AU691811B2 - Antibodies - Google Patents

Description

9k P A WO 94/29351 PCT/GB94/01290 FIELD OF THE INVENTION This invention relates to altered antibodies, to pharmaceutical, therapeutic and diagnostic compositions containing said antibodies; to processes for preparing said compositions; to methods of therapy and diagnosis using said antibodies, to a method of modulating the function of cell surface associated antigens using said, antibodies; to DNA sequences coding for said antibodies; to cloning and expression vectors containing DNA sequences coding for said antibodies; to host cells transformed with said vectors and to processes for preparing said antibodies.

BACKGROUND OF THE INVENTION In orrl",r for an antibody to be effective therapeutically it is desirable that it ac :,vas the required physiological effect without producing any significant advarse toxic effects. Such toxic effects may be mediated, for example, via complement fixation.

Antibody when bound to its cognate antigen can link to and activate the complement cascade. Complement consists of a complex series of proteins. The proteins of the complement system form two interrelated enzyme cascades, termed the classical and alternative pathways, providing two routes to the cleavage of C3, the central event in the complement system. The sequence of events comprising the classical complement pathway is recognition, enzymatic activation, and membrane attack leading to cell death. The recognition unit of the complement system is the Cl complex. The Cl complement protein complex is a unique feature of the classical complement cascade leading to C3 conversion. Complement fixation occurs when the Clq subcomponent binds directly to immunoglobulin antigen immune complex. Whether or not complement fixation occurs depends on a number of constraints. For example, only certain subclasses of ir. munoglobulin can fix complement even under optimal conditions. These are IgG1, lgG3 and IgM in man and IgG2a, igG2b and IgM in mice.

-I

WO 94/29351 PCT/GB94/01290 2 The C1q molecule is potentially multvalent for attachment to the complement fixation sites of immunoglobulin. The CH 2 domain of IgG and probably the CH 4 domain of IgM contain binding sites for Clq.

Fc bearing cells also play a role in enhancing the effect of the immune response by binding to and opsonising, phagocytosing or killing target cells coated with antibody of the relevant class. Three igG binding receptors (FcyR) have been described for murine and human leukocytes. FcyRI has high binding affinity for monomeric IgG, while FcyRII and FcyRill have low affinity for mono igG and interact mainly with antigen complexed IgG. The presence of Fc receptors confers on these immune cells the ability to mediate a number of effector mechanisms important in the effector phase of the humoral response.

The gamma 1 isotype of human IgG, like lgG3, binds to FcRI and, when complexed with its cognate antigen, activates complement and binds to FcRII and FcRIII. Conversely, human lgG2 and lgG4 are relatively inactive isotypes; both fail to activate the classical complement pathway and lgG4 binds weakly to FcRI [Burton, D R and Woof, J M (1992) Adv. Immunol. 1., 1. Lucisano Valim, Y M and Lachmann, P J. (1991) Clin. exp. Immunol, S4, 1].

Localisation of amino acid residues of IgG that interact with FcRI in the

CH

2 domain of human IgG is well established [Woof, J M get~ (1986) Molec. Immunol. 319. Lund, J g.W (1991) J. Immunol, 147, 2657; Canfield, S M and Morrison, S L (1991), J. exp. Med. 1ZJ, 1483; Chappel, S M ~tl, (1991) Proc. Natl. Acad. Sci. B3, 9036; Chappel, S M e.a/ (1993), J. Biol. Chem 26., 25124; Alegre, M-L .eLa (1992) J. Immunol, 1A4, 3461]. Amino acid sequence comparisons of the CH 2 domains of antibodies from different species and subclasses that bind well to FcRI suggested that a region at the N-terminal end of CH 2 comprising residues Leu 234 Ser 239 (using the Kabat Eu numbering system [Kabat, E A et ag, (1987) Sequences of proteins of Immunological interest. US Dept. of Health and Human Services, Bethesda, MD, USA]) is critical for interaction with FcRI. The motif Leu 234, Leu 25, Gly 236, Gly 237, Pro 238, Ser 239, ispresent in all IgG isotypes wi.n high affinity for FcRI [Woof, J M etal r WO 94/29351 PICT/GB94/01290 3 (1986), Molec. Immunol. 23 319]. Domain exchanges between Ig's with different Fc effector functions have demonstrated the importance of CH 2 for FcRI binding [Canfield, S M and Morrison, S L (1991), J. exp. Med. 1Z3.

1483; Chappel, S M .eL.a (1991) Proc. Natl. Acad. Sci. 8, 9036; Chappel, S M .eala (1993), J. Biol. Chem 2M. 25124] in particular the residue 235.

Replacement of the Leu residue at position 235 with a Glu residue reduces the affinity of lgG3 for FcRI by 100 fold [Lund, J geal (1991) J. Immunol, 147, 2657; Canfield, S M and Morrison, S L (1991), J. exp. Med. 1.1, 1483]. The same Leu 235 to Glu change when performed on an IgG4 variant of OKT3 [Alegre, M-L eal (1992) J. Immunol, 14., 3461] abolished its FoRI binding and, consequently, its mitogenic properties.

Although the sequence requirements for FcRIII binding has been less extensively studied, Sarmay .tLa [(1992) Molec. Immunol. 29 633] have, identified the CH 2 domain residues 234 to 237 as important for lgG3 binding to all three Fc receptors. The relative importance of each residue differs with each Fc receptor with 235 and 237 being most important for FcRIII mediated cell killing.

In contrast, another Fc mediated function, Clq binding and subsequent complement activation, appears to require the carboxyl terminal half of the

CH

2 domain [Tao, M Canfield, S and Morrison, S L(1991) J. Exp.

Med. 173, 1025]. Morrison's group, following sequence analysis of polymorphisms in the CH 2 domain of human IgGs also identified the importance of the C-terminal region of CH 2 With a Pro to Ser change at 331 in IgG1 they abolished complement fixation and reduced Clq binding [Tao, M H LW (1993), J. Exp. Med. 178, 661]. Using inter- and intradomain switch variants of CAMPATH-1, Greenwood BeLal. (1993) [Eur. J.

Immunol. 2, 5 1098] further endorsed the importance of the C-terminal end of CH 2 Complement fixation could be restored to human IgG4 with just the carboxyl terminal of CH 2 from residue 292 of IgG1 and not the Nterminal half or any other domain. Duncan Winter (1988) [Nature, 332, 21] identified a motif in CH 2 of Glu 318, Lys 320 and Lys 322 of the mouse IgG2b isotype. Changing any of these residues abolished Clq binding, as did the use of competitive peptides of sequences in this region. However, the Clq motif residues are also found in antibodies that do not fix sm~- I- WO 94/29351 PCT/GB94/01290 4 complement suggesting that these residues may well be necessary but not sufficient for complement activation.

We have found that amino acid residues necessary for Clq and FcR binding of human IgG1 are located in the N-terminal region of the CH 2 domain, residues 231 to 238, using a matched set of engineered antibodies based on the anti-HLA DR antibody L243. Changing the leucine 235 in the CH 2 region of IgG3 and lgG4 to glutamic acid was already known to abolish FcRI binding, we have confirmed this for IgG1 and also found a concomitant abolition of human complement fixation with retention of FcRIll mediated function. Changing the glycine at 237 to alanine of IgG1 also abolished FcRI binding and reduced complement fixation and FcRill mediated function. Exchanging the whole region 233 to 236, with the sequence found in human lgG2 abolished FcRI binding andcomplement fixation and reduced FcRIl mediated function of IgG1. In contrast, a change in the previously described Clq binding motif, from lysine at 320 to alanine had no effect on igG1-mediated complement fixation.

The proposed site Leu 234 Leu 235 Gly 236 Gly 237 Pro 238 Ser 239, is present in all IgG isotypes with high affinity for FcyRI. Recent mutagenesis experiments on lgG3 antibodies have introduced point mutations in this region and the ability of the mutants to interact with FcyRI has been examined [Lund etal (1991) J. Immunol 147. 2657-2662]. The most marked effect is seen at position 235 where replacement of the naturally occurring Leu residue with a Glu residue produces an Ig with a >100-fold decrease in affinity for FcyRI.

Our observation of the effect of this alteration at residue 235 on the ability of the antibody to fix complement was highly surprising. Earlier protein engineering studies had introduced mutations at various positions in order to locate the Clq-binding site on IgG [Duncan Winter (1988) Nature, 32, 738-740]. The binding site for Clq was localised to three side chains, Glu 318, Lys 320 and Lys 322 of the mouse lgG2b isotype. Residues Glu 318, Lys 320 and Lys 322 are conserved in all the human IgGs, rat IgG2b and IgG2c, mouse lgG2a, IgG2b and lgG3, guinea pig IgG1 and rabbit IgG.

1M Further experiments showed that the affinity of human Clq for mutant mouse IgG2b antibodies in which residue 235 was mutated was unaffected i.e. it was in the same range of values as that obtained with the wild type.

Although the fact that altering residue 235 of the C, 2 region of IgG is known to abolish FcyRI binding as we too observed, this concomitant substantial reduction in complement fixation has not been reported or suggested elsewhere and was completely unexpected.

SUMMARY OF THE INVENTION The invention provides a method of treating diseases in which antibody therapy leads to undesirable toxicity due to antibody mediated complement fixation comprising administering an altered antibody wherein one or more amino acid residues in the N-terminal region of the CH 2 domain of said antibody are altered characterised in that the ability of said antibody to fix complement is altered as compared to unaltered antibody.

In a preferred embodiment the altered antibody binds to one or more cellular Fc receptors especially FcRIII and excluding FcRI i.e. the antibody does not bind :"significantly to FcRI, and more preferably binding to FcRI is abolished.

Accordingly in a further aspect the invention provides an altered MHC specific S antibody wherein one or more amino acid residues in the N-terminal region of the 20 CH 2 domain of said antibody are altered characterised in that the ability of said antibody to fix complement is altered, as compared to unaltered antibody.

In a tirther preferred embodiment the invention therefore provides an altered antibody wherein one or more amino acid residues in the N-terminal region of the

CH

2 domain of said antibody are altered characterised in that the ability of said antibody to fix complement is altered as compared to unaltered antibody and said altered antibody binds to one or more cellular Fc receptors especially FcRIII and does not bind significantly to FcRI.

MAW IIF #19346 RSI 31 March 1998 II I I WO 94/29351 PCT/GB94/01290 6 The constant region of the antibodies to be altered according to the invention may be of animal origin and is preferably of human origin. It may also be of any isotype but is preferably human IgG and most preferably human IgG1.

In a preferred embodiment of the invention the amino acid residue(s) which is altered lies within amino acid positions 231 to 239, preferably within 234 to 239.

In a particularly preferred embodiment of the invention the amino acid residue(s) which is altered lies within the motif Leu 234 Leu 235 Gly 236 Gly 237 Pro 238 Ser 239.

In a most preferred embodiment the amino acid residue(s) which is altered is either Leu 235 and/or Gly 237.

DETAILED DESCRIPTION OF THE INVENTION As used herein the term 'altered' when used in conjunction with the ability of an antibody to fix complement most usually indicates a decrease in the ability of antibody to fix complement compared to the starting antibody. By choosing appropriate amino acids to alter it is possible to produce an antibody the ability of which to fix complement is substantially reduced such as for example by altering residue Leu 235. It is also possible to produce an antibody with an intermediate ability to fix complement by, for example altering amino acid residue Gly 237.

As used herein the phrase 'substantially reduce complement fixation' denotes that human complement fixation is preferably 30%, more preferably 20% and most preferably <10% of the level seen with the starting wild type unaltered antibody.

The term 'significantly' as used with respect to FcRI binding denotes that the binding of antibody to FcRI is typically 20%, and is most preferably of that seen with unaltered antibody.

I

lb-- s IL- WO 94/29351 PCT/GB94/01290 7 The altered antibodies of the invention preferably bind to FcRIII as measured by their ability to mediate antibody dependent cellular cytotoxicity (ADCC) at a concentration no greater than ten times that of the wild type unaltered antibody.

The proteins encoded in the Major Histocompatibility Complex region of the genome are involved in many aspects of immunological recognition. It is known that all mammals and probably all vertebrates possess basically equivalent MHC systems and that immune response genes are linked to the MHC.

In man the major histocompatibility complex is the HLA gene cluster on chromosome 6. The main regions are D, B, C and A. The D region contains genes for class II proteins which are involved in cooperation and interaction between cells of the immune system. Many diseases have been found to be associated with the D region of the HLA gene cluster. Studies to date have shown associations with an enormous variety of diseases, including most autoimmune diseases (see for example, European Patent No. 68790). European Patent No. 68790 suggests controlling diseases associated with a particular allele of certain regions of the MHC such as the HLA-D region in humans by selectively suppressing the immune response(s) controlled by a monoclonal antibody specific for an MHC-class II antigen.

We have found that by altering an MHC-class II specific antibody at position 235 in the N-terminal region of the CH 2 domain it is possible to produce an antibody which fully retains its immunosuppressive properties but which has substantially reduced toxicity invitr and is tolerated in_ iQ.

In a further preferred embodiment the invention provides an MHC specific antibody wherein one or more amino acid residues in the N-terminal region of the CH 2 domain of said antibody are altered characterised in that the ability of said antibody to fix complement is altered as compared to unaltered antibody.

-"I

I~ WO 94/29351 PCT/GB94/01290 8 In a preferred embodiment the invention provides an MHC specific monoclonal antibody characterised in that said antibody has been altered at position 235 of the N-terminal region of the CH 2 domain.

In some instances such as with MHC specific monoclonal antibodies it may be desfrable that the alteration in the N-terminal region of the CH 2 domain of the antibody while altering the ability to fix complement additionally inhibits the binding to FcRI receptors.

The antibodies are preferably specific for MHC-class II antigens and due to the alteration of one or more amino acid residues in the N-terminal region of the CH 2 domain will not bind significantly to FcRI.

In a further preferred embodiment the altered antibodies of the invention or.

for use according to the invention are directed against an MHC class II antigen characterised in that said antibody has been altered at position 235 of the N-terminal region of the CH 2 domain.

In a particularly preferred embodiment, the altered antibodies of the invention or for use according to the invention are directed against an MHC class II antigen characterised in that said antibody has been altered at position 235 of the N-terminal region of the CH 2 domain and the ability of said antibody to fix complement is altered as compared to unaltered antibody and said altered antibody binds to one or more cellular Fc receptors especially FcRIII and does not bind significantly to FcRI.

In a further aspect the invention provides a method for producing an altered antibody with altered ability to fix complement comprising altering one or more amino acids in the N-terminal region of the CH 2 domain of said antibody, altering the ability of said antibody to fix complement as compared with unaltered antibody.

As used herein the term 'altered antibody' is used to denote an antibody which differs from the wild type unaltered antibody at one or more amino acid residues in the N-terminal region of the Cj2 domain of the Fc region of the antibody. The alteration may for example comprise the substitution I _I WO 94/29351 I'CT/GB94/01290 9 or replacement of the starting wild type antibody amino acid by another amino acid, or the deletion of an amino acid residue.

The residue numbering used herein is according to the Eu index described in Kabat eLa/ [(1991) in: Sequences of Proteins of Immunological Interest, Edifion. United States Department of Health and Human Services.] In human IgG1 and IgG3 antibodies the naturally occurring amino acid at position 235 of the N-terminal region of the CH 2 domain is a leucine residue. The alterations at position 235 of replacing leucine by glutamic acid or alanine have been found particularly effective at producing a potent immuno-suppressive antibody with minimal toxicity in itro and which is tolerated in vivo.

The alteration at position 237 of replacing glycine by alanine has been found to produce an antibody with an intermediate ability to fix human complement, i.e. the complement fixation level is approximately 15-80%, preferably 20-60%, most preferably 20-40% of that seen with the starting wild type unaltered antibody.

The residue(s) could similarly be replaced using an analogous process to that described herein, by any other amino acid residue or amino acid derivative, having for example an inappropriate functionality on its side chain. This may be achieved by for example changing the charge and/or polarity of the side chain.

The altered antibodies of the invention may also be produced for example, by deleting residues such as 235, or by, for example, inserting a glycosylation site at a suitable position in the molecule. Such techniques are well known in the art, see for example the teaching of published European patent application EP-307434.

The altered antibodies of the invention may also be produced by exchanging lower hinge regions of antibodies of different isotypes. For example a G1/G2 lower hinge exchange abolished complement fixation and is a further preferred embodiment of the invention. This is described in rmM WO 94/29351 I('T/G I94/(1129(0 more detail in the accompanying examples. The G1/G2 lower hinge exchange results in an antibody with altered residues in the 231 to 238 region of the N-terminal region of the CH 2 domain wherein one or more residues may be altered and/or deleted.

In a particularly preferred embodiment of the invention the antibody is a human IgG1 antibody directed against an MHC class II antigen.

In a further aspect the invention provides a method of modulating the function of cell surface associated antigens avoiding complement mediated toxicity comprising administering an altered antibody wherein one or more amino acid residues in the N-terminal region of the CH 2 domain of said antibody are altered characterised in that the ability of said antibody to fix complement is altered as compared to unaltered antibody.

In a preferred embodiment of this aspect of the invention said altered antibody is able to bind to one or more cellular Fc receptors especially FcRIII while binding to FcRI is significantly reduced.

Examples of such cell surface antigens include for example adhesion molecules, T-cell receptor, CD4, CD8, CD3, CD28, CD69, MHC Class I, MHC Class II and The invention also includes therapeutic, pharmaceutical and diagnostic compositions comprising the altered antibodies according to the invention and the uses of these products and the compositions in therapy and diagnosis.

Thus in a further aspect the invention provides a therapeutic, pharmaceutical or diagnostic composition comprising an altered antibody according to the invention, in combination with a pharmaceutically acceptable excipient, diluent or carrier.

The invention also provides a process for the preparation of a therapeutic, pharmaceutical or diagnostic composition comprising admixing an altered WO 94/29351 W CT)/G 194/01290 11 antibody accurding to the invention together with a pharmaceutically acceptable excipient, diluent or carrier.

The antibodies and compositions may be for administration in any appropriate form and amount according to the therapy in which they are employed.

The altered antibodies for use in the therapeutic, diagnostic, or pharmaceutical compositions, pr for use in the method of treatment of diseases in which antibody therapy leads to undesirable toxicity due to antibody mediated complement fixation are preferably MHC specific antibodies most preferably specific for MHC Class II antigens, and most preferably have specificity for antigenic determinants dependent on the DRx chain.

The therapeutic, pharmaceutical or diagnostic composition may take any suitable form for administration, and, preferably is in a form suitable for parenteral administration e.g. by injection or infusion, for example by bolus injection or continuous infusion. Where the product is for injection or infusion, it may take the form of a suspension, solution or emulsion in an oily or aqueous vehicle and it may contain formulatory agents such as suspending, preservative, stabilising and/or dispersing agents.

Alternatively, the antibody or composition may be in dry form, for reconstitution before use with an appropriate sterile liquid.

If the antibody o, composition is suitable for parental administration the formulation may contain, in addition to the active ingredient, additives such as: starch e.g. potato, maize or wheat starch or cellulose or starch derivatives such as microcrystalline cellulose; silica; various sugars such as lactose; magnesium carbonate and/or calcium phosphate. It is desirable that, if the formulation is for parental administration it will be well tolerated by the patient's digestive system. To this end, it may be desirable to include in the formulation mucus formers and resins. It may also be desirable to improve tolerance by formulating the antibody or compositions in a capsule which is insoluble in the gastric juices. it may also be I IC-~~CI~*RII~ Inram~~l- 'I/G 1194/01290 12 preferable to include the antibody or composition in a controlled release formulation.

If the antibody or composition is suitable for rectal administration the formulation may contain a binding and/or lubricating agent, for example polymeric glycols, gelatins, cocoa-butter or other vegetable waxes or fats.

The invention also provides methods of therapy and diagnosis comprising administering an effective amount of an altered antibody according to the invention to a human or animal subject.

The antibodies and compositions may be for trnminisiration in any appropriate form and ,mount according to the therapy in which they are employed. The dose at which the antibody is administered depends on the nature of the condition to be treated and on whether the antibody is being, used prophylactically or to treat an existing condition. The dose will also be selected according to the age and conditions of the patient. A therapeutic dose of the antibodies according to the invention may be, for example, preferably between 0.1-25mg/kg body weight per single therapeutic dose and most preferably between 0.1-10mg/kg body weight per single therapeutic dose.

Immunological diseases which may be treated with the antibodies of the invention include for example joint disease such as ankylosing spondylitis, juvenile rheumatoid arthritis, rheumatoid arthritis; neurological disease such as multiple sclerosis; pancreatic disease such as diabetes, juvenile onset diabetes; gastrointestinal tract disease such as chronic active hepatitis, celiac disease, ulcerative colitis, Crohns disease, pernicious anaemia; skin diseases such as psoriasis; allergic diseases such as asthma and in transplantation related conditions such as graft versus host disease, and allograft rejection. Other diseases include those described in European Patent No. 68790.

The altered antibodies of the invention may also be useful in the treatment of infectious diseases e.g. viral or bacterial infections and in cancer immunotherapy.

-I I I IZRII~R~P PIR~LIIII-~i WO 94/29351 1()1/01290 13 As used herein the term 'antibody' is used to cover natural antibodies, chimeric antibodies and CDR-grafted or humanised antibodies. Chimeric antibodies are antibodies in which an antigen binding site comprising the complete variable domains of one antibody is linked to constant domains derived from another antibody. Methods for carrying out such chimerisation procedures are described in EP 120694 (Celltech Limited), EP 125023 (Genentech Inc and City of Hope), EP 171496 (Res. Dev.

Corp. Japan), EP 173494 (Stanford University) and WO 86/01533 (Celltech Ltd). CDR grafted or humanised antibodies are antibody molecules having an antigen binding site derived from an immunoglobulin from a non-human species and remaining immunoglobulin-derived parts of the molecule being derived from a human immunoglobulin. Procedures for generating CDRgrafted or humanised antibodies are described in WO 91/09967 (Celltech Ltd), WO 90/07861 (Protein Design Labs. Inc) and WO 92/11383 (Celltech, Ltd).

In further aspects the invention also includes DNA sequences coding for the altered antibodies according to the invention; cloning and expression vectors containing the DNA sequences, host cells transformed with the DNA sequences and processes for producing the altered antibodies according to the invention comprising expressing the DNA sequences in the transformed host cells.

According to a further aspect of the invention there is provided a process for producing an altered antibody of the invention which process comprises: a. producing in an expression vector an operon having a DNA sequence which encodes an antibody heavy or light chain.

b. producing in an expression vector an operon having a DNA sequence which encodes a complementary antibody light or heavy chain.

c. transfecting a host cell with both operons, and d. culturing the transfected cell line to produce the antibody molecule WO 94/29351 PICT/G194/01290 14 wherein at least one of the expression vectors contains a DNA sequence encoding an antibody heavy chain in which one or more amino acid residues in the N-terminal region of the CH 2 domain of said antibody has been altered from that in the corresponding unaltered antibody.

As will be readily apparent to one skilled in the art, the alteration in the Nterminal region of the CH 2 domain may be made using techniques such as site directed mutagenesis after the whole altered antibody has been expressed. To express unaltered antibody the DNA sequences should be expressed following the teaching described above for altered antibody.

The DNA sequences preferably encode a humanised antibody; a CDRgrafted heavy and/or light chain or a chimeric antibody.

The cell line may be transfected with two vectors, the first vector containing the operon encoding the light chain-derived polypeptide and the second vector containing the operon encoding the heavy chain derived polypeptide. Preferably the vectors are identical except in so far as the coding sequences and selectable markers are concerned so as to ensure as far as possible that each polypeptide chain is equally expressed.

Alternatively, a single vector may be used, the vector including a selectable marker and the operons encoding both light chain- and heavy chainderived polypeptides.

The general methods by which the vectors may be constructed, transfection methods and culture methods are well known pe.S.e. Such methods are shown, for instance, in Maniatis Btal Molecular Cloning, Cold Spring Harbor, New York 1989 and Primrose and Old, Principles of Gene Manipulation, Blackwell, Oxford, 1980.

The altered antibody according to the invention is preferably derived from the anti-MHC antibody L243, which has been deposited at the American Type Culture Collection, Rockville, Maryland USA under Accession number ATCC HB55, and is most preferably a chimeric or a CDR-grafted derivative II ILM r WO 94/29351 PCT/G 94/01290 thereof. L243 was previously described by Lampson and Levy [J.

Immunol. (1980) 125, 293].

The standard techniques of molecular biology may be used to prepare DNA sequences coding for the altered antibodies according to the invention. Desired DNA sequences may be synthesised completely or in part using oligonucleotide synthesis techniques. Site-directed mutagenesis and polymerase chain reaction (PCR) techniques may be used as appropriate. See for example "PCR Technology Principles and Applications for DNA Amplification" (1989), Ed. H. A. Erlich, Stockton Press, N.Y. London. For example, oligonucleotide directed synthesis as described by Jones e.i1 [Nature, 321, 522 (1986)] may be used. Also oligonucsotide directed mutagenesis may be used as described by Kramer ltal [Nucleic Acid Res. 12 9441 (1984)].

Any suitable host cell/vector system may be used for the expression of the DNA sequences coding for the altered antibody. Bacterial e.g. .Eli and other microbial systems may be used. Eucaryotic e.g. mammalian host cell oxpression systems may also he used such as for example COS cells and CHO cells [Bebbington, C R (1991) Methods 2. 1'.6-145], and myeloma or hybridoma cell lines [Bebbington, C R (1992) Bio/Technology .LQ 169-175].

Where the altered antibody is derived from L243 CHO based expression systems are preferably used.

Assays for determining FcRIII binding indirectly via ADCC assays and for determining complement fixation and Clq binding are well known in the art, and are described in detail in the following examples.

Immune function/immunosuppression by antibodies may be assayed using techniques well known in the art including for example: Mixed Lymphocyte Responses and T-cell antigen recall responses to Tetanus Toxoid. These assays are described in detail in the following examples.

WO 94/29351 PCT/G 194/01290 16 The inolntion is illustrated in the following non-limiting examples and with reference to the following figures in which: Figure 1 shows: Figure 2 shows: Figure 3 shows: Figure 4 shows: Figure 5 shows: Figure 6 shows: Figure 7 shows: Figure 8 shows: Figure 9 shows: Figure 10 shows: Figure 11 shows: a map of plasmid pMR15.1 a map of plasmid pMR14 the nucleotide sequence and predicted amino acid sequence of L243 heavy chain the nucleotide and amino acid sequences of clone 43, clone 183 clone 192 the nucleotide sequence and predicted amino acid sequence of L243 light chain a map of plasmid pGamma 1 a map of plasmid pGamma 2 the nucleotide sequence of hinge and CH2 region of human C-gamma 1 Antigen binding potency of L243 human isotype series G1 G4[L235E] G1 [L235A] G4 G1 [G237A] 9 100% G1 [K320A] FcRI binding of L243 isotype series G1 G4 G1 [G237A] G1 [L235A] G2 G1/G2-L-hinge G1 [L235E] G4 [L235E] G1[K320A] human complement fixation by L243 isotype series G1 G4 G1 [G 237A] G1 [L235A] G2 G1/G2 L-hinge G1 [L235E] G4 [L235E] G1 (K320A] binding of human Clq to L243 human isotype series G1 Cells alone G4 [L235E] Cells+ Clq G1 [L235E] G1 [K320A] G4 Figure 12 shows: L I WO 94/29351 r:I/;14019 11CIII/G115,14/01290 Figure 13 shows: Figure 14 shows: Figure 15 shows: Figure 16 shows: human complement fixation by 1243 isotype Gi 1~ G 1[1-235AJ -4k- G2 G4 [1235E] as- G1 [IL235E] G1 [K320A] guinea pig complement fixation by L243 isotype Gi -1 G1 [IL235A] -s-G2 4-G4 [L235E] -a-G1 [1-235E] 4- G1 [K320A] rabbit complement fixation by L243 isotype 4- Gi 1~ Gi [1-235A] 4-G2 G4[1-235E] G1 [L235E] [K320A] FoRIII binding of 1243 isotype series by ADC G 1i- G4 G1 (K320A] G1 [G237A] G2 -s-G1IG2 L-hinge -a-G1[1-235A] G4 [1235E] G1 [L235E] L243 Isotype Series Inhibition of TT recall response -4G2 G4- Scyclosporin S medium control L243 Isotype Series Inhibition of TT recall response wa hGl -1 G1/G2 L-hinge hGl [1235E] G4 [L235E] medium control cyclosporin 1243 Isotype Series Inhibition of Mixed Lymphocyte Reaction.

a- hGl1 G1/G2 L-hinge 4- hGI11235E] G4[1-235E] Figure 17 shows: Figure 18 shows: Figure 19 shows: WO 94/29351 IPCT/G 1194/01290 Figure 20 shows: Figure 21 shows: Figure 22 shows: Figure 23 shows: cyclosporin A medium control L243 Isotype Series inhibition of TT response G1 G1 [L235A] G1 [G237A] O Cyclosporin A Medium control L243 Isotype Series Inhibition of Mixed Lymphocyte Reaction G1 [L235E] G1 [L235A] G1 0/ Cyclosporin A Medium control the nucleotide and amino acid sequence of VI region in L243-gL1 shows the nucleotide and amino acid sequence of VI region of L243-gL2 the nucleotide and amino acid sequence of Vh region of L243-gH a graph of the results of a competition assay for L243 grafts vs FITC-chimeric L243 cH cL cH gL1 K- gHcL gHgL1 a graph of a Scatchard analysis for L243 gamma 4 cHcL Kd 4.1nM gH gL1 Kd 6.4nM gHgL2 Kd 9.6nM a graph of FcRIII binding of chimeric, grafted and grafted [L235E] L243 as measured by ADCC Chimeric G1 wt Chimeric G1 [L235E] Figure 24 shows: Figure 25 shows: Figure 26 shows: Figure 27 shows: WO 94/29351I CI 1/029 1ICT/G1194/01290 Figure 28 shows: Figure 29 shows: GraftGl1 wt Graft G1 (1-235E] a graph of immunosuppressive activity of CDR grafted I1243 measured by MLR I" Graft G1lwt Graft G1 [1-235E] SCyclosporin Chimeric GI wt -s-Chimeric G1 [L235E] Z~Medium Control a graph o~f CDR grafted L243 and grafted [1235E] 1243 TT recall response -s-Graft Gl1wt -I-GraftiG1 [L235E] SCyclosporin -~-Chimeric G1 wt -e-Chimeric G1 [1235E] I. Medium Control a graph of complement mediated cytotoxic potency of CDR grafted 1243 and CDR grafted [L235E] 1243 Chimeric G1 wt Chirnerc Gi [1-235E] Graft Gl1wt Graft G1 (L235E] Figure 30 shows: DETAILED 12ESCRIPTJONOE SPECIFIC EMBODIMENTS, EXA MPESI EXUD21 Gene Clonin and Expression RNA preparation from 1243 hybridomfia-g11k Total RNA was prepared from 3 x l0exp7 1243 hybridoma cells as described below. Cells were washed in physiological saline and dissolved in RNAzoI (0.2ml per l0exp6 cells). Chloroform (Q.2ml per 2ml

I

L WO 94/29351 I'C(/Gi; 1194/01290 homogenate) was added, the mixture shaken vigorously for 15 seconds and then left on ice for 15 minutes. The resulting aqueous and organic phases were separated by centrifugatlon for 15 minutes In an Eppendorf centrifuge and RNA precipitated from the aqueous phase by the addition of an equal volume of isopropanol, After 15 minutes on ice, the RNA was pelleted by centrifugation, washed with 70% ethanol, dried and dissolved in sterile, RNAase free water. The yield of RNA was 350 j.g.

Amino acid sequence of the L243 lighbtchain.

The sequence of the first nine amino acids of the mature L243 light chain was determined to be NH2-DIQMTQSPAS.

PCR cloning of L243 Vh and VI The cDNA genes for the variable regions of L243 heavy and light chains were synthesised using reverse transcriptase to produce single stranded cDNA copies of the mRNA present in the total RNA, followed by Polymerase Chain Reaction (PCR) on the cDNAs with specific oligonucleotide primers.

a) cDNA synthesis cDNA was synthesised in a 20pl reaction containing the following reagents: n.mM Tris-HCI PH8.3, 75mM KCI, 10mM dithiothreitol, 3mM MgCl2, 0.5mM each deoxyribonucleoside triphosphates, units RNAsin, 75ng random hexanucleotide primer, 2p.g L243 RNA and 200 units Moloney Murine Leukemia Virus reverse transcriptase. After incubation at 42 0 C for 60 min the reaction was terminated by heating at 950C for 5 minutes.

b) PCR.

Aliquots of the cDNA were suYtlcted to PCR using combinations of primers for the heavy and ligirt chains. The nucleotide sequences of the 5' primers for the hesvy and light chains are shown in Tables 1 and 2 respectively. These sequences, all of which contain a restriction site starting 6 nucleotides from their 5' ends, followed by the sequence GCCGCCACC to allow optimal translation of the resulting mRNAs, an initiator codon and a further 20 WO 94/29351 IN'( i191/ 01290 21 nucleotlde, are a compilation based on the loader poptlde sequences of known mouse antibodies [Kabat olLaI (1991) In Sequences of Proteins of Immunologlcal Interest, 5th Edition United States Department of Health and Human Services].

The 3' primers are shown in Table 3. The light chain primer spans the V C junction of the antibody and contains a restriction site for the enzyme Spll to facilitate cloning of the VI PCR fragment. The heavy chain 3' primers are a mixture designed to span the J C junction nf the antibody. The first 23 nucleotides are identical to those found at the start of human C gamma 1, 2, 3 and 4 genes and include the Apal restriction site common to these human isotypes. The 3' region of the primers contain a mixed sequence based on those found in known mouse antibodies [Kabat E A, Wu, Perry H M, Gottesman K S, and Foeller L; In: Sequences of Proteins of Immunological Interest, 5th Edition, US Department of Health and Human Services (1991)].

The combinations of primers described above enables the PCR products for Vh and VI to be cloned directly into the appropriate expression vector (see below) to produce chimeric (mouse human) heavy and light chains and for these genes to be expressed in mammalian cells to produce chimeric antibodies of the desired isotype.

Incubations (20 Il) for the PCR were set up as follows. Each reaction contained 10 mM Tris-HCI pH 8.3, 1.5 mM MgCl2, 50 mM KCI, 0.01% w/v gelatin, 0.25 mM each deoxyribonucleoside triphosphate, 1 6 pmoles 5' primer mix (Table 6 pmoles 3' primer, 1 Il cDNA and 0.25 units Taq polymerase. Reactions were incubated at 950C for 5 minutes and then cycled through 940C for 1 minute, 550C for 1 minute and 720C for 1 minute. After 30 cycles, aliquots of each reaction were analysed by electrophoresis on an agarose gel. Reactions containing 5' primer mixes B1, B2, B3 and B5 produced bands with sizes consistent with full length VI fragments while reaction B9 produced a fragment with a size WO 94/2935 'I'"/(;lIl94/0 1290 22 oxpootod of a Vh goel, Tho band produced by the 13 prlmoro wat not followed up ao provioun ronull iriad shown thfil thli brind oorregopondo to a light chain preudogono produod by the hybridomia coll, a) MI u.la~u.~lalauLg.aLt BJ.r.mnea uia DNA fragments produced in reactions B2, B3 and B5 were digested with the oni,/mos BotB1 and 8pl1, conoontratod by othnnol preolpitation, eloctrophorosod on a 1.4 agarose gel and DNA bands In the range of 40G base pairs recovered. These were cloned by Ilgatlon into the vector pMR15.1 (Figure 1) that had been restricted with BstB1 and Spll. After ligatlon, mixtures were transformed Into E. coll LM1035 and plasmlds from the resulting bacterial colonies screened for inserts by digestion with BstB1 and Spl1. Representatives with inserts from each ligation were analysed further by nucleotide sequencing.

In a similar manner, the DNA fragments produced in reaction B9 and digested with Hindlll and Apal were cloned into the vector pMR14 (Figure 2) that had been restricted with Hindlll and Apal.

Again, representative plasmids containing inserts were analysed by nucleotide sequencing.

d) Nucleotide sequence analysis Plasmid DNA (pE1701 and pE1702) from two isolates containing Vh inserts from reaction B9 was sequenced using the primers R1053 (which primes in the 3' region of the HCMV promoter in pMR14) and R720 (which primes in the 5' region of human C gamma 4 and allows sequencing through the DNA insert on pMR14). The determined nucleotide sequence and predicted amino acid sequence of L243 Vh in pE1702 is given in Figure 3. The nucleotide sequence for the Vh insert in pE1701 was found to be identical to that in pE1702 except at nucleotide 20 (A in pE1701) and nucleotide 426 (A in pE1701). These two differences are in the signal peptide and J regions of Vh respectively and indicate that the two clones WO 94/29351 P(T"I G 94/(01290 23 examined are Independent isolates arising from the use of different primers from the mixture of oligonucleotides during the PCR stage.

To analyse the light chain clones, sequence derived from priming with R1053 was examined. The nucleotlde sequence and predicted amino acid sequence of the VI genes arising from reactions B2 (clone 183), B3 (clone 43 and B5 (clone 192) are shown in Figure 4. Comparison of the predicted protein sequences shows the following: i) clones 182, 183, 43 and 45 all code for a VI gene which, when the signal peptide is removed, produces a light chain whose sequence is identical to that determined by amino acid sequence analysis for L243 light chain (see above).

ii) clones 182 and 183 contain a VI gene that codes for a signal peptide of 20 amino acids, while the VI gene in clones 43 and results from priming with a difarent set of oligonucleotides and has a leader sequence of only 15 amino acids.

ili) Clone 192 does not code for L243 VI. Instead, examination of the database of antibody sequences [Kabat, 1991] indicates that clone 192 contains the VI gene for MOPC21, a light chain synthesised by the NS1 myeloma fusion partner used in the production of the L243 hybridoma.

iv) Clones 182 and 183 are identical except at nucleotide 26 (T in clone 182, C in clone 183). This difference can be accounted for by the use of different primers in the PCR and indicates that clones 182 and 183 are independent isolates of the same gene.

The nucleotide sequence and predicted amino acid sequence of the complete VI gene from clone 183 is shown in Figure Construction of human gamma 1 and gamma 2 isotypes.

The L243 Vh gene was subcloned on a Hindll Apal fragment into pGamma 1 and pGamma 2, vectors containing the human C gamma 1 and C gamma 2 genes respectively (Figures 6 and 7).

I-

WO 94/29351 P"I'I; 1194/01290 24 th~lman Isal.e mutants PCR mutagenesis was used to change residue 235 in human C gammal contained in the vector pGamma 1 from leucine to either glutamic acid or to alanine and to change residue 237 from glycine to alanine. The lower hinge region of human C-gamma 1 was also replaced by the corresponding region of human C-gamma 2. The following oligonucleotides were used to effect these changes: i) L235E change R4911 5' GCACCTGAACTCGAGGGGGGACCGTCAGTC3' R4910 5'CCCCCCTCGAGTTCAGGTGCTGAGGAAG3' II) L235A change R5081 5'GCACCTGAACTCGCAGGGGGACCGTCAGTC3' R5082 5'GACTGACGGTCCCCCTGCGAGTTCAGGTGC3' lls) G237A change R5088 5'GCACCTGAACTCCTGGGTGCACCGTCAGTC3' R5087 5'GACTGACGGTGCACCCAGGAGTTCAGGTGC3' IV) Exchange of lower hinge regions R4909 5'GCACCTCCAGTGGCAGGACCGTCAGTCTTCCTC3' R4908 5'CGGTCCTGCCACTGGAGGTGCTGAGGAAGAG3' Other oligonucleotides used in the PCR mutagenesis are: R4732 5'CAGCTCGGACACCTTCTCTCCTCC3' R4912 5'CCACCACCACGCATGTGACC3' R4732 and R4912 prime between nucleotides 834 and 858 and between nucleotides 1156 and 1137 respectively in human C gamma 1 (Figure 8).

The general strategy for the PCR mutagenesis was as follows. For each amino acid change, two rounds of PCR were used to generate' DNA fragments containing the required substitutions. These fragments were WO 94/29351 PICT/GB94/01290 then restricted with the enzymes Bgl II and Sty1 and used to replace the corresponding fragments containing the wild type sequence in the pGamma 1 vector, (Figure 6).

For the first round PCR, reactions (20 pLI) were prepared containing the following reagents 10 mM Tris HCI pH 8.3, 1.5 mM MgCl2, 50 mM KCI, 0.01% gelatin, 0.25 mM each deoxyribonucleoside triphosphate, 50 ng pGamma 1 DNA, 0.4 unit Taq polymerase and 6 pmoles of each of the primer. The following combinations of primers were used: R4911 /R4912, R4910 R4732, R5081 /R4912, R5082 R4732, R5088 R4912, R5087 R4732, R4909 R4912, R4908 R4732.

After 30 cycles through 940C for 1 minute, 5500 for 1 minute and 720C for 1 minute, the reactions were extracted with chloroform, the newly synthesised DNA precipitated with ethanol, dissolved in water and electrophorersed on a 1.4 agarose gel. Gel slices containing the DNA fragments were excised from the gel, the DNA recovered from the agarose using a "Mermaid" kit (from Stratech Scientific Ltd., Luton, England) and eluted into 20pj. sterile water.

Second round PCR was in a 100 Lp. reaction containing 10 mM Tris HCI pH 8.3, 1.5 mM MgCI2, 50 mM KCI, 0.01 gelatin, 0.25 mM each deoxyribonucleoside triphosphate, 2 units Taq polymerase, 1/20 of each pair of DNA fragments from the first round reaction and 30 pmoles of each of R4732 and R4912. After 30 cycles, see above, the reactions were extracted with phenol chloroform and precipitated with ethanol.

Fragments were digested with Bgll 1 and Styl, electrophoresed on a 1.4 agarose gel and DNA bands of 250 base-pairs recovered from gel slices as previously described.

II w i.

I 111111111~1" WO 94/29351 PCT/GB94/01290 26 These Bgl II Sty1 fragments were ligated in a 3 way ligation to the 830 base-pair Sty1 EcoR1 fragment, containing the C terminal part of the

CH

2 domain and the entire CH 3 domain of human C gamma 1, and the Bglll EcoR1 vector fragment from pGammal (see Figure After transformation into LM1035, plasmid minipreps from resulting colonies were screened for the presence of the Bgl II Sty1 fragment and representatives of each taken for nucleotide sequence analysis. From this, plasmids containing the desired sequence were identified and, for future reference, named as follows: pGamma1[L235E] containing glutamic acid at residue 235, pGarnma1[L235A] containing alanine at residue 235, pGammal [G237A] containing alanine at residue 237, pGammal [gl-,g2] containing the C-gamma 2 lower hinge region.

The above plasmids were each restricted with Hindi 11 and Apal and the Hindi 11 Apal fragment containing L243 Vh inserted to produce the following plasmids: L243Gammal[L235E] L243Gammal[L235A] L243Gammal[G237A] L243Gamma [g1->g2] a) Production of chimeric L243 antibody Antibody for biological evaluation was produced by transient expression of the appropriate heavy and light chain pairs after co-transfection into Chinese Hamster Ovary (CHO) cells using calcium phosphate precipitation.

On the day prior to transfection, semi confluent flasks of CHO-L761 cells were trypsinised, the cells counted and T75 flasks set up each with 10exp7 cells.

On the next day, the culture medium was changed 3 hours before transfection. For transfection, the calcium phosphate precipitate was C nrr-~irn WO 94/29351 WO 9/2931 'cc;1CA 94/01 29() 27 prepared by mixing 1.25 ml of 0.25M CaCI2 containing 50 .tg of each of heavy and light chain expression vectors with 1.25 ml of 2xHBS (16.36 gm NaCI, 11.9 gm HEPES and 0.4 gm Na2HPO4 in 1 litre water with the pH adjusted to 7.1 with NaOH) and adding immediately into the medium on the cells. ,After 3 hours at 37 C in a C02 incubator, the mediuri and precipitate were removed and the cells shocked by the addition of 15 ml glycerol in phosphate buffered saline (PBS) for 1 minute. The glycerol was removed, the cells washed once with PBS and incubated for 48 96 hours in 25 ml medium containing 10 mM sodium butyrate. Antibody was purified from the culture medium by binding to and elution from protein A Sepharose anld quantitated using an Ig ELISA (see below).

b) ELISA For the ELISA, Nunc ELISA plates were coated overnight at 40C with a F(ab)2 fragment of a polyclonal goat anti-human Fc fragment specific antibody (Jackson lmmuno-research, code 109-006-098) at 5 g/ml in coating buffer (15mM sodium carbonate, 35mM sodium hydrogen carbonate, pH6.9). Uncoated antibody was removed by washing 5 times with distilled water. Samples and purified standards to be quantitated were diluted to approximately 1 glg/ml in conjugate buffer (0.1M Tris-HCI 0.1M NaCI, 0.2% v/v Tween 20, 0,2% w/v Hammersten casein). The samples were titrated in the microtitre wells in 2-fold dilutions to give a final volume of 0.1 ml in each well and the plates incubated at room temperature for 1 hr with shaking. After the first incubation step the plates were washed 10 times with distilled water and then incubated for 1 hr as before with 0.1 ml of a mouse monoclonal anti-human kappa (clone GD12) peroxidase conjugated antibody (The Binding Site, code MP135) at a dilution of 1 in 700 in conjugate buffer. The plate was washed again and substrate solution (0.1 ml) added to each well. Substrate solution contained 150 kl N,N,N,N-tetramethylbenzidine (10 mg/ml in DMSO), 150 ILl hydrogen peroxide (30% solution) in 10 ml 0.1M sodium acetate/sodium citrate, pH6.0. The plate was developed for 5 -10 minutes until the absorbance at 630nm was approximately 1.0 for the top standard.

Absorbance at 630nm was measured using a plate reader and the concentration of the sample determined by comparing the titration curves with those of the standard.

I IIP WO 94129351 WO 94/9351 C'I'/G 8194/01290 28

IAADLE-

Oligonucteotide primers for the 5' region of mouse heavy chains.

OH 1: 5'ATGAAATGCAGCTGGGTCAT(G,C)TTCTT3' 0H2: 5'ATGGGATGGAGCT(A,G)TATCAT(C,G)(C,T)TCTT3' 0H3: 5'ATGAAG(A,T)TGTGGTTAAACTGGGTTTT3' CF14: 5'ATG(G,A)ACTTTGGG(T,C)TCAGCTTG(G,A)T3' 5'ATGGACTCCAGGCTCAA11TAGTITTIT3' 0H6: 5'ATGGCTGTC(C,T)T(G,A)G(G,O)GCT(G,A)CTCTTG3' 0H7: 5'ATGG(G,A)ATGGAGC(G,T)GG(G,A)TC1TT(A,C)TCTT3' 0H8: 5'ATGAGAGTGCTGATTCTTTTGTG3' 0H9: Ii'ATG(C,A)TTGGGTGTGGA(A,C)(,TTGCTATT3' OH 10: 5'ATGGGCAGACTTACATTOTOATTCCT3' CHi 1: 5'ATGGATTTTGGGCTGATFFITTATTG3' OH 12: 5'ATGATGGTG11"AAGTCTTGTACCT3' Each of the above primers has the qsequence 5'GCGCGCAAiGCTTGCCGCCACO3' added to its 5' end.

WO 94/29351 PrG34O 9 IIC'I*/GB9,1/01,290 29 IAM-E-2 Otiaonucleotide primers for the 5' reion of mouse ight chains CLi: 5'ATGAAGTTGCCTGTTAGGCTGTTGGTGCT3' CL2: 5'ATGGAG(T,A)CAGACACACTCCTG(T,C)TATGGGT3' CL3: 5'ATGAGTGTGCTCACTCAGGTCCT3' CL4: 5'ATGAGG(G,A)CCCCTGCTCAG(A,lTT(C,T)TTGG3' 5'ATGGAT1T(T,A)CAGGTGCAGATT(TA)TCAGCTT3' 0L6: 5'ATGAGGT(T,G)C(T,C)(T,C)TG (T,C)T(G,C)AG(T,C)T(T, C)CTG (A,G)G3' CL7: 5'ATGGGC(T,A)TCAAGATGGAGTCACA3' OLS: 5'ATGTGGGGA(T,C)CT(G,T)1TT(T,C)C(A,C)(A,C)T1TTCA AT3' 0L9: 5'ATGGT(G,A)TCC(T,A)CA(G,C)CTCAGTTCCTT3' 5'ATGTATATATGTTGTTGTCTA1C3' CLi 1 5'ATGGAAGCCCCAGCTCAGCTTCTCTT3' Each of the above primers has the sequence 5'GGACTGTTCGAAGCCGCCACC3' added to its 5' end.

M

WO 94129351 WO 94/9351C1/CB194/()1290 -Oligonucleotide p2rimers for the 2' ends of-moue Vh and VI cgenes.

Light chain CLI 2) 5'GGATACAGTTGGTGCAGCATCCGTACGTTT3' Heavy chain R2155 5'GCAGATGGGCCCTTCGTTGAGGCTG(A,C)(A,G)GAGAC(G,T,A)GTGA3' TALE~i 4 Primer Mixtures for PCL4 Bi :CL2.

B2 0L6.

B3 CL8.

B4 CL4, CL9.

CLi, CL3, 0L5,CL7, CLi0, CLi1.

B6 CH6.

B7 CH7.

B8 CH2, CH4.

B9 :CHi, CH3, CH5, CH8, CH9, CH10, CHil, CH12.

Il r- -i WO 94/29351 PCT/GB94/01290 31 Biological prooerties of engineered L243 The aim of the following experiments was to separate the immunosuppressive effects of anti-MHC-ll antibodies from possible toxic consequences of their use. In the process we hope to demonstrate which Fc effector functions are necessary for immunosuppression.

ANTIGEN BINDING POTENCY BY INHIBITION ASSAY The principle of this experiment is that antibodies that have the same binding will compete equally well with a labelled antibody for their cognate antigen. Any changes in the antigen binding potency of the engineered L243 antibodies will be revealed in this system.

Murine L243 (lgG2a) was labelled with fluorescein (FITC) using standard, techniques. All dilutions, manipulations and incubations were done in Phosphate Buffered Saline (Gibco UK) containing 0.1% Sodium Azide (iPH UK) and 5% Foetal Calf Serum (Sigma UK). Serial dilutions of engineered antibodies in 100p.l in RB polystyrene tubes (2052 12x75mm Falcon UK) were premixed with a constant amount in 100p.l (at a previously determined optimal concentration) of the labelled antibody on 5x10 4 indicator cells (JY B lymphoblastoid cell line bearing high levels of HLA- DR). Cells and antibody were incubated together at 4C for washed twice and binding revealed using a Fluorescence Activated Cell Scanner (FACS Becton Dickinson). After appropria.e analysis, median fluorescence intensity is plotted against antibody concentration.

Reaulft As expected, none of the changes in the Fc portion of the molecule affected antigen binding ability (Figure 9).

ASSESSMENT OF FCy RI BINDING.

The ability of the engineered variants of L243 to bind to FcgRI was measured. The principle of this experiment is that antibodies will bind to cells through Fc receptors and the affinity of this interaction is determined by the subclass and hence the structure of the Fc of the antibody. The I I -I C*L1--rr i WO 94/29351 PCT/G 94/01290 32 assay is based on the ability of the engineered antibodies to compete for binding with FITC labelled murine IgG2a to IFNy stimulated U937 cells.

U937 (myelomonocytic) cells, when incubated with 500p./ml IFNy (Genzyme UK) for 24 hours, expresses high levels of FcgRI, as assessed by CD64 binding and monomeric IgG2a binding, low levels of FcyRII and no FcyRIII.

U937 cells are washed extensively in DMEM containing 25mM HEPES (Gibco UK), incubated for 2 hours at 370C in RPMI 1640 (Gibco UK) and then washed again in DMEM containing 25mM HEPES (Gibco UK) to remove bovine IgG bound to Fc receptors. Serial dilutions of engineered antibodies were prepared in 50pl in Phosphate Buffered Saline (Gibco UK) containing 0.1% sodium azide in V-bottom 96 well microtitre plates' (ICN/Flow UK) and were incubated with 5x104 U937 cells in 50pI for at 4 0 C. 501l of FITC labelled IgG2a antibody was then added to all wells, at a previously determined optimal concentration, for a further 90min at Cells were washed once in the microtttre tray, transferred to RB polystyrene tubes (2052 12x75mm Falcon UK) washed once again and binding was revealed using a Fluorescence Activated Cell Scanner (FACS Becton Dickinson). After appropriate analysis, median fluorescence intensity is plotted against antibody concentration.

Changes in the N-terminal region of the CH2 domain of IgG1 had profound effects on binding to FcRI (Figure 10). As expected, wild type IgG1 bound well to FcRI, lgG4 about 10 times less well and lgG2 did not bind at all.

We have confirmed that the Leu 235 to Glu change in human lgG4 reduced its low FcRI binding to nothing and that the same change in IgG1 completely abolishes FcRI binding. Ala at 235 reduced (by about 100 fold) but did not ablate FcRI binding. Changing Gly "237 to Ala of IgG1 also abolished FcRI as did exchanging the whole region 233 to 236, with the sequence found in human lgG2. The G1[K320A] change had no effect on FcRI binding.

Il I I C WO 94/29351 I'CT/G 94/01290 33 ANTIBODY DEPENDENT COMPLEMENT MEDIATED CYTOTOXIC[TY.

The ability oi the engineered variants of L243 to fix human complement was assessed using the technique of antibody dependent complement mediated cytotoxicity.

The principle of the experiment is that antibodies will mediate complement lysis of target cells bearing their cognate antigen if the Fc of the antibody is able to interact with the components of the (usually classical) complement cascade. The critical interaction is with the Clq molecule.

The source of complement in these experiments is human venous blood freshly drawn into endotoxin free glass bottles which is then allowed to clot' at 370C for 1 hour. The clot is detached from the glass and then incubated at 400 for 2 hours to allow it to retract. The clot is then removed and the serum separated from the remaining red cells by centrifugation at 1000g.

Once prepared, the serum can be stored for up to one month at -200C without noticeable deterioration of potency but is best used fresh.

All manipulations, dilutions and incubations are done in RPMI 1640 medium (Gibco UK) containing 2mM Glutamine (Gibco UK) and 10% foetal calf serum (Sigma UK). Target cells (JY B lymphoblastoid cell line bearing high levels of HLA-DR) are labelled with 1mCi Na 51 Cr for 1 hour at room temperature, agitated every 15 min. The cells are then washed three times, to remove free radiolabel, and resuspended at 2x10 6 /ml. Serial antibody dilutions are prepared in duplicate in V-bottom 96 well microtitre plates (ICN/Flow UK) in 25gl. Control wells containing medium only are also prepared to establish the spontaneous release of label giving the assay background. Target 5 1 Cr labelled JY cells are added to all wellq in The same number of JY cells are also added to wells containing 2% Triton x100 in water to establish the 100% release value. Target cells and antibody are incubated together and, after 1 hour at room temperature, serum as a source of complement is added to all wells (except the 100%) for a further 1 hour at room temperature. 100l of EDTA saline at is then added to stop any further cell killing, the microtitre plates are IL II i~ WO 94/29351 PCT/G 1394/0 1290 34 centrifuged at 200g to pellet the Intact cells and 100pl supernatant Is removed and counted in a gamma counter.

Percent cell lysis is calculated by subtracting the background from all values and then expressing them as a percentage of the adjusted maximum release. Replicates vary by less than Percent cell lysis is then plotted against antibody dilution.

Results The ability of L243 to fix human complement was not affected by all the changes made in the N-terminal region of the CH 2 domain, residues 233 to 237 (Figure 11). Wild type IgG1 mediated potent killing with 600ng/ml giving half maximum cell killing (64% maximum). IgG2 and lgG4 caused no cell killing even at 20pg/mi. The Gly to Ala at 237 gave an intermediate.

level killing (20% maximum killing at 2p.g/ml). Exchanging the whole lower hinge region with the sequence found in human IgG2 failed to cause lysis even at 20p.g/ml. Changes at 235 in IgG1 had unexpectedly profound effects on human complement fixation. Changing the Leu 235 to Glu abolished complement lysis (no killing at 20p.g/ml). Ala at 235 permitted low levels of killing. In contrast, a change in the previously described Clq binding motif [Duncan A R and Winter G (1998), Nature, 332 from Lys to Ala at 320 effected no change from the IgG1 Aild type killing maximum cell killing and half the cells dead with 600ng/ml).

DIRECT BINDING OF Ciq Measurement of the direct binding of human Clq to different engineered variants of L243 was established to confirm that complement mediated cytotoxicity was due to activation of the classical pathway.

Purified human Clq (Sigma UK) was directly labelled with fluorescein isothiocyanate (FITC Sigma) using conventional methods. All dilutions, manipulations and incubations were done in Phosphate Buffered Saline (Gibco UK) containing 0.1% Sodium Azide (BDH UK) and 5% Foetal Calf Serum (Sigma UK). 5x10 4 indicator cells (JY B lymphoblastoid cell line bearing high levels of HLA-DR) were coated with the different engineered antibodies by incubating at saturating concentrations for 1 hour at 4°C in WO 94/29351 I'CT/GB94/01290 RB polystyrene tubes (2052 12x75mm Falcon UK). After washing, serial dilutions of FITC labelled Clq in 100pl were added and were incubated together for a further 30 min at 400. After washing, binding of Clq was revealed using a Fluorescence Activated Cell Scanner (FACS Becton Dickinson). After appropriate analysis, median fluorescence intensity is plotted against Clq concentration.

Results Direct binding of human Clq to the L243 human isotype series confirmed the results with complement mediated cytotoxicity (Figure 12). Labelled human Clq bound well to wild type IgG1, when bound to JY cells, and bound poorly to lgG4. Equilibrium dissociation constants were determined essentially as described by Krause etal [Behring Inst. Mitt. BZ 56 (1990)] and were 1.2 x10-7M and 1.5 xlO0 8 M for lgG4 and IgG1 respectively.' These values compare favourably with those obtained for the mouse antibodies IgG1 and lgG2a which have similar functions [Leatherbarrow and Dwek (1984), Molec. Immunol. 2L. 321]. The Leu 235 to Glu change in IgG1 reduced the binding of Clq to the same level as lgG4. In contrast, a change in the previously described Clq binding motif [Duncan A R and Winter G (1988) Nature 332 21], from Lys to Ala at 320 had no effect on Clq binding. The Leu 235 to Glu change in igG4 did not alter wild type binding.

Rabbit and Guinea Pig complement The G1[L235E] and G1[L235A] modifications behaved differently when rabbit or guinea pig serum was used as a source of complement instead of human. With rabbit C' they caused the same level of lysis as the wild type G1. With guinea pig they caused 40% and 49% plateau level killing, respectively, compared with 80% killing by the IgG1 wild type. The 235 change only affects human complement binding indicating that rabbit and guinea pig complement interact differently with human IgG1 (see Figures 13-15).

ANTIBODY DEPENDENT CELL MEDIATED CYTOTOXICITY.

The ability of the engineered variants of L243 to bind to FcgRIII was assessed using antibody dependent cell mediated cytotoxicity (ADCC).

I M WO 94/29351 PCT/G; 119.1/01290 36 The principle of the experiment Is that antibodies will mediate lysis of target cells bearing their cognate antigen if the Fc of the antibody is able to Interact with Fc receptor bearing effector cells capable of cytotoxicity. The critical interaction is between antibody Fc and cellular Fc receptors.

Effector cells are prepared fresh for each experiment. Human venous blood is drawn into endotoxin free tubes containing heparin. Peripheral blood mononuclear cells (PBMC) are prepared by density gradient centrifugation according to the manufacturers instructions (Pharmacia).

PBMC are adjusted to 1x10 7 cells/ml in RPMI 1640 medium (Gibco UK) containing 2mM Glutamine (Gibco UK) and 10% foetal calf serum (Sigma UK), in which all manipulations, dilutions and incubations are done.

lb Target cells (JY B lymphoblastoid cell line bearing high levels of HLA-DR) are labelled with 1mCi Na 5 1C r for 1 hour at room temperature, agitated every 15 min. The cells are then washed three times, to remove free radiolabel, and resuspended at 2x10 6 /ml. Serial antibody dilutions are prepared in duplicate in sterile U-bottom 96 well microtitre plates (Falcon UK) in 25p.l. Control wells containing medium only are also prepared to establish the spontaneous release of label giving the assay background.

Target 5 1 Cr labelled JY cells are added to all wells in 101±l. The same number of JY cells are also added to wells containing 2% Triton xl00 in water to establish the 100% release value. Target cells and antibody are incubated together and, after 30min at room temperature, 2511 effector cells are added to all wells (except the 100%) for a further 4 hours at 37oC.

100pl of EDTA saline at 4oC is then added to stop any further cell killing, the microtitre plates are centrifuged at 200g to pellet the intact cells and 100p.l supernatant is removed and counted in a gamma counter.

Percent cell lysis is calculated by subtracting the background from all values and then expressing them as a percentage of the adjusted maximum release. Replicates vary by less than Percent cell lysis is then plotted against antibody dilution.

WO 94/29351 'C ll91/01290 37 Bseault Not all the changes made In the N-terminal region of the CH2 domain, residues 233 to 237, affected FcRIII mediated function (Figure 16 and Tables 5 and L243 IgG2 was unable to mediate peripheral blood mononuoloar ceil oytotoxlolty (ADCC) of HLA-DR positive JY lymphoblastold cell at concentrationo up to 100y/ml, IgG4 caused a low lovol of ADOC (20% maximum killing at ly/ml) which could be abrogated by the Lou 235 to Glu ohango. Wild typo IgG1 was a potont modlator of coil killing giving 50% cell death at 6ng/ml antibody. Gly to Ala at 237 reduced the IgG1 wild type killing to the level seen with IgG4. Exchanging the w'hole lower hinge region with the sequence found In human igG2 gave Intermediate levels of killing with 6Ong/ml needed for 50% cell death. In contrast, changes at 235 In IgG1 had minimal effect on ADCC.

Changing the Leu 235 to Ala gave levels of killing comparable with the G1 wild type (9ng/ml for 50% cell death)) and changing the Leu 235 to Glu reduced ADCC a little (40ng/ml for 50% cell death). A change In the previously described C1q binding motif, from Lys to Ala at 320 had no effect on the ability of IgG1 to mediate ADCC.

ExVyIQ T cell function experiments were performed where an Interaction between MHC-II and the T cell receptor was an obligatory requirement for T cell activation. The 1243 Isotype series was tested In mixed lymphocyte reactions, which measures both naive and memory T cell activation, and recall responses to tetanus toxoid which only measures a memory T cell response.

jMixe LB ryhocte Reaction.

The immunosuppressive potency of engineered variants of L243 was assessed using a mixed lymphocyte reaction.

The principle of the experiment is that when leucocytes from one individual are mixed with those of another which express different HLA alieles, they will recognise each other as foreign and will become activated. This activation is dependent, primarily, on interactions between the CD3/TcR I I b I WO 94/29351 IICTIO, 11911/0O12900 complex on T cells and the MHC-ll molecule on antigen presenting cells.

Antibodies that bind to MHC-il are known to inhibit this reaction.

Leucocytes are prepared fresh for each experiment. Human venous blood from two individuals is drawn into endotoxin free tubes containing heparin.

Periphbral blood mononuclear cells (PBMC) are prepared by density gradient centrifugation according to the manufacturers instructions (Pharmacia). PBMC are adjusted to 2x106 cells/ml in RPMI 1640 medium (Glbco UK) containing 2mM Glutamine (Gibco UK), 100/ml/100lg/ml Penicillin/ Streptomycin (Gibco) and 10% foetal calf serum (Sigma UK), in which all manipulations, dilutions and incubations are done. PBMC from one individual are irradiated with 3000 rads. These cells will be stimulate a response from the other individual.

Serial antibody dilutions are prepared in triplicate in sterile U-bottom 96 well microtitre plates (Falcon UK) in 100,l. Control wells containing medium only and optimal Cyclosporin (Sandimmun®, Sandoz) levels (100nM) are also prepared to establish the maximum response and maximum inhibition, respectively. Equal numbers of irradiated stimulators and responders are mixed together and 10Q0ul are added to each well.

Wells of stimulator alone and responders alone are also set up as contr.

The experiment is incubated at 370C in 100% humidity and 5%CO2 L, a days. Response is measured by assessing proliferation during the liat 18 hours of culture by incubation with 1l/CVwell 3 H-Thymidine (Amersham UK), harvesting on to glass filter mattes and counting using a beta counter.

Results are plotted as CPM against antibody concentration. Replicates vary by less than T cell Recall Response to Tetanus ToxoId The ability of the engineered variants of L243 to suppress a secondary response was assessed using a recall response to Tetanus Toxoid.

The principle of the experiment is that T lymphocytes from an individual previously immunised with Tetanus Toxoid (TT) will respond to TT when re-exposed .xJljMe This activation is dependent on the interaction I I II r~YP IY -rmb WO 94/29351 I'ICGB9401290 39 between the CD3/TcR complex on T cells and the MHC-II molecule on cells which process and present the antigen. Antibodies that bind to MHC- II are known to inhibit this reaction.

Lymphocytes are prepared fresh for each experiment. Human venous blood is drawn into endotoxin free tubes containing heparin. Peripheral blood mononuclear cells (PBMC) are prepared by density gradient centrifugation according to the manufacturers instructions (Pharmacia).

PBMC are adjusted to 2x10 6 cells/ml in RPMI 1640 medium (Gibco UK) containing 2mM Glutamine (Gibco UK), 100pIml/100.g/ml Penicillin/ Streptomycin (Gibco) and 10% foetal calf serum (Sigma UK), in which all manipulations, dilutions and incubations are done.

Serial antibody dilutions are prepared in triplicate in sterile U-bottom 96' well microtitre plates (Falcon UK) in 100il. 50la1 containing an optimal concentration of TT, previously determined by experimentation, is added to all wells. Control wells containing medium only or Cyclosporin (Sandimmun, Sandoz) (100nM) are also prepared to establish the maximum response and maximum inhibition, respectively. 50p.l PBMC are then added to each well. The experiment is incubated at 370C in 100% humidity and 5%C02 for 7 days. Response is measured by assessing proliferation during the last 18 hours of culture by incubation with 1 CiVwell 3 H-Thymidine, harvesting on to glass filter mattes and counting using a beta counter.

Results are plotted as CPM against antibody concentration. Replicates vary by less than Besultt (Figures 17-21) There were no significant or qualitative differences between the effects of the L243 human isotype series between the MLR and TT response.

Maximal inhibition was achieved with G1, G1[L235E] and G1[L235A].

Approximately two orders of magnitude more of G2, G4 and G1[G237A] was required to give similar levels of inhibition. The G1/G2 L hinge exchange mutant was intermediate in immuno-suppresser potency. There was no correlation between complement fixation or FcRI binding and .1 I WO 94/29351 I('T/I('II /0o (1290 39 between the CD3/TcR complex on T cells and the MHC-II molecule on cells which process and present the antigen. Antibodies that bind to MHC- II are known to inhibit this reaction.

Lymphocytes are prepared fresh for each experiment. Human venous blood is drawn into endotoxin free tubes contain' ig heparin. Peripheral blood mononuclear cells (PBMC) are prepared by density gradient centrifugation according to the manufacturers instructions (Pharmacia).

PBMC are adjusted to 2x10 6 cells/mi in RPMI 1640 medium (Gibco UK) containing 2mM Glutamine (Gibco UK), 100pl/mV100pg/ml Penicillin/ Streptomycin (Gibco) and 10% foetal calf serum (Sigma UK), in which all manipulations, dilutions and incubations are done.

Serial antibody dilutions are prepared in triplicate in sterile U-bottom 96' well microtitre plates (Falcon UK) in 100l. 501ul containing an optimal concentration of TT, previously determined by experimentation, is added to all wells. Control wells containing medium only or Cyclosporin (Sandimmun, Sandoz) (100nM) are also prepared to establish the maximum response and maximum inhibition, respectively. 50p.l PBMC are then added to each well. The experiment is incubated at 370C in 100% humidity and 5%C0 2 for 7 days. Response is measured by assessing proliferation during the last 18 hours of culture by incubation with iCVwell 3 H-Thymidine, harvesting on to glass filter mattes and counting using a beta counter.

Results are plotted as CPM against antibody concentration. Replicates vary by less than RBeults (Figures 17-21) There were no significant or qualitative differences between the effects of the L243 human isotype series between the MLR and TT response.

Maximal inhibition was achieved with G1, G1[L235E] and G1[L235A].

Approximately two orders of magnitude more of G2, G4 and G1[G237A] was required to give similar levels of inhibition. The G1/G2 L hinge exchange mutant was intermediate in immuno-suppresser potency. There was no correlation between complement fixation or FcRI binding and ~cl IIIILL b_ C il WO 94/29351 PCl/CG 1194/01290 immuno-suppression, G1 binding well to FcRI and fixing complement and G1[L235E] doing neither, but both giving good immunosuppression. But, there was good correlation with FcRIII binding. Human G1 and G1[L235E] interact with FcRIII and give good immunosuppression. The G1/G2 L hinge is intermediate in FcRIII binding and immuno-suppression. In contrast, the G237A mutation in human G1, in agreement with published observations, reduces FcRIII binding. This antibody gave poor immunosuppression. (Table Table 6 shows a number of L243 isotype mutants.

Conclusion We have found that amino acid residues necessary for Clq and FcR binding of human IgG1 are located in the N-terminal region of the CH2 domain, residues 231 to 238, using a matched set of engineered, antibodies based on the anti-HLA DR antibody L243. Changing the leucine 235 in the CH2 region of IgG3 and lgG4 to glutamic acid was already known to abolish FcRI binding, we have confirmed this for IgG1 and also found a concomitant abolition of human complement fixation with retention of FcRIII mediated function. Changing the glycine at 237 to alanine of IgGi also abolished FcRI binoing and reduced complement fixation and FRIIIl mediated function. Exchanging the whole region 233 to 236, with the sequence found in human lgG2 abolished FcRI binding and complement fixation and reduced FcRIII mediated function of IgG1. In contrast, a change in the previously described Clq binding motif, from lysine at 320 to alanine had no effect on IgG1-mediated complement fixation.

The effect of these changes in IgG1 on FcRI binding are similar to published observations using lgG3 and lgG4 [Lund J ~Ltl J. Immunol.

1991. 147, 265; and Alegre M-L etl, J. Immunol. 1992. J48, 3461] with changes at 235 and 237 in the lower hinge/N-terminal CH 2 region markedly reducing FcRI binding. The simi.a:ities between these three isotypes strongly suggests that they interact with FcRI in a similar way.

We have found residues necessary for FcRIII binding of human IgG1 within the lower hinge/N-terminal end of the CH 2 region. Modification at 237 and L-l I WO 94/29351 PCT/l(; 1194/01290 41 exchanging the lower hinge for IgG2 residues caused low and intermediate levels, respectively, of FcRIII mediated killing. These effects are similar to those reported by Sarmay aLl [Molec. Immunol. 1992. 22 633] for human igG3. Ir, contrast to Sarmay tal using lgG3, our changes at residue 235 of IgG1 had little effect on FcRIII binding.

Greenwood LW [Eur. J. Immunol. 1993. 23, 1098], using inter and intra domain switch variants between IgG1 and IgG4, identify residues in IgG1 necessary for FcRIII binding in the C-terminal half of the CH 2 domain beyond 292. This indicates that the residues we have identified within the lower hinge/N-terminal end of the CH 2 region are necessary but not sufficient for FcRIII effector function mediated through binding of human IgG1.

IgG1 variants with changes at 235 failed to mediate lysis with human complement and did not bind purified human C1q. We also found that an IgG1 molecule containing a change at 320 gave complement mediated killing equivalent to the IgG1 wild type. Residues, Glu 318, Lys 320 and Lys 322 were identified by protein engineering studies as necessary in mouse IgG2b for Clq binding [Duncan, A R and Winter G, Nature, 1988.

322., 21]. The same study also demonstrated that the 235 change in mouse IgG2b left unchanged its affinity for human Clq [Duncan, A R and Winter G, Nature, 1988. 322, 21]. The apparent contradiction between these observations is probably due to differences in Clq contacts between human IgG1 and mouse IgG2b.

We found that most changes in the lower hinge/N-terminal end of the CH 2 domain affect Clq binding. The G1/G2 lower hinge exchange abolished complement fixation and the change at 237 also 'educes it significantly. In contrast, Greenwood etLal [Eur. J. Immunol. 1993. 23. 1098], found residues necessary for human complement fixation in the C-terminal half of the CH 2 domain. Tao .Lta[J. exp. Med. 1993. 178, 661] also identify the C-terminal half of the CH 2 domain as necessary for complement fixation.

They are able to separate Clq binding from complement mediated lysis.

IgG1 with a Pro to Ser change at 331, in the C-terminal half of the CH 2 domain, is able to bind human Clq as well as the wild type but is unable to WO 94/2931S P(T/G194/01290 '12 activate complement. This predicts that the amino acids that we have identified within the lower hinge/N-terminal end of the CH 2 region are necessary for Clq binding and that the C-terminal residues are necessary for the binding and activation of the antibody dependent complement cascade beyond Clq.

WO 94/29351 CI/B/02( ITTKNI94/01290 43 Summary of 1243 IsotypD Series 1943 RI RIIt Clq MLR TT G2 ii G4 Gi G 1L235E G 1L235A i G 1G237A mm WO 9,1/29351 PUGQ1/( I29( Hurm lsgy Mtnts f LQm NAME 235 235 237 320 235 231-238

L

L

G

K

L

APELLGGP

E

A

A

A

E

AP-PVAGP

G 1 L235E] Gi1 [L235A] G 1 [G237A] G 1 [K320A] G4[L235E] Gl1/G2L-hinge WO 9'4/29351 ~iG19/19 [ICT/O 1194/0 1290

IABILZ

.SImmlrljyofL243 Isotype Series L243Ba I~b C.mpigntc G2 >10 100000 >20/0 G4 1.2 lOOQ0ex >20/0 G4[L235E] >10 100000 >20/0 Gi1 0.13 5 0-6/65 G1/G2Lh >10 500 >20/0 G1EL235E] >10 40 >20/0 Gi1 [1-235A] 5.0 9 >20/0 G 1[G237A] >10 lOOQ0ex 2.0/20 G 1[K320A] 0.1 10 0.6[1) a) mg/mi antibody necessary for 50% inhibition of binding of FITOlabelled mouse lgG2a antibody to U937 cells.

b) nglml antibody necessary for half maximal cell killing in ADOC. (ex) extrapoiated value.

c) mg/mI antibody necessary for half maximal cell killing by human complement and percent plateau cell killing.

WO 94/29351 l'CT/G1194/01290 48 L243 is a mouse monoclonal antibody raised against human MHC Class II.

The nucleotide and amino acid sequences of L243 VI and Vh are shown in Figures 5 and 3 respectively. The following examples describe the humanisation of the L243 antibody (CDR grafting).

CDR grafting of L243 light chain Alignment of the framework regions of L243 light chain with those of the four human light chain subgroups [Kabat, Wu, Perry, H.M., Gottesman, K.S. and Foeller, C. 1991, Sequences of Proteins of Immunological Interest, Fifth Edition] revealed that L243 was most homologous to antibodies in human light chain subgroup 1. Consequently, for constructing the CDR grafted light chain, the framework regions chosen corresponded to those of the human Group 1 consensus sequence. A' comparison of the amino acid sequences of the framework regions of L243 and the consensus human group I light chains is given below and shows that there are 21 differences (underlined) between the two sequences.

Analysis of the contribution that any of these framework differences might have on antigen binding (see published International patent application No.

W091/09967) identified 4 residues for investigation; these are at positions 45,49,70 and 71. Based on this analysis, two versions of the CDR grafted light chain were constructed. In the first of these, L243-gL1, residues 45,49,70 and 71 are derived from the L243 light chain while in the second, L243-gL2, all residues are human consensus.

Light chain Comparisons Hu group 1 consensus DIQMTQSPSSLSASUGORUTITC L243 IQMTQSPASLSUSUGETUTITC 4 4 9 Hu Group 1 consensus WYQQKPGKRPKLLIY L243 RQ lP~GKPQLLUF WO 94/29351 9'/293~I I ''i~/G1194/01I290~ Hu Group 1 consensus o upsRFSGSGTDFTLTISSLQPEDFATYYC *GUPSRFSOSOSGTIQYSLK IHSLQSEOFGOYYC L2'13 Hu Group 1 consensus FGQGTKUEIKI FGGGTHLEIKR L243 The construction of L243-gLl is given below in detail. The following oligonucleotides were used in the Polymerase Chain Reactions (POR) to introduce changes into the framework regions of the chimneric light chain: R5043 R5044 R50'45 R5046 R5047 R5048 5' GTRGGRGRCCGGGTCACCATCRCRTGTCGRGCRR3 V 5 'CTGRGGRGCTTTTCCTGGTTTCTGCTGATRCCRTGCTARR3' 5 'RRRCCRGGRRRRGCTCCTCRGCTCCTGRTCTTTGCTGCATC3' 5 'CTTCTGGCTGCRGGCTGGRGRTRGTTAGGGTRTRCTGTGTGCC3' 5 'CTTCRGCCTGCRGCCRGRTTTTGCTRCTTRTTRCTGTCAR3' 5 'GGGCCGCTACCGTRCGTTTTRGTTCCRCTTTGGTGCCTTGRCCGRR3' Three reactions, each of 20 p.1, were set up each containing 10 mM Tris- HCl pH 8.3, 1.5 mM MgCl2, 50 mM KCl, 0.01 w/v gelatin, 0.25 mM each deoxyribonucleoside triphosphate, 0.1 pgg chimeric L.243 light chain DNA 6 pmoles of R5043/R5044 or R5045/R5046 or R5047/R5048 and 0.25 units Taq polymerase. Reactions were cycled through 9400 for 1 minute, 550C for 1 minute and 7200 for 1 minute. After 30 cycles, each reaction was analysed by electrophoresis, on an agarose gel and the POR fragments excised from the gel and recovered using a Mermaid Kit.

Aliquots of these were then subjected to a second round of PCR. The reaction, 100 p.1, contained 10 mM Tris-HOI pH 3.3, -1.5 mM MgCI2, 50 mM KCI, 0.01% w/v gelatin, 1/10 of each of the three PCR fragments from the first set of reactions, 30 pmoles of R5043 and R5048 and 2.5 units Taq polymerase. Reaction temperatures were as above. After the POR, the WO 94/29351 CICT/G 1194/01290 mixture was extracted with phenol chloroform and then with chloroform and precipitated with ethanol, The ethanol precipitate was recovered by centrifugatlon, dissolved in the appropriate buffer and restricted with the enzymes BstEII and Spll. The resulting product was finally electrophoresed on an agarose gel and the 270 base pair DNA fragment recovered from a gel slice and ligated into the vector pMR15.1 (Figure 1) that had previously been digested with the same enzymes.

The ligation mixture was used to transform E. coli LM1035 and resulting colonies analysed by PCR, restriction enzyme digests and nucleotide sequencing. The nucleotide and amino acid sequence of the VI region of L243-gL1 is shown in Figure 22.

Construction of CDR grafted light chain L243-gL2 L243-gL2 was constructed from L243-gL1 using PCR. The following oligonucleotides were user to introduce the amino acid changes R1053 5'GCTGACAGACTRACRGRCTGTTCC3' R5350 GC3' R5349 TRAC3' R684 5'TTCRRCTGCTCRTCAGAT3' Two reactions, each 20 were set up each containing 10 mM Tris-HCI pH 8.3, 1.5 mM MgCl2, 50 mM KCI, 0.01% w/v gelatin, 0.25 mM each deoxyribonucleoside triphosphate, 0.1 p.g L243-gL1, 6 pmoles of R1053/ R5350 or R5349/R684 and 0.25 units Taq polymerase. Reactions were cycled through 940C for 1 minute, 550C for 1 minute and 720C for 1 minute. After 30 cycles, each reaction was analysed by electrophoresis on an agarose gel and the PCR fragments excised from the gel and recovered using a Mermaid Kit.

i WO 94/29351 N'T/G094/0(41012900 Allquots of these were then subjected to a second round of PCR. The reaction, 100 pl, contained 10 mM Trls-HCI pH 8.3, 1.5 mM MgCI2, 50 mM KCI, 0.01% w/v gelatin, 1/5 of each of the PCR fragments from the first set of reactions, 30 pmoles of R1053 and R684 and 2.5 units Taq polymerase.

Reaction temperatures were as above. After the PCR, the mixture was extracted with phenol chloroform and then with chloroform and precipitated with ethanol. The ethanol precipitate was recovered by centrifugation, dissolved in the appropriate buffer and restricted with the enzymes BstEll and Spll. The resulting product was finally electrophoresJ on an agarose gel and the 270 base pair DNA fragment recovered from a gel slice and ligated into the vector pMR15.1 (Figure 1) that had previously been digested with the same enzymes.

The ligation mixture was used to transform E. coli LM1035 and resulting colonies analysed by PCR, restriction enzyme digests and nucleotide sequencing. The nucleotide and amino acid sequence of the VI region of L243-gL2 is shown in Figure 23.

CDR grafting of L243 heavy chain CDR grafting of 1243 heavy chain was accomplished using the same strategy as described for the light chain. L243 heavy chain was found to be most homologous to human heavy chains belonging to subgroup 1 and therefore the consensus sequence of the human subgroup 1 frameworks was chosen to accept the L243 heavy chain CDRs.

A comparison of the framework regions of the two structures is shown below where it can be seen that L243 differs from the human consensus at 28 positions (underlined). After analysis of the contribution that any of these might make to antigen binding, only residues 27,67,69,71,72 and were retained in the CDR grafted heavy chain, L243-gH.

Heayw chain comparisons 2 7 Hu Group 1 consensus QUQLUQSGREUKKPGRSUKUSCKASGYTFT L243 QQLUQSGPELKKPGETUKISCKRSGFTFT

-I

WO 9d/29351 1ICT/Gy 094/01290 Hu Group 1 consensus W~URQRPGQGLE.1lG L243 WUKQAPGKGLKWMG 6 6 77 7 7 912 Hu Ordup 1 consensus RUTITRIJTSTSTRYM1ELSSLRSEOTAUYYCR L243 RFRFSLETSRSTRYtL INL KNEDTRKYFCR Hu Group I consensus JOGGLUTUSS L243 L4GQGTTLTUSS Construction of CDR grafted heavy chain. 1.243_gH L243gH was assembled by subjecting overlapping oligon uc!eot ides to POR in the presence of the appropriate primers. The following oligonucleotides, were used in the POR: R300J4 :5 'GGGGGGRRGCTTOCCGCCRCCRTGG3' R3005 5 TCCCCCCGGCCCTTTGRAGC8G3' R4902 RGRGGTGRRGRRGCCTGGRGCRTCT03' R4903 CROGGGRCTCGRGTGGR3' R4904 ORGACRTCTOCRTCTRCRGCRTRCRT3' R4905 :5'RGCRGTGTACTCTGTCR63GACATTRCAGCRGTGGTRCCTR CROGGTTCGRCTRCTOGGGACRGOGG3' R4897 'TGIGAGTGCRCTCCTGTTGTCRCRGACRGGRIGARCRGGRACRCC CARACCACTCCRTGGTGGCGGCRRGCTTCCCCCC3' R4898 CTTCRCRiGATGCTCCAGGCTTCTTCR3' M499 TCCCHTCCRCTCGRGTCCCTGTCCAG3' ~0*b3~ n--nl ;m~i--rr WO 94/29351 I A'C; B94/01290 51 R4900 GCTCCRTGTRTGCTGTRGRTGCAGRT3' R4901 CCCTGTCCCCRGTRGTCGRR3' The assembly reaction, 50 pl, contained 10 mM Tris-HCI pH 8.3, 1.5 mM MgCI2, 50 mM KCI, 0.01% w/v gelatin, 0.25 mM each deoxyribonucleoside triphosphate,1 pmoie of each of R4897 R4905, 10 pmoles of each of R3004 and R3005 and 2.5 units Taq polymerase. Reactions were cycled through 94 C for 1 minute, 55 C for 1 minute and 72 C for 1 minute. After cycles, the reaction was extracted with phenol/chloroform then with chloroform and precipitated with ethanol. After centrifugation, the DNA was dissolved in the appropriate restriction buffer and digested with HindllI and Apal. The resulting fragment was isolated from an agarose gel and ligated into pMR14 (Figure 2) that had previously been digested with the same enzymes. pMR14 contains the human gamma 4 heavy chain constant region and so the heavy chain expressed from this vector will be a gamma 4 isotype. The ligation mixture was used to transform E. coli LM1035 and resulting bacterial colonies screened by restriction digest and nucleotide sequence analysis. In this way, a plasmid containing the correct sequence for L243gH was identified (Figure 24).

Construction of Gamma 1 versions of chimeric and CDR grafted L243 heavy chain Human Gamma 1 versions of L243 heavy chains were constructed by transferring the variable regions of both the murine and the CDR grafted heavy chains as Hindlll to Apal fragments into the vector pGammal (Figure This vector contains the human Gamma 1 heavy chain constant region.

Evaluation of activities of CDR grafted genes The activities of the CDR grafted genes were evaluated by expressing them in mammalian cells and purifying and quantitating the newly synthesised antibodies. The methodology for this is described next,

-I

I* WO 94/29351 PCT/GB94/01290 52 followed by a description of the biochemical and cell based assays used for the biological characterisation of the antibodies.

a) Gene Expression in CHO cells Chimeric and CDR grafted L243 was produced for biological evaluation by transient expression of heavy and light chain pairs after co-transfection into Chinese Hamster Ovary (CHO) cells using calcium phosphate precipitation as described above for production of chimeric L243.

Antibody concentration was determined using a human Ig ELISA (see below).

b) ELSA For the ELISA, Nunc ELISA plates were coated overnight at 4oC with a F(ab)2 fragment of a polyclonal goat anti-human Fc fragment specific antibody (Jackson Immuno-research, code 109-006-098) at 5 ig/ml in coating buffer (15mM sodium carbonate, 35mM sodium hydrogen carbonate, pH6.9). Uncoated antibody was removed by washing 5 times with distilled water. Samples and purified standards to be quantitated were diluted to approximately 1 pg/ml in conjugate buffer (0.1M Tris-HCI 0.1M NaCI, 0.2% v/v Tween 20, 0,2% w/v Hammersten casein). The samples were titrated in the microtitre wells in 2-fold dilutions to give a final volume of 0.1 ml in each well and the plates incubated at room temperature for 1 hr with shaking. After the first incubation step the plates were washed 10 times with distilled water and then incubated for 1 hr as before with 0.1 ml of a mouse monoclonal anti-human kappa (clone GD12) peroxidase conjugated antibody (The Binding Site, code MP135) at a dilution of 1 in 700 in conjugate buffer. The plate was washed again and substrate solution (0.1 ml) added to each well. Substrate solution contained 150pl N,N,N,N-tetramethylbenzidine (10 mg/ml in DMSO), 150pl hydrogen peroxide (30% solution) in 10 ml 0.1M sodium acetate/sodium citrate, The plate was developed for 5 -10 minutes until the absorbance at 630nm was approximately 1.0 for the top standard. Absorbance at 630nm was measured using a plate reader and the concentration of the sample determined by comparing the titration curves with those of the standard.

I- lllllli~-i- rr~; WO 94/29351 PCT/GB94/01290 53 c) Competition Assay The principle of this assay is that if the antigen binding region has been correctly transferred from the murine to human frameworks, then the CDR grafted antibody will compete equally well with a labelled chimeric antibody for binding to human MHC Class II. Any changes in the antigen binding potency will be revealed in this system.

Chimeric L243 was labelled with fluorescein (FITC) using the method of Wood et al [Wood,T., Thompson, S and Goldstein, G 1965, J. Immunol 225-229 and used in the competition assay described above.

Figure 25 compares the ability of combinations of L243 heavy and light chains to compete with FITC-labelled chimeric L243 for binding to JY cells.

All combinations were effective competitors although none of those containing CDR grafted heavy or light chains were as effective as the chimeric antibody itself. Thus, the combinations cH/gL1, gH/cL and gH/gL1 were 89%, 78% and 64% respectively, as effective as chimeric L243 in this assay.

d) Determination of Affinity constants by Scatchard Analysis L243 antibodies were titrated from 10pg/ml in PBS, 5% fetal calf serum, 0.1% sodium azide in 1.5-fold dilutions (150R.I each) before incubation with 5x10 4 JY cells per titration point for 1 hour on ice. The cells were previously counted, washed and resuspended the same medium as the samples. After incubation, the cells were washed with 5ml of the above medium, spun down and the supematant discarded. Bound antibody was revealed by addition of 100p of a 1/100 dilution of FITC conjugated antihuman Fc monoclonal (The Binding Site; code MF001). The cells were then incubated for 1 hour on ice and then the excess FITC conjugate removed by washing as before. Cells were dispersed in 250p.l of the same buffer and the median fluorescence intensity per cell was determined in a FACScan (Becton Dickinson) and calibrated using standard beads (Flow Cytometry standards Corporation). The number of molecules of antibody bound per cell at each antibody concentration was thus established and used to generate Scatchard plots. For the purpose of calculation, it was WO 94/29351 ICT/GB94/01290 54 assumed that the valency of binding of the FITC conjugate to L243 was 1:1 and that the F/P ratio was 3.36 (as given by the manufacturer).

A Scatchard plot comparing the affinities of chimeric L243 (cH/cL), L243gH/L243-gL1 and L243-gH/L243-gL2 is shown in Figure 26. Chimeric L243 was found to have an apparent Kd of 4.1 nM while the CDR grafted antibodies containing gL1 and gL2 light chains had apparent Kd of 6.4nM and 9.6nM respectively. The difference in Kd values of the antibodies with the two CDR grafted light chains reflects the contribution made by residues 45,49,70 and 71 that had been retained, in L243-gL1, from the parent light chain.

e) Antibody dependent cell mediated cytotoxicity The ability of chimeric and CDR grafted L243 to mediate antibody dependent cell cytotoxicity (ADCC) was compared as described previously.

The principle of the experiment is that antibodies will mediate lysis of target cells bearing their cognate antigen if the Fc of the antibody is able to interact with Fc receptor bearing effector cells capable of cytotoxicity.

A comparison of the activities of chimeric (cH/cL) and CDR grafted (gH/gL1) L243 human gamma 1 isotypes in the above assay is shown in Figure 27. Both antibodies were effective mediators of cell lysis with maximal activity being achieved at antibody concentrations of less than 100 ng/ml. There was no significant difference between the activities of the two antibodies.

f) Immune function tests Ex.&vi T cell function experiments were performed where an interaction between MHC-ll and the T cell receptor was an obligatory requirement for T cell activation. Chimeric and CDR grafted L243 antibodies were compared in mixed lymphocyte reactions, which measures both naive and memory T cell activation, and in recall responses to tetanus toxoid which only measures a memory T cell response.

WO 94/29351 PCT/1GB94/01290 1) Mixed Lymphocyte reaction as described above The principle of the experiment is that when leucocytes from one individual are mixed with those of another individual which express different HLA alleles, they will recognise each other as foreign and the lymphocytes will become activated. This activation is dependent primarily on interactions betweeh the CD3/TcR complex on T cells and the MHC Class II molecule on antigen presenting cells. L243 is known to inhibit this reaction.

When an MLR was carried out to compare the effectiveness of the Gamma 1 isotypes of chimeric and CDR grafted L243 as inhibitors of T cell activation, no significant differences were observed between the two antibodies (Figure 28). Greater than 90% inhibition of the MLR was observed using 100 ng/ml of either antibody.

2) T cell recall response to Tetanus toxoid The ability of chimeric and CDR grafted L243 to suppress a secondary response was assessed using a recall response to Tetanus toxin. The principle of the experiment is described above.

The results of an experiment comparing the ability of human gamma 1 isotypes of chimeric and CDR grafted L243 to inhibit the response to TT is shown in Figure 29. Both antibodies were effective inhibitors of the T cell response to TT and produced titration curves that were indistinguishable.

EXAMPLE 4 The ability of CDR grafted L243 with the alteration at position 235 i.e.

L[235E] to mediate antibody dependent cell cytoxicity (ADCC) was measured essentially as described in the previous examples. The results are shown in Figure 27.

Similarly the CDR grafted L243 [L235E] antibody was tested in a mixed lymphocyte reaction and in recall response to tetanus toxoid essentially as described in the previous Examples. The results are provided in Figures 28 and 29.

WO 94/29351 "CT/G1194/01290 56 The ability of the CDR-grafted L243 antibody [L235E] to fix human complement was assessed using the technique of antibody dependent complement mediated cytotoxicity as described in the previous Examples.

The results are shown 'n Figure

Claims (9)

1. An altered MHC specific antibody wherein one or more amino acid residues in the N-terminal region of the Cj2 domain of said antibody are altered characterised in that the ability of said antibody to fix complement is altered as compared to unaltered antibody.

2. An antibody according to Claim I which binds one or more cellular Fc receptors and does not bind significantly to FcRI.

3. An antibody according to Claim 1 or 2 wherein the amino acid residue which is altered lies within amino acid positions 231 to 239.

4. A method of modulating the function of cell surface associated antigens avoiding complement mediated toxicity comprising administration of an altered antibody wherein one or more amino acid residues in the N-terminal region of thp CH 2 domain of said antibody are altered characterised in that the ability of said antibody to fix complement is altered as compared to unaltered antibody ond also said antibody.

A method according to Claim 4 wherein said altered antibody is able to bind one or more cellular Fc receptors especially FcRIII while binding to FcRI is significantly reduced.

6. An altered antibody according to any one of claims 1 to 3 derived from the 20 anti-MHC antibody L243 (ATCC

7. A method of treating diseases in which antibody therapy leads to undesirable toxicity due to antibody mediated complement fixation comprising administering an altered antibody wherein one or more amino acid residues in the N-termir:al region of the CH 2 domain of said antibody are altered characterised in that the ability of said antibody to fix complement is altered as compared to unaltered antibody. t MAW Ii#')9346 RS> 1I Mnrch 1998 58

8. An altered antibody according to any one of claims 1 to 3 or 6 substantially as hereinbefore described.

9. A method according to any one of claims 4, 5 or 7 substantially as hereinbefore described. DATED: 31 March 1998 CARTER SMITH BEADLE Patent Attorneys for the Applicant: CELLTECH THERAPEUTICS LIMITED S* ft f ft o°* 0= 31 March 1998