney, D. R. etal:, Antibody, Imrrrunocanj. Radiapharm. 2: 1-13 (1989)).
Future BsFtab')2 destined for clinical applications are likely to be constructed from antibodies which are either human or at least "humanized" (Riechmann, L. etal., Nature 332: 323-327 (1988) to reduce their immunogenicity (Hale, G. et al.; Lancet i: 1394-1399 (1988)).
Recently a facile route to a fully humanized BsF(ab')x fragment designed for tumor immunatherapy has been demonstrated (Shalaby, M. R. et al., J.
Exp: Mea! 175: 217-225 (19921): This approach involves separate E. coli expression of each Fab' arm followed by traditional directed chemical coupling in vitro to form the BsF(ab')2. Orie aim of the BsF(ab')a was a humanized version (Carter, P. et al.; Proc. Natl. Acad Sci. USA ( 1992a) and -., Cartel, P., et al., BiolTechnology 10: 163-167 (1992b)) of the marine ~5 monoclonal Ab 4D5 which is directed against the p185"~R2 product of the p~rotooncogene HER2 (c-erbB-2D (Fendiy, B. M. et al.. Cancer Res. 50: 1550-1558 (1989)). The humanization of the antibody 4D5 is shown in Example 1 of this application. The second arm was a minimalistically humanized anti-CD3 antibody (Shalaby etal:'supra) which was created by installing the CDR
loops from the variable domains of the marine parent monoclonal Ab UCHT1 (Beverley; P: C. L. and Callard; R. E., Eur. J. lmmunol. 11: 329-334 (1981 )) into the humanized anti-p185"ERZ antibody. The BsF(ab')2 fragment containing the most potent humanized anti-CD3 variant (v1 ) was demonstrated by flow ' cytometry to bind specifically to a tumor target f . .. ,:~.~ . .v.,.,,. ".;....,. , ., : ~ ~,..:....'.,. :. . w.
. , WO 9Z/22653 ~ ~ tl ~3 ~ ~ ~ PCH'/US92/05126 .., ~I
overexpressing p185"ERZ and to hurnan peripheral blood mononuclear cells carrying CD3. in addition, Bs Flab' )z v1 enhanced the cytotoxic effects of activated human CTL 4-fold against SK-BR-3 tumor cells overexpressing p185'°~Z. The example descries efforts to improve the antigen binding affinity of the humanized anti-CD3 arm by the judicious recruitment of a smelt number of additional murine residues into the minimalistically humanized anti-CD3 variable domains.
MATERIALS AND METH~DS
C~nstruction of mutations in the anti-CD3 variable region ,genes.
The construction of genes encoding humanized anti-CD3 variant 1 (v1 ) variable light (V~) and heavy (V") chain damains in phagemid pUC119 has been described (Shalaby et al. supra): Additional anti-CD3 variants were generated using an efficient site-directed mutagenesis method 4Carter, P., Muta9enesis: a~ practical approach, dM. J. McPherson, Ed.), Chapter 1, IRL
Press, Oxfiord, UK ( 1991 )) using mismatched oligonucleotides which either install or remove unique restriction sites. f?ligonucleotides used are listed below using lowercase o indicate the targeted mutations. Corresponding ~24 coding changes are denoted by the starting amino acid in one letter code followed by the residue numbered acdording to Kabat, E. A. etal., SeQuences of Proteins of Immunplogicat Jnterest, 5'" edition, National Institutes of Health; Bethesda, MD; USA 11991 ); then the replacement amino acid and ~; 'f finally the identity of the anti-CD3 variant:
HX 11; 5' GTAGATAAATCCtctAACACAGCCTAtCTGCAAATG 3' (SEC~:ID. NO. 11 ) VH K75S, v6;
HX12; 5' GTAGATAAATCCAAAtctACAGCCTAtCTGCAAATG 3' 6SEQ.tD. NO. 12) V" N76S; .v7;
HX13, 5' GTA~ATAAATCCtcttctACAGCCTAtCTGCAAATG 3' i (SEQ.ID. NO. 13) V" K75S:N76S' v8;
X14, 5' CTTATAAAGGTGTTtCcACCTATaaCcAgAaatTCAA
GGatCGTTTCACgaTAtcCGTAGATAAATCC 3' (SEQ.ID.ND. 14) V" T5'?S:A6DN:D61 Q:S62K:V63F:G65D, v9;
LX6, 5' CTATACCTCCCGTCTgcatTCTGGAGTCCC 3' (SEQ.ID. NO. 15) ,.. .,~ '..:~:~ . :..' :'.~.., . :y., . . . ~.~.. . ,.
. a . :"f. . .
P
~cr/us~zeomz6 , .
WO lB/~~653 V~ E55H, v11.
Oligonucleotides HX11, HX12 and HX13 each remove a site for BspMl, whereas LX6 removes a site for Xhol and HX14 installs a site for Eco~3V
(bold). Anti-CD3 variant v10 was constructed from v9 by site-directed mutagenesis using oligonucleotide. HX13, Mutants were verified by dideoxynucleotide sequencing. (Sariger, F. et al., ProG. IVatl. .4cad. Sci.
PISA
74: 5463-5467 (1977)).
E, coli expression of Fab' fragrr~ents The expression plasmid, pAK19, for the co-secretion of tight chain and heavy chain Fd' fragment of the most preferred humanized anti-p185"ERz variant, HuMAb4D5-8; is described in Carter et al., 199~b, supra. Briefly, the Fab' expression unit is bicistronic with both chains under the transcriptional cantrof of the phoA promoter. Genes encoding humanized V~
and V" domains are precisely fused on their 5' side to a gene segment encoding the heat-stable enterotoxin II signet sequence and on their 3°
side to human k, C~ and IgG1 C"1 constant domain genes, respectively. The C"1 gene is immediately followed by a sequence encoding the hinge sequence CysAlaAla and followed by a b~cteriophage ~I to transcriptiona) terminator.
Fab' expression plasrv~ids for chimeric and humanized anti-CD3 variants (v1 to v4, Shalaby et at.aupra; v6 to v1 ~, this study) were created from pAK19 by precisely repBacing anti-p185"~Z V~ and V" gene segments with those .~
,~, encoding mursne and ct~rresponding humanized variants of the anti-CD3 ~5 antibody, respedtively; by sub-cloning and site-directed mutagenesis. The Fab' expressign ptasmid for the most potent humanized anti-CD3 variant identified in this study (v9) is designated pAK2~. The anti-p185"~2 Fab' fragment seas secreted fr~m ,~. coli K12 strain 25F2 containing plasmid pAK19 grown for 32 to~40 hr at, 37' C in an aerated 10 liter fermentor. The final cell densigy vvas 120-150 ODSSO and the titer of soluble and functional anti-pi 85"E'~2 Fab' was 1-2 g/titer as judged by antigen binding ELISA
(Carter et al., 199~b, supra). Anti-CD3 Fab' variants were secreted from E coli containing c~rresponding expression plasmids using very similar fermentation protocols. The highest expression titers of chimeric and O 92J22ba3 ~.~ o ~ ~ ~ ~ P~1~S92I051~b humanized anti-GD3 variants ware 2ta0 mgAiter arid 700 mgliiter, rsspeGtively, as judgBd bY total immunoglobulin LISA.
Cansrructlon of esF(ab'ls fragments Fab' fragments ware directly recovered from E. toll fermentation pastes in th0 free thiol form (Fab'-SH) by affinity pu~ificatiar~ on Streptococcal protein G at pH 5 in the presence of ~DTA (Carter et al., 199~b supra).
Thioether linked BsFtab'iz fragments (antnpf $5'~RZ J anti-CD3) were constructed by the procedure of Glannie'et al. supra with the following modifications. Anti-p 1 i35"~ Fob'-SH in t fl0 rriM 'Cris acetate. 8 mM SDTA
(pH ~,O) was reacted with O.t vol of 40 mM N.N'-1,2-phenylenedimatemide (o-PDM) in dinn~athy) lormamide for -'I.5 hr at 20 -C. ~xcass o-PDM was removed by protein G vurifiaation of the Fab' maioimide derivative (Fob'-mal) followed by buffer exchange into 24 rnM sodium acetate, 5 mM I:DTA (pH
5.~) Icoupllng buffer) using o~ntriprep-'~0 concentrators (AmiGOn). The total eoncentration of Fab' variants was estimated from the measured absorbents at 280 nm (Hu~b4D5-8 Fab° a°."' ° '1.56, Carter et al., 1992b, supra).
The free_ti~141 content of Fab' preparations was estimated by reaction with 5, 5'-dithiobis(2-nitrobenzoic ,acid) as desoribed by Creighton, T. E..
protein structura~ a praciical~apprvach, IT. E. Creighton, Ed.). Chapter 7, IRL Press.
Oxford, UK ( 1990), Fquimolar amounts of anti-p185"a" Fob'-ma) (assumivg ' quantitative, reaotlon of Fab'-SH with 4-PDM) and each anti-Ci7S Fab'-SH
variant werecoupled together at a combined concentration of 1 to 2.5 mgJrnl in the Goupl4rtg buffer for to to 48 hr at, 4 'C. The coupling reaction was ~'' ad ju~te~ td 4 mM cysteine ax IpH~ 7.0 and ~ incubated for 15 min at 20 ' C to reduce .any ~Wa~rted disulfide-linked Flab' )= formed. These reduction Gorrditians are sufficient to redi~Ge inter-heavy, chin disulfide bonds with virtually no' reduction of the disulfide between light and heavy chains. Any free thiol~ generated were their blacked with a0 mM iodaacetamida.
BsF(a~'Iz wss isolated f~bm the Coupling reaction by S 9 DO-HR IPharmaoia) 5iz9 eXC,lusion cnr'"omatogrsPhy -I2-5 crrr x 1 d0 Gm) in the presence of P~3S.
The BsFtab'l,.,samplqs ,were passed through a 0.2 mm filter flesh frozen in .
liquid nitrogen aryl stored at ~~-'~Q:'~~
*_.~rade.mark TOTRL P.20 5.& 19/12/20(71 16:23 X416 368 1645 - -_ Oreceived WO'~2/22653 ~ ~ ~ ~ ~ ~~ ~ P~f/US92fOS126 ,-FIoHr cyto~etric analysis of Flab' i?binding to Jurkat cells The Jurkat human acute T cell leukemia cell line was purchased from the American Type Culture Collection (Rockviile, MDD (ATCC TIB 152D and Brawn as recommended by the ATCC. Aliquots of 10~ Jurkat cells were incubated with appropriate concentrations of BsFfab'D2 tanti-p185"ER2 / anti-GD8 variantD or control mono-specific'~anti-p185"ER2 F(ab')2 in PBS plus 0.1 °~
~w/vD bovine serum albumin and.10 mM sodium azide for 45 min at 4 "C.
The cells were washed and than incubated with fluorescein-conjugated goat anti-human F(ab'DZ tOrganon Teknika, West Chester, PAD for 45 min at 4 ° C.
Cells were washed and analyzed on a FAGScan tBecton Dickinson and Go., Mountain View, CA1. Cells (8 x 10~D were acquired by list mode and Bated by forward light scatter versus side light scatter excluding dead cells and debris.
RESULTS
Design of humanized anti-CD3 variants The most potent humanized anti-CD3 variant previously identified, v1, differs from the marine parent antibody, UCHT1 at 19 out of 107 amino acid residues within V~ and at 37 out of 122 positions within VH ~Shalaby et ~l.,su~raD 1992D: Here we recruited back additional marine residues into anti-CD3 v1 in an attempt to improve the binding affinity for CD3. The strategy chosen was a compromise between minimizing both the number of additional marine «sidues recruited and the number of anti-CD3 variants to be analyzed. We focused our attentions on a few CDR residues which were ~rigina9ly kept as human sequences in our minimalistic humanization regime.
Thus human residues in VH CDR2 of anti-GD8 v1 were replaced en bloc with their marine counterparts to give anti-CD3 v9:
T57S:A60N:D61 Q:S62K:V63F:G65D (Fig. 5D. Similarly, the human residue E55 in VL CDR2 of anti~CD3 v1 was replaced with histidine from the marine anti-GD3 antibody to generate anti-CD3 v11. In addition, V" framework region tFRD resialues 75 and 76 in anti-CD3 v1 were also replaced with their marine counterparts to creqte anti-CD3 v8: K75S:N76S. VH residues 75 and 76 are located in a loop close to V,~ C~R 1 and CDR2 and therefore might ~.,.,,-.;.: .. .., . .
,...~,~V~ 92/22653 ~ ~ ~ ~ ~ ~ ~ PCT/US92/OS126 sS
influence antigen binding. Additional variants created by combining mutations at these three sites are described below.
Preparation of BsF(ab°,12 fragments Soluble and functional anti-p185"~R~ and anti-CD3 Fab' fragments were recovered directly from corresponding E. Coli fermentation pastes with the single hinge cysteine predominantly in the free thioi form (75-100 % Fab'-SH) by affinity purification on Streptococcal protein G at pH
5 in the presence of EDTA (Carter et al., 1992b, supra). Thioether-linked BsF(ab')2 fragments were then constructed by directed coupling using o-PDM
as described by Glennie et al., supra. One arm was always the most potent humanized anti-p185"ERZ variant, HuMAb4D5--8 (Carter etal., 1992a, supra) and the other either a chimeric or humanized variant of the anti-CD3 antibody. Anti-p185"E"2 Fab'-SH was reacted with A-PDM to form the rnaleimide derivative (Fab'-mal) and then coupled to the Fab'-SH
for each anti-CD3 variant. F(ab')2 was then purified away from unreacted Fab' by size exclusion chromatography as shown for a representative preparation (BsF(ab')~ v8) in data not shown. The F(ab')Z fragment represents ~- 54~ of the total amount of antibody fragments (by mass) as judged by integration of the chromatograph peaks.

SDS-PAGE analysis of this BsF(ab')a v8 preparation under non-reducing conditions gave one major band with the expected mobility (M, -- 96 kDD as well as several very minor bands (dada not shown). Amino-terminal sequence ; ,~

analysis of the major band after efectroblotting on to polyvinyiidene difiuoride 26 membrane Mat~udaira, P:, J. .Biol. Chew. 282: 10035-10038 (198'7) gave the expected nnixed sequence from a stoichiometric 1:1 mixture of light and heavy chains (V~ l V": D/E, I/~J, QIQ, M/L, TIV, Q/E, S/S) expected for BsF(ab')2. The amino terminal region of both light chains are identical as are both heavy chains and correspond to donsensus human FFi sequences.
We have previously demonstrated that F(ab')2 constructed by directed chemical coup9in~ carry both anti-p185"~''2 and anti-CD3 antigen specificities (Shalaby et aL; supra). The level of contamination of the BsF(ab')2 with monospecific ~(ab' )2 is likely o be very low since mock coupling reactions with either anti-p185"E"2 Fib'-mal or anti-CD3 Fab'-SH alone did not yield detectable 2~~0~~;~~
CVO 92/22653 PC'I'/US92/d5126 - -w.
8 (~
quantities of Flab' )Z. Furthermore the coupling reaction was subjected to a mild reduction step followed by alkylation to remove trace amounts of disulfide-finked Flab' )2 that might be present. SDS-PAGE of the purified F(ab' )2 under reducing conditions gave two major bands with electrophoretic mobility and amino terminal sequence anticipated for free light chain and thioether-linked heavy chain dimers.
Scanning LASER dsnsitometry of a o-PDM coupled F(ab')Z preparation suggest that the minor species together represent --10°~ of the protein.
These minor contaminants were characterized by amino terminal sequence analysis and were tentatively identified on the basis of stoichiometry of light and heavy chain sequences and their electrophoretic mobility (data not shown). These data are consistent with the minor contaminants including imperfect Ftab' )z in which the disulfide bond between light and heavy chains is missing in ane or both arms, trace amounts of Fab' and heavy chain thioether-linked to tight chain.
finding of BsF(ab'Jz to Jurkat cells Binding of BsFtab')2 containing different anti-CD3 variants to Jurkat cells (human acute'T cell IeukemiaD was investigated by flow cytometry (data not shown). BsF(ab')a v9 binds much more efficiently to Jurkat cells than does our starting molecule; BsF(ab°)a v1, and almost as efficiently as the chimer'sc BsFtaka')2. Installation of additional marine residues into anti-CD3 v9 to create v10 (V" K75S:N76S) and v'l2 (V,~ K75S:N76S plus V~ E55H) did s: P
not farther improve, binding of corresponding BsF(ab°)2 to Jurkat cells. Nor did recruitment of these s°reurine residues into anti-CD3 v1 improve Jurkat binding: V~, K75S (v6): d/~ N76S tv7), V~, K75S:N7CS tvB), V' E55ii tv11 ) tnot shown). BsF(ab')2 v9 was chosen for future study since it is amongst the most efficient variants in binding to Jurkat cells and contains fewest marine residues in the humanized anti-CD3 arm. A monospecific anti-p1 B5~'~RZ Flab' )a did not show significant binding to Jurkat cells consistent with the interaqtion being mediated through the anti-CD3 arm.
DISCUSSI~N
A minima(istic strategy was chosen to humanize the anti-p1 B5"E~2 V1'~CD 92/22653 ~ ~ ~ ~ ~ ~ ~ PCTlUS92l0512G
(Carter efi al., 1992a, supra) and anti-CD3 arms (Shalaby et al., supra) of the BsF(ab')2 in this study in an attempt to minimize the potential immunogenicity of the resulting humanized antibody in the clinic. Thus we tried to install the minimum number of marine CDR and FR residues into the context of consensus human variable domain sequences as required to recruit antigen-binding affinity and biological properties comparable to the marine parent antibody. Molecular modeling was used firstly to predict the marine FR
residues which might be important to antigen binding and secondly to predict the marine CDR residues that might not be required. A small number of humanized variants were then constructed to test these predictions.
Our humanization strategy was very successful for the anti-p185"ERz antibody where one out of eight humanized variants (HuMAb4D5-8, IgG 1 ) was identified that bound the p185"~RZ antigen - 3-fold more tightly than the parent marine antibody (Carter at al., 1992a, supra). HuMAb4D5-8 contains a total of five marine FR residues and nine marine CDR residues, including V"
CDR2 residues 60-65, were discarded in favor of human counterparts. In contrast, BsF(ab')2 vl containing the most potent humanized anti-CD3 variant out of four originally constructed (Shalaby etal., supra) binds J6 cells with an affinity (/Cd) of 140 nM which is -- 70-fold weaker than that of the _ corresponding chimetic lBsF(ab')2.
Here we have rest~red T cell binding of the humanized anti-CD3 close to that of the chimeric variant by replacing six human residues in V" CDR2 with their marina aounte~parts: T57S:A60N:D61 C~:S62K:V6SF:G65D (anti,; $
CD3 v9; Fig. 5). It.appears rinore 6ikely that these marine residues enhance antigen binding indirectly by influencing the conformation of residues in the N-tarmirfal part of V" CDR2 rather than by directly contacting antigen.
Firstly, only N-terminal residues in V" CDR2 (50-58) have been found to contact antigen ih one or more of eight crystallographic structures of antibody/antigen complexes (Kabat et al., supra; and Mian, 1. S. at al., J.
lVfot. viol. - 2'! 7: 183-151 ( 1991 ): Fig. 5). Secondly, molecular modeling suggests that residues in the C-terminal part of V" CDR2 are at least partially buried tFig. 5). BsF(ab')a v9 binds to SK-BR-3 breast tumor cells with equal efficiency ~o BsF(ab°)2 v1 and chimeric BsF(ab')Z as anticipated since the anti-p185"ERa arm is identical in all of these molecules (Shalaby et al., supra, not W~O 92!22653 ~ ~ ~ ~ ~ ~j PCf'/US92/05126 shown).
Our novel approach to the construction of BsF(ab')2 fragments exploits an E. coli expression system which secretes humanized Fab' fragments at gram per titer titers and permits their direct recovery as Fab'-SH (Carter et al., 1992b, supra). Traditional vdirected chemical coupling of Fab'-SH
fragments is then used to form BsF(ab')2 in vitro (Brennan et al.,supra; and Glennie et al., supra). This route to Fab'-SH obviates problems which are inherent in their generation from intact antibodies: differences in susceptibility to proteolysis and nonspecific cleavage resulting in heterogeneity, low yield as well as partial reduction that is not completely selective for the hinge disulfide bonds. The strategy of using ~ co/i-derived Fab'-SH containing a single hinge cysteine abolishes some sources of heterogeneity in BsF~ab')2 preparation such as intra-hinge disulfide formation and contamination with intact parent antibody whilst greatly diminishes others, eg. formation of Flab' )3 fragments.
BsF(ab'l2 fragments constructed here were thioether-linked as originally described by Gteenie et a/., supra with future in vivo testing of these molecules in mind-. Thioether bonds, unlike disulfide bonds, are not susceptible'ta cleavage by trace amounts of thiot, which led to the proposal 2Q that thioether-linked Flab' )2 may be more stable than disulfide-(inked F(ab' )2 in vivo (Glennie et al., supra): This hypothesis is supported by our preliminary pharmecokinetic experiments in normal mice which suggest that thioether-finked BsF(ab')Z vl has a 3- fold longer plasma residence time than ~; ~
BsF(ab')2 v.1 iinked'by a single disulfide bond. Qisulfide and thioether-linked chime~ic BsF(ab')Z,were found to be indistinguishable in their efficiency of cell binding and in their retafgeting of CTL cytotoxicity, which suggests that o-PDM directed cou~ating does not compromise binding of the BsF(ab°)2 to either antigen (not shown). Nevertheless the nature of the linkage appears not to be critical since; a disulfide~lir~ked BsF(ab')2 (marine anti-p185"ERZ
i marine anti-CD3) was 'recently shown by others (Nishimura et al., lnt. J.
Cancer 50: 800-804 (192) to have potent anti-tumor activity in nude mice.
Our previous study (Shalaby et al.; supra! together with this one and that of Nishimura~ T. et al., supra improve the potential for using BsF(ab')a in targeted immunotherapy of ~pl g5"ERZ_overexpressing cancers in humans.

. . . ..
. :, . . ° y . . ', . . . .
W(~ 92/22653 ~ ~, Q ~ o ~~ 9 PCT/LJS92/05126 ~~
EXAMPLE 4. Humanization of an anti-CD18 antibody A marine antibody directed against the leukocyte adhesion receptor ~-chain (known as the H52 antibody) was humanized following the methods described above. Figures 6A and 6~ provide amino acid sequence comparisons for the marine and humanized antibody light chains and heavy chains.

... , : .. . .. , ;.
!~O 92/22653 ~ ~ Q ~ ~ y ~ ~ Q Pt.T/U~92/05126 SEQUENCE LISTING
(1) GENERAL INFORMATION:
(i) APPLICANT: Genentech, Inc.

(ii) TITLE OF INVENTION: Imamunoglot~ulin Variants (iii) NUMBER OF SEQUENCES: 25 ~ . , , (iv) CORRESPONDENCE ADDRESS:

(A) ADDRESSEE: Genentech, Inc.

(B) STREET: 460 Point San Bruno Blvd (C) CITY: South San Francisco '95 (D) STATE: California (E) COUNTRY: USA

(F) ZIP: 94080 (v) COMPUTER READABLE FORM:

(A} MEDIUM TYPE: 5.25 inch, 360 K'~ floppy disk (B) COMPUTER: IBM PC compatible (C) OPERATING SYSTEM: PC-DOS/MS-DOS

(D) SOFTWARE: patin (Genentech) (vi) CURRENT APPLICATION DATA:

(.A) APPLICATION NUMBER:

(B) FILING DATE:

(C) CLASSIFICATION:

3G1 (vii) PRIOR AP,PLICAT~ON DATA:

(A) APPLICATION NUMBER: 07/715272 (B) APPLICATION DATE: 14-JUN-1991 (viii) ATTORNEY/AGENT INFORMATION:

(A) NAME: Adler, Carolyn R.

(B) REG1STRATION NUMBER: 32,324 (C) REF'ERENCE/DOCKET NUMBER: 709P1 , r; .~

(ix)- TELECOMMUNICATION TNFORMATION:

~() (A) TELEPkIONE: 415/225-2614 (B) TELEFAX: 415/952-9881 (C} TELEX: 910/371-7168 (2) INFORMATION
FOR SEQ
ID N0:1:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 109 amino acids (B} TYPE: amino acid (D) TOPOLOGY: linear ~ ~ ~
~ ~ '~
~

' .. ~'Vb'Q 92/22b53 T/US92/05125 P(.

(xi) SEQUENCE DESCRIPTION: SEQID
N0:1:

Asp Ile Gln Thr GlnSer ProSer SerLeu SerAla SerVal Met GIy Asp Arg Thr IleThr CysArg AlaSer GlnAsp ValAsn Val Thr Ala Va1 Trp TyrGln GlnLys ProGly LysAla ProLys A1a Leu Leu Ile Ser AlaSer FheLeu GluSer GlyVal ProSer Tyr Arg Fhe Ser Ser ArgSer GlyThr AspFhe ThrLeu ThrIle GIy Ser Ser Leu Pro GluAsp PheAla ThrTyr TyrCys GInGln G1n His Tyr Thr Pro ProThr PheGly G1nGly ThrLys ValGlu Thr 95 100 lOS

I1e Lys Arg Thr 2~a 109 ' (2) INFORMATION FOR
SEQ
ID
N0:2:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: acids 120 amino (B) TAPE: amino acid (D) 3,'OPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQID
N0:2:

Glu Val Gln Val GIuSer GlyG1y G1yLeu ValG1n FroGly Leu 1 5 10 15 ,~,~
~e GIy Ser Leu Leu SerCys AlaAla SerGly FheAsn IleLys Arg Asp Thr Tyr His TrpVa1 ArgGln AlaPro GlyLys G1yLeu Ile ~4~ Glu Trp Va1 Arg IleTyr ProThr AsnGly TyrThr ArgTyr Ala Ala Asp Ser Lys GIyArg PheThr IleSer A1aAsp ThrSer Val Lys Asn Thr Tyr LeuGln MetAsn SerLeu ArgAla GluAsp Ala 9~V(~ 92/22653 ~ ~ ~ ~ ~ .j ~ P~'/1JS92/05126 Thr Ala Val Tyr Tyr Cys Ser Arg Trp Gly Gly Asp G1y Phe Tyr Ala Met Asp Val Trp Gly Gln Gly Thr Leu Va1 Thr Val Ser Ser (2) INFORMATION FOR SEQ ID N0:3:
1~ (i) SEQUENCE CHARACTERISTICS:
(Aj LENGTH: 109, amino acids (B) TYPE: amino acid (D) TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID N0:3:
Asp Ile G1n Met Thr Gln Ser Pro Ser Ser Leu Ser Ala Ser Val 24 Gly Asp Arg Val Thr Ile Thr Cys Arg Ala Ser Gln Asp Val Ser Ser Tyr Leu Ala Trp Tyr Gln Gln Lys Pro Gly Lys A1a Pro Lys Leu Leu Ile Tyr Ala A1a Ser Ser Leu Glu Ser Gly Val Pro Ser Arg Phe Ser Gly Ser Gly Ser Gly Thr Asp Phe Thr Leu Thr I1e Ser Ser Leu Gln Pro Glu Asp Phe Ala Thr Tyr Tyr Cys G1n Gln 35 Tyr Asn Sex Leu Pro Tyr Thr Phe Gly GIn Gly Thr Lys Ua1 Glu Ile Lys Arg Thr 109 ' (2) INFORMATION FOR SEQ ID N0:4:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 120 amino acids (B) TYPE: amino acid (D) TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID N0:4:
50 Glu Val Gln Leu Val Glu Ser Gly Gly Gly Leu Val G1n Pro Gly GIy Ser Leu Arg Leu Ser Cys Ala Ala Ser Gly Phe Thr Phe Ser ....' ,, , '..' ":'. . , ,.~ , .; ~.~'~... ..,;...~ , ,. ..... ' ... v ',:.'.~
. ...... ...
W~ 92/22653 ~ ~ ~ ~ ~ ~ ~ ~CI'/U~92/~5126 q 43 Asp Tyr Ala Met Ser Trp Val Arg Gln Ala Pro Gly Lys Gly Leu 35 . 40 45 Glu Trp Val Ala Val Ile Ser Glu Asn Gly Gly Tyr Thr Arg Tyr Ala Asp Ser Val Lys Gly Arg Phe Thr Ile Ser Ala Asp Thr Ser 65 ?0 ?5 1A Lys Asn Thr Ala Tyr Leu Gln Met Asn Ser Leu Arg Ala Glu Asp Thr Ala Val Tyr Tyr Cys Ser Arg Trp Gly Gly Asp Gly Phe Tyr 95 '100 105 A1a Met Asp Val Trp Gly G1n Gly Thr Leu Val Thr Val Ser Ser 2~0 (2) INFORMATIONFOR SEQ
ID N0:5:

( i) SEQUENCECHARACTERISTICS:

(A) LENGTH: acids 109 amino (B) TYPE: amino acid (D) TOPOLOGY:
linear (x i) SEQUENCEDESCRIPTION:SEQ ID
N0:5:

Asp Ile Val Thr Gln His LysPhe MetSer ThrSer Val Met Ser 3~ 1 5 10 15 Gly Asp Arg Ser Ile Cys LysAla SerGln AspVal Asn Val Thr ~5, Thr Ala Va1 Trp Tyr Gln LysPro G1yHis SerPro Lys Ala Gln 35 . 40 45 Lsu Leu Ile Ser Ala Phe ArgTyr ThrGly ValPro Asp Tyr Ser . 50 55 60 Arg Phe Thr Asn Arg Gly ThrAsp PheThr PheThr Ile Gly Ser ~,5 ?0 75 Ser Ser Val Ala Glu Leu A1aVal TyrTyr CysGln Gln Gln Asp ~5 80 85 90 His Tyr Thr Pro Pro Phe GlyGly GlyThr LysLeu Glu Thr Thr 95 ~ 100 105 50 Ile Lys Arg Ala V1~~ 92/22653 ~ ~ ~ ~ '~ ~ . F'C°T/6.JS92/AS126 {2) INFORMATION FOR SEQ ID N0:6:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 120 amino acids (B) TYPE: amino acid (D) TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID N0:6:
Glu Val Gln Leu Gln Gln Ser Gly Pro Glu Leu Val Lys Pro Gly Ala Ser Leu Lys Leu Ser Gys Thr Ala Ser G1y Phe Asn Ile Lys ~ 25 30 Asp Thr Tyr Ile His Trp Val Lys G1n Arg Pro Glu Gln Gly Leu Glu Trp Ile Gly Arg Ile Tyr Pro Thr Asn Gly Tyr Thr Arg Tyr Asp Pro Lys Phe Gln Asp Lys A1a Thr Ile Thr A1a Asp Thr Ser Ser Asn Thr Ala Tyr Leu Gln Val Ser Arg Leu Thr Ser Glu Asp Thr Ala Val Tyr Tyr Cys Ser Arg Trp G..y Gly Asp G1y Phe Tyr Ala Met Asp Tyr Trp Gly Gln Gly Ala Ser Val Thr Val Ser Ser 110 11.5 120 (2) INFORMATION FOR SEQ ID N0:7:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 27.bases ' (B) TyPE:wucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID N0:7:

(2) INFORMATION FOR SEQ ID N0:8:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 31 bases (B) TYPE: nucleic acid ~t'.~~p ~ . '=J . . . . ' i '~ ~s .~,......, . .. ,...~.:~;'.: .~ , ,:;~. ,m~,:~,.'. ' , .,, ' ~.: . ~ :
' :.'." . ~ . ';. ..,.. .
'WO 92122653 210 3 f~ ~ 9 p~T/US92/OS126 °I 5 (C) STRANDEDNESS: single (D) TOPOLOGX: linear (xi) SEQUENCE DESCRIPTION: SEQ ID N0:8:

(2) INFORMATION FOR SEQ ID N0:9:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 22 bases (B) TYPE: nucleic acid (Cj STRANDEDNESS: single (D) TOPOLOGY: linear (~ci) SEQUENCE DESCRIPTION: SEQ ID N0:9:

(2) INFORMATION
FOR SEQ
ID N0:10:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 34 bases 34 (B) TYPE: nucleic acid (C) STRANDEDNESS: single (~y) TOPOLOGY: linear (~ci) SEQUENCE ~3ESCRIPTION: SEQ ID N0:10:

TGAGGAGACG GTGACCGTGG TCCCTTGGCC CCAG 34 , 4i (2) INFORMATION
EOTt SEQ
ID NO:11 (i) SEQUENCE CHARACTERISTICS:

(~) LENGTH: 36 bases ~5 (B) 'TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO:11;

;GTAGATAAAT ~CTCTAACAC AGCCTATCT~ CAAATG 36 YS;,x r.'. ''; :.~.'~ ..,.; ., . !.,'.'..... ;.:.: ~ ,;~:. .;'..~:".. :.,. ,~i -.."~ ;... ' ''.'.".
.7 ::~ ., . .
~''~ a- .:r. . ,t.-, t ...
!< ~,~.~.. '~ ~'%'~ .., '' .Y.....~ ..;9-, v , '.'.. .,.~. ., '' ' ~ .., :;, .. " .' .~, ;, ,;~ ;., 2~.fl~f~~~
WO 92122653 V '~' '~. PC.°T/US92/0~126 (2) INFORMATION FOR SEQ ID N0:12:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 36 bases (B) TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear, (x~.) SEQUENCE DESCRIPTION: SEQ ID N0:12:
' (2) INFORMATION F'OR SEQ TD NO:13:
(i) SEQUENCE CHARACTERISTICS;
(A) LENGTH: 36 bases (B) TYPE: nucleic acid (C) STRANDEDNESS; single (D) TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID N0:13:

GTAGATAAAT CCTCTTCTAC AGCCTATCTG CAAA1'G 36 (2) INFORt~tATION
FOR SEQ
LD N0:14:

(i) SEQUENCE CHARACTERISTICS:

(~,) ' ~;EN~TH: 68 bases ~5 (B) TYPE: nucleic acid (C) STRANDEDNESS: single (D) T0P0LOGY: liner ,; r~

(xi) SEQUENCE'DESCRIPTION: SEQ ID N0:14:

CTTATAAAGG TGTTTCCACC TATAACCAGA AATTCAAGGA TCGTTTCACG

$5, ATATCCGTAG ATAAATCC 68 (2) INFORMATION
FOR SEQ
ID NO:15:

50 (i) SEQUENCE CHARACTERISTICS:

(Al LENGTH: 30 bass (g) TYPE: nucleic arid (C) STRANDEDNESS: singly (D) TOPOLOGY: linear k 'n. ,. ~ ..- ~,, .:. ~,' . ~ ~ .~ . , ,~., ,,~... -y... ~ ::. ,. ..... . ...
7 r..~:....,. . ~::; .mt.~ , ..:-... . .. ' . ~. ~ ,. . .... ~.~..'. ,... ~ ~.
-.w . ..
P~'/LJS92/05126 (xi)~SEQUENCE DESCRIPTION: SEQ ID N0:15:

(2) INFORMATION FOR SEQ ID N0:16:
( i) SEQUENCE CHE~RACTERISTICS
(A) LENGTH: 107 amino acids (B) TYPE: amino acid (D) TOPOLOGY: linear (xi) SEQUENCE DESGRIPTION: SEQ ID N0:16:
Asp Ile Gln Met Thr Gln Thr Thr Ser Ser Leu Ser Ala Ser Leu Gly Asp Arg Va1 Thr Ile Ser Cys Arg Ala Ser Gln Asp Ile Arg Asn Tyr Leu Asn Trp Tyr Gln Gln Lys Pra Asp G1y Thr Va1 Lys Leu Leu Ile Tyr Tyr Thr Ser Arg Leu His Ser G1y Val Pro Ser Lys Phe Ser Gly Ser Gly Ser Gly Thr Asp Tyr Ser Leu Thr Ile ' 3~ 65 70 75 Ser Asn Leu Glu Gln Glu Asp Ile A1aThr TyrPhe CysG1n Gln Gly Asn Thr Leu Pro Trp Thr Phe AlaGly GlyThr LysLeu Glu 95 100 , 1,05 Ile Lys (2) INFORMATION FOR SEQ ID N0:17:

(i,) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 107 amino acids (B) APE: amino acid (D) TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID,N0:17:

5~: Asp Ile Gln Met Thr Gln Ser Pro SprSer LeuSer AlaSer Val Gly Asp Arg Val Thr Ile Thr Cys ArgAla SerGln AspIle Arg . 20 25 30 WU 92/2263 ~ ~ ~ ~ PGT/LJ892/05126 ~ ~~ ~

B

Asn Tyr TrpTyr GlnGln LysPro GlyLys AlaPro Lys Leu Asn Leu Leu TyrThr SerArg LeuG1u SerGly ValPro Ser Ile Tyx Arg Phe SerGly SerGly ThrAsp TyrThr LeuThr I1e Ser Gly Ser Ser ProGlu AspPhe AlaThr TyrTyr CysGln Gln Leu Gln G1y Asn ProTrp ThxPhe GlyGln GlyThx LysVal Glu Thr Leu 95 '100 105 Ile Lys (2) INFORMATION SEQ
FOR ID
N0:18:

(i) SEQUENCE CHARACTERISTICS:

(A) acids LENGTH:
1.07 amino (B) amino TYPE: acid (D) linear TOPOLOGY:

(xi) SEQUENCE SE~QID
DESCRIPTION: N0:18:

Asp Ile ThrGln SexPro SerSer LeuSer AlaSer Val Gln Met Gly Asp Theale ThrCys ArgAla SerGln SerIle Ser ArE
Val Asn Tyr TrpTyr GlnG1n LysPro GlyLys AlaPro Lys Leu Ala 35 40 ~5 Leu Leu Ala'A1aSerSer LeuG1u SerGly ValPro Ser Ile Tyr 50 55 60 ~..a 4~ Arg Ptae SerGly SerGly ThrApp PheThr LeuThr Iie Ser Gly Ser Ser ProGlu AspPhi A1aThr TyrTyr GysG1n Gln Leu Gln Tyr Asn ProTxp ThrPhe GlyGln GlyThr LysVal Glu Ser Leu iii LyS

~.~~

V~iJ 92/22653 ~ ~ ~ c~! ~ ~ ~ PCTlUS92/~5126 (2) INFORMATION FOR SEQ ID N0:19:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 129 amino acids (B) TYPE: amino acid (D) TOPOL(?GY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID N0:19:
Glu Val Gln Leu Gln Gln Ser Gly Pro Glu Leu Val Lys Pro Gly Ala Ser Met Lys Ile Ser Cys Lys Ala Ser Gly Tyr Ser Phe Thr ~ 25 30 Gly Tyr Thr Met Asn Trp Val Lys Gln Ser His Gly Lys Asn Leu Glu Trp Met Gly Leu Ile Asn Pro Tyr Lys Gly Val Ser Thr Tyr 20 50 55 fi0 Asn Gln Lys Phe Lys Asp Arg Phe Thr Ile Ser Lys Ala Thr Leu 25. Thr Val AspwLys Ser Ser Ser Thr Ala Tyr Leu Met Glu Leu Leu Asn Ser Leu Thr Ser Glu Asp Sex A1aVal TyrTyr CysAla Arg Ser Gly Tyr Tyr G1y Asp Ser Asp ~rpTyr PheAsp ValTrp Gly Ala Gly Thr Thr Val Thr Val Ser Ser (2),INFORMATION FOR SEQ ID ~:r N0:20:

(i),SEQUENCE CHARACTERISTICS:

(A) LENGTH: '122 amino acids (B) TYPE: amino acid (D) TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID
N0:20:

Glu Val Gln Leu Val Glu Ser Gly GlyGly LeuVa1 GlnPro Gly Gly Ser L~u Arg Leu Ser Cys -AlaAlaSer GlyTyr SerPhe Thr Gly Tyr~Thr Met Ann Trp Val Arg GlnAla ProGly LysGly Leu ~V~O X2/22653 ~ ~ ~ ~ ~ ~ ~ ~"~i'/11S~32/OS126 Glu Trp Val Ala Leu Ile Asn Pro Tyr Lys Gly Val Ser Thr Tyr Asn Gln Lys Phe Lys Asp Arg Phe Thr IIe Ser Val Asp Lys Ser Lys Asn Thr Ala Tyr Leu Gln Met Asn Ser Leu Arg Ala G1u Asp Thr Ala Va1 Tyr Tyr Cys Ala Arg Ser Gly Tyr Tyr Gly Asp Sar Asp Trp Tyr Phe Asp Val Trp Gly Gln G1y Thr Leu Val Thr Val Ser Ser (2) INFORMATION FOR SEQ
ID N0:21:

~0 ( i} SEQUENCE CkIARACTERISTICS:

(A) LENGT~1: 122 aminoacids (B) TYPE: amino acid (D) TOPOLOGY: linear (xi) SEQUENCE SEQ ID
DESCRIPTION: N0:21:

Glu Val Gln Leu Val Glu G1y GlyGly LeuVal GlnPro Gly Ser Gly Ser Leu Arg Leu Ser Ala AlaSer GlyPhe ThrPhe Ser Cys Ser Tyr Ala Met Ser Trp Arg GlnAla ProGly LysGly Leu Val G1u Trp Val Ser Val Ile Gly AspGly G1ySer ThrTyr Tyr Ser 40 Ala Asp Ser Val Lys Gly Phe ThrIle SerArg AspAsn Ser Arg Lys Asn Thr Leu Tyr Leu Met AsnSer LeuArg AlaGlu Asp Gln Thr A.la VaI Tyr Tyr Cys Arg G1yArg ValGly TyrSer Leu Ala Ser Gly Leu Tyr Asp Tyr GIy GlnGly ThrLeu ValThr Val Trp Ser Ser .

~v~ ~z~zzss~ 2 ~. 0 3 (~ ~ 0 ~c°°rius~zios'z6 (2) INFORMATION FOR SEQ ID N0:22:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 454 amino acids (B) TYPE: amino acid (D) TOPOLOGY: linear (xi) SEQUENCE DESCRT:?TION: SEQ ID N0:22:
Gln Val G1n Leu Gln Gln Ser G1y Pro Glu Leu Val Lys Pro G1y Ala Ser Val Lys I1e Ser Cys Lys Thr,Ser G1y Tyr Thr Phe Thr Glu Tyr Thr Met His Trp Met Lys Gln Ser His Gly Lys Ser Leu Glu Trp Ile Gly Gly Phe Asn Pro Lys Asn G1y Gly Ser Ser His Asn Gln Arg Phe Met Asp Lys Ala Thr Leu Ala VaI Asp Lys Ser Thr Ser Thr A1a Tyr Met Glu Leu Arg Ser Leu Thr Ser Glu Asp Ser Gly Ile Tyr Tyr Cys Ala Arg Trp Arg Gly Leu Asn Tyr Gly Phe Asp ValArg TyrPhe AspVal TrpGly A1aGly ThrThr Val Thr Val SerSer AlaSer ThrLys GlyPro SerVal PhePro Leu AIa Pro SerSex Lys.Ser ThrSer GlyGly ThrAla AlaLeu Gly :140 145 150 Cys Leu ValLys AspTyr PhsPro GluPro ValThr Va1Ser Trp Asn Ser GlyAla LeuThr SerGly ValHis ThrPhe ProAla Val Leu G1n SerSer GlyLeu TyrSer LeuSer SerVal ValThr Val a 185 ~ 190 195 Pro Ser SerSer LeuGly ThrGln ThrTyr IleCys AsnVal Asn His Lys ProSer AsnThr LyrsVal AspLys LysVal GluPro Lys :,''... ...... ..... ,...= . ..:::.:;
i3'~ 92/22653 ~ ~ ~ ~ ~ ~ ~ PCT/US92105126 (~ Z
Ser Cys Asp Lys Thr His Thr Cys Pro Pro Cys Pro Ala Pro Glu Leu Leu Gly Gly Pro Ser Val Phe Leu Phe Pro Pro Lys Pro Lys Asp Thr Leu Met Ile Ser Arg Thr Pro Glu Val Thr Cys Val Val Val Asp Val Ser His~,Glu Asp Pro Glu Va1 Lys Phe Asn Trp Tyr 27~5~ 280 285 Val Asp GIy Val Glu Val His Asn Ala Lys Thr Lys Pro Arg Glu Glu Gln Tyr Asn Ser Thr Tyr Arg Val Val Ser Val Leu '1'hr Val Leu His Gln Asp Trp Leu Asn Gly Lys Glu Tyr Lys Cys Lys Val Ser Asn Lys Ala Leu Pro Ala Pro Ile Glu Lys Thr Lle Ser Lys 2~ Ala Lys Gly Pro Arg Pro GlnVal TyrThr LeuPro Pro Gln Glu Ser Arg Glu Met Thr Asn GlnVal SerLeu ThrCys Leu Glu, Lys Val Lys Gly Ty~cPro Asp 'IleAla ValGlu TrpGlu Ser Phe Ser Asn Gly Gln Glu Asn Tyr LysThr ThrPro ProVa1 Leu Pro Asn Asp Ser Asp Ser Phe Leu TyxSer LysLeu ThrVal Asp ~-.e Gly Phe Lys Ser Arg GlriGln Asn Val~'heSerCys SerVal Met Trp Gly 4~5 430 435 Hia Glu Ala His Asn Tyr ~'hrG1n LysSex LeuSer Leu Leia His Ser Pro Gly , Lys (2) TNFORMATIONFOR SEQ :
ID
N0:23 ( j.) SEQUENCECHARACTERISTICS:

(A)' LENGTH: 57 aminoacids (~) TYPE: ami no acid (D)'TOPOLOGY: linear '~V~ 92/22653 ~ ~ ~ PC1'/U~92/0512~6 p o3 (xi) SEQUENCE DESCRIPTION: SEQ ID N0:23:
His His Gln Val Gln Leu Gln Gln Ser Gly Pro Glu Leu Val Lys Pro Gly Ala Ser Val Lys Ile Ser Cys Lys Thr Ser Gly Tyr Thr Phe 1'hr Glu Met Gly Trp Ser Cys Ile Ile Leu Phe Leu Val Ala Thr Ala Thr Gly Val His Ser Glu Val Gln Leu Val Glu Ser G1y Gly Gly Leu Val Gln Pro Gly Gly Ser Leu Arg Leu Sex Cys Ala Thr Ser Gly Tyr Thr Phe Thr Glu Tyr Thr Met His Trp Met Arg Gln Ala Pro Gly Lys Gly Leu Glu Trp Val Ala G1y Ile Asn Pro Lys Asn Gly Gly Thr Ser His Asn Gln Arg Phe Met Asp Arg Phe 110 las 120 Thr Il.eSerVal AspLys SerThr SerThr A1a TyrMet GlnMet 3G Asn Ser LeuArg AlaGlu AspThr AlaVal Tyr TyrCys AlaArg Trp Arg GlyLeu AsnTyr GlyPhe AspVal Arg TyrPhe AspVal Trp Gly GlnGly ThrLeu Va1Thr ValSer Ser AlaSex ThrLys 170 1.75 180 G1y Pro SerVal Phe-Pro LeuAla PrnCys Ser ArgSer ThrSer Glu Sar ThxAla AlaiLeu GlyCys LeuVal Lys AspTyr PhePro 45- Glu Pro ValThr ValSer TrpAsn SerGly Ala LeuThr SerGly Val His ThrPhe ProAla ValLeu GlnSer Ser GlyLeu TyrSer Leu Ser SerVal ValThr Val'ThrSerSer Asn PheGly ThrGln ;:.r F:.:.., P ~
J.,,~, ~.~i,. :.., y 7 !3 >

Gu.
tG 7. .:: k 7 .
.. ~x I .:fin 1 , 4 4 ~0.' .A.-a., . ..>t d~'. ~ r 1 ' .~ 1 ~~.4r ~. , 7f , ,,. > . . ~, .' a"
:,:4" ...
:~ ;. , , F,.., . ...:;~ .;'~',t...y, ;.:. ,";'.~,;: ' .. ;.,.~.'~~o,> , ' m ", ..y'.. .., ' ,.. ' ..'...~~ ~ . .,. .~ , ',~;~ .... ,, ,. ~.:.;
r, . , ' ". . ~ ~ .;,.;~.~ . . ",. v:.,:. , . ;: , , . . ,..,. . ;. ,.. . ~
.~'~.. . ~.:.~.~ . :~ ~ ~ ' ~ .: , ~..,~. ..:.~ .,.
WCD 92/22653 ~ ~ ~ J ~ ~ ~ PC.TlUS92/05126 Ion Thr Tyr Thr Cys Asn Val Asp His Lys Pro Ser Asn Thr Lys Val Asp Lys Z'hr Val Glu Arg Lys Cys Cys Val Thr Cys Pro Pro Cys Pro Ala Pro Glu Leu Leu Gly G1y Pro Ser Val Phe Leu Phe Pro Pro Lys Pro Lys Asp Thr Leu Met Ile Ser Arg Thr Pro Glu Val 305 ~ 310 315 Thr Cys Val Val Val Asp Val Ser His Glu Asp Pro Glu Val Lys 320 ~ 32S 330 Glu Cys Pro Pra Cys Pro Ala Pro Pro Val A1a G1y Pro Ser Val Phe Leu Phe Pro Pro Lys Pro Lys Asp Thr Leu Met Ile Sex Arg Thr Pro Glu Val Thr Cys Val Val Val Asp Val Ser His Glu Asp Pro Glu Val Gln Phe Asn Trp Tyr Val Asp Gly Met Ghu Val Hi.s 380 385 ~390 Asn Ala Lys Thr Lys Pro Arg Glu Glu Gln Phe Asn Ser Thr Phe Arg Val Val Ser Val Leu Thr Val Val His Gln Asp Trp Leu Asn Gly Lys Glu Tyr Lys Cys Lys Va1 Ser Asn Lys Gly Leu Pro Ala Pxo Tle Glu Lys Thr Ile Sex Lys Thr Lys Gly Gln Pro Arg Glu 440 44S 450 ''~
Pro Gln Val Tyr Thr Leu Pro Pro Ser Arg Glu Glu Met Thr Lys Asn Gln Va1 Ser Leu Thr Cys Leu Va1 Lys Gly Phe Tyr Pro Ser Asp Ile Ala Val.Glu Trp Glu Ser Asn Gly Gln Pro Glu Asn Asn Ty~ Lys Thr Thr Pro Pro Met Leu Asp Ser Asp Gly Ser Phe Phe Leu Tyr Sir Lys Leu Thr Val Asp Lys Ser Arg Trp Gln G1n Gly ". :.
~V~ 92/22653 ~ ~ ~ ~ ~ ~ e~ PCT/LJS92i~15126 Asn Val Phe Ser Cys Ser Val Met His Glu Ala Leu His Asn His Tyr 'Thr Gln Lys Ser Leu Ser Leu Ser Pro Gly Lys (2) INFORMATION FOR SEQ ID N0:24:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 214 amino acids (B) TYPE: amino acid (D) TOPOLOGY: linear (xi) SEQUENCE DESGRIPTION: SEQ ID N0:24:
Asp Va1 G1n Met Thr GIn Thr Thr Ser Ser Leu Ser Ala Ser Leu Gly Asp Arg VaI Thr Ile Asn Cys Arg AlawSer GIn Asp Ile Asn Asn Tyr Leu Asn Trp Tyr GIn Gln Lys Pro Asn Gly Thr Val Lys Leu Leu IIe Tyr Tyr Thr Ser Thr Leu His Ser Gly Va1 Pro Ser Arg Phe Ser Gly Ser G1y Ser Gly Thr Asp Tyr Ser Leu Thr Ile Ser Asn Leu Asp Gln Glu Asp Ile Ala Thr Tyr Phe Cys Gln Gln Gly Asn Thr Leu Pro Pro Thr Phe Gly Gly Gly Thr Lys Val Glu Ile Lys Arg Thr Val Al:a Ala Pro Ser Val Phe Ile Phe Pro Pro , 4~0 Ser Asp Glu Gln Leu Lys Ser Gly Thr AIa Ser Val Val Cys Leu Leu Asn Asn Phe Tyr Pro Arg Glu Ala Lys Val Gln Trp Lys Val Asp Asn Ala Leu GIn Ser G1y Asn Ser Gln Glu Ser Val Thr GIu Gln Asp Ser Lys Asp Ser Thr Tyr Ser Leu Ser Ser Thr Leu Thr 50 170 ' 175 180 Leu Ser Lys AIa Asp Tyr Glu Lys His Lys Val Tyr Ala Cys Glu ~~, r .,,...,.~. , ~ ,.. ~. ...::~., ., ,..: ,,:;... , . :~', '~';. . ' ::' ..
,~::. .....~'~.'y.:, ~,... ., ., ,~,.~:.:.. . .,~ ~,.... ' .;~. .. .
........:. . ;....
CVO 92/22653 ~ ~ ~ ~ ~ ~ ~ P~(:'1'/US92105126 . -~ bb Val Thr His Gln Gly Leu Ser Ser Pro Val Thr Lys Ser Phe Asn Arg Gly Cys Glu (2) INFORMATION D
FOR SEQ N0:25:
I

(i) SEQUENCE
CHARACTERISTICS:

(A) acids LENGTH:

amino (P) amino TYPE: acid (D) linear TOPOLOGY:

(~i) SEQUENCE SEQ ID
DESCRIPTION: N0:25:

Met G1y Ser CysIle IleLeu PheLeu ValAla ThrAla Thr Trp Gly Val Ser AspIle GlnMet ThrGln SexPro SerSer Leu His Ser Ala Val GlyAsp ArgVal ThrIle ThrCys ArgAla Ser Ser Gln Asp Asn AsnTyr LeuAsn TrpTyr GlnGln LysPro Gly Ile Lys Ala Lys LeuLeu IleTyr TyrThr SerThr LeuHis Ser Pro Gly Val Ser ArgPhe SerGly SerGly SerGly ThrAsp Tyr Pro Thr Leu Ile SerSer LeuGln ProGlu AspPhe AlaThr Tyr Thr ~5 95 100 105 Tyr Cys Gln GlyAsn ThrLeu ProPro ThrPhe GlyGln Gly Gln Thr Lys G~LuIleLys ArgThr ValAla AlaPro SexVal Phe Val Ile Phe Pro SerAsp GluGln LeuLys SerGly ThrAla Ser Pro '6Ia1 Val Leu LeuAsn AsnPhe TyrPro ArgGlu AlaLys Val Cys Gln Trp Val AspAsn AlaLeu GlnSer GlyAsn SerGln Glu Lys Ser Va1 G1u GlnAsp SerLys AspSer ThrTyr SerLeu Ser Thr u~r~ :,, ,,.
. ~'~ ~zrzzs~3 ~ ~ o ~ ~ ~ ~ Pf'I'/1J~92/05126 lob Ser Thr heu Thr Leu Ser Lys Ala Asp Tyr Glu Lys His Lys Val . 200 205 210 Tyr Ala Cys Glu Val Thr His Gln G1y Leu Ser Ser Pro Val Thr Lys Ser Phe Asn Arg Gly Glu Cys

Claims (129)

WE CLAIM:

1. A method for making a humanized antibody comprising non-human, import Complementarity Determining Region (CDR)amino acid residues and human Framework Region (FR) amino acid residues, comprising the steps of:
(a) obtaining the amino acid sequences of an import variable domain and of a VH subgroup III consensus human variable domain;
(b) identifying CDR amino acid sequences in the import and the human variable domain sequences;
(c) substituting import CDRs for the corresponding human CDRs;
(d) aligning the amino acid sequences of a FR of the import antibody and the corresponding FR of the consensus variable domain;
(e) identifying import antibody FR residues in the aligned FR sequences that are non-homologous to the corresponding consensus variable domain residues;
(f) determining if the non-homologous import amino acid residue is expected to have at least one of the following effects:
(1) non-covalently binds antigen directly;
(2) interacts with a CDR; or (3) participates in the VL-VH interface;
(g) for any non-homologous import antibody amino acid residue which is expected to have at least one of these effects, substituting that residue for the corresponding amino acid residue in the consensus variable domain FR
sequence; and (h) preparing a humanized antibody which binds antigen, wherein the humanized antibody comprises an amino acid sequence determined according to the above steps.

2. The method of claim 1, having an additional step of determining if any such non-homologous residues are exposed on the surface of the domain or buried within it, and if the residue is exposed, retaining the consensus residue.

3. The method of claim 1, having the additional steps of searching the import variable domain sequence for glycosylation sites, determining if any such glycosylation site is expected to affect the antigen binding or affinity of the antibody, and if so, substituting the glycosylation site into the consensus sequence.

4. The method of claim 1, having the additional steps of searching the consensus variable domain sequence for glycosylation sites which are not present at the corresponding amino acid in the import sequence, and if the glycosylation site is not present in the import sequence, substituting the import amino acid residues for the amino acid residues comprising the consensus glycosylation site.

5. The method of claim 1, having an additional step which comprises aligning import antibody and consensus variable domain FR sequences, identifying import antibody FR
residues which are non- homologous with the aligned consensus FR sequence, and for each such non- homologous import antibody FR residue, determining if the corresponding consensus variable domain residue represents a residue which is highly conserved across all species at that site, and if it is so conserved, preparing a humanized antibody which comprises the consensus amino acid residue at that site.

6. The method of claim 1, wherein the corresponding consensus residues are selected from the group consisting of 4L, 35L, 36L, 38L, 43L, 44L, 46L, 58L, 62L, 63L, 64L, 65L, 66L, 67L, 68L, 69L, 70L, 71L, 73L, 85L 87L, 98L, 2H, 4H, 24H, 36H, 37H, 39H, 43H, 45H, 49H, 58H, 60H, 67H, 68H, 69H, 70H, 73H, 74H, 75H, 76H, 78H, 91H, 92H, 93H, and 103H.

7. A method for making a humanized antibody comprising non-human Complementarity Determining Region (CDR) amino acid residues and human Framework Region (FR) amino acid residues, comprising providing an import, non-human antibody variable domain amino acid sequence having CDR
amino acid residues and FR amino acid residues; obtaining the amino acid sequence of a VH subgroup III consensus human antibody variable domain having CDR amino acid residues and FR amino acid residues; substituting non-human CDR amino acid residues for human CDR amino acid residues in the consensus human antibody variable domain; substituting an amino acid residue for the consensus amino acid residue at at least one of the following sites:
4L, 35L, 36L, 38L, 43L, 44L, 46L, 58L, 62L, 63L, 64L, 65L, 66L, 67L, 68L, 69L, 70L, 71L, 73L, 85L, 87L, 98L, 2H, 4H, 24H, 36H, 37H, 39H, 43H, 45H, 49H, 58H, 60H, 67H, 68H, 69H, 70H, 73H, 74H, 75H, 76H, 78H, 91H, 92H, 93H.
and 103H; and preparing a humanized antibody which binds an antigen, wherein the humanized antibody comprises an amino acid sequence determined according to the above steps.

8. The method of claim 7, wherein the substituted residue is the residue found at the corresponding location of the non-human antibody.

9. A humanized antibody variable domain having a functional antigen binding region, said humanized antibody variable domain comprising non-human Complementarity Determining Region (CDR) amino acid residues incorporated into a VH subgroup III consensus human antibody variable domain, and further comprising an amino acid substitution at a site selected from the group consisting of:
4L, 35L, 36L, 38L, 43L, 44L, 46L, 58L, 62L, 64L, 65L, 66L, 67L, 68L, 69L, 70L, 71L, 73L, 85L, 87L, 98L, 2H, 4H, 24H, 36H, 37H, 39H, 43H, 45H, 49H, 58H, 60H, 68H, 69H, 70H, 73H, 74H, 75H, 76H, 78H, 92H, and 93H.

10. The humanized antibody variable domain of claim 9, wherein the substituted residue is the residue found at the corresponding location of the non-human antibody from which the non-human CDR amino acid residues are obtained.

11. The humanized antibody variable domain of claim 9, wherein no human (FR) Framework Region (FR) residue other than those set forth in the group has been substituted.

12. A method for making a humanized antibody comprising introducing Complementarity Determining Region (CDR) amino acid residues from an import antibody variable domain into a VH subgroup III consensus human antibody variable domain.

13. A humanized antibody variable domain having a functional antigen binding region, said humanized antibody variable domain comprising non-human Complementarity Determining Region (CDR) amino acid residues incorporated into a VH subgroup III consensus human antibody variable domain and further comprising a non-human import Framework Region (FR) residue, wherein the non-human import FR residue introduces a glycosylation site which affects the antigen binding or affinity of the humanized antibody variable domain.

14. A humanized antibody which binds the HER2 receptor with an affinity of about 4.7 nM Kd or better affinity and comprises a heavy chain variable domain which comprises non-human import antibody Complementarity Determining Region (CDR) amino acid residues incorporated into a VH subgroup III consensus human variable domain.

15. A humanized variant of a non-human parent antibody, wherein the humanized variant comprises non-human Complementarity Determining Region (CDR) amino acid residues and human Framework Region (FR) amino acid residues, and:
(a) binds the HER2 receptor with an affinity of about 4.7 nM Kd or better affinity:
(b) mediates specific cell lysis of SK-BR-3 calls in the presence of IL-2 activated human peripheral blood lymphocytes at least about four fold more effectively than the non-human parent antibody; and (c) mediates Antibody Dependent Cellular Cytotoxicity (ADCC) selective for cell types which overexpress p185HER2 at least about two fold more effectively than for cell types which express low levels of p185HER2.

16. The humanized variant of claim 15 which binds the HER2 receptor with an affinity of about 0.82 nM Kd or better affinity.

17. The humanized variant of claim 16 which binds the HER2 receptor with an affinity of about 0.10 nM Kd.

18. The humanized variant of claim 15 which inhibits proliferation of SK-BR-3 cells incubated for 96 hr with the antibody.

19. The humanized variant of claim 18 wherein the antibody inhibits proliferation of SK-BR-3 cells to about 66% of untreated control or greater inhibition.

20. A humanized variant of a non-human parent antibody, wherein the humanized variant comprises non-human Complementarity Determining Region (CDR) amino acid residues and human Framework Region (FR) amino acid residues, and;
(a) binds the HER2 receptor with an affinity of about 4.7 nM Kd or better affinity; and (b) comprises a Framework Region (FR) amino acid substitution at a site selected from the group consisting of 73H, 78H, 93H and 66L, utilizing the numbering system set forth in Kabat.

21. The humanized variant of claim 20 which consists of about 1 to about 5 FR substitutions.

22. The humanized variant of claim 20 which comprises a FR substitution at site 73H.

23. The humanized variant of claim 20 which comprises a FR substitution at site 78H.

24. The humanized variant of claim 20 which comprises a FR substitution at site 93H.

25. The humanized variant of claim 20 which comprises a FR substitution at site 66L.

26. The humanized variant of claim 20 which further comprises a FR substitution at site 71H.

27. The humanized variant of claim 26 which comprises FR
substitutions at sites 71H, 73H, 78H, 93H and 66L.

28. A humanized variant of a non-human parent antibody, wherein the humanized variant comprises non-human Complementarity Determining Region (CDR) amino acid residues and human Framework Region (FR) amino acid residues; binds the HER2 receptor with better affinity than the non- human parent antibody; and comprises Framework Region (FR) amino acid substitutions at sites 71H, 73H, 78H, 93H and 66L, utilizing the numbering system set forth in Kabat.

29. An antibody which binds an antigen and comprises non-human heavy chain variable domain Complementarity Determining Region (CDR) amino acid residues which bind said antigen and VH subgroup III consensus human variable domain Framework Region (FR) amino acid residues; and further comprises non-human light chain variable domain CDR amino acid residues which bind said antigen.

30. The antibody of claim 29, further comprising VL kappa subgroup I consensus human variable domain FR amino acid residues.

31. A humanized antibody variable domain comprising non-human Complementarity Determining Region (CDR) amino acid residues which bind an antigen incorporated into a human antibody variable domain, and further comprising a Framework Region (FR) amino acid substitution at a site selected from the group consisting of: 4L, 38L, 43L, 44L, 58L, 62L, 65L, 66L, 67L, 68L, 69L, 73L, 85L, 98L, 2H, 4H, 36H, 39H, 43H, 45H, 69H, 70H, 74H and 92H, utilizing the numbering system set forth in Kabat.

32. The humanized variable domain of claim 31 wherein the substituted residue is the residue found at the corresponding location of the non-human antibody from which the non-human CDR amino acid residues are obtained.

33. The humanized variable domain of claim 31 wherein no human Framework Region (FR) residue other than those set forth in the group has been substituted.

34. The humanized variable domain of claim 31 wherein the human antibody variable domain is a consensus human variable domain.

35. The humanized variable domain of claim 31 wherein the residue at site 4L has been substituted.

36. The humanized variable domain of claim 31 wherein the residue at site 38L has been substituted.

37. The humanized variable domain of claim 31 wherein the residue at site 43L has been substituted.

38. The humanized variable domain of claim 31 wherein the residue at site 44L has been substituted.

39. The humanized variable domain of claim 31 wherein the residue at site 58L has been substituted.

40. The humanized variable domain of claim 31 wherein the residue at site 62L has been substituted.

41. The humanized variable domain of claim 31 wherein the residue at site 65L has been substituted.

42. The humanized variable domain of claim 31 wherein the residue at site 66L has been substituted.

43. The humanized variable domain of claim 31 wherein the residue at site 67L has been substituted.

44. The humanized variable domain of claim 31 wherein the residue at site 68L has been substituted.

45. The humanized variable domain of claim 31 wherein the residue at site 69L has been substituted.

46. The humanized variable domain of claim 31 wherein the residue at site 73L has been substituted.

47. The humanized variable domain of claim 31 wherein the residue at site 85L has been substituted.

48. The humanized variable domain of claim 31 wherein the residue at site 98L has been substituted.

49. The humanized variable domain of claim 31 wherein the residue at site 2H has been substituted.

50. The humanized variable domain of claim 31 wherein the residue at site 4H has been substituted.

51. The humanized variable domain of claim 31 wherein the residue at site 36H has been substituted.

52. The humanized variable domain of claim 31 wherein the residue at site 39H has been substituted.

53. The humanized variable domain of claim 31 wherein the residue at site 43H has been substituted.

54. The humanized variable domain of claim 31 wherein the residue at site 45H has been substituted.

55. The humanized variable domain of claim 31 wherein the residue at site 69H has been substituted.

56. The humanized variable domain of claim 31 wherein the residue at site 70H has been substituted.

57. The humanized variable domain of claim 31 wherein the residue at site 74H has been substituted.

58. The humanized variable domain of claim 31 wherein the residue at site 92H has been substituted.

59. An antibody comprising the humanized variable domain of claim 31.

60. An antibody which binds p185 HER2 and comprises a humanized antibody variable domain, wherein the humanized antibody variable domain comprises non-human Complementarity Determining Region (CDR) amino acid residues which bind p185 HER2 incorporated into a human antibody variable domain, and further comprises a Framework Region (FR) amino acid substitution at a site selected from the group consisting of:
4L, 38L, 43L, 44L, 46L, 58L, 62L, 65L, 66L, 67L, 68L, 69L, 73L, 85L, 98L, 2H, 4H, 36H, 39H, 43H, 45H, 69H, 70H, 74H, 75H, 76H, 78H and 92H, utilizing the numbering system set forth in Kabat.

61. The antibody of claim 60 wherein the substituted residue is the residue found at the corresponding location of the non-human antibody from which the non-human CDR amino acid residues are obtained.

62. The antibody of claim 60 wherein no human Framework Region (FR) residue other than those set forth in the group has been substituted.

63. The antibody of claim 60 wherein the human antibody variable domain is a consensus human variable domain.

64. The antibody of claim 60 wherein the residue at site 4L has been substituted.

65. The antibody of claim 60 wherein the residue at site 38L has been substituted.

66. The antibody of claim 60 wherein the residue at site 43L has been substituted.

67. The antibody of claim 60 wherein the residue at site 44L has been substituted.

68. The antibody of claim 60 wherein the residue at site 46L has been substituted.

69. The antibody of claim 60 wherein the residue at site 58L has been substituted.

70. The antibody of claim 60 wherein the residue at site 62L has been substituted.

71. The antibody of claim 60 wherein the residue at site 65L has been substituted.

72. The antibody of claim 60 wherein the residue at site 66L has been substituted.

73. The antibody of claim 60 wherein the residue at site 67L has been substituted.

74. The antibody of claim 60 wherein the residue at site 68L has been substituted.

75. The antibody of claim 60 wherein the residue at site 69L has been substituted.

76. The antibody of claim 60 wherein the residue at site 73L has been substituted.

77. The antibody of claim 60 wherein the residue at site 85L has been substituted.

78. The antibody of claim 60 wherein the residue at site 98L has been substituted.

79. The antibody of claim 60 wherein the residue at site 2H has been substituted.

80. The antibody of claim 60 wherein the residue at site 4H has been substituted.

81. The antibody of claim 60 wherein the residue at site 36H has been substituted.

82. The antibody of claim 60 wherein the residue at site 39H has been substituted.

83. The antibody of claim 60 wherein the residue at site 43H has been substituted.

84. The antibody of claim 60 wherein the residue at site 45H has been substituted.

85. The antibody of claim 60 wherein the residue at site 69H has been substituted.

86. The antibody of claim 60 wherein the residue at site 70H has been substituted.

87. The antibody of claim 60 wherein the residue at site 74H has been substituted.

88. The antibody of claim 60 wherein the residue at site 75H has been substituted.

89. The antibody of claim 60 wherein the residue at site 76H has been substituted.

90. The antibody of claim 60 wherein the residue at site 78H has been substituted.

91. The antibody of claim 60 wherein the residue at site 92H has been substituted.

92. A humanized antibody variable domain comprising non-human Complementarity Determining Region (CDR) amino acid residues which bind an antigen incorporated into a consensus human variable domain, and further comprising an amino acid substitution at a site selected from the group consisting of:

4L, 38L, 43L, 44L, 46L, 58L, 62L, 65L, 66L, 67L, 68L, 69L, 73L, 85L, 98L, 2H, 4H, 36H, 39H, 43H, 45H, 69H, 70H, 74H, 75H, 76H, 78H and 92H, utilizing the numbering system set forth in Kabat.

93. A humanized antibody which lacks immunogenicity compared to a non-human parent antibody upon repeated administration to a human patient in order to treat a chronic disease in that patient , wherein the humanized antibody comprises non-human Complementarity Determining Region (CDR) amino acid residues which bind an antigen incorporated into a human antibody variable domain, and further comprises an amino acid substitution at a site selected from the group consisting of:
4L, 38L, 43L, 44L, 46L, 58L, 62L, 65L, 66L, 67L, 68L, 69L, 73L, 85L, 98L, 2H, 4H, 36H, 39H, 43H, 45H, 69H, 70H, 74H, 75H, 76H, 78H and 92H, utilizing the numbering system set forth in Kabat.

94. A humanized variant of a non-human parent antibody which binds an antigen and comprises a human variable domain comprising the most frequently occurring amino acid residues at each location in all. human immunoglobulins of a human heavy chain immunoglobulin subgroup wherein amino acid residues forming Complementarity Determining Regions (CDRs) thereof comprise non-human antibody amino acid residues, and further comprises a Framework Region (FR) substitution where the substituted FR residue: (a) noncovalently binds antigen directly; (b) interacts with a CDR; (c) introduces a glycosylation site which affects the antigen binding or affinity of the antibody; or (d) participates in the VL-VH interface by affecting the proximity or orientation of the VL and VH regions with respect to one another.

95. The humanized variant of claim 94 which binds the antigen up to 3-fold more in the binding affinity than the parent antibody binds antigen.

96. A humanized antibody heavy chain variable domain comprising non-human Complementarity Determining Region (CDR) amino acid residues which bind antigen incorporated into a human antibody variable domain, and further comprising a Framework Region (FR) amino acid substitution at a site selected from the group consisting of: 24H, 73H, 76H, 78H, and 93H, utilizing the numbering system set forth in Kabat.

97. The humanized variable domain of claim 96 wherein the substituted residue is the residue found at the corresponding location of the non-human antibody from which the non-human CDR amino acid residues are obtained.

98. The humanized variable domain of claim 96 wherein no human Framework Region (FR) residue other than those set forth in the group has been substituted.

99. The humanized variable domain of claim 96 wherein the human antibody variable domain is a consensus human variable domain.

100. The humanized variable domain of claim 96 wherein the residue at site 24H has been substituted.

101. The humanized variable domain of claim 96 wherein the residue at site 73H has been substituted.

102. The humanized variable domain of claim 96 wherein the residue at site 76H has been substituted.

103. The humanized variable domain of claim 96 wherein the residue at site 78H has been substituted.

104. The humanized variable domain of claim 96 wherein the residue at site 93H has been substituted.

105. The humanized variable domain of claim 96 which further comprises an amino acid substitution at site 71H.

106. The humanized variable domain of claim 96 which further comprises amino acid substitutions at sites 71H
and 73H.

107. The humanized variable domain of claim 96 which further comprises amino acid substitutions at sites 71H, 73H and 78H.

108. An antibody comprising the humanized variable domain of claim 96.

109. A humanized variant of a non-human parent antibody which binds an antigen, wherein the humanized variant comprises Complementarity Determining Region (CDR) amino acid residues of the non-human parent antibody incorporated into a human antibody variable domain, and further comprises Framework Region (FR) substitutions at heavy chain positions 71H, 73H, 78H and 93H, utilizing the numbering system set forth in Kabat.

110. A humanized antibody variable domain comprising non-human Complementarity Determining Region (CDR) amino acid residues which bind an antigen incorporated into a human antibody variable domain, and further comprising a Framework Region (FR) amino acid substitution where the substituted FR residue:

(a) noncovalently binds antigen directly;
(b) interacts with a CDR; or (c) participates in the V L-V H interface by affecting the proximity or orientation of the V L and V H regions with respect to one another, and wherein the substituted FR
residue is at a site selected from the group consisting of: 4L, 38L, 43L, 44L, 58L, 62L, 65L, 66L, 67L, 68L, 69L, 73L, 85L, 98L, 2H, 4H, 24H, 36H, 39H, 43H, 45H, 69H, 70H, 73H, 74H, 76H, 78H, 92H and 93H, utilizing the numbering system set forth in Kabat.

111. The humanized variable domain of claim 110 wherein the substituted residue is the residue found at the corresponding location of the non-human antibody from which the non-human CDR amino acid residues are obtained.

112. The humanized variable domain of claim 110 wherein no human Framework Region (FR) residue other than those set forth in the group has been substituted.

113. A humanized antibody comprising the light chain variable domain sequence in SEQ ID NO: 17.

114. A humanized antibody comprising the heavy chain variable domain sequence in SEQ ID NO: 20.

115. A humanized anti-CD3 antibody comprising the amino acid sequences in SEQ ID NO: 17 and SEQ ID NO: 20.

116. A variable domain of a humanized antibody heavy chain comprising the variable domain sequence within SEQ
ID NO: 23.

117. A variable domain of a humanized antibody light chain comprising the variable domain sequence within SEQ
ID NO: 25.

118. A humanized anti-CD18 antibody comprising the heavy chain variable domain sequence within SEQ ID NO: 23 and the light chain variable domain sequence within SEQ ID
NO: 25.

119. A polypeptide comprising the following amino acid sequence:

GVPSRFS
GSX2SGTDFTLTISSLQPEDFATYYCQQHYTTPPTFGQGTKVEIK, wherein X1 is E or Y and X2 is R or G.

120. A polypeptide comprising the following amino acid sequence:
EVQLVESGGGLVQPGGSLRLSCMSGFNIKDTYIHWVRQAPGKGLEWVARIYPTNGYT
RYADS

SS, wherein X3 is A or R, X4 is T or D, X5 is A or L, X6 is S
or A and X7 is V or Y.

121. The polypeptide of claim 119 comprising the amino acid sequence:
DIQMTQSPSSLSASVGDRVTITCRASQDVNTAVAWYQQKPGKAPKLLIYSASFLYSG
VPSRFSG
SRSGTDFTLTISSLQPEDFATYYCQQHYTTPPTFGQGTKVEIK.

122. The polypeptide of claim 120 comprising the amino acid sequence:
EVQLVESGGGLVQPGGSLRLSCAASGFNIKDTYIHWVRQAPGKGLEWVARIYPTNGY
TRYADS

VKGRFTISADTSKNTAYLQMNSLRAEDTAVYYCSRWGGDGFYAMDYWGQGTLVTVSS

123. A humanized antibody comprising the heavy chain variable domain sequence:
EVQLVESGGGLVQPGGSLRLSCAASGFNIKDTYIHWVRQAPGKGLEWVARIYPTNGY
TRYADSV

wherein X1 is V or Y, and wherein the humanized antibody binds p185HER2 more tightly than the non-human import antibody from which the CDR amino acid residues of the humanized antibody are derived.

124. The humanized antibody of claim 123 further comprising the light chain variable domain sequence:
DIQMTQSPSSLSASVGDRVTITCRASQDVNTAVAWYQQKPGKAPKLLIYSASFLYSG
VPSRFSG
SRSGTDFTLTISSLQPEDFATYYCQQHYTTPPTFGQGTKVEIK.

125. A humanized anti-HER2 antibody which comprises the light chain variable domain amino acid sequence:
DIQMTQSPSSLSASVGDRVTITCRASQDVNTAVAWYQQKPGKAPKLLIYSASFLYSG
VPSRFSGS
RSGTDFTLTISSLQPEDFATYYCQQHYTTPPTFGQGTKVEIK
and the heavy chain variable domain amino acid sequence:
EVQLVESGGGLVQPGGSLRLSCAASGFNIKDTYIHWVRQAPGKGLEWVARIYPTNGY
TRYADSV
KGRFTISADTSKNTAYLQMNSLRAEDTAVYYCSRWGGDGFYAMDYWGQGTLVTVSS.

CLAIMS:

126. A humanized IgG1 antibody comprising the variable domains of Claim 125.

127. A humanized IgG1 antibody comprising human gamma 1 heavy chain constant domains, human light chain constant domains, and the variable domains of Claim 125.

128. A humanized IgG1 antibody comprising human gamma 1 heavy chain constant domains, human kappa 1 light chain constant domains, and the variable domains of Claim 125.

129. A humanized IgG1 antibody comprising human gamma non-A allotype heavy and light chain constant domains and the variable domains of claim 125.