WO1990008160A1

WO1990008160A1 - Growth factor receptor-like peptides without tyrosine-kinase activity

Info

Publication number: WO1990008160A1
Application number: PCT/GB1990/000083
Authority: WO
Inventors: Michael Joseph Ezra Sternberg; William John Gullick; Cyrus Chothia
Original assignee: Imperial Cancer Research Technology Ltd.
Priority date: 1989-01-20
Filing date: 1990-01-19
Publication date: 1990-07-26

Abstract

A peptide which lacks tyurosine kinase activity, having a binding portion of at least 8 contiguous amino acid residues adapted to form an intra-membrane alpha-helix and being capable of reducing tyrosine kinase activity in a target cell, the binding portion having a sequence identical to the sequence at positions P0 to P7 of a natural (wild type) or mutant growth factor receptor/tyrosine kinase portein having the motif set out below or having a sequence which differs at not more than 3 positions from the sequence at positions P0 to P7 of such a protein, wherein the motif comprises a sequence of 5 contiguous amino acid residues designated: P0XXP3P4, wherein the residues X are the same or different and each is any naturally occurring L-amino acid residue, P0 is a residue selected from Ala, Gly, Ser, Thr, Pro, Val, Ile and leu, P3 is a residue selected from Ala, Gly, Met, Val, Ile and Leu and P4 is a residue selected from Ala, Gly, Thr, Pro and Val.

Description

GROWTH FACTOR RECEPTOR-LIKE PEPTIDES WITHOUT TYROSI E-KINASE ACTIVITY

The present invention relates to novel peptides and their use in the treatment or diagnosis of cancers. Certain types of cancer, such as some breast, neck, face and neural cancers, are known to be associated with over-expression or mutation of various growth factor receptors having tyrosine kinase activity. The growth factor receptors in question are typically located in the cell membrane and have portions exposed at the external and internal surfaces of the membrane (respectively the growth factor receptor domain and tyrosine kinase domain) linked by a transmembrane region. It has previously been reported that a point mutation in the transmembrane region of the rat neu protein converts the protein into a constitutively active oncogenic tyrosine kinase and that this results in uncontrolled cell growth. However the exact mechanism of activation of the tyrosine kinase, whether in response to binding of a growth factor to the receptor domain or as a result of such a point-mutation, is currently the subject of much controversy.

From an analysis of the amino acid residue sequence of the transmembrane portions of a number of growth factor receptor/tyrosine kinase proteins, the present inventors have identified a common motif. The motif is considered to be involved in a dimeriεation reaction between the transmembrane regions of the normal proteins which is promoted by ligand binding and that this reaction is believed to occur to a greater extent and in- the absence of ligand binding, in the mutant proteins and to be responsible for inducing the oncogenic tyrosine kinase activity. Most of the growth factor receptor/tyrosine kinase proteins involved are monomeric, hence the use herein of the terms "di er" and "dimerisation". However, some of the growth factor receptor/tyrosine kinase proteins are themselves dimers or multimers and, in these cases, the invention is concerned with the association of the native di- or multimers into larger conglomerates; insulin receptor is an example of such naturally multimeric proteins.

This information is exploited by designing peptides (hereafter referred to as "inhibitors") adapted to become located in the cell membrane and inhibit ^• protein:protein dimerisation and hence cell growth. Such peptides may, therefore, be useful in the prevention or treatment of cancers associated with over-expression or mutation of growth factor receptor/tyrosine kinase proteins and resultant uncontrolled cell division. These inhibitors may also be used to control normal cell growth processes where this is desirable, for instance to inhibit cell growth in dividing tissues such as gut-lining or bone marrow during radio- or chemo-therapy.

The invention will now be further described with reference to the figures of the drawings in which: Fig.l. shows the sequences of 20 growth factor receptor/tyrosine kinase proteins in the respective transmembrane -regions.

Fig.2. shows a computer generated space-filling model of the interactions between portions of the transmembrane alpha helix regions of two neu protein molecules incorporating the respective motif residues. Fig. 3. shows the sequence of the minigene and corresponding transmembrane model protein encoded thereby as described in Example 3 below.

Referring to Fig.l, the amino acid residue sequence is given (in the international 1-letter code) of the transmembrane region (the hydrophobic transmembrane portion is bounded by bold vertical bars) of each of 20 growth factor receptor/tyrosine kinase-type proteins. The residues involved in the motif, identified in 18 of the proteins, are indicated by boxes.

The motif comprises a sequence of 5 amino acid residues designated P°XXP³P⁴ wherein the nature of the three residues at positions 0, 3 and 4 (i.e, P° , P³ and P⁴ respectively) is critical whereas the two residues at positions 1 and 2 (each indicated by "X") are variable*.

In the following discussion of growth* f-actor receptor/tyrosine kinase proteins and peptides related thereto the following convention is used to designate amino acid residues:

The first residue of the motif is designated P° (this corresponds with Ala 661 in the neu protein). Residues towards the N-terminus of the peptide are referred to by negative superscripts, eg. P^"4 is 4 places towards the N-terminus from P° (corresponding to Thr 657 in the neu protein) , whereas residues towards the C-terminus have positive superscript position indicators, eg. P³ and P corresponding to the remainder of the motif (i.e, Val 664 and Gly 665 in the neu protein). In the growth factor receptor/tyrosine kinase proteins the N-terminal is generally extracellular and the C-terminal is intracellular thus P^~ positions of the peptides would usually lie towards the outside of the cell membrane and P⁺ positions towards the inside.

Further occurrences of the motif in certain of the sequences are denoted by background shading at P . The arrow points to Val 664 (V) in neu to denote the position of the mutations that lead to tyrosine kinase activation.

The proteins are: neu - cellular oncogene product from rat neuroblastoma cells; c-ErbB-2- human version of neu (also known as HER2) ; EGFR-H - human epidermal growth factor receptor; DER - Drosophila gene product related to EGFR; v-ErbB - oncogene product from AEV-H strain of avian erythroblastosis virus; PDGFR-A and PDGFR-B and A- and* B- for s of mouse platelet-derived growth factor receptor; CSF1R - colony-stimulating factor type 1 receptor (also known as c-Fms); v-Fms - oncogene product from McDonough feline sarcoma virus; c-Kit - human homologue of oncogene product Hardy-Zuckerman 4 feline sarcoma virus; Ret - human cellular oncogene product; INS.R- human insulin receptor; IGF1R - human insulin-like growth factor-1 receptor; DILR - Drosophila gene product related to INS.R; Ros-H human homologue of oncogene product from UR2 avian sarcoma virus; Ros-C chicken version of Ros-H; 7LESS - Drosophila sevenless gene product; Met - human MNNG-induced oncogene product; Trk -^' human colon carcinoma oncogene product; Ltk - mouse leukocyte tyrosine kinase.

In the known natural proteins, the possible amino acid residues observed at each of the three critical positions are as follows:

Position P_° (amino acid residues having a small side chain)

Ala Gly Ser Thr Pro

Position P³ (amino acid residues having an aliphatic side chain)

Ala Val He Leu Position P⁴ (amino acid residues with the smallest side chains)

^"Ala Gly

From the frequency of occurrence of the residue types in the hydrophobic region, the probability for the residues described above occurring at P° , P³ and P⁴ is 0.311, 0.680 and 0.169 respectively. Thus the probability (P) of observing the pattern at one location is 0.0358. There are 20 locations where the pattern can occur within the typical 24 residue hydrophobic segment, so the pattern is expected to occur at least once in one sequence with a probability of 1 -(1-P)²⁰=0.52, (Rothbard, J.B. & Taylor, W.R. EMBO J. 7, 93-100 (1988)). The likelihood of observing this pattern 18 or more times in 20 sequences is thus 0.0005. This strongly suggests that the motif does not occur by chance.

It is position P³ where the point mutatάon (Val 664 becomes Glu) arises in the oncogenic variant of the neu protein. Other known oncogenic mutations in the neu protein at P³ (though of lesser activity than the Val 664 to Glu mutation) are Val to Asp or Gin. From a knowledge of the physical properties (eg. size and shape) and chemical properties (eg. hydrophilic/hydrophobic nature and charge structure) of the permitted residues above, it can further be predicted that the amino acid residues Val, He and Leu may appear at P° , Gly and Met may appear at P and Val, Ser, Thr and Pro may appear at P⁴.

Without wishing to be bound by theory, it is postulated that the hydrophobic transmembrane region of the growth factor receptor/tyrosine kinase proteins adopts the conformation of an alpha-helix such that the residues P , P³ and P⁴ are all exposed on one face of the helix and interact with a face of the corresponding transmembrane helix of a second protein, leading to dimerisation. An inhibitor having a similar or identical motif (but lacking the tyrosine kinase activity of the oncogenic protein) may also adopt an alpha-helical conformation within the membrane and bind to the dimerisa_tion site of the oncogenic protein, thus inhibiting protein:protein dimerisation or disrupting the dynamic equilibrium between free and dimerised protein, thereby preventing transformation of the cell, and so the inhibitor may "switch off" the tyrosine kinase activity in transformed cells.

The present invention therefore provides an inhibitor which lacks tyrosine kinase activity, ^:has a binding portion of at least 8 contiguous amino acid residues adapted to form an intra-membrane alpha-helix and is capable of reducing tyrosine kinase activity in a target cell, the binding portion having a sequence identical to the sequence at positions P° to P⁷ of a natural (wild type) or mutant growth factor receptor/tyrosine kinase protein having the motif as hereinbefore defined or having a sequence which differs at not more than 3 positions from the sequence at positions P° to P⁷ of such a protein.

Preferably the binding portion is at least 9 amino acid residues in length, corresponding to the sequence at positions P° to P or P^~ to P ; more preferably it is at least 10 residues in length and corresponds to positions P^-1 to P⁸ and yet more preferably the binding portion contains 12 or 13 residues corresponding to P^~4 to P⁷ or P⁸. The inhibitors may be larger than this; additional sequences at the N- and/or C-termini of the binding portion may correspond with such sequences in the growth factor receptor/tyrosine kinase protein to be inhibited but this is not essential and other sequences may be selected to improve the desirable properties of the inhibitors. Those inhibitors which include the sequence at position P^~4 to P^-1 in the binding portion are permitted to differ at not more than one position in the region P^~4 to P^"1 inclusive from the sequence of the natural or mutant protein which is to be inhibited. The motif-binding portion of inhibitors of* the invention may be located at any suitable position' within the inhibitor, for instance at or near either of the N- and C-termini or at any position therebetween.

The positioning of the modifications and the nature of permitted modifications will be discussed below.

The inhibitors according to the invention comprise the residues of the 20 commonly occurring natural L-alpha amino acids and may also incorporate the residue of one or more rare L-alpha amino acids or non-naturally occurring L-alpha amino acids having' similar properties to the naturally occurring ones. Suitable non-naturally occurring or rare amino acids are well known in the art. Additionally, the inhibitors of the invention may incorporate the residues of D-alpha amino acids at any positions outside the transmembrane alpha-helix region. Furthermore, the inhibitors may be glycosylated at any positions outside the transmembrane region.

In its broadest sense then, the present invention encompasses any peptide of 8 or more alpha amino acid residues having the features set out above including polypeptides at least up to the size of the mature proteins (for instance up to or even more than about 1000 amino acid residues) whose dimerisation is to be inhibited, although clearly it would be undesirable for the inhibitors to have tyrosine kinase activity. Typical large inhibitors of the invention would include the whole or a substantial portion of the transmembrane and growth factor receptor domains (there is unlikely to be any advantage to be gained by- having a substantial but incomplete part of the^; growth factor receptor domain) of the native protein and possibly a part of the tyrosine kinase domain. Having regard to the potential antigenicity of large polypeptides, it is presently preferred that the inhibitors of the present invention are from 8 to 30 or even up to 40 amino acid residues in length or that substantially the whole of the inhibitor sequence corresponds to that of the natural or mutant protein to be inhibited, preferably including glycosylation as found in the mature^'protein.

The amino acid peptide sequence of the binding portion may be^" modified relative to that in the vicinity of the motif of the protein to be inhibited in order to improve various properties of the inhibitor peptide. For instance, modifications might be made to improve the potency of the peptide inhibitor or to reduce the likelihood of digestion by proteolytic enzymes (such as pepsin, trypsin and alpha-chymotrypεin) in the gaεtro-inteεtinal tract or elsewhere. The sequences susceptible to proteolysiε and thus to be avoided are well known in the art. Further reasonε for modification would be to facilitate εyntheεiε of the inhibitor by chemical or recombinant DNA techniqueε, to avoid binding to proteinε other than the target protein, εuch as serum proteins, for instance albumin, or to improve solubility as_. an aid to administration of the peptide.

Referring to Fig. 2, the alpha-helix of one neu protein is shown mainly in light grey (the "left" helix) whereas that of the second neu protein is εhown-mfainly in dark grey (the "right" helix). In the left helix, Gly 665 is dark grey. In the right helix Gly 665 is in light grey. Fig 2a shows neu with Glu 664 in the right helix in black forming the proposed hydrogen bond to the carbonyl oxygen (black) of Ala 661 of the left helix. The exact conformation for the side chains cannot be modelled and the Figure just shows that a hydrogen bond could be formed. Fig 2b shows neu with Val 664 in the right helix in black . packing against Gly 665 in the left helix. The view was chosen so that the carbonyl oxygen is visible. n the membrane, the helices will be roughly oriented by +_ 25° to the perpendicular through the membrane surface. It is possible that some or all of the hydrophobic transmembrane region adopts a 3₁₀ helical conformation rather than an α-helix as has been observed in the crystal structure of alamethicin [Fox, R.O. & Richards, F.M. Nature (London) 300, 325-330 (1982)]. Modelling suggested that the 3₁₀ conformation around the motif is not compatible with inter-helical hydrogen bonding and the steric conεtraintε on residues P° and P⁴. Outside the motif, the 3_lfJ helical conformation is possible. As this structure has a higher rise per residue than an α-helix, some proteins having shorter hydrophobic transmembrane sectionε might adopt thiε conformation in order to traverεe the membrane. However in globular proteins, only short sections of the chain adopt the 3₁₀ conformation and so it is unlikely that long sections of transmembrane regions will be a 3_{1 0} helix.*

The transmembrane regions lie roughly*parallel as opposed to antiparallel with the two helix axes oriented at about -50° . The packing is approximately symmetric with the diad axis of symmetry lying along the helix interface along the averaged N- to C- direction of the helix axes. The most probable interaction involves the carboxyl group P³ (eg. Glu 664 in the neu protein) in one helix (helix A on the right in Fig 2a) forming a hydrogen bond with the main chain carbonyl oxygen of P° (eg. Ala 661 in the neu protein) in the other helix (helix "B"). The carbonyl oxygen of P° (Ala 661) still forms the main-chain/main-chain hydrogen bonding within each α-helix. A second and symmetric hydrogen bond is formed between P (Glu 664) in helix B with the oxygen of P° (Ala 661) in helix A. Whilst it is possible that parts of the transmembrane region will adopt a different conformation, namely that of a 3₁₀ helix, the motif region and adjacent residues are expected to be in an alpha-helical conformation and so the following conεideration of εequence modification will not be affected if regionε outside the motif and adjacent residues are not in an alpha-helix.

TABLE 1

Peptide Suitabilityy oofr Position positions for modification

Key: N «= not exposed in di er

Y = exposed in dimer and thus suitable for modification. Y(?) indicates a position which is only partially exposed in the dimer and which might be modified but where modification is less favoured. (Y) indicates a motif position and one preferably not modified.

NB Modification at P° to P⁴ and possibly also at P^~ and/or P might alter the conformation of the alpha helix at or near the motif-binding portion. Modificationε at theεe positions are therefore less favoured.

Modifications of the protein εequence may be incorporated into the inhibitor peptideε provided that εuch modificationε ^"are not deleterious to the inhibitor:protein binding interaction which will mimic the protein:protein interaction. In general, poεitionε involved in the binding εurface of the inhibitor εhould therefore be conserved whereas positions which are exposed on the non-binding εurface will be available for modificationε. Variations at the motif positions P° , P³ and P⁴ might be made by subεtituting another of the permitted amino acid reεidueε listed above or those predicted as also being posεible motif reεidueε. Such variationε may enhance or diminiεh the activity of an inhibitor and trial and error teεting would be required to aεcertain the desirability of εuch variationε. However it iε conεidered that either Ser or Thr could be introduced at either or both the_. P° and P⁴ positions without destroying the activity of the peptide as an inhibitor; thiε is discussed further below.

Water-solubility iε important for adminiεtration and delivery via body fluids of the inhibitors of' the present invention particularly in relation to the shorter inhibitors which correspond with only a part or the whole of the relatively hydrophobic transmembrane region of the target proteins. However, thiε requirement muεt be balanced againεt the simultaneous requirement that the inhibitor or a portion thereof muεt be capable of insertion into the cell membrane in order to interact with the transmembrane domain of the target protein. The water solubility of individual inhibitor peptides may be improved by various techniques which will be apparent to those skilled in the art. It is presently preferred to enhance water-solubility by inclusion of hydrophilic but uncharged residues. Modification by subsituting Ser or Thr is preferred because at positions after the third residue from the N-terminal of an alpha helix the hydroxyl side chain of Ser and Thr can form an intra-helical hydrogen bond to a preceding main chain carbonyl oxygen (thus the hydrogen bonding requirement of this polar group can be satisfied when the peptide is in a hydrophobic environment) whilst in an aqueous environment the hydroxyl side chain of Ser or Thr can form a hydrogen bond to the water and therefore increase the solubility of the peptide.

Ser or Thr residueε may be included at any suitable position, including P° and/or P⁴ (This is an exception to the rule set out above). Particularly where the peptide is relatively short (30 residues or less) at least one and preferably a total of at least two and up to four Ser and Thr residues may be introduced in the P^~4 to P⁸ region to enhance water solubility.

For longer peptides, additional hydrophilic residues may be required to achieve a desirable degree of solubility.

The ability of the inhibitors to enter the cell membrane iε alεo of concern. In any polypeptide, the amino- (N-) terminus has a positively charged amino group (-NH₂ ⁺) and the carboxyl- (C-) terminuε haε a negatively charged carboxyl (COO^~) group. If the inhibitor is designed εuch that either or both the termini lie within ^'the hydrophobic region of the membrane, the presence of ' such charged groups is likely to make insertion of the inhibitor into the membrane energetically unfavourable. This can be overcome by acetylating the terminal amino group and/or converting the terminal carboxyl group into an amide using conventional methods of peptide chemistry.

An additional consideration is that in any alpha-helix, the first three main chain amide (-NH) groups (i.e, those at the N-terminus of the helix) do not form intra-helical hydrogen bonds. Neither do the last three main-chain carbonyl (C=0) groups (i.e. those at the C-terminus of the helix). Theεe main chain NH and 00 groupε are polar in character and thiε polarity iε likely to reduce the affinity of the peptide for the. hydrophobic membrane. One way to decrease the polarity of the N-terminus of the alpha-helix is to replace the first, second and/or third residues at the N-terminus with an* amino acid, such as Pro, in which the side chain 'forms a covalent link with the amide nitrogen thereby removing a polar group. The use in any or all of the last three positions (of the alpha helix) of residueε of non-naturally occurring L-alpha amino acid derivatives or analogues in which the carbonyl group iε maεked or replaced may also be contemplated.

An alternative and preεently preferred εtrategy to avoid the problem of polar main chain groups at the termini of the transmembrane alpha helical peptide would be to use an inhibitor of sufficient length to extend beyond the hydrophobic section of the cell membrane at either or both of its termini. When the motif binding portion is correctly positioned in relation to a protein to be inhibited the location of the hydrophobic residues and of polar and charged residues in the εequence of the target protein can be used as a guide as to the likely location of the ends of the hydrophobic transmembrane region in a manner shown by the location of the vertical bars in Fig. 2.

From a consideration of free energy of association it may be argued that, in order to favour dimerisation or aggregation of the inhibitor with the target protein in cell membrane, the most effective inhibitors will include modifications at P³ to include residues capable of hydrogen-bonding (eg. Glu, Asp, Gin or Asn) to main chain helix carbonyl group of P° in the other helix. This is particularly the case with shorter inhibitors which lack all or part of a growth factor binding domain and/or a part of the (inactivated) tyrosine kinase domain. The reason for selecting the P³ position for such modification iε that thiε .position is involved in the transforming mutations of the neu protein (Val 664 changes to Glu, Asp, Gin or Aεn) and is important in the protein:protein dimerisation interaction. Thus the inhibitor would have at P³ the appropriate residue (eg. Glu, Asp, Gin or Asn) corresponding to the sequence of the oncogenic mutant protein rather than the residue corresponding to the P³ residue of the wild type protein (asεum^'ing theεe are different) or a reεidue "εelected to hydrogen bond to the P° reεidue of the mutant. The inhibitorε would thuε bind with increased specificity to the mutant protein and be more effective in blocking the dimerisation.

As discuεεed earlier, it may be deεirable to inhibit growth factor receptor dimeriεation in order to control normal cell growth. In thiε εituation, it may be uεeful for the inhibitor to have at P³ a potential hydrogen-bonding reεidue which can bind to the main chain carbonyl of P of the other helix (i.e. the wild-type growth factor receptor).

A further type of modification which may be convenient iε that wherein one or more Tyr residues are introduced to enable radio-labelling (for instance with radio-iodine) for use in tracer experiments.

Modifications have been discuεεed above in connection with the sequence in the vicinity of the motif-binding region of the inhibitors. The modified sequences may be extended, at either or both their N- and C-termini, with further sequenceε which are .preferably εi ilar to or the same as the sequenceε of the correεponding regionε of the proteinε to be inhibited. The sequences of theεe N- and C- terminal extenεions are less critical and the posεibilitieε for modification are correεpondingly greater.

As mentioned above, it iε important that the inhibitors of the invention lack tyrosine kinase activity otherwise they might exacerbate the condition to be treated. With short inhibitor peptides this cannot occur. With longer ones, especially those similar to the native (wild type or mutant) protein, the tyrosine kinase active sites should be inactivated if present. Inactivating modifications will be readily apparent to those skilled in the art or can be identified by trial and error, for instance by site-directed mutagenesis and screening for tyrosine kinase activity. Methods for detecting any tyrosine kinase activity, or lack of εuch activity, are well known.

In one preferred embodiment the preεent peptides contain a εequence (the motif-binding portion) identical to that of all five residues forming the motif of the growth factor receptor/tyrosine kinase protein to be inhibited. In other embodiments the inhibitor peptides contain a sequence substantially homologous to the motif of the protein to be inhibited. Conveniently at least one of the resi'dueε in the motif-binding portion of the inhibitor is identical to the corresponding P° , P³ or P⁴ residue in the motif of the protein to be inhibited. More preferably at least two and yet more preferably all three of the residues in the inhibitor are identical to the corresponding P° , P³ and P⁴ residues of the motif of the protein to be inhibited.

Tables 2 and 3 below give the sequences of the motif region of two human proteins which are asεociated with cancer and of peptideε deεigned^' aε inhibitors thereof according to the present invention. Such peptideε form preferred embodiments of the invention.

TABLE 2 EGF-R

* See Table 1

NB. International 1-letter codes are used for all peptides sequences S/T; indicates either Ser or Thr residue at this position. TABLE 3 c-erbB-2

Peptide Expoεure* Wild Inhibitor Peptideε Poεition Type 1 2 3

* See Table 1

NB International 1-letter codeε are uεed for all peptide sequenceε; S/T indicates either Ser or Thr residue at thiε position. The ability of candidate peptides ostenεibly falling within the scope of invention to form intra-membrane alpha helices and their ability to inhibit tyrosine kinase activity may be tested in vitro by conventional techniques. Two methods for screening for inhibition of rat neu protein dimerisation are as follows:

Three cell lines are used. The first lacks (or at least has very low levels of expresεion of) any native (wild type or mutant) growth factor receptor/tyrosine kinase protein of the type under investigation and acts as a control. The second cell line has growth factor receptor/tyrosine kinase expresεed at a high level but in a non-tranεforming manner and the third haε moderate levelε of expression, this time of a transforming mutant of the growth factor receptor/tyrosine kinase protein, or is otherwise transformed by dimerisation of a growth factor receptor/tyrosine kinase protein. Such cell lines can be produced ^b initio by recombinant DNA techniques or are publicly available, an example being (1) "regular" NIH3T3 (mouse) cells (which have low levels of mouse neu* protein), (2) NIH3T3 cells transfected with an expresεed normal rat neu gene and (3) NIH3T3 cells transfected with an expresεed oncogenic mutant rat neu gene.

In the first assay, samples of the live cells are challenged with a solution of the test peptide and cultured. The control (1) and "normal" (2) cell lines will continue to grow as usual if the peptide does not interfere with normal cell procesεes whereas the mutant (3) cell line will grow but cease to show the characteristic morphology of transformed (cancerous) cells if the peptide inhibits the dimerisation.

In a second asεay, cell membranes are prepared from samples of each of the three cell lines and the tyrosine kinase activity of each iε meaεured in the abεence and presence of the test peptide. However thiε technique cannot be used with any peptide containing tyrosine residues as these would be preferentially phosphorylated.

In the preεence of a peptide of the invention which inhibitε mutant neu protein, the tyroεine kinaεe activity of the cell line (1) will be εubεtantially unchanged whereas that of at leaεt cell line (3) will be conεiderably reduced, for inεtance at least 5 fold and preferably 10 fold. The tyrosine kinase activity of cell line (2) may or may not be reduced depending upon the selectivity of the peptide for inhibiting dimerisation of oncogenic rather than normal protein. The different reactions of cell lines (2) and (3) may be used to identify peptides with high selectivity where thiε iε desired.

These procedures may be used in analogous manner to screen for inhibition of other growth factor receptor/tyrosine kinase oncogenic proteins.and may alεo be adopted in connection with the diagnoεtic methodε deεcribed below.

Inhibitorε according to the preεent invention may be produced by conventional techniqueε. Relatively short inhibitors are readily produced by d_e novo chemical synthesis by well known techniques. Longer inhibitors may be produced by coupling shorter fragmentε or by recombinant DNA techniques. Glycosylation where required may be achieved by conventional techniques. Processes for producing the inhibitors form a further aspect of the invention.

In further aspectε the invention provides an inhibitor as hereinbefore defined for use in a method of diagnosis practiced on the human or animal body or in a method of treatment of the human or animal body by therapy or surgery. The invention further provides use of an inhibitor protein as hereinbefore defined for the preparation of a medicament for use in treating or preventing cancer or controlling cell growth by inhibition of a growth factor receptor/tyrosine kinase protein.

The invention further provides a method for the treatment or diagnosiε of cancer in a human or animal having cancer associated with a growth factor receptor/tyrosine kinase protein or for control of cell growth in a patient in need thereof comprising administering an effective, non-toxic amount of an inhibitor as hereinbefore defined.

In a particular aspect of the treatment method according to the present invention, the amino acid sequence of the growth factor receptor/tyrosine kinase oncogenic protein responsible for the patient's cancer or for cell growth which is to be controlled will be determined, for instance from the patient's chromosomal DNA sequence or by other techniques known in molecular biology and an inhibitor according to the present invention will be selected from ^"a collection of εuch peptides or designed and εyntheεised εpecifically in order to inhibit the particular target protein of the individual patient.

Depending upon the εite of the tumour to be treated, inhibitors according to the invention will usually be administered in the form of pharmaceutical compoεitionε compriεing the inhibitor and a pharmaceutically acceptable carrier or diluent therefor. Such compoεitionε form a further aspect of the invention. Suitable carriers and diluents include water and pharmaceutically acceptable organic solventε and oily media. The compoεitionε may contain optional acceεεory ingredients εuch as bufferε, isotonic agents, preservatives, anti-oxidantε and the like and may be preεented in unit- or multi-dosage forms, ready-for-use or for reconstitution, for instance by addition of water for injection. The inhibitors and compositionε may be administered by conventional routes, for inεtance orally for treatment of the gastrointeεtinal ' tract or when formulated to ensure assimilation of the inhibitor before its digestion by proteolytic enzymes in the gastrointeεtinal tract or elεewhere. For treatment of other tiεεueε the inhibitorε may be adminiεtered intravenouεly or, in particular caεes, by injection into, or close to the treatment εite (eg. the εite of the tumour). Specialiεed vectors such aε lipoεomeε and targetting moieties such as anti-tumour antibodies, and . particularly anti-growth factor receptor antibodies, may be used to direct delivery of the inhibitors to particular regions of the^' body or to the tumour site.

The inhibitors of the present invention will typically be administered to normal adults of about 75kg in quantities of from about O.lmg to 500mg, preferably 0.5mg to lOO g, more preferably 1 to 50mg at intervals of from a few hours to a few days. The exact dose will be determined depending upon the body weight, state of health, the degree of inhibition required and the amount of target tissue (severity of the tumour), therapeutic index and rapidity of clearance of the inhibitor.

In diagnosis a labelled inhibitor may be used for localisation of a tumour within a patient's body or detection of transformed cells in sampleε removed therefrom. Labelled inhibitorε form a further aεpect of the invention; labels may for instance be radio-isotopes or radio-opaque moieties, enzyme labels or fluorescent labels. Labelled inhibitors are formed by conventional methods*.

The present invention further provides 'a diagnostic method comprising contacting a tissue sample or cells from a potential or suspected cancer patient with an inhibitor of the invention and detecting any binding of the inhibitor to growth factor receptor/tyrosine kinase protein. Such a procedure may be used to detect the presence of mutant protein in the • sample or an excess of wild type protein relative to normal levels thereof. Labelled inhibitor may be used to demonstrate the binding or this may be asεayed for instance using the tyrosine kinase activity testε outlined above.

The invention will be further illuεtrated by the following Exampleε:

EXAMPLE 1

The probable mechanism by which ligand binding to the extracellular domain of a growth factor receptor causeε activation of itε cytoplaεmic tyroεine kinase (TK) domain is that binding promotes receptor dimerisation¹. The rat neu protein can be converted² into a constitutively active oncogenic TK by a carcinogen induced point-mutation in the transmembrane region (Val 664 -> Glu)³. A previous suggestion⁴ was that dimerisation of the transmembrane α-helices is stabiliεed by hydrogen bondε between the protonated Glu εide chainε. Molecular modelling now suggests that hydrogen bonds between the protonated Glu in one helix with a main-chain carbonyl oxygen in the other stabilises dimerisation. The same packing arrangement could occur in the untransformed molecule (Val 664). In both structures, the α-helical packing is mediated by two small chains (Ala 661 and Gly 665). A similar sequence motif iε observed in all but two of the 20 transmembrane α-helices of the TK family ^~ . Peptides that prevent α—helix dimerisation might inhibit TK activity of a specific growth factor receptor. Point mutations⁸ have been introduced into neu around Val 664. Glu 664, Gin 664 and Asp 664 had respectively high, moderate and low transforming activity. Other mutation^'s at 664 and at 663 or 665 were not transforming. Introduced Glu and Asp mutations⁹ at the equivalent position in c-erbB-2 (the human equivalent of neu) lead to TK activation and transformation. One explanation is that Glu (and Asp) side chains will be protonated in the membrane's hydrophobic environment and can form a hydrogen bond with the other helix stabiliεing dimeriεation.

The stereochemistry of neu dimerisation waε investigated. One possibility is that the two helices do not pack but the Glu side chains are partially or fully extended and form carboxyl-carboxylate hydrogen bonds¹⁰. Alternatively, the α-helices could pack and this was examined using molecular graphicε and space-filling models.

The transmembrane helix packing in the dimer generally is symmetric. The most probable interaction involves the carboxyl group of P³ , (Glu (664)) in one helix (helix A on the right in Fig 2a) forming a hydrogen bond with the carbonyl oxygen of P° (Ala 661) in the other helix (B). The carbonyl oxygen of Ala 661 still forms the main-chain/main-chain hydrogen bonding for the α-helix. Such bifurcated hydrogen bonding¹¹ is common within α-helices with the hydroxyl side chains or Ser and Thr and can occur for Glu. A second and symmetric hydrogen bond is formed between Glu 664 in B with the oxygen of Ala 661 in A. The helical axes lie at about -50° which is a favourable orientation for α-helix packing . This angle would provide some separation for dimerisation of the extracellular and of the intracellular domains. In the ^"untransformed neu protein (Fig 2b) the same helix/helix arrangement would pack Val 664 (denoted position P ) in helix A against Gly 665 (P⁴ ) in helix B and would require that the side chain at 661 (P° ) is not large.

TK activity can therefore be explained stereochemically. For the transforming Glu mutant, there is a substantial free energy of dimerisation when two inter-helical hydrogen bonds are formed leading to high activity. In the untransformed protein, the same geometry for helix dimerisation can occur but this iε only εtabiliεed by van der Waals packing and so the dimerisation free energy will be lesε than for the Glu mutant with the conεequential lower, constitutive, level of activity. It iε suggeεted that the Gin and Aεp mutationε progreεεively form weaker inter-helical hydrogen bondε with the conεequential lower levelε of activity. A εimilar relationship between helix packing and activity would apply to c-erbB-2 and the introduced mutations⁹.

The generality of this helix model in other TK growth factor receptors was suggested from the presence of a related sequence εpecific motif at P° which requireε a small εide chain (Gly, Ala, Ser, Thr or Pro); P³ an aliphatic εide chain (Ala, Val, Leu or He) and P⁴ , only the smallest εide chains (Gly or Ala). The presence of the motif suggestε that helix aεεociation might be a mechaniεm for'TK activation in many or all of these receptors. Indeed in the EGF receptor, th introduced mutation¹³ of Val to Glu at P³ leads to higher activity in the presence of EGF. In di eric receptors, (e.g. insulin and insulin-like growth factor 1), this heli association would be between a pair of dimers leading to the formation of tetramers or even higher orders of association¹⁴. It is not clear why thiε pattern is absent in v-fms¹⁵ and the drosophila EGF-receptor homologue¹ .

In the 18 sequences the pattern does not occur a the same position in relation to the C-terminus of the non polar region (Fig. 1.) although it tends to occur in the N-terminal half of the region. Thus the exact position of the motif within the helix need not be conserved between receptors. This might explain why transmembrane regions cannot be interchanged between different family members an maintain TK activity¹⁷.

The probablity that the pattern occurred by chance in 18 out of the 20 sequenceε is only 0.0005. However, the pattern might be a feature of all transmembrane regions. 20 sequenceε¹⁸ were εelected arbitrarily for proteins with one transmembrane region. The pattern did not occur in eight proteins, occurred once in four, twice in five and three times in two and five times in one protein. The multiple occurrence of the pattern is probably due to a run of Ala and Gly on one helix face that might mediate helix packing in dimeric molecules

Modelling also conεidered ά-helix packing with Glu-Glu hydrogen bondε for neu. The carboxyl εide chainε would have to be εandwiched in the helix interface which is an unlikely structure¹². With this packing, Val at 664 in one helix would not pack sufficiently close to Gly 665 (P ) in the other to restrict P⁴ to be Gly or Ala and explain the motif.

The generality of helix association for the family can be examined by introducing Glu, Gin or Asp mutations at P³ that might increaεe TK activity (e.g. EGF receptor). Indeed in all 17 reported DNA εequenceε, a εingle base change at P3 can lead to these mutationε and εcreening might identify naturally occurring tranεforming proteins. Residueε other than Gly or Ala at P⁴ εhould prevent packing and reduce activity. In addition, peptideε with the εequence of the tranεmembrane region of one member of the tyroεine kinaεe family, particularly thoεe with a tranεforming mutation involved in hydrogen bonding, might be able to inhibit εpecifically the activity of that molecule by forming a non-productive complex thuε limiting receptor activity. Such inhibitors may represent a novel therapeutic strategy for cancer cells, such as in breast cancer^20,21, whose transformed state may be .dependent on unregulated growth factor receptor activity.

EXAMPLE _2

We have εynthesised several peptideε from the tranεmembrane region of the rat neu protein either containing the naturally occurring Val residue at position P³ or altered by εubstituting the activating amino acid Glu at the same position. In initial studies two sequences were chosen which are (with the exception of the above change at P³ ) derived directly from the natural molecule. Their sequenceε are:

Pep.l F I I A T V E G V L L F L I Pep.2 F I I A T V V G V L L F L I

Neither peptide waε apparently εoluble in water.

Uεing the conεiderationε of increasing solubility as discuεsed above to choose suitable residueε, we then made two more peptideε. Their sequences are:

Pep.3 A S P V T Y I I A T V E G V L S S L Pep.4 A S P V T Y I I A T V V G V L S S L

Peptide 3 was found to be readily εoluble in water. Peptide 4 was apparently insoluble probably because insufficient hydrophilic amino acids had been incorporated.

We have now synthesiεed two further peptideε modified with additional changes in an attempt to promote both to become water εoluble. The peptideε have been deεigned using the theoretical predictions discusεed above to choose suitable residues which can changed without predicted detriment to the products ability to interact with and inhibit the neu protein. Their sequenceε are:

Pep.5 A S P V T S S I A T V E G V S L F S Pep.6 A S P V T S S I A T V V G V S L F S

Both peptideε were found to be water soluble at neutral and acidic pH, although peptide 6 was lesε εoluble than peptide 5.

The serieε of amino acid εubεtitutionε in the above peptideε exemplifieε the approach of introducing amphipathic residues at chosen points in the εequence to increase the peptideε' water εolubility.

We next examined whether the peptideε 5 and 6 retained their ability to diεsolve in lipid solutionε. Using the Langmuir through εystem²⁵ both peptides dissolved in phospholipid monolayers when introduced into the underlying aqueouε phaεe. Thiε establisheε that the εubεtitutionε have rendered the peptideε amphipathic in that they are now εoluble both in water and in phoεpholipidε and may therefore in principle be adminiεtred in aqueous solutionε intravenouεly and yet diεsolve and accumulate in tumour cell membrane. Example 3_

Nue mini gene

We have constructed an erbB-2 minigene to enable us to test the polypeptides using the wild type transmembrane domain and to provide a vehicle to assay the relative efficiency of altered εequenceε. Oligonucleotideε were synthesised repreεenting minimal signal peptide, extracellular, transmembrane, intracellular and C-terminal peptide sequences of erbB-2. The oligonucleotides were annealled in pairs and the five resulting double stranded sequenceε were ligated together. The εequence of thiε minigene iε given in Fig. 3, and incorporateε the following features: i) a 5' Sal 1 restriction εite and 3' Kpn 1 reεtriction εite for cloning into a eukaryotic expression vector; ii) unique engineered Xma 1 and Eco Rl restriction εiteε bounding the tranεmembrane domain for facile directional inεertion of teεt tranεmembrane sequenceε; iii) an alanine reεidue inεtead of a cyεteine at poεition 8 within the proceεεed minigene extracellular domain to prevent any chance intermolecular diεulphide bridge formation.

The minigene iε inεerted in from of an appropriate eukaryotic promoter and tranεfected into tranεformed cellε expreεsing wild type neu receptor. Promoters which allow a controlled induction of expression are used, for example the dexamethasone inducible MMTV LTR and the heat and heavy metal inducible heat shock (Hsp 70) promoter. Stable tranεformantε are έelected by cotranεfection of the inducible minigene with a hygromycin resistance pla^'smid. The reεulting cell lineε are induced to express the minigene and monitored by Western blotting of membrane preparations using the 21N antibody against the C-terminal peptide present in the minigene encoded protein; two band should be visible representing the wild type neu receptor and the mature minigene product (c.a. 8κd) . Cell lines that appear to express the minigene well are compared by culturing in the presence or absence of inducer and examining their phenotypic appearance and mitotic index to ascertain if the presence of the minigene product can inhibit the growth of the cells through non-productive dimerization with the wild type neu-receptor.

Once this asεay for the minigene iε eεtabliεhed, the transmembrane sequence can be swapped for those representing the synthesised εoluble peptideε, enabling their relative efficiency in the assy to be estabilished. In this way an appropriate compromise between peptide solubility and biological function can be achieve'd.

REFERENCES

1. Yarden, Y. & Schleεεinger, J. Biochemistry 26, 1434-1442. (1987. 2. Bargmann, C.I. & Weinberg, R.A. Proc. Nat. Acad. Sci., U.S.A. 85, 5394-5398 (1988).

3. Bargmann, C.I., Hung, M-C & Weinberg, R.A. Cell 45, 649-657 (1986).

4. Gullick, W.J. in Hormoneε and their Action, Part II (eds Cooke, B.A. King, R.J.B. & van der Molen, H.J.) 349-360 (Elsevier, Amsterdam, 1988).

5. Hankε, S.K. Quinn, A.M. & Hunter, T. Science 241, 42-52 (1988).

6. Ben-Neriah, Y & Bauskin, A.R. Nature (London) 333, 672-676 (1988).

7. Matsui T., Heidaran, M. , Mike T., et al. Science 243, 800-804 (1989).

8. Bargmann, C.I. & Weinberg, R.A. EMBO J. 7, 2043-2052 (1988).

9. Segatto, 0., King, C.R., Pierce, J.H., Di Fiore, P.P. & Aaronεon, S.A. Mol. Cell. Biol. 8, 5570-5574 (1988).

10. Sawyer, L. & James, M.N.G. Nature (London) 295, 79-80 (1982).

11. Baker, E.N. & Hubbard, R.E. Prog Biophys. Mol. Biol. 44, 97-179 (1984).

12. Chothia, C, Levitt, M. & Richardson, D.J. Mol. Biol. 145 215-250 (1981).

13. Kashles, 0., Szapary, D., Bellot, F. Ullrich, A., Schlesεinger, J. & Schmidt, A. Proc. Nat. Acad. Sci. USA 85, 9567-9571 (1988).

14. Anderson, A.S. Nature (London) 337, 12 (1989). 15. Hampe, A., Gobet, M. , Sherr, C.J. & Galibert, F. Proc. Nat. Acad. Sci., USA 81, 85-89 (1984).

16. Livneh, E. , Glazer, L., Segal, D., Schlessinger, J. & Shilo, B. -Z, Cell 40, 599-607 (1988).

17. Escobedo, J.A., Barr, P.J. & Williams, L.T. Mol. Cell. Biol. 8, 51.26-5131 (1988).

18. George, D.G., Barker, W.C. & Hunt, L.T. Nucl. Acid. Res. 14, 11-15 (1986).

19. Traverε, P., Blundell, T.L., Sternberg, M.J.E. & Bodmer, W.F. Nature (London) 310, 235-238 (1984).

20. Slamon, D.J., Clark, G.M., Wong, S.G., Levin, W.J., Ullrich, A & McGuire, W.L. Science 235, 177-182 (1987).

21. Sainεbury, J.R.C., Farndon, J.R., Needham, G.K., Malcolm, A.J. & Harriε, A.L. Lancet i, 1398-1402 (1987).

22. Rothbard, J.B & Taylor, W.R. EMBO J. 7, 93-100 (1988).

23. Barlow, D.J. & Thornton, J.M.J. Mol. Biol. 201, 601-619 (1988).

24. Fox, R.O & Richardε, F.M. Nature (London) 300, 325-330 (1982).

25. Langmuir, J.J., Che . Phyε., ^, 757-776 (1933).

Claims

1. A peptide which lackε tyroεine kinase activity, having a binding portion of at least 8 contiguou amino acid residues adapted to form an intra-membrane alpha-helix and being capable of reducing tyrosine kinase activity in a target cell, the binding portion having a sequence identical to the sequence at positions P to P o a natural (wild type) or mutant growth factor receptor/tyrosine kinase protein having the motif set out below or having a sequence which differs at not more than poεitions from the sequence at positionε P° to P⁷ of εuch protein, wherein the motif comprises a sequence of 5 contiguous amino acid residueε deεignated

P°XXP³P⁴

wherein the reεidueε X are the same or different and each is any naturally occurring L-amino acid residue

P° is a residue selected from Ala, Gly,' Ser, Thr,

Pro, Val, He and Leu,

P³ is a residue selected from Ala, Gly, Met, Val,

He and Leu and P⁴ is a residue selected from Ala, Gly, Thr, Pro and Val.

2. A peptide according to claim 1 and having or containing a sequence selected from: ATGMVGAX ; ATGXVGAXL; PSIATGXVGAXL; ^"PSIATGXVFAXLLLLV;

VSAWGXXL; PLTSIVSAWGXXL; PLTSIVSAWGXXLVWXG;

ASPVTYIIATVEGVLSS ; ASPVTYIIATWGVLSSL; ASPVTSSIATVEGVLSFS and ASPVTSSIIATWGVLSFS;

wherein the residues X are the same or different and each iε Ser or Thr.

3. A peptide according to claim 1 or claim 2 for uεe in a method of diagnoεis practiced on the human or animal body or in a method for treatment of the human or animal body.

4. Use of a peptide according to claim.1 or claim 2 in the preparation of a medicament for use in a method for treating or preventing cancer or controlling cell growth by inhibition of a growth factor receptor/tyrosine kinase protein.

5. A method for treating or diagnosing cancer in a human or animal having cancer asεociated with a growth factor receptor/tyrosine kinase protein or for control of cell growth iri a patient in need thereof comprising administering an effective, non-toxic amount of a peptide according to claim 1 or claim 2.

6. A pharmaceutical composition comprising a peptide according to claim 1 or claim 2 and a pharmaceutically acceptable diluent or carrier therefor.

7. A diagnostic method comprising contracting a tisεue sample or cells from a potential or suεpected cancer patient with a peptide according to claim 1 or claim 2 and detecting any binding of the inhibitor to growth factor receptor/tyrosine kinase protein.

8. A process for producing a peptide according to claim 1 or claim 2 comprising chemically linking amino acid residues together in appropriate sequence.

9. A process for producing a peptide according to claim 1 or claim 2 comprising expresεing a functional DNA sequence encoding the peptide in a cell-free system or by culturing cells harbouring such a sequence and recovering the peptide.

10. Nucleic acid encoding a peptide according to claim 1 or claim 2.