CA2180834A1 - Antibodies to egf receptor and their antitumour effect - Google Patents

Abstract

This invention relates to the development of agents and methods for treating tumour cells. More particularly, it is concerned with antibodies against the epidermal growth factor receptor (EGFR), which have been found to exhibit an antitumour effect, inducing terminal differentiation in tumour cells, especially those which are characterised by over-expressing EGFR. Included in the invention are these antibodies and their functional equivalents, mutants and derivatives. The antibodies may be murine (rat or mouse), but preferably have framework regions and or constant regions based on or derived from human antibodies. The present invention also discloses that fragments of these antibodies (eg Fabs) can retain the properties of the complete antibody. The antibodies are useful, e.g. in treating or detecting tumour cells and in targeting therapies.

Description

21 8~834 woss/20o4s P~ r;ll8 ANTIBODIES TO EGF RE~;~LOR AND T~EIR A~LL~ . . EFFECT
Field o~ the Tnv~ntion This invention relates to the development of agents and methods for treating tumour cells. More particularly, it is rnnrPrnGd with antibodies against the epidermal growth factor receptor ~EGFR), which have been found to exhibit an antitumour ef f ect .
Backqround to the Invention There: is increasing evidence to suggest that polypeptide growth factors and their receptors are involved not only in the regulation of normal cell proliferation and differentiation but also, when aberrantly expressed, in the pathogenesis of certain types of human malignancy. The epidermal growth factor receptor (EGFR) and its ligands is one such example. This receptor is a 170kD tr~n~ Ldlle glycoprotein with tyrosine kinase activity which transmits the mitogenic action of the EGF family of growth factors including EG~, TGFc~ and amphiregulin. The binding of these ligands to the e~ternal domain of the EGF receptor initiates a number of early and delayed responses in the target cells leading ul timately to DNA synthesis and cell divi s ion .
Over-_xpression of the EGFR has been reported in a number of human malignancies inrlllflinr; cancer of the breast, brain, bladder, head and neck, pancreas and lung (1,9-12).
~igh levels of expression of this receptor have also been associated with poor survival in some of these patients (1,12,13) . In addition, the histological and biological examination of human tumour biopsies and cell lines has shown that uv~L~L~ssion of the EGF receptor is often ~ccnmr~ni ed by the production of one or two of its ligands (TGF~ and/or EGF) by the same tumours, suggesting that an autocrine loop may be responsible for growth of tumours of this type (14-17) . Furthermore, since the ligand-induced activation of such cells acts primarily via receptors on the cell surface rather than intracellularly, such a system may form a suitable target for monoclonal antibody directed therapy (1,18-24).
We have described recently the production of 21 rat monoclonal ~ntihQflip~ raised against five distinct epitopes on the external domain of the human EGF receptor using as immunogen LICR-LON-HN5, a so,uamous cell carcinoma of head and neck, MDA-MB 468, a breast carcinoma cell line or A431 an epidermoid carcinoma cell line (1,25,26). Our aim was to obtain as diverse a population ( isotype/epitope) of antibodies as possible from which the best m~b or combination of m~bs for rlinir~l and diagnostic application could: be selected.
Of these antibodies ICR64 (IgGl, ) directed against epitope D and ICR16 (IgG2a) and ICR62 (IgG2b) both directed against epitope C were (in that order) the most effective at SUBSTiTlJTE SHEET (RULE 26) i ~
wossl20o~s ' 2 1 8 0834 ~ ,S1t 118 inh;h1t;ng ligand binding and the growth in vitro of 6~uamous cell carcinomas overexpre5sing the EGFR. However ICR62 was the most effective of the three antibodies in inducing the regression of xenografts of such tumours growing in athymic mice.
summarY of the Invention ~ere we describe the results of an ;mmllnnl~; stological p7~;~m;n;ltion of the events occurring in tumours during regres6ion where athymic mice bearing established xenogra$ts had been undergoing m~b therapy. Our aim was to investigate (a) the mPrh~ln; Sm by which antibodies to the EGFR inhibited tumour growth and (b) to determine if Yiable tumour cells were present in the residual tumour nodules following treatment with antibody, and if so to ~ torm;nP
if loss of antigen expression was a significant factor in the escape of these tumour cells.
The results of initial experiments suggested that prolonged exposure of xenografts of the ~N5 tumour (a squamous cell carcinoma) which overexpresses the EGF receptor to monoclonal antibody ICR62 (IgG2b) induces a complete regression of the xenograft if antibody treatment c1 ~ Pd at the time of tumour implantation. Even where treatment was delayed until tumours were established, ICR62 induced complete or almost complete regression of the tumours (39) .
~istological examination of the tumours r~m;~;n;ng at the end of the experiment showed that, while few viable tumour cells could be detected, numerous keratinised areas were observed, suggesting that only differentiated tissues rema ined ( 3 9 ) .
The applicants have now investigated the possibility that receptor blockade induces terminal differentiation (ie the c-P~ r reversion to normal phenotype) of squamous cell carcinomas which overexpress the EGF receptor.
Accordingly, in a first aspect, the present invention provides antibodies to EGF receptor, and fragments thereof, for therapeutic use. Tn~-l11tlP~l in this aspect of the invention are those antibodies identified herein, and their functional equivalents, mutants and derivatives. The antibodies may be murine (rat or mouse), but preferably have framework regions and/or constant regions based on or derived from human antibodies. The antibodies may be complete immunoglobulin molecules, but they may be Ig fragments, for example monovalent or divalent Ig entities such as Fab fragments, single chain Fv molecules etc.
In a further aspect, the present invention discloses the use of antibodies to the EGFR, or fragments thereof, in the preparation of a medicament for inducing terminal differentiation in tumour cells, especially those which are characterised by over-expressing BGFR.
SU~STITUTE SHEET (RULE 26 W0 95/20045 2 1 8 0 8 3 4 y~ 8 In a further aspect, the present invention provides DNA
encoding the above antibodies, expression vectors comprising that DNA and host cells transformed with the eXpression vectors In a further aspect, the present invention provides pharrnaceutical compositions comprising one or more of the above antibodies, for example selected from antibodies ICR16, ICR62 and ICR64. The amounts of the ~nt;ho~;es used in the compositions can be detPrmin~rl by the skilled person, and will be typically in the range 1 to 300mg.
Optionally, the the above ~nt;ho~ or fragements can be conjugated to a label, toxin or drug.
Alternatively, the antibodies could be used to target an active agent administered in a precursor form to cells, for conversion to the active f orm by an activating agent produced in, or targeted to, the cells to be treated. This type of approach is sometimes known as ADEPT (see for example, EP-A-415731 and WO 90/07936) Also included are peptide or mimetic molecules which mimic the antibody binding to EGFR and likewise induce terminal differentiation of turnour cells. The present invention also include6 the use of the above antibodies in the design or synthesis of these mimetics.
Although the EGF receptor antibodies r- t;nnPrl above are primarily of interest in the treatment of tumour cells, they may also find appl;l-~t;nn~ in the treatment of other disorders, eg arthritis, psoriasis, atherosclerosis, SLE, inflammation or other proliferative disorders.
~odi f ied Z~n t; bo~3; e~
The ~nt;ho~;es described above can altered in a variety of ways using rl~l ` in~nt DNA technology coupled with advances in the f ield of monoclonal antibody and protein engineering. This has enabled access to a large selection of antibodies and antibody fragments with different properties and structures to natural ~ntiho~; es .
The production of monoclonal i~nt;hotl;e5 is well established in the art. Monoclonal ~nt;ho~q;es can be subjected to the techniques of recnmh;n~nt DNA technology to produce other antibodies or chimeric molecules which retain the specif icity of the original antibody . Such techniques may involve introducing DNA f~nnnr1;n~ the lm~-lnn~lobulin variable region, or the complementarity determining regions 50 (CDRs), of an antibody to the constant regions, or constant regions plus fl ~uLh regions, of a different immunoglobulin. See, for instance, EP-A-184187, GB
2188638A or EP-A-239400. A hybridoma producing a monoclonal antibody may be subject to genetic mutation or other changes, which may or may not alter the binding specificity of antibodies produced.
SUBSTITUTE S~EET (RULE 26) Wo 9sl2004s 2 18 0 8 3 4 r~ 118 As an alternative or supplement to immlln;~ing a mammal with a peptide, an antibody specific for a protein may be obtained from a recomhin~ntly produced library of expressed immunoglobulin variable domains, eg using lambda bacteriophage or filamentous bacteriophage which display functional immunoglobulin binding domains on their surfaces; for instance see W092/01047. The library may be naive, that is constructed from sequences obtained from an organism which has not been immunised with the target, or may be one constructed using seS~uences obtained f rom an organism which has been exposed to the antigen of interest (or a fragment thereof ) .
Antibodies may be modifled in a number of ways. Indeed the 1~ term ~antibody" should be construed as covering any specific binding substance having an binding domain with the re~auired specificity. Thus this term covers antibody fragments, derivatives, functional e~auivalents and homologues of ~nt;ho~l;es, including any polypeptide comprising an immunoglobulin binding domain, whether natural or synthetic. Chimaeric molecules comprising an immunoglobulin binding domain, or e~uivalent, fused to another polypeptide are therefore inrlll~l. Cloning and expression of chimaeric antibodies are described in EP-A-0120694 and EP-A-0125023.
It has been shown that fragments of a whole antibody can perform the function of binding antigens. Examples of binding fragments are (i) the Fab fragment consisting of VL, VH, CL and CH1 domains; (ii) the Fd fragment consisting of the VH and CH1 domains; (iii) the Fv fragment consisting of the VL and VH domains of a single antibody; (iv) the d~b fragment (Ward, E.S. et al., Nature 341, 544-546 (1989) ) which consist6 of a VH domain; (v) isolated CDR regions;
(vi) F(ab')2 fragments, a bivalent fragment comprising two linked Fab ELC~. tr~ (Vii) single chain Fv molecules (scFv), wherein a VH domain and a VL domain are linked by a peptide linker which allows the two domains to associate to form an antigen binding site (Bird et al, Science~, 242, 423-426, 1988; Huston et al, PNAS USA, 85, ~879-5883, 1988); (viii) bispecific single chain Fv dimers (PCT/US92/09965) and (ix) "diabodies", multivalent or multispecific fragments constructed by gene fusion (WO94/13804; P. Holliger et al Proc. ~atl. Acad. Sci. USA
90 6444-6448, 1993).
Diabodies are multimers of polypeptides, each polypeptide comprising a first domain comprising a binding regi.on of an immunoglobulin light chain and a second domain comprising a binding region of an; ~3l obulin heavy chain, the two domains being linked (eg by a peptide linker) but unable to associate with each other to form an antigen binding site:
antigen binding sites are formed by the association of the first domain of one poIypeptide within the multimer with the second domain of another polypeptide within the multimer (WO94/13804).
SU8SmUTE SHEET (RULE 26~

woss/2004s 2180834 r~ ,,s~ 118 Where bispecific antibodies are to be used, these may be conventional bispecif ic antibodies, which can be manufactured in a variety of ways (E~olliger, P. and Winter G. Current Opinion Biotechnol. 4, 446-449 (1993) ), eg prepared chemically or from hybrid hybridomas, or may be any of the bispecific antibody fragments mentioned above.
It may be preferable to use scFv dimers or diahodies rather than whole antibodie~. Diabodies and scFv can be constructed without an Fc region, using only variable domains, potentially reducing the effects of anti-idiotypic reaction. Other forms of bispecific antibodies include the single chain "Janusins" described in Tr~lln.ork~r et al, Bmbo Journal, 10, 3655-3659, (1991).
Bispecific diabodies, as opposed to bispecific whole antibodies, are also particularly useful because they can be readily constructed and expressed in E. coli . Diabodies (and many other polypep~ides such as antibody fragments) of ~L)LU~Liate binding specificities can be readily selected using phage display (Wl~94/13304) from libraries. If one arm of the diabody is to be kept constant, for instance, with a specificity directed against antigen X, then a library can be made where the other arm is varied and an antibody of ~l,L~,~Liate specificity selected.
Previously, bispecifi c antibodies incorporating a sperific;ty for the T-cell co-receptor CD3 have been shown to inhibit tumour growth (Titus , J . A. et al ., J . Immunol .
138, 4018-4022 (1987) ) and to cure lymphoma (Brissinck J.
et al, J. Immunol. 174, 4019-4026 (1991) ) .
It may be desirable to "humanise" non-human (eg murine) antibodies to provide antibodies having the antigen binding properties of the non-human antibody, while minimising the immunogenic response oE the ~nt;horliesl eg when they are used in human therapy. Thus, humanised ~nt;horl~es comprise fLd~ Lk regions derived from human immunoglobulins (acceptor antibody) in which residues from one or more Compl ::lry determining regions (CDR' s) are replaced by residues from CDR's of a non-human species (donor antibody) such as mouse, rat or rabbit antibody having the desired properties, eg specificity, affinity or capacity. Some of the fL~ uL~ residues of the human antibody may also be replaced by correspo~lding non-human residues, or by residues not present in either donor or acceptor antibodies . These modif ications are made to the further ref ine and optimise the properties of the antibody .
Brief DescriT~tion of the D~awi~r~
Fi~r,ures 1 and 2 show f low cytometric analysis of DNA in nuclei obtained from ~N5 cells following four days incubation in vitro in DMEM-29,iFCS alone or in DMBM-2%F- ~
cr,nt~;n;ng anti-EGFR m?~bs (156nM) or BGF (lOnM) . F, 1 shows DNA histograms; Fig 2 shows the percentage of cells in each phase.
SUBSTITU rE 5HEE7' (RULE 26 wo 95/20045 ` - / 2 1 8 0 8 3 4 r ~ r l 18 Figures 3 and 4 show the inf luence of antibodies to the EGFR or their Fab fragments on the binding of 1'5I-EGF ~Fig 3) or l2sI-TGF~Y tFig 4) to the human bladder carcinoma cell l ine EJ .
Figures 5 and 6 show Scatchard plots of l25I-EGF binaing to EJ cells in the absence or presence of monovalent and divalent m~bs ICR9 (Fig 5) and ICR62 ~Fig 6) .
Figures 7-9 show the effect of treatment with ~nt;hnrl;~ to the EGFR or their Fab fragments on the growth in vitro of HN6 cells (Figs 7), other head and neck carcinoma cell lines overexpressing the EGF receptor ~Fig 8), and the TGFo!
induced proliferation of quiescent human foreskin fibroblasts (Fig 9).
Figure 10 shows titration curves showing the inhibition of binding of 125I-EGF to BJ cells by serum taken at the times indicated. A, patient 8 (20mg); B, patient lO (40mg); or C, patient 13 (lOOmg) . The starting concentration of the ICR62 standard was lOug/ml.
Figure 11 shows the development of human anti-rat antibodies in the serum of patient number 9 f ollowing injection of 20mg ICR62 shown by binding to intact ICR62 (A), Fab ICR62 (B) and scFv ICR62 (C) .
Figure 12 the development of ~ human anti-rat antibodies in the serum of patient number 11 following injection of 40mg of ICR62 shown by binding to intact ICR62 (A), Fab ICR62 (B) but not scFv ICR62 (C) .
The ~equ~nc~e listinqs show the DNA and deduced amino acid sequences of the variable regions of the heavy and light chains of ICR62 and ICR64, with the CDRs indicated in square brackets.
Detailed Descri~tion 4 0 2:xDeriments P~rt 1. Investiqation of cell lines in vitro and of human tumour xenoqrafts in vivo tl--~inq or followinq treatment with several an~;h~;es to the EGFR5 Materials and methods Cell lines.
The head and neck carcinoma cell line LICR-LON-ErN5 and the breast carcinoma cell line MDA-MB 468 were cultured routinely in Dulbecco' s modified Eagle' s medium' (DMEM) supplemented with 10% foetal calf ' serum (FCS) and the antibiotics penicillin, streptomycin and neomycin. For growth inhibition assays the concentration of FCS in medium was reduced to 2~ to minimi~e the effects of any' growth SUBSTITUTE 5HEET (RULE 26) W0 95/20045 r~ 8 f actors present .
Mnn~lcl~n ~1 Antibodies.
The preparation of rat monoclonal antibodies to the extracellular domain of the human EGFR has been described previously (25,26) . ICR16 (IgG2a) was raised against the receptor on the squamous cell carcinoma ~N5 while m~bs ICR62 (IgG2b), ICR61 (IgG2b), and IC~64 (IgGl) were raised against the receptor on the breast carcinoma cell line MDA-MB 468. z~nt;h~ P~ ICR16 and ICR62 bind to epitope C
and antibodies ICR61 and ICR64 bind i to another distinct epitope D on the external domain of the BGF receptor.
Isotype matched control antibodies included ALN/11/53 (IgG2a) and 11/160 (IgG2b) which are directed against a specific antigen on the rat sarcoma lISN (27), or RCI/4/74 (IgG1) an antibody directed against an idiotopic determinant on ICR16 (llnrllhl; ~hPd data) . Monoclonal antibody to cytokeratin 10 (RKSE-60) was obtained from EuroPath Ltd, Cornwal].. Mouse monoclonal antibody to involucrin ( 2 8 ) was a gif t f rom Dr Fiona Watt ( Imperial Cancer Research Fund, London).
Treatment of mice bearing human tumour xenogra~ts with J~nt;ho~ to the EGFR.
Xenografts of human tumours were established in athymic (nu/nu) mice and treated with antibodies as described previously (29, 39) . Briefly, the protocols used were as 3 0 f ol lows .
A) MDA-MB 468 xenoqrafts. Three groups of four mice were inoculated in both flanks with 5 x 106 tumour cells. On the day of tumour inoculation (day 0) one group of mice was injected i.p with ICR62, the second with ICR16 and the third with control antibody. Treatment with antibody (20~Lg/dose) was ~ r~ntinllf~cl for a further 4 consecutive days and thereaf ter three times weekly until day 18 (total dose 0.44mg/mouse). Animals were observed for up to 100 days when the experiment was terminated and the tumour nodules L~ ;n1ng were excised, weighed and fixed for histological minzlt; on (see below) .
B) HN5 xenoqrafts Since treatment of athymic mice with antibody initiated at the time of tumour inoculation results in complete and permanent regression of these tumours (39), ~N5 tumour xenografts were set up as described above but tl~e commencement of treatment with antibodies to the EGFR was delayed until the tumours had reached a mean diameter of about 0 5cm. In this experiment groups of four to five mice were treated with antibodies ICR61 or ICR64, or with combinations of ICR61 plus ICR62 or ICR64 plus ICR62; pairs which do not compete for binding to the EGFR. In each case a second group of mice was treated with control antibodies or saline. Treatment with antibodies ~200,ug/dose) was for five consecutive days and SUBSTlTlITE 5HEET (RULE 26) Wo95/2004S 21 80834 r~. s~- 118 thereaf ter three times weekly until the day indicated in each experiment (total dose 2.2mg/mouse). Animals were observed for up to lOO day6 or killed when the tumours reached a mean diameter of O 8 -1. Ocm.
~istological exA~ination of human tumour Y~n~ Afts _ollowing treatment with Ant;ho~
Two protocols were used = for examination of tumour specimens.
For routine histological ~m~n~tinn, tumours were excised and samples were fixed in Methacarn and then embedded in paraffin. Four micron sections were cut and stained with h;lPm;ltmcylin and eosin (X & E) .
All ; nhi ~tochemical studies of human tumour xenografts were perf ormed by the indirect method using immunoperoxidase-conjugated F(ab')2 rabbit anti-rat Ig as the second reagent (Star 51, Serotec Ltd, Oxford). This antibody had been preadsorbed to remove all reactivity with mouse immunoglobulins. Samples of tumour tissue were snap f rozen in isopentane precooled to liquid nitrogen temperature. Sections of 511m thickness were cut on a cryostat and mounted on glass slides that had been coated previously with 19~ aminopropyltriethoxy- saline (Sigma A3648) in acetone. Slides were dried at 37C (30 minutes) fixed in formol-calcium I~or 5 minutes, and immersed in icé
cold choloroform/acetone (1:1) for 5 minutes. After three washes with PBS, endogenous peroxidase was blocked by immersing the sections for 10 minutes in PBS cnnt~ining 396 H2O,. The sections were incubated first with rat antibodies to the EGFR (1-20/lg/ml, 90 minutes at room temperature) then, after three washes in PBS, with 100~1 of a 1/100 dilution of ; nrl~roxidase conjugated F (ab' ), rabbit anti-rat IgG (in PBS-0.596 BSA) for 90 minutes at room temperature. After washing twice with PBS and twic~e in DDW
bound peroxidase was v; ~ e~ using diamino-benzidine (DAB) [lOOmg DAB ~n lOOml of 0.1 M Tris buffer pH7.2 lOOml H,O, 66~1 H,O,~ . After 5 minutes incubation the séctions were washed twice in DDW, counterstainéd in Mayer~ s h;~ lm for 1 minute, dehydrated by passing through graded alcohols then mounted in DPX.
In some cases, where xenograf ts had been treated with antibody to the EGFR, staining of the frozen sections was performed as above but without incubation in the~ primary antibody in order to visualize the presence of any l~ i ni ng therapeutic mAb .
Flow cytometric analysis of tumour cells following treatment with antibody.
XN5 cells (7.5 x 105) were seeded into 25cm2 Nunc flasks (Gibco Europe ~td, Scotland) cnnt~ining 15ml DMEM ~lus 2~
FCS. Monoclonal antibody (25~g/ml), EGF (lOnM) or the SU8STITUTE SHEET (RULE 26) 21 ~0834 wo gs/~oo-~s ~ r~ 118 equivalent amount of medium only was then added and the cultures were incubated for four days at 37C. Flow cytometric analysis of the nuclei prepared ~rom these cells was perf ormed essentially as described by Ormerod et al (30) . Brie~ly, a suspension o~ single cells from each flask was prepared in 200 ~Ll of PBS and this was followed by the vigorous addition of 2ml ice cold 70g6 ethanol-30~ PBS. The cells were incubated for at least 30 minutes at 4C
harvested by centrifugation and r~fi~lcp~onfl~ in 700111 PBS.
Following the addition of 100~11 of RNAse (lmg/ml, Sigma) and 200~L1 of propidi.um iodide (10011g/ml, Sigma) the suspension was incubated first at 37C for 30 minutes and then on ice for 90 minutes. The nuclei were analysed using an Ortho Cytofluorograph 50H equipped with a Spectra-Physics argon-ion laser producing 200mW at 488nm and an Ortho 2150 comp~lter system (30).
T -f~ re8cent 8tFlining of cytokeratin 10 and involucrin .
5x10~ HN5 cells were plated onto glass cover slips in 24 well plates cnnt~;n;n~ lml DMEM plus 2~FCS. Following overnight ;n~lhz~ti nn at 37C, specific or control antibodies (25 llg/ml) or medium alone were added to the cultures and the cells were incubated for a further three to four days at 37C. ~fter two washes with PBS, the cells were fixed for 5 minutes in ice-cold methanol then washed by ;nrllh~t;nn in PBS ~or 30 minutes. Mouse anti - cytokeratin 10 or anti- involucrin m~b diluted in PBS-0.596 BSA was added and the coverslips were ;ncllh~t~rl at 4C for 1 hour. After three washes, bound primary antibody was detected using fluorescein-conjugated sheep anti-mouse Ig (Amersham International) . The coverslips were mounted in Hydromount: glycerol (1:1) and ~y;im;n~tl for green fluorescence using a 2eiss Axiovert 100 microscope.
Resul ts Eis~lo~irAl examination of regressing tumours.
& E ~tainin~
We have shown previollsly that when athymic mice bearing xenografts of the breaFt carcinoma MDA-MB 468 were treated from day 0-18 with a total dose of 44011g of ICR16 (IgG2a) half of the tumours regressed completely but small static nodules persisted at the remaining sites when the experiment was terminated on day 100 (3S) . Xowever, the present experiments showed that similar treatment with m~b ICR62 (IgG2b) resulted in complete eradication of all tumours. Histological examination of H&E stained sections of the tumour nodules remaining f ollowing treatment with ICR16 showed that a few areas cnnt~;n~d apparently viable tumour cells amongst the largely necrotic zones. The necrotic cells also showed substantial loss of cytorl ~Fm; c staining following treatment with ICR16 compared with the SUBSTITIJTE SHEET tRULE 26) , . . , . , , . . , , .. _ _ WO95/20045 . 2 1 80834 E~,LI-,.,.~.'C_il8 progressively growing tumours treated with control ant ibody .
We have also P~m;nP~ HN5 tumour xenoyrafts following treatment with one, or a cnm~;n~t;on of two anti-EGFR mAbs that bind to two distinct epitopes on the PYtPrn~l domain of the human EGF receptor. When athymic mice bearing HN5 xenografts were treated from day 7-24 with a combination of ICR62 plus ICR64 (total dose 2.2mg/mouse), on termination of the experiment ~day 79) complete control was observed at

2/10 sites and the tumour nodules Ll ;n;ng were still undergoing regression (29). Histological examination of these tumour nodules showed the presence of few if any viable cells and the lesions were composed largely of scar tissue6 compared to tumours that were growing in animals treated with the control antibody.
Similarly, when HN5 tumour xe~ografts were treated from day 5-24 with antibody ICR61, 2/8 tumours had disappeared completely when the experiment was terminated on day 75 ~29) . Examination of the nodules rPm~;n;ng also showea that few if any viable cells were present and the lesions were similar in appearance to those described above.
An ~ & E stained section of an HN5 nodule Ll ;n;ng after treatment from day 7-24 with a combination of ICR61 plus ICR62 ~total dose 2 . 2mg/mouse) showed that zones of keratinisation and apparent squamous differentiation were present. Indeed, keratinized areas could be found in all of the tumour lesions L~ ;n;ng in mice that had undergone treatment with the specif ic antibodies used either alone or in combination . These f indings suggest that squamous dif f erentiation of HN5 was ~an important consequence of treatment with EGFR- specif ic ~nt; hgfl; es .
T. .. ~ lnq~ st~i~ina An integral part of this study was to determine if any viable tumour cells l~ ; nPd af ter treatment with anti-EGFR
antibody that t t~nt;nllPrl to express high levels of the EGFR
or if loss of this antigen could contribute to their escape. While none of the rat mAbs used in this study bind to the EGFR in formalin-fixed paraffin embedded sections, they are all effective in staining the - ` LCLLIt:S
of cryopreserved tissues. Strong membrane reactivity was obtained when frozen sections of HN5 tumour xenografts were stained with specific antibody. No staining was obtained if treatment with specific antibody was omitted or when the sections were pretreated with control antibody.
We have shown previously that when athymic mice bearing xenografts of the HN5, A431 or MDA-MB 468 tumours were treated with antibody ICR62, the tumours regressed more rapidly, (and in most cases completely), compared with the same tumours treated with mAbs ICR16 or ICR64 (39) . We have carried out an i ~h; stochemical investigation of SUBSTITUTE SHEET (RULE 26) W095120045 2 1 80834 F~,l, .v.s,r 118 regressing tumours to determine the reasons f or the greater ef f ic~iency of antibody ICR62 in vivo . First, athymic mice bearing HN5 xenograftfi were treated for a short period ~days 0-4) with either ICR16 or ICR62. On day 7 tumours were excised, cryosections were prepared and stained with peroxidase conjugated F(ab')2 rabbit anti-rat IgG. The uniform staining of t11e tumour cell membranes with the second antibody, shows that all of the tumour cells were coated with specif ic rat antibody at this time . The total area of tumour remaining was smaller in the mice treated with ICR62 (mean tumour tl; -t,~r 27.591i of control) however, compared with those treated with ICR16 (mean tumour diameter 8696 of control) . Also, the ICR62 treated tumours showed a more extensive host cell infiltrate surrounding the remaining viable tumour cells.
Finally, we have l~x;qm; nf~1 the residual nodules at day 101 after mice bearing established xenografts of the HN5 tumour were treated with ICR64 from day 7-24. Despite the fact that the last treatment with antibody was given 77 days previously, staining of the sections with peroxidase conjugated anti-rat Ig showed that significant amounts of rat m~b 1 ; nf~rl associated with the dead cells and the keratinized areas where tumour destruction was complete.
Since the second antibody used was a F (ab' ), preparation that had been pre-adsorbed against mouse Ig the staining could not have been due to non-specific binding to mouse Ig or Fc-receptor bearing cells. In contrast, the small areas of viable tumour did not stain, suggesting either that ICR64 had not reached ~hese locations or that it had been lost from the cell membranes during cell proliferation. To ~rmi nf~ if these cells still overexpressed the EGFR, t nncen~t;ve sections were treated first with ICR64 and then with the peroxidase conjugated second antibody. Here, the nests of viable cells stained positively suggesting that loss of antigen was not a significant factor in the escape of these cells from antibody treatment.
}~N5 cells treated with an~;ho~ to the EGFR undergo t~;nAl differentiati~n.
The finding of keratin whorls in tumour nodules L. ;n;ng af ter antibody treatment was of particular interest . We have shown previously that, at rnn~pntrations above 5nM, m~bs ICR16 ICR62 and ICR64 inhibit completely the growth of HN5 celis cultured in medium ~nnt~ining 2~6 FCS (25,26) .
To investigate the possibility that terminal differentiation could be a pathway for tumour cell inactivation we have ~lDt~rmi nf~d the cell cycle characteristics of gro~th arrested cells and ~ m;n~rl them for expression of the differentiatio~ markers involucrin (31-33) and cytokerati~l 10 (34,35).
Growth arrested ~5 cells a- ~te in G0/G1.
Flow cytometric analysis of nuclei following treatment of SUBS 111 ~ITE SHEET tRULE 26) Wo 95/2004s 2 1 8 0 8 3 4 ~ . 118 XN5 cells with either antibodies to the EGFR (156nM ICR16 or ICR62) or with the ligand EGF (lOnM) is illustrated in Figures 1 and 2. After treatment with doses of anti-EGFR
m~bs which inhibited completely the growth of XN5~ cells, the numbers of cells in S and G2 /M had decreased substantially compared with the controls growing in medium alone, and most of the cells were arrested in GO/G1.
Treatment with EGF at a ~hnrontration (lOnM) which also inhibited (but not completely) the growth of XN5 cells also resulted in a decrease in the percentage of cells in both S and G2/M phases of cell cycle but to a lesser extent than f ollowing treatment with the m~bs . We did not see any evidence for DNA fragmentation in a pre-G1 peak indicative of apoptosis (figure 1).
Growth arrested ~5 cellg svnthesize m:~rkers of tGrm;nAl dif f erenti~tion .
When XN5 cells were incubated for four days with 156nM of the anti-EGFR mAbs most of the cells were found to express the differ--nt;~t'nn marker cytokeratin 10 as visualised by ;mmllnoflu~rescence with m~b RKSE-60. The proportion of positive cells varied with the anti-EGFR antibody used, the most effective being ICR64, and by day 4 the majority of the treated cells were strongly positive for this differPnt;At;--n marker. In addition the majority of cells expressing cytokeratin 10 were larger than the cytokeratin negative cells. Similar results were obtained when the treated XN5 cells were stained with antibody to another differentiation marker, involucrin. ~ ~
Essentially identical results were obtained when X~5 cells were treated with ICR16 or ICR62 and P~Am; nPd f or CK10 and involucrin expression (data not shown). Xowever, neither differentiation marker was expressed by cells that had been grown to near confluence in medium alone or with control antibody although these cultures c~ntA;nPh approximately ten times more cells than in the wells treatea with antibodies to the EGFR.
Further inve~tiqation of the ~ h;lln;-l~n of action of the ~Inl ;ht~ n~l their Fab Ez ~ tc Antibody ICR62, which binds to epitope C on the EGFR
inhibits (a) the binding of liyand to the receptor, (b) the EGF and TGF~ induced proliferation of human fibroblasts and (c) the growth of a range of human tumour cells that uv~L~I~ss the EGFR (26) . Antibody ICR9 binds to epitope A on the EGF receptor, and (a) increases the binding of ligand (EGF and TGFo~) to the EGF receptor, (b) PnhAnrp~ the EGF induced proliferation of human fibroblasts, and (c) stimulates the growth of tumours over-expressing the EGF
receptor (25). Xere, our aim was to discover if the monovalent Fab fragments o~= these m~bs were as effective as the bivalent m~bs in inducing these ef f ects or i~
bivalent antibody was essential for their activity SUBSTll'lJTE S~IEET (RULE 26) _ _ _ _ _ w095120045 21 80834 r~ [lls ~2,3,5,7,44,45) .
Methods Re_erring to Figures 3 and 4; monoclonal antibodies ICR9 and ICR62, and their Fab fragments, were prepared as described previously (25,26,36) . Fab fragments of ICR9 and ICR62 were prepared by papain digestion and purified by gel and affinity chromatography. Analysis by SDS-PAGE indicated t~at the Fab preparations were free of cnntAm;nAting intact antibody (data not shown) . Human rec inAnt EGF and TGFa (Collaborative Research, Waltham, Mass) were labelled with Iodine-125 (Nal2sI, Amersham International) using IODO-GEN, to a speci_ic activity of 10~Ci/llg as described previously (25). A competitive RIA was used to determine the effect of the mAbs and their Fab fragments on the binding of l2sI-EGF
and l2sI-TGFa to the EGFR orl EJ cells as described (26) . The rat m~bs 11/160 and ALN/11/53 directed against an irrelevant antigen (27) were used as controls.
Referring to Figs 5 and 6; l2sI-EGF was prepared as described above. EJ cells were seeded into 96 well plates (2x10 cells/well) in DMEM contA;n;nrj 1096 FCS. Following 48 hours ;nrllhAt~on at 37C, the cells were washed twice with ice cold binding buffer (DMEM, 15mM Hepes rnntA;n;ng 0.196 BSA) and ;nr-lh~ted with 20111 of ICR9, Fab ICR9, ICR62 or Fab ICR62 (2011g/ml) or medium alone for 30 minutes on ice prior to addition of 30~1 of l2sI-EGF (0.86-104nM) . Following a further 5 hours ;ncllh~t;nn on ice the cells were washed three times with binding buffer lysed and the bound radioactivity was determined. Non-specific binding was corrected by subtracting the counts bound in the presence of a 100-fold excess of cold EGF and was less than 6% of the total counts bound. Each value is the mean of triplicate samples.
Referring to Figures 7 and 8; HSC-l, HSC-2, HSC-4, Ca9-22, LICR-LO~-HN5 and LICR-LON-HN6 cell lines (6,9) were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10~ foetal cal~' serum (FCS) and the antibiotics penicillin, streptomycin and neomycin. To investigate the effect of treatment with intact m~bs or their Fab fragments on the growth of the tumour cells in vitro about 5x103 cells in 100~Ll of DMEM rnnt~;n;ng 2% FCS were seeded into each well of a 96-well plate. After incubation for 4 h at 37DC, 100~1 aliriuots of dilutions of m~bs or Fab fragments were added to triplicate wells and the cultures were incubated at 37C. When the cells in the control wells (cnntAin;ng medium alone) were nearly confluent, all cells were fixed, stained with methylerle blue and the A62~ det~rm;n~ as described previously (26) .
Refe~ring to Figure 9; DE532 cells (Flow Laboratories) seeded at 4 x 10i cells/ml in EMEM-10% FCS into 24-well plates were grown to cnnfl ~lonr~ then the medium was replaced with EMEM-1~6FCS. After 48 hours in this medium, SUBSTlTUrE SHEEr (RULE 26~

W095120045 : ' 2 1 8 0834 P~ 8 50~1 alir~uots of mZ~b or Fab ~ragment (25~g/ml) and/or TGFu (5ng/ml) were added to triplicate well6 and the cells were incubated overnight at 37C then pulsed for 6 hours with 2~Ci/well of 3H-thymidine. The acid insoluble radioàctivity incorporated into DNA was determined in a liquid scintillation counter.
Photomicrographs of HN5 tumour cells were taken followin treatment for four days with 156nM of ICR62, ICR62 Fab fragment or control antibody 11/160 (27), and viewed alternatively by phase contrast illumination or by immunofluorescent staining for expression of cytokeratin 10 .
5 x 10~ HN5 cells were plated onto glass cover sli s and placed in 24 well plates cnnt~in;ng lml DMEM-29d FCS7well.
Following overnight incubation at 37C, specific or~control antibodies (25~1g/ml) were added to the cultures and the cells were incubated for a further four ;days at 37C.
After two washes with PBS, the cells were fixed for 5 minutes in ice-cold methanol ther. washed by inrllh~t;on in PBS for 30 minutes. Mouse anti-cytokeratin 10 (RKSE-60 EuroPath Ltd Cornwall) diluted 1/40 in PBS-0.5% BSA was added and thé coverslips were ;nrllh~t,~fl at 4C for 1 hour.
Af ter three washes, bound primary antibody was detected using fluorescein-conjugated sheep anti-mouse Ig (Amersham International Ltd). The coverslips were mounted in ~ydl~ t:glycerol (1:1) and ~m;n~fl for ~ green fluorescence using a Zeiss Axiovert 100 microscope Result8 When the Fab preparations of ICR9 and ICR62 were tested f or their ef f ects on ligand binding they were f ound to enhance (ICR9) or inhibit (ICR62) the binding of both li;I-EGF and I-TGFu to the receptor on the bladder carcinoma cell line EJ with similar effiripnr;~o~ to the bivalent parental IgGs.
At the same r~nc~ntration, a control rat antibody (ALN/11/53), directed against an unrelated antigen (27) was without effect on the binding of ligand. Scatchard analysis of these data showed that treatment of EJ cells at 4C with monovalent or bivale~t ICR9 promoted the:binding of l2;I-EGF to the receptor by increasing the affinity of the receptor f or ligand ( Figure 5 ) . The Kd f or binding of l2sI-EGF was 2.1 x 10~ M~l in medium alone and 7 x lo6 M-l or 1. 8 x 109 M-l in the presence of bivalent or monovalent ICR9 respectively. On the other hand, monovalent and bivalent ICR62 inhibited the binding of EGF completely (Figure 6).
We then r~mr~red the monovalent Fab fragments with the parental bivalent antibodies for their effects on the growth of the head and neck carcinoma cell line HN6, which expresses about 1. 2 x 106 EGFR/ce~l (26) . The results illustrated in Figure 7 show that at concentrations above 0.6nM both the Fab fragment and intact ICR9 stimulated the growth of HN6 cells. The Fab fragment of ICR62, llkewise, SUBSmUTE SHEET (RULE 26) W095/20045 21 8 0834 r~ 8 was found (Figure 7) to be as effective as the bivalent m~b at inhibiting the growth in vitro of XN6 cells. Moreover, the monovalent ICR62 inhibited the growth of f ive other head and neck tumour cell lines (XN5, XSC-1, XSC-2, XSC-4 and Ca9-22) ~6, 9) which I-)Vt'L~ SS the BGF receptor ( f igure 8 ) . Like the intact ICR62, monovalent Pab fragments of this antibody also inhibited completely the TGFa~ induced proliferation of quiescent human foreskin fibroblasts (Figure 9) . Furth~- ~ t`l the Fab of ICR62 was as effective as the bivalent antibody in ;nrlllr;n r the differ~nt;Atinn of XN5 cells. The finding that Fab fragments of the rat m~bs ICR9 and ICR62 were as effective as the intact molecule in ; n~91lr1 ng the respective changes in the function of the EGF receptor was unexpected since other investigators had shown that Fab fragments of murine antibodies were either poorly active or ineffective compared to the intact antibody (3,5,7,8,44,45).
Discus~ion The results of our studies (Part 1 above) have shown that the nodules l~ ;n;ng at up to 82 days following the last treatment with antibodies to the EGF receptor consisted largely of necrotic and keratinized areas with few viable cells present. In another study (37) it was reported that treatment of an established human colorectal tumour xenograft with high doses of the mouse anti-EGFR m~b 225 (2mg twice weekly) resulted in complete regression of these tumours by 2-3 weeks (37). This study also showed that by day 7 most of the cells were necrotic and after 14-21 days most tumour cells had been replaced by connective tissues (37) . In the present investigation, the presence of keratinized areas was of particular interest and this f inding pointed to a previously undescribed ef f ect of treatment with antibodies to the EGFR; namely that prolonged receptor blrrkA~l~ can induce terminal differentiation of the squamous carcinoma cells. We have investigated this function of antibodies to the EGFR by screening the anti-EGFR m~b treated cells for expression of the terminal differentiation markers involucrin, a 92 kD
cytoplasmic precursor of rnrn;f;~tl envelope (31-33) and cytokeratin 10 expressed during terminal differlont;At;rn in squamous epithelia (34-35).
We haYe shown that during treatment with antibodies to the EGFR the XN5 cells were undergoing terminal differentiation and the majority of the treated cells expressed the differentiation markers involucrin and cytokeratin 10.
Furthermore, we also noted that cells expressing these differ~nt;At;rn markers were of larger size as previously shown for cultured humall keratinocytes (32). The results o~
our flow cytometric analysis of XN5 cells following treatment with anti-EGFR m~bs showed that, as would be anticipated for differentiated cells, they were arrested in the G0/G1 phases of the cell cycle. Taken together these results suggest, f or the f irst time to our knowledge, that SUBSmUTE SHEET (RULE 26) Wo ss/2004s 2 1 8 0 8 3 4 P.~ 118 antibodies against the EGF receptor which block growth factor-receptor interactionr may inhibit the growth~of BGFR
overexpressing tumours by in~ ;ng differentiation. In agreement with our f indings, Rodeck and colleagues have shown that growth inhibition of A43 1 cells by a mouse antibody to the EGFR (m~b 425 ), results in a decrease in the percentage of cells in S and G2/M phases of the cell cycle and a increase in the percentage of cells in=.G0/G1 phases of the cell cycle (20). A similar mode o~ action was also suggested recently by Bacus and colleagues (38) for the effect of growth ;nh;h;tr~ry antibodies directed against the external domain of the HER-2/c-erbB-2 receptor, which like EGFR belongs to the type I growth factor receptor family (4). This study also provided evidence for a relationship between the in vivo antitumour actiYity of anti-HER-2/c-erbB-2 m~bs and their capacity to induce differentiation of breast cancer cells in vitro (38).
Two other interesting findings came from the ; rlh; stochemical staining of the tumour specimens.
Firstly, it was clear that the increased rate of tumour regression seen during treatment with ICR62 (IgG2b) was due to rapid loss of tumour cells and, in addition, more infiltrating host cells were observed ffull~)ullding the tumour foci. As discussed above, although not the most effective inhibitor of growth in vitro, antibody ICR62 was the most effective inducer of the regression of three different human carcinomas grown as xenografts in ~athymic mice (39). These results point to a role for host immune effector functions in vivo. Indeed, it is well do~1lm~ntP~
that rat antibodies of the IgG2b isotype, like murine IgG2a and human IgG1, are the most effective at mediating ADCC
with Fc-receptor bearing effector cells and in activating the complement cascade (20, 40) . Secondly, we were surprised to find that in the nodules Ll ;n;n~ following treatment of the HN5 tumour with ICR64, cell mem.branes in the necrotic areas still retained antibody that had been given 77 days earlier. This unexpected finding po;ints to the stability of the immune complexes formed and is consistent with the results of experiments in vitro which indicate that the immune complexes f ormed on binding of these antibodies to HN5 cells are stable and are not rapidly internalised or shed from the cells . The~ small num.ber of viable cells remaining, which did not stain with the anti-rat Ig reagent had not lost their expression of the EGFR since they wére positive when restained using ICR64 as the first reagent. We conclude that antigenic modulation was not a significant factor in the es=cape of viable cells from antibody treatment and, providing antibody access was satisfactory, these cells might be susceptible to further treatment with antibody.
Conclu io~s. . =
On the basis of our data we conclude that the best antibodies for clinical application will be those which are SUBS~TUTE S~IEET (RULE 26) Wo gsl2004~ 2 1 8 0 8 3 4 F~~ 18 directed at the correct epitope and are also the mos~
effective at intlll~;n~ terminal differentiation and recruiting and activating host immune effector functions ~43 ) -The results of the further experiments (Part 2 above) showed that, af ter f ou~- days incubation in the presence o$
156nM monovalent or bivalent ICR62 the majority of HN5 cells expressed the terminal differentiation marker cytokeratin 10 (34) . Control cultures grown in the absence of specific antibody did not express cytokeratin 10, although by f our days the cell monolayers were almost conf luent .
The f inding that blockade of the EGF receptor by ICR62 Fab can induce the terminal differPnt;~ n of HN5 cells is of some interest and may have clinical application. To our knowledge this is the f irst report to show that smaller molecules such as the monovalent Fab fragment of an anti-EGFR antibody can be as effective as bivalent antibody in (a) blocking ligand binding, (b) preventing the growth of tumours uv~L~Lt:ssing the EGF receptor, (c) directing term.inal differentiation in such ce~ls. The possibility that smaller molecules than intact antibody can be employed to induce these effects could be significant for clinical application since the rate of extravasation of the smaller molecules should improve uptake into tumour, although clearance f rom the blood will be increased .
The r---h~ni Fm by which EGF activates the EGF receptor has not yet been resolved and two models (intramolecular or intermolecular activation) have been proposed. The intramolecular model suggests that binding of EG~ to the EGF ~ receptor induces a conf ormational change in the extracellular domain of a receptor monomer which is propagated across the plasma ' cl~e to the cytoplasmic kinase domain leading to its activation (46,49). The intermolecular model (2,3,48) on the other hand, proposes that receptor dimeri zation following ligand binding triggers phosphorylation by the cytoplasmic tyrosine kinase. On the basis of the second model it was expected that bivalent antibody would block activation by preventing receptor dimeri2ation (3,7,8,44). However, the results of a recent investigation using antibody 13A9 (47), whose activity has been sho~n to take place at a distance from the EGF binding site s~ggest that this is not the case. In this study, Carraway and Cerione (47) have shown that treatment of A431 cells with this antibody prevents EGF
receptor aggregation f ollowing binding of EGF but does not prevent activation via the receptor as judged by several parameters including receptor autophosphorylation, EGF-stimulated changes in cytosolic free [Ca'7~ and mitogenesis. They provide convincing evidence to suggest that conformational changes in the receptor, not receptor aggregation, is the signal generated following ligand b inding .
SUBSTllrUTE SHEE r (RULE 2~i) WO s~/20045 2 1 8 0 8 3 4 r~ 8 Our results would support this view. With bivalent and monovalent ICR9, it is clear that binding of this antibody causes a marked increase (three to nine-fold) in the affinity of the receptors for EGF. Since these experiments were carried out at 4C it is unlikely that this result could be due receptor aggregation. Furthermore, ~because ICR9 competes with neither the EGFR ligands nor the antibodies that block ligand binding we conclude that it acts at a distance i.e. it induces a conformational change at the ligand-binding site. On this basis we propose that, some at least of the antibodies which prevent ligand binding do so by altering the conformation of the receptor to reduce its affinity for ligand. This concept is leant support by the observation that only some of the rat antibodies which block binding of ligand to the EGF
receptor are able to displace bound ligand i . e some of the antibodies must act at a distance from the ligand binding site .
The possibility that scFv or constrained peptides based on the complementarity detPrm;n;n~ regions of immunoglobulins can be developed f or ef f ectively targeting the EGFR on squamous cell carcinomas is an exciting prospect especially if such drugs can induce tPrmin~l differPnti~tinn~ of the tumour cells. The possibility also exists for developing non-peptide mimetic molecules, which mimic the action of these binding proteins or peptides, and have a similarly effective interaction with the EGFR. The anti-EGFR
antibodies herein can be used to screen for such non-antibody molecules which have a similar effect on :tumour cells .
The accompanying sequence listings show within square brackets the complementary detPn[nininr~ regions (CDRs) for the heavy and light chains of ~ntiho~iies ICR62 and ICR64, together with their surrounding fL ..~lh regions (FRs).
These CDRs can be grafted intD the FRs of human antibodies, optionally with retention of selected murine residues in the FRs, according to well est~hli~llptl principles discussed above(see for example Winter GB-A-2188638, EIarris et al WO92/04381 and Queen ét al, PNAS (1989), 86:10025-33) to create.corresponding h~ niqPtl ~3nt;ho~1;P-q which may be more C-~L)LU~Liate for the treatment of humans.
~1 ;n;rJ~l Study IJsing ICR62 Mrn-lrl~-n~-l Antibody For clinical application, it is important to recognise that the EGF receptor is also expressed on some normal human tissues but the level of exp~ression is much lower :than on the corresponding tumour cells (9-11). However, in our experience tumours which express 105 or less receptors are much less sensitive to antibody treatment than are the EGFR
overexpressing tumour cells (1,26) . Indeed, Mendelsohn and colleagues in a preclinical study with rll;mp~n7ee5, have reported that treatment with a total dose of 650mg of antibody produced no toxicity (41) . In addition, in a ~ W095/20045 2180834 P.~ 51~ 118 hase I clinical trial where rAtiPntc with lung carclnoma were treated with do~es of up to 300mg of the mouse antibody 225 they observea no untoward effects in the patients (41,42).

m~b ICR62 was then tested in a Phase I clinical study in p;':lt; PntC with cancer of the head and neck ~11 pts) or Lung ( 9 pts ) cancers . A maj ority of the patients in this study had undergone previous treatment (i.e, surgery, radiotherapy, chemotherapy, and/or~ immunotherapy) . The aims of this clinical study were to (a) monitor patients for any possible signs of toxicity, (b) investigate if therapeutic antibody had localised specifically to tumour cells in metastatic lesions, (c) tl~tPrm;nP the level of ICR62 rPm~;n;n~ in circulation following treatment and (d) discover if the patien~ s mounted a human anti-rat antibody (HAR~) response.
Materials and Methods Patient Selection and 'rreatment Patients were c~nf~ Pl-ed eligible for inclusion in this trial who had i) inoperable histologically or cytologically confirmed diagnosis of sguamous cell carcinoma of the lung or head and neck, ii) immunocytoAhPm; f'Al 1 y proven expression of the EGF receptor, iii) an ~COCG performance status of 0-2, iv) no known history of allergy or atopy, v) no immunological therapy within the previous 4 weeks, vi) no significant abnormalities of renal, hepatic, or bone marrow function ( haemoglobin, 10g/dl, white count, 3x10'/1, platelet ~ 120, creatinine < 130, liver enzymes and bilirubin < x2 normal ) .
Details of tumour grade and the previous treatments that the patients had received are summarised in Table l Patients were skin tested (lOug intr~Pr~~lly) lhr before antibody was administered and no patient was found to give an adverse response. Antibody in phosphate buffered saline pH 7.4 was given intravenously as a single bolus injection over a period of 30-60 minutes. Groups of three patients were treated with 2.5, 10, 20, or 40mg of ~CR62 and a further eight patients were given 100mg of this antibody.
All patients were evaluated for signs of toxicity. Also, blood samples were taken from these patients before and at inter~als af ter treatment with ICR62 so that the serum could be monitored f or the level of m~b ICR62 in circulation and tested for the presence of human anti-rat antibodies (HARA) . In some of the patients, given antibody doses of 40mg or 100mg, biopsies were taken from accessible metastatic lesions, 24 hours following treatment with ICR62 and Pl~minPd for the localization of ICR62 to the tumour cells .

w0 95/20045 2 1 8 0 8 3 4 r~ 118 Preparation of mAb ICR62 All antibody for clinical use was prepared according to the g ~ l in~s prepared by the CRC-MRC Jolnt Committee.
Hybridoma cells were grown in Dulbecco's r ';f;P~l Eagle's medium (DMEM) rrntAln;ng 396 or 5% fetal calf serum of North American origin and antibiotics a6 described previously (27), in either in a Verax Type 1 Bioreactor or as bulk cultures in roller bottles. Supernatants were harvested under aseptic conditions and then precipitated with (NH4) 2SO4 at 45% saturation. Us~ng autoclaved reagents, column packings and rnntA;n~'rs, the precipitates were dissolved in water and dialysed against 0 . 0175M phosphate buffer pH 6.6. After centrifugation in a Beckman 45Ti rotor at 30, 000g to remove insoluble '''tor; Al, the dialysate was fractionated by passage through a column of Whatman DE52 cellulose equilibrated with and eluted with 0 . 0175M phosphate buffer pH 6 . 6 . The flow through fractions rnntA;n;ng the purified (~95~ m~b) ICR62 were bulked and dialysed against five changes of sterile ~_ phosphate buffered saline (PBS) . After filter ster; 1; ~Ati nn the preparation was aliquoted, frozen and stored at -20C until use.
Determination Of the level of ICR62 in Circulation The amount of ~ free m~b ICR62 present in serum was det~rm;n~9 by its ability to inhibit the binding of EGF or TGF to the bladder carcinoma cell line EJ as described previously (27). Doubling dilu2t~Lons of serum (50/11) 2w5ere mixed with an equal volume of 1 I-EGF (4xlO4cpm) or 1 I-TGF (4xlO4cpm) . Standards rnntA;n;ng known rnnrF~ntrations of ICR62 were set up in the same way. Aliquots of 90~Ll of each mixture were then transfered to monolayers of EJ cells grown to confluency in 96 well plates. After ;nrllhAt;nn for one hour on ice the cells were washed three times, then lysed in lM NaOH rnntA;ning lg~ sarkosyl and the bound radioactivity was determined in a Hydragamma spectrometer (Oakfield Instruments Ltd., Oxford).
Determination of Human Anti-Rat (HARA) ~pnn~e Polyvinyl chloride 96-well plates (Dynatech Labs, Virginina), were coated with rat antibody by incubation overnight at 4C with 50111/well of a stock sloution~
(10~g/5ml of PBS) of. ICR62, ICR62 Fab, or ICR62 scFv. The plates were washed three times with PBS cnntAininr 0.5~ BSA
and then incubated for 2 hours with 200111/well of PBS-0.596 BSA to block the remaining sites. After a further three washes with PBS rnntA;n;ng 0,596 BSA, ~lmlhl;nj dilutions of the pAt; Pntc sera made in PBS-0 .5~ BSA were added in duplicate to the wells and the plates were i nrl~hAtGcl for 1 hour at ambient temperature . Af ter washing the plates three times with PBS-0.5~6 BSA, human antibodies bound to ~C~62 or its fragments were detected by the addition of l~I-labelled rabbit anti-human F(ab')2 After incubation wossnoo~s 21 80834 r~l. . 118 for 1 hour at ambient temE~erature, the plates were washed three times with PBS- 0 . 5% BSA and then cut into individual wells and the bound radioactivity determined in a Hydragamma coun~er.
T nh; qtnrh~m; ~try All ; lnnhl qtochemical studies were performed by the indirect method using sheep ;lnt;hoA;es to rat F~ab')2 conjugated to horseradish peroxidase (Amersham Tnt~rn~t; onal) .
To determine expression of the EGFR, tumour biopsies were snap f rozen in lirluid nitrogen then mounted in OCT medium and sections Of 511m thickness cut. One section was stained with ~aematoxylin & Bosin (F[&E) to determine the nature of the tumour sample . A second sample was f ixed in acetone at 4C for 10 minutes, then after washing briefly in PBS the section was incubated with m~b ICR62 (lOO~lg/ml, for 1 hour Following washing in PBS for 5 minutes, the sections were incubated with a 1:100 ~l;lllt;nn of sheep anti-rat F(ab')2 conjugated to horse radish peroxidase (Amersham) for 45 minutes. Peroxidase staining was demonstrated by incubating the sections for 10 minutes in a solution cnnt~;ning 0.0596 ~ m;nnhenzidine (Sigma), 0.1~ hydrogen peroxide (Merck) and 0.07~ imidazole (Merck) . After washing in running tap water f or 5 minutes, the sections were counterstained in Mayer' s Haematoxylin (HD Suppliers) for ~30 seconds. Finally the sections were dehydrated, cleared and mounted.
To monitor the lor~l;c~t;nn of ICR62 to tumour cells in patients that were given doses of 40mg or 100mg of antibody, biopsies were taken from metastatic sites, 24 hours after the anti}~ody was administered, then frozen sections were cut and stained with sheep anti-rat antibody alone. To determine the proportion of cells that had bound therapeutic antibody serial sections taken f rom the same biopsies were stainéd first with ICR62 followed by the peroxidase conjugated sheep anti rat F (ab~ ) 2 so that all EGFR expressing cells were stained.
Resul ts .
Ef~ect o~ treatment on patients Twenty patients with so,uamous cell carcinoma of the head and neck or lung whose tumours were found to overexpress the EGF receptor wçre recruited into this dose esr;ll~tinn trial to assess the toxicity of thç rat m~b ICR62 (Table 1 ) . All had extensive disease and, with the exception of four~ patients (see Table 1), had received previous treatment (s) for their disease e.g. surgery, radiotherapy and/or chemotherapy (Table 1).
Patients were admittecL to hospital on the day before the w0 95/20045 , 2 1 8 0 8 3 4 P~ l8 administration of antibody and their biochemical and haematological parameters were monitored Following intravenous inj ection of the antibody some of the pat ent PYhih;tP~l mild rigors and fever or hypotension (Table 1).
The symptoms which were not related to the dose of antibody administered were readily controlled. In no case were any severe toxicities observed. The p~t1 PntS were released from hospital 24 hours after the injection of antlbody and they were seen at weekly intervals f or up to slx weeks .
None of the patients reported any untoward effects of their treatment during the follow-up period.
Blood levels of ICR62 We have investigated if mAb ICR62 could be detected in the serum of patients given different doses of this antibody by monitoring the inhibition of binding of EGF or TGF to the EGFR by sera taken at intervals following treatment. We have been unable to detect any free antibody in the serum, taken at 4 hours or 24 hours post-treatment, of oatients that were given 20mg or less of ICR62 (Pigure la) .
However, significant amounts of circulating ICR62 could be detected, in the blood of patients given doses of 40mg or lOOmg of ICR62 (Figure lb & lc) . In addition, the level of ICR62, in;ng in circulation was found to be highest in patients recieving lOOmg of ICR62. In one patient about half o~ the antibody administered was in circulation on day 3 but no f ree mAb could be detected in the blood at 7 days post ICR62 treatment.
In; ected ICR62 binds to the EGFR on tumour cell me~branes Biopsies were taken of metastatic lesions f rom six ratiPnt~ 24 hours after treatment of the p~t;Pnt~ with 40mg or iOOmg of ICR62. Frozen sections were P~---;nPd for the presence of cell bound antibody by staining directly with a peroxidase-conjugated second antibody reagent. Two of the biopsies were found to consist largely of necrotic material and were discarded but four yielded well defined regions cf~nt~in;ng infiltrating tumour. The sections illustrated in figure 2a show that 24 hours after treatment of two p~t;ont~ with 40mg of ICR62, the antibody had localized to the membranes of tumour cells in the metastatic sites and the proliferating cell6 at the periphery of the tumour were strongly stained. When a sequential section was stained with ICR62 prior to the addition of the sheep anti-rat reagent, it was found that cells at the interior of the tumour which also expressed the EGFR had not bound the therapeutic antibody administered 24 hours earlier (figure 2b) . ~owever, when frozen sections of biopsies obtained from metastases of two patients treated with lOOmg of ICR62 were Pl~m; nPr~ 7 it was clear that the m~b had penetrated further into the metastatic lesions (Figures 3a & 3b) . Again, very good membrane staining o~ the proliferating cells at the tumour periphery was observed.

WO9S/20045 21 8 0834 r~ 8 Developmert of human ~nt;ho~ q to ICR62 We have investigated the immunogenicity of m~b ICR62 in these patients by determining if human anti-rat antibodies (HARA) were present in the serum of patients given different doses of ICR62. Also, we have investigated if the human anti-rat response included anti-idiotypic antibodies by detPrm;n;n~ if the ~nt;ho~lies in the patients serum bound to scFv fragments of ICR62. Of the twenty patients treated with ICR62, human anti-rat antibodies were detected in the blood of only four patients (numbers 9, 10, 16 and 19 see Table 1).
Of these only two patients given 20mg (No 9) or 40mg (No 11) produced anti-idiotypic ~nt;hofl;Pq which bound to scFv ICR62 ( Table 1, Figure 4) . Sera from two out of the eight patients treated with 100mg of ICR62 (patients 16 and 19) crntA;nPd ~nt;ho~l;Ps~ directed against detPrm;n~nts on the constant region since they bound to the Fab and intact antibody but not to the ICR62 scFv (Table 1, Figure 5) . In one patient (No 12), the results obtained with the serum taken before treatment indicated that antibodies which bound to rat IgG were present in the sera bef ore ICR62 was administered but tests were not done to determine i~
binding was specific or due to the presence of ~to~nt;ho~l;es such as rheumatoid factor.
Discus~ ion .
Overexpression of the EGFR acrr~r;ln;P~l by production of the EGF family of ligands has been found to occur in a wide range of human malign~nr; P~:: and this ~hPnl has been correlated with a poorer prognosis in these p~t;Pntq.
During the past fourteen years a number of mouse monoclonal antibodies have been raised against epitopes on the external domain of the human EGFR and these have been used not only to investigate growth factor-receptor interaction and the mPrh~n; qm(s) of activation of the BGF receptor system but also for diagnostic and therapeutic applications in cancer. Several of the mouse antibodies have undergone clinical evaluation in Phase I and Phase II studies in patients with head and neck, lung or brain cancers including mAb EGFRl (52,53), m~bs 225 and 528 (42,54), m~b 425 (EMD 55900 E Merk, 55-59) and m~b RG83852 (60) . The aim of these studies like the one presented here, was to determine whether tréatment of cancer patients with anti-EGFR mAbs produced lif e threatening toxicities by binding to the EGFR expressed by normal tissues including liver and skin. The results of these studies have shown that mouse 5 0 antibodies to the EGF~ can be given saf ely to patients without toxicity. For example, Divigi and colleagues have treated p~t;Pnt~ with advanced sriuamous cell carcinoma of the lung with single doses of up ~ to 300mg of m~b 225, including 4mg of indium lll-labeled 225, without any sign of toxicity (42) . Furthermore, with doses of 40mg or greater they were able to image presumed sites of wo95/200~5 ` 21 80834 r~ C~ 118 metastasis greater or equal to lcm in diameter (42).
We selected mAb ICR62 for a phase I clinical study since it was the most effective of a large number of rat m~bs we had made against the human EGFR, at; n~llr; ng the regression of xenografts of head and neck, wlva and breast carcinomas grown in athymic mice (39 50 51). The results of this Phase I rl; n; r~ study has indicated that m~b ICR62 can be given safely, at doses up to 100mg, to cancer patients without any life thr~AtFn;ng toxicities. This finding is in agreement with the results of studies with mouse Ant;hor~ to the EGFR described above. Furthermore, in patients that had been given doses of 40 or 100mg o~ICR62, significant amounts of therapeutic m~b were detected in the blood of these patients at 4 and 24 hours post ICR62 treatment. Moreover, in these r~t;Pntc, good lorAli7At;nn of ICR62 was observed in metastatic lesions at 24 hours following treatment and penetration of antibody was greater in the tumours of patients treated with 10 Omg of the m~b .
Importantly, the most accessible cells, i.e. those nearest the blood supply, are likely to be the most actively dividing cells and these were the most strongly stained.
It is the prol;f~rAt;on of these cells which must be arrested. The m~;ntA;nAnr~ of receptor blockade may require repeated treatment with antibody. Certainly, our experience using the xenograft model (39,50,51) points to the need to rA;ntA;n suff;r;~ntly high blood levels for long enough to a) block EGFR ~unction, b) recruit to the tumour and activate host effector cells and c) induce t~rm;nAl differPnt;At;nn. The data on the blood half-life of ICR62 obtained in this investigation suggest that twice weekly doses of 100mg may be sufficient to n~;ntA;n a high enough level of this antibody for therapeutic activity. In the present invest;rj~t;on we were able to biopsy metastatic sites at only a single timepoint so we have no inforr-t;nn cnnr~rn;ng the stability of the antibody at the tumour cell surf ace or of the ef f ects of treatment on the recruitment and activation of host immune effector cells.
When sera were tested for anti~odies to rat; n~lnhlll;n, we found that only four out of the twenty patients had responded following a single dose of m~b ICR62 . Of ~ these, only two were directed against the idiotype of ICR62.
These results suggest that m~b ICR62 was not as immunogenic as the mouse antibodies to the EGFR that have been used previously in clinical studies. For example, Divigi and colleagues found all 19 lung cancer patients treated with a single dose of 1 to 300mg of m~b 225 developed human anti-mouse antibodies (42) . Stasieki and colleagues have also found that a single infusion or multiple infusions at monthly intervals of mPb EMD 55900 in glioma patients elicited human anti-mouse antibodies (57). On the other hand, these authors reported that following multiple infusions of glioma r;~t;~ntc: with mAb EMD 55900 at shorter ~5 intervals (three times/week, during 4 weeks or -longer), human anti-mouse antibodies were not detectable in the woss/2004s 2180834 r~ rl l8 serum of these patients (57) . The immunogenicity of anti-EGFR m~os may however, be reduced by either production of chimearic or humanized ver8ion of the i~ntiho~lies or by using differe~t antibodies to the EGFR for the second and third treatment.
In summary, the results of this Phase I dose escalation study has indicated that the rat m~b ICR62, which acts as an EGF a~d TGF antagonist, a) can be adminstered safely in patients with squamous cell carcinoma, b) localizes efficently to metastatic sites in p~,t;~ntS with squamous cell carcinoma, and ~_) may therefore be useful in the treatment of a signif i~ant number of cancer patients whose tumours overexpress the EGF receptor. From the results of our preclinical studies in athymic mice, we have noted that to ~chieve tumour destruction repeated adminstration of EGFR m~bs may be required. For this reason, a phase II
clinical study is planned using multiple doses of m~b ICR62 .

W095120045 26 2180834 r~l,. 1 118 TAB. E 1 rA7.70~ WF Ll N Aa~A7 r~ ~77 OF r~V~UI DAnor DOI~ LWC ~ DISWI.F~ DA7LOF NALA
N. srAD~ r~r5A77aN V~ DL~7NODZ 77~A711~r 77~7ur~5 ~r~AD~ DON D~
U IIN I~ IW7 L ~- I L~ .~ rl L~ U
~n~,7 L F NN 7~ I~FI IW7 Lun L~r~ ~a ID NC
F NN ~ o~a ~ U ILn Ll.~ rD . ~--U NN . ~ W~ ILLn 1~ rD . 1 : 1'"10~ v ~, Il NNn I 1,r71 Im~ ~J
r ~ D.
L ~ NN - ) ~IJ n lw) 1~ L~ 10.~ N V ~ NC 1 ~n7 11 L ~ ~ n ~ r~ rD LL~
~, u L ~ u ~.n ~ r rD L~7 . J 7 yl 1~ ~D
. F L J- ~ u~ Ar~ rD ~<.~
~C U NN ~J I L~ n ~ T ~D rD ~:~L-I
IL F NN a ~ n ~ ~L~n 41~ v ~ NC LOS~7.7 "' F~7LU~ V
~L U ~ U .-1 L~ L~ ~eq NC
7 ~D
D. u L 7~ 1-.---1 ~ ~5 I D.Li~ lo~ r r~C ~ _U A~
U ~ 7 bl NN L. I ~ L~ O~ c rD . 1~ ~L~) L a I L~.Ln ' L L4V~ ~U~ CII" m rD 7-.~L-) U NN )~~ ~L~n --rs7b7~ D . O~
w7 ~O rD L
~L U L O~ LC~LO 'n ~ ~--IO NC
F N~ L7.n ~ ~WO 111 V. IX~ _ C 7 Ar .
C ~n~ G7 N .__ ~ F L 1~ o I~U ~7 :LF~ v rD
I; U L 7~ .U ~.tU ~ . NC ~ A' SUBSmUTE SHEET (RULE 26) Wo ss/20045 2 1 8 0 8 3 4 F~ 8 Re~ere ces ~
1. Modjtahedi, H. and Dean C. The receptor for EGF and its ligands: Expression, prognostic value and target f or therapy in cancer (Review) . Int . J . Oncol ., 4, 1994, (in press).
2. Ullrich,A. and Schlessinger,J. Cell. 61, 203-212 ( 1990 ) .

3. Spaargarer"M., Defize,L.H.K., Boonstra,J. and de Laat,S.W. J. Biol. Chem., 266, 1733-1739 (1991) .

4. Prigent,S.A. and Lemoine,N.R. The type 1 EGFR-related family of growth factor receptors and their ligands.
Progress in Growth Factor Research, 4: 1-24 1992.

5. Aboud-Pirak,E., I~urwitz,E., Pirak,M.E., Bellot,F., Schlessinger,J. and Sela,M. J.Natl.Cancer.Inst., 80, 1605-1611 (1988) .

6 . Kamata, N., Kazuhi ro, C., Rikimaru, K., Horikoshi, M., Enomoto, S. and Kurokir T. Cancer Res., 46, 1648-1653 (1986) .
20 7. Sato,J.D., Kawamoto,T., Le,A.D., Mendelsohn,J.
Polikoff,J. and Sato,J.H. Mol.Biol., Med. 1, 511-514 ( 1983 ) .
8. Fan,Z., Masui,H., Altas,I. and Mendelsohn,J. Cancer Res. 53, 4322-4328 (1993) .
9. Cowley,G., Smith,J.A., Gusterson,B., Hendler,F., and Ozanne,B. The amount of EGF receptor is elevated on squamous cell carcinomas. Cancer Cells, 1: 5-10, 1984.
10 . Hendler, F.J. and Ozanne B.W. Human squamous cell lung cancers express increased epidermal growth factor receptors. J. Clin. Invest., 74: 647-651, 1984.
11. Ozanne, B., Richa~^ds,C.S., Hendler,F., Burns,D. and Gusterson, B . Overexpression of the EGF receptor is a hallmark of squamous cell carcinomas. J.Pathol., 149:
9-14, 1986.
12. Gullick,W.J. Prevalence of aberrant expression of the Pr; ~Prm~l growth factor receptor in human cancers .
British Medical Bulletin, Vol 47: No.1 87-98 1991.
13. Harris,AL, Nicholson,S, Sainsbury,R, Wright,C and Farndon,J.R. Epidermal growth factor receptor and other oncogenes as prognostic markers. J. Natl.
Cancer Inst. Monogr. 11: 181-87, 1992.
14. Sporn,M.B. and Roberts,A.B. Autocrine growth factors and cancers. Nature 313: 745-747, 1985.
15. Di Marco, E., Pie~-cé, J. H., Fleming, T. P., Kraus, M.
H., Molloy, C. J., Aaronson, S. A. and DiFore, P. P.
Autocrine interaction between TGFa and the EGF-receptor: quantitative requirements for induction of the malignant phenotype. Oncogene, 4: 831-838, 1989.
16. Yoshida,K., Kyo,E., Tsuda,T., Tsujino,T., Ito,M. and Tahara, E . EGF and TGFa, the ligands of hyperproduced EGFR in human esophageal carcinoma cells act as a autocrine growth factor. Int.J.Cancer 45: 131-~5, 199 0 .
17 . Morishigie, K . I ., Kurachi, H ., Amemiya, K . Fuj ita, Y., 5 5 Yamamot o , T ., Miyake , A . and Tan i z awa , O . Evi denc e f or the involvement of TGFa and EGFR autocrine mechanism wo95l20045 ' - - 21 8 0834 r~ l18 in primary human ovarian cancers i~ vitro. Cancer:
Res., 51: 5322-5328, l991.
18. Masui,H., Kawamoto,T. Sato,J.D., Wolf,B., Sato,G. and Mendelsohn,J. Growth inhibition of human tumor cells in athymic mice by anti-~or~ 1 growth factor receptor monoclonal antibodies. Cancer Research. 44:
1002-1007 1984.
l9. Van de Vijvert,M., Kumar,R. and Mendelsohn,J. Ligand-induced activation of A431 cell epidermal growth f actor receptors occur8 primarily by an autocrine pathway that acts upon receptors on the surface rathe than intracellularly. J. Biol. Chem., 266: 7503-7508, 1991 .
20. Rodeck,U., Herlyn,M., Herlyn,D., Molthoff,C.
Atkinson, B ., Varello, M ., Steplewski, Z . and Koprowski, X . Tumor growth ~ t; nn by a monoclonal antibody to the epidermal growth factor receptor-Immunological r~~ tPtl and effector cell-irdependent effects. Cancer Research. 47: 3692-3696, 1987.
21. Tm;ln;c:ll;,R, Yamaguchi,K., Kuranami, M., Kyo, E.
Hozumi,T. and Abe,K. Inhibition of= growth of human lung adenocarcinoma cell lines by anti-transforming growth factor-tY monoclonal antibody. J. Natl. Cancer Inst ., 81 : 220-223 , 1989 .
22. Fendly,B.M., Winget,M., Hudziak,R.M., Lipari,M.T.
Napier,M.A. and Ullrich,A. Characterization of muriné
monoclonal ;~nt;ho~ reactive to either the human epidermal growth f actor receptor or HER2 /neu gene product. Cancer Res., 50:, 1550-1558, 1990.
23. Sato,H. and Sato,J.D. Growth factor receptor monoclonal antibodies and cancer immunotherapy. J.
Nat. Cancer Inst. 81: No21 1600-1601, 1989.
24 . Mendelsohn, J . Epidermal growth factor receptor as a target f or therapy with antireceptor monoclonal antibodies. J. Natl. Cancer Inst. Monogr. 13: 125-131, 1992 .
25. Modjtahedi,H., Styles,J., Box,G., Eccles,S.
Gusterson,B., and Dean, C. Antitumour activity of rat Mabs to the human receptor f or EGF . I~, Mutant Oncogenes: Targets for Therapy? Eds. Epenetos'A.A. and Lemonie,N.R. Chapman & Hall, London pp35-47, 1993.
26. Modjtahedi,H., Styles,J.M., and Dean,C.J. The human EFG receptor as a target for cancer therapy: six new rat m~bs against the receptor on the breast carcinoma MDA-MB 468. Br. J. Cancer, 67: 247-253, 1993.
27. Dean,C.J., Styles,J.M., Gyure,L.A., Peppard,J., Hobbs,S.M., Jackson,E., and Hall,J.G. The production of hybridomas from the gut associated lymphoid tissue of tumour bearing rats. I. Mesenteric nodes as a source of IgG producing cells. Clin.exp.Immunol., 57:
358-364, 1984.
28. Huason, D. L., Weiland, K. L., l~ooley, T. P., Simon, M. and Watt, F. M. Characterisation of ~ eight monoclonal antibodies to involucrin. Hybridoma, ll:
367-379, 1992.
29. Modjtahedi, H., Eccles, S. Box, G, Styles, J. and Dean W0951~004S 21 80~34 P~ rrl18 C.J. Antitumour activity of combinations of antibodies directed against dlf f erent epitopes on the extracellular domain of the human EGF receptor. Cell BioPhysics 21/22: 1993 (in press) 30. Ormerod,M.G. Anal~sis of DNA. I~, Flow Cytometry, Ed M. G. Ormerod Oxf~rd University Press, New York, 69-87, 1990.
31. Banks-Schlegel, S and Green, H. Involucrin synthesis and tissue assembly by keratinocytes in natural and cultured human ep- thelia. J. Cell Biol ., 90: 732-737, 19 81 .
32. Watt F. M. and Green, H. Involucrin synthesis is corrélated with cell size i~ human epidermal cultures.
J. Cell Biol, 90: 738-742, 1981.
33 Walts, A. E. and Said, J. W. Involucrin, a marker of sS~uamous and urothelial dif f erentiation . An ~h; ~tochemical study on its distribution in normal and neoplastic tissues. J. Path., 145: 329-340, 1985 .
34. Moll, R., Franke, W. W., Schiller, D. L., Geiger, B.
and Krepler, R. The catalog ~f human cytokeratins:
patterns of expression in norrnal epithelia, tumors and cultured cells. Cell, 31: 11-24, 1982.
35. Lane, E. B. and Z~lPl~An~lpr~ C. M. Use of keratin antibodies in tumor diagnosis. Cancer Biology, 1: 165-179, 1990.
36. Léger, O.J.P. and Dean, C.J. Mol. Immunol. 28, 673-680, (1991).
37. Tachikawa, T., Boman, B., Mendelsohn,J., and Masui, H., Tumour cytotoxity by anti-epidermal growth factor receptor EGFR monoclonal antibody mAb.
Proc .Amer .Assoc . Cancer Res . 33: 345, 1992 .
38. Bacus,S.S., Stancovski,H., Huberman,E., Chin,D., Hurwitz, E ., Mill s, G . B ., Ullrich, A ., Sela, M . and Yarden, Y. Tumor-inhibitory monoclonal antibodies to the HER-2/neu receptor induce differPnt;At;~n of human breast cancer cells, 52: 2580-2589, 1992.
39. Modjtahedi, H., Eccles, S.A., Box, G., Style, J., and Dean C., Immunotherapy of human human tumor xenografts overexpressing the EGF receptor with rat antibodies that block gro~th f actor receptor interaction .
Br.J.Cancer 67: 254-261, 1993.
40. Bruggemann, M., ~-eale, C., Clark M., Bindon C., and Waldmann H., A matched set of rat/mouse chimeric Antihots;ps:~ J.Immunol. 142:3145-3150, 1989.
41. Mendelsohn, J., The epidermal growth factor receptor as a target f or therapy with antireceptor monoclonal antibodies. Seminars in Cancer Biology 1:339-344, 1990 .
- 50 42. Divigi, C.R., Welt, S., Kris, M., Real, F.X., Yeh, S . D . J ., Gral l a , R ., Merchant , B ., S chweighart , S ., Unger, M., Larson, S.M., and Mendelsohn, J. Phase I
and Imaging Trial of Indium 111-Labeled Anti-Epidermal Growth Factor Receptor Monoclonal Antibody 225 in ~At;PntC with Squamous Cell Lung Carcinoma. J. Natl.
Cancer Inst. 83:97-104, 1991.

Wo9s/20045 ' ~ 2180834 r~ ril8 43. Pierce, G.B., and Speers, W.C., Tumours as Caricatures of the Process of Tissue Renewal: Prospects for Therapy by Directing Dif ferentiation . Cancer Res .
48:1996-2004, 1988.
44. Schreiber,A.B., Libermann,T.A., Lax, I ., Yarden,Y. and Schlessinger,J. J. Biol. Chem. 258, 846-853 (1983) .
45. F~rnAn~ Pol,J.A. J.Biol.Chem., 260, No 8, 5003-5011 (1985) .
46. Koland, J.G and Cerione, R.A. J.Biol.Chem., 263,2230-2237 (1988) 47. Carraway, K.L., and Cerione R.A J.Biol.Chem., 268, 23860-23867 (1993) 48. Yarden,Y., and Schlessinger,J. Biochemistry 26, 1434-1442 (1987).
49. Weber, W., Bertics, P.J., and Gill G.N. J.Biol.Chem., 259, 14631-14636 (1984).
50. Dean CJ, Modjtahedi ~I, Eccles S, et al: Immunotherapy with An~;ho-lies to the EGF receptor. Int J Cancer 8 :103-107 1994.
51. Modjtahedi ~I, Eccles S, Sandle J, et al:
Differentiation or immune destruction: two pathways for therapy of squamous cell carcinomas with antibodies to the Pr; dGrmAl growth factor receptor.
Cancer Researc~. 54:1695-1701, 1994.
52. Soo KC, Ward M, Roberts KR, et al., RA-q;oi~lnoscintigraphy of s~lAm~ carcinomas of the head and neck. Head and Neck Surgery 9:349-352, 1987.
53. Kalofonos ~IP, Pawlikowska TR, Xemingway A, et al:
Antibody guided diagnosis and therapy of brain gliomas using radiolabeled monoclonal antibodies against epidermal growth factor receptor and ~l A~ l'n~Al alkaline phosphatase. J Nucl Med, 30:1636-1645, 1989.
54. Baselga J, Scott A, Pfister D, et al: Comparative pharmacology in Phase I and imaging trials utilizing anti-f~r;ti~r~Al growth factor receptor ~ ~(EGFR) monoclonal ~nt;ho-l;es (M~bs) labeled with 131I or lllIN. Proc Annu Meet Am Soc Clin Oncol, 12 :A368, 1993.
55. Brady LW, Miyamoto C, Woo DV, et al. Malignant astrocytomas treated with iodine-125 labelled monoclonal antibody 425 against epidermal growth factor receptor: a phase II trial. Int J R;l~ ;A~ n Oncology Biol Phys, 22:225-230, 1991.
56. Magdelenat ~I, Delattre JY, Mady E, et al: A phase I
study of the anti-EGFR monoclonal antibody 425 in patients with malignant gliomas. J Tumor Marker Oncol 6(4) :60, 1991.
57. Stasiecki P, K~ hF~;~fl JT, Westphal M, et al: The development of ~AMA in glioma ~At;on~: can be suppressed by administration of murine MAb in short time intervals. Proc Amlu Meet Am Assoc Cancer Res, 34 :A2835 ,1993 .
58. Bilzer T, Stassiecki P, Vega F, et al:Immunotherapy of malignant gliomas with the anti-EGFR monoclonal antibody. Proc Annu Meet Am Assoc Cancer Res, 34:A2877, 1993.

WO 95/20045 2 1 8 0 8 3 4 P~ ,,5,~ 118 59. Dadparvar S, Krishna L, Miyamoto C, Brady LW, Brown SJ, Bender H, Slizofs}~i WJ, Eshleman J, Chevres A, Woo DV: Indium-111-labeled anti-EGFr-425 scintigraphy in the tlf~tf~ti nn o~ malignant gliomas . Cancer Suppl 73:884-889, 1994.
60. Shin DM, Perez-Soler R, Donato NJ, et al: Tumor l~ri~ growt~ factor receptor (EGFR) saturation and proliferating cell nuclear antigen ~PCNA) expression in patients treated with anti-EGFR monoclonal antibody (MoAb) RG83852 (iII Phase I clinical study) . Proc Annu Meet Am Soc Clin Oncol, 12 :A940, 1993 .

W0 95/20045 2 1 8 0 8 3 4 P~ l i8 Se~Iuerlce Listi~gs ICR62 ATG GGA TGG ATC TGT ATC ATC m CTT GTG GCA ACA GCT ACA GGT GGC CAC TCC CAG GTC
Met Giy Trp lle Cys lle lle Phe Leu Val Ala Thr Ala Ti~r Gly Gly His Ser Gin Val AAC CTA CTG CAG TCT GGG GCT GCA CTC GTG AAG CCT GGG GCC TCT GTG AAG TTG TCT TGC
Asn Leu Leu Gin Ser Gly Aia Aia Leu Vai Lys Pro Gly Aia Ser Vai Lys Leu Ser Cys AAA GGT TCT GGT TTT ACA TTC ACT rGAC TAC AAG ATA CAClTGG GTG AAG CAG AGT CAT GGA
Lys Gly Ser Gly Phe Thr Phe Thr LASP Tyr Lys lle His~ Trp Val Lys Ghn Ser His Gly AAG AGC CTT GAG TGG ATT GGG TAT TTT AAT CCT AAC AGT GGT TAT AGT ACC TAC AAT GAA
Lys Ser Leu Glu Trp lle Gly Tyr Phe Asn Pro Asn Ser Gly Tyr Ser Tilr Tyr Asn Glu AAG TTC AAG AGClAAG GCC ACA TTG ACT GCA GAC AAA TCC ACC GAT ACA GCCTAT ATG GAG
Lys Phe Lys Ser ~Lys Aia Thr Leu Thr Aia Asp Lys Ser Thr Asp Thr Ala Tyr Met Glu _ CTT ACC AGT CTG ACA TCT GAG GAC TCT GCA ACC TAT TAC TGT ACA AGA CTA TCC CCA GGG

Leu Thr Ser Leu Ti)r Ser Glu Asp Ser Ala Thr Tyr Tyr Cys Thr Arg Leu Ser Pro Gly GGT TAC TAT GTT ATG GAT GCC TGG GGT CAA GGA GCT TCA GTC ACT GTC TCC TCA GCC CAA

Gly Tyr Tyr Vai Met Asp Aia Trp Gly Gin Gly Aia Ser Vai Thr Val Ser Ser Ala G~n ACA ACA GCC CCA TCT GTC TAT CCA CTG GCT CCT GGA

Thr Thr Ala Pro Ser Val Tyr Pro Leu Aia Pro Cly SUBSTITUTE SHEET (RULE 26) ~ W095/~004S 21 8 0834 r~ s ATG ATG GCT CCA GTC CAG CTC TTA GG~G CTG CTG CTG ATT TGG CTC CCA GCC ATG AGA TGT
Met Met Ala Pro Val Gln Leu Leu Gl~ Leu Leu Leu lle Trp Leu Pro Ala Met Arg Cys 90 . 120 GAC ATC CAG ATG ACC CAG TCT CCT TCA TTC CTG TCT GCA TCT GTG GGA GAC AGA GTC ACT
Asp lle Gln Met Thr Gln Ser Pro Ser Phe Leu Ser Ala Ser Val Gly Asp Arg Val Thr ATC AAC TGC AAA GCA AGT CAG AAT AIT AAC AAT TAC TTA AAClTGG TAT CAG CAA AAG CTT
lle Asn Cys Lys Ala Ser Gln Asn n~ Asn Asn Tyr Leu AsnlTrp Tyr Gln Gln Lys Leu GGA GAA GCT CCC AAA CGC CTG ATA TAT AAT ACA AAC AAT TTG CAA ACAlGGC ATC CCA TCA
Gly Glu Ala Pro Lys Arg Leu lle Tyr Asn Thr Asn Asn Leu Gln Thr 1Gly lle Pro Ser AGG TTC AGT GGC AGT GGA TCT GGT ACA GAT TAC ACA CTC ACC ATC AGC AGC CTG CAG CCT
ArB Phe Ser Gly Ser Gly Ser Gly Thr Asp Tyr Thr Leu Thr De Ser Ser Leu Gln Pro GAA GAT m GCC ACA TAT TTC TGC rT~G CAG CAT AAT AGT TTT CCC ACGlTTT GGA GCT GGG
Glu Asp Phe Ala Thr Tyr Phe Cys ~RU Gln His Asn Ser Phe Pro ThrlPhe Gly Ala Gly ACC AAG CTG GAA CTG AAA CGG GCT GAT GCT GCA CCA ACT GTA TCT ATC TTC CCA CCA TCC
Thr Lys Leu Glu Leu Lys Arg Ala Asp Ala Ala Pro Thr Val Ser De Phe Pro Pro Ser AAA TCG
Lys Ser SUBSTmlTE SHEET (RULE 26J

WO 95/20045 - 2 1 8 0 8 3 4 r~ 118 Sec~uence Listings ICR64 ATG AAG TTG TTG CTA AAC TGG GTT TTT CTA GTA ACA CTT TTA AAT GGT ATT CAG TGT GAG
Met Lys Leu Leu Leu Asn Trp Val Phe Leu Vai Thr Leu Leu Asn Gly lle Cin Cys Glu GTG AAG CTG GTA GAA TAT GGA GGA GGT TTG GTG CAG CCT GGG CTT CTC TCA GAC CTC TCC
Vai Lys Leu Val Glu Tyr Gly Gly Gly Leu Vai Ghn Pro Gly Leu Leu Ser Asp Leu Ser TGC GAA GCT TCT GGA TTC ATG TTC AGT~GAT TTC TTC ATG GAG TGG ATC CGA CAG GCT CCA
Cys Glu Aia Ser Gly Phe Met Phe Ser iAsp Phe Phe Met Glu Trp lle Arg Ghn Ala Pro GGG AAA GGA CTG GAG TGG ATT GCA rGCA AGT AGA AAC AAA GCT AAC GAT TAT TCA GCA GTG
Gly Lys Gly Leu Glu Trp 11e Aia iAia Ser ArB Asn Lys Ala Asn Asp Tyr Ser Ala Val TAC AGT GCA TCT GTG AAG GAC CGA TTC ACC ATC TCA AGG GAT TCT CAC AAA AGC ATC CTC
Tyr Ser Ala Ser Val Lys Asp Arg Phe Ti]r lle Ser Arg Asp Ser His Lys Ser Iie Leu 330: 360 TAT CTT CAG ATG AAC ACA CTC AAA CCT GAG GAT ACT GCC ATT TAT TAC TGT GCA AGArGAT
Tyr Leu Gin Met Asn Thr Leu Lys Pro Glu Asp Thr Aia lie Tyr Tyr Cys Ala Arg LASP

TAT TAC CAA AGT GGC TTG TTT GAT TAC TGG GGC CAA GGA GTC ATG GTC ACA GTC TCC TCA
Tyr Tyr Ghn Ser Gly Leu Phe Asp Tyr Trp Gly Gin Gly Val Met Vai Thr Val Ser Ser -SUBSTITUTE SHEET(RULE 26) W095/20045 21~30834 ~ r 118 ATG GGT GTG CCC ACT CAG CTC CTG GG~ TTG TTG CTG CTG TGG ATA ACA GAT AGA ATA TGT
Met Gly Val Pro Thr Gln Leu Leu Gly Leu Leu Leu Leu Trp lle Thr Asp Arg lle Cys 90 : 120 GAC ATC CAG ATG ACA CAG TCT CCA GCT TCC CTG TCT GCA TCT CTG GGA GAA ACT GTC ACC
Asp lle Gln Met Thr Gln Ser Pro Ala Ser Leu Ser Ala Ser Leu Gly Glu Thr Yal Thr ATC GAA TGT CTA GTA AGT GAA GAC ATT TAC AGT AAT TTA GCGlTGG TAT CAG CAG AAG CCA
lle Glu Cys Leu Val Ser Glu Asp lle Tyr Ser Asrl Leu Ala~ Trp Tyr Gh~ Gln Lys Pro GGG AAA TCT CCT CAG CTC CTG ATC TAT GAT GCA AGT AGC TTG CAA GAT GGG GTC CCA TCA
Gly Lys Ser Pro Gln Leu Leu lle Tyr Asp Ala Ser Ser Leu Gln Asp Gly Val Pro Ser CGG TTC AGT GGC AGT GAA TCT GGC ACA CAG TAT TCT CTC GAG ATC AAC AGC CTG CAA TCT
Arg Phe Ser Gly Ser Glu Ser Gly Thr Ghn Tyr Ser Leu Glu lle Asn Ser Leu Gln Ser GAA GAT GCC GCG ACT TAT TTC TGT CAA CAG CAT CAT GAT TAT CCT CGG ACG TTC GGT GGA
Glu Asp Ala Ala Thr Tyr Phe Cys Gln Gln His His Asp Tyr Pro Arg Thr Phe Gly Gly -GGC ACC AAG CTG GAA TTG AAA
Gly Thr Lys Leu Glu Leu Lys SUBSTITUTE SHEET (RULE 26)

Claims (9)

CLAIMS:

1. Use of the antibodies to EGF receptor, or fragments thereof, in the preparation of a medicament for inducing terminal differentiation in tumour cells.

2. The use according to claim 1 wherein the antibodies are antibodies ICR62, ICR64 or ICR16, or fragments thereof.

3. The use according to claim 1 or claim 2 wherein the antibodies are humanised, having framework regions based on or derived from a human antibody and having complementary determining regions (CDRs) derived from a non-human antibody.

4. The use according to claim 3 wherein the antibody has complementary determining regions (CDRs) derived from the ICR62VH, ICR62VK, ICR64VH or ICR64VK sequences shown in the accompanying sequence listing.

5. The use according to any one of the preceding claims wherein the cells are characterised by over-expressing EGF
receptor.

6. The use according to any one of the preceding claims wherein the tumour cells are bladder, brain, head, neck, pancreas, lung, breast, or ovary tumour cells.

7. The use according to any one of the preceding claims wherein the medicament comprises from 1 to 300mg of the antibodies.

8. The use according to any one of the preceding claims wherein the antibodies or fragments of antibodies are conjugated to a cytotoxic compound, drug or toxin, or a label.

9. A method of using an EGF antibodies ICR62, ICR64 or ICR16, or a fragment thereof in the design or synthesis of peptide or non-peptide compounds having the property of inducing terminal differentiation in tumour cells