WO1999009416A2

WO1999009416A2 - Method of drug selection

Info

Publication number: WO1999009416A2
Application number: PCT/GB1998/002504
Authority: WO
Inventors: Henry R. Wolfe
Original assignee: Nycomed Imaging As
Priority date: 1997-08-20
Filing date: 1998-08-20
Publication date: 1999-02-25
Also published as: JP2001516029A; AU8816898A; GB9717652D0; AU8817398A; WO1999009417A2; WO1999009417A9; WO1999009417A3; WO1999009416A3; NO20000811L; NO20000811D0; CA2300903A1; EP1005655A2

Abstract

Method of selecting a candidate drug compound comprising generating from a starting oligopeptide a plurality of oligopeptides having single site amino acid modifications which lead to a change in the ratio of binding affinity of said oligopeptide to a receptor to biological reponse of said binding, identification of components in the oligopeptide responsible for secondary structure and replacement of said components with non-natural amino acids, construction of combinatorial libraries based on the use of sets of amino acids for the modification sites identified and the construction of further libraries based upon oligopeptides produced in the first library having the desired binding activity. Drug candidates are selected from the identified members of the libraries.

Description

Method of drug selection

The present invention relates to a method of selecting a candidate drug compound having affinity for biological receptors.

Cell surfaces, e.g. the surfaces of erythrocytes, endothelial cells, tumour cells, etc have characteristic binding sites for biological ligands. Frequently such binding sites are the result of the secondary structure, i.e. the folding conformation, of cell surface protons, and binding to these triggers a response by the cell. One example is the ST receptors which are found only on the apical brush border membranes of the cells lining the intestinal tract of placental mammals.

A variety of bacteria, such as Eschericia coli, Vibrio cholerae, Citrobacter freundii and Yersinia enterocolitica, which may infect the mammal gut produce homologous peptide toxins which bind to ST receptors and trigger a cascade of biochemical processes eventually leading to fluid secretion into the intestinal lumen and hence diarrhoea. These ST enterotoxins are a major cause of infectious diarrhoeal disease in developing countries, the fourth leading cause of mortality and morbidity in the pediatric population worldwide. These enterotoxins typically contain 18 or 19 amino acid residues, are stable to proteases and maintain their bioactivity even after incubation at 100°C for 15 minutes. Examples of such heat stable ST enterotoxins are listed in Table 1 below:

Table 1 Source Structure E.Coli STa Asn- hr-Phe-Tyr-Cys-Cys-Glu-Leu-Cys-Cys-

Asn-Pro-Ala-Cys-Ala-Gly-Cys-Tyr E.Coli STh Asn-Ser-Ser-Asn-Tyr-Cys-Cys-Glu-Leu-Cys- Cys-Asn-Pro-Ala-Cys-Thr-Gly-Cys-Tyr Yersinia ST Ser-Ser-Asp-Trp-Asp-Tyr-Cys-Cys-Asp-Leu-

Cys-Cys-Asn-Pro-Ala-Cys-Ala-Gly-Cys-Tyr Vibrio ST Ile-Asp-Cys-Cys-Glu-Ile-Cys-Cys-Asn-Pro- Ala-Cys-Phe-Gly-Cys-Leu-Asn

Guanylin Pro-Gly-Thr-Cys-Glu-Ile-Cys-Ala-Tyr-Ala-

Ala-Cys-Thr-Gly-Cys Uroguanylin Asn-Asp-Asp-Cys-Glu-Leu-Cys-Val-Asn-Val- Ala-Cys-Thr-Gly-Cys-Leu

The naturally occurring ST enterotoxins have a relatively complicated secondary structure, due for example to the presence of multiple disulphide bridges. Thus E.Coli STa has disulphide bridges between the Cys residues at positions 5 and 10, 6 and 14 and 9 and 17. Such cell receptor binding oligopeptides are of interest both for therapeutic and diagnostic purposes.

Thus for example an oligopeptide capable of binding to a cell surface receptor may be coupled to a therapeutically or diagnostically effective moiety and serve as a biological vector to deliver that moiety to sites possessing such cell surface receptors. This is particularly interesting in the field of radiopharmaceuticals where a radionucleotide capable of detection in a nuclear imaging technique, e.g. scintigraphy, SPECT or PET, may be caused to accumulate at body sites having the target receptors and so allow such sites to be detected and if desired mapped. Where the receptor occurs solely or predominantly on undesired cells, e.g. tumour cells, a cytotoxic dose of radiation can likewise be delivered to the site of concern using radiation emitting vector-bound radionuclides .

In the case of the heat stable ST enterotoxins mentioned above, the ST receptors occur naturally only in the intestinal lumen and are found elsewhere in the body only as a result of metastases of colon cancers. Parenteral administration of a radionuclide-tagged ST oligopeptide can be used to detect and treat such metastases (see US-A-5518888 and W095/11694) .

Generally however the use of native oligopeptides as vectors for therapeutic or diagnostic moieties is problematic, for example because the native oligopeptide will bring about its natural consequences on binding at the cell surface receptors.

It is therefore well-known to seek to develop peptidic or non-peptidic analogs of native cell surface receptor binding peptides, in particular substances which are antagonists rather than agonists, i.e. which bind to a receptor without triggering the natural consequences of a native ligand binding to the receptor. In the case of the ST enterotoxin, an ST antagonist might be administered orally so as to suppress diarrhoea.

Analogs of the native ligands may also be sought which have greater stability in vivo, to labelling conditions, or to storage, etc. or which have a longer blood pool residence time in vivo.

The generation of a candidate receptor - (ant) agonist may typically proceed from modification of a native ligand, by generation of a non-peptidic analog of the native ligand, by selection from a random library of biological oligomers, or by generation of a chemical structure which fits both geometrically and functionally into the receptor pocket where this has been elucidated. In the latter case, drug candidate development will generally involve computer aided molecular design (CAM- D) .

All of these techniques have significant drawbacks - thus while candidate structures may be identified by CAM-D, it is relatively time-consuming to refine the selection process if the identified optimum ligand structure turns out to be difficult to synthesize, unacceptably toxic, insufficiently stable in vivo or even insufficiently selective. The use of randomly generated combinatorial libraries, e.g. of oligopeptides, oligonucleotides or oligosaccharides, is associated with problems of deconvolution and may not serve to identify materials with suitable distributions, stabilities and selectivities without extensive experimentation .

Modification of known native ligand structures generally proceeds via single amino acid alterations, e.g. deletions, insertions or exchanges. In a large oligopeptide, the number of possible single amino acid modifications is immense and the technique is accordingly highly time-consuming.

The generation of a purely non-peptidic structural and functional analog of a known native ligand may be accompanied by the same problems as mentioned above for CAM-D, namely refinement beyond the initial candidate may involve complex chemistry and toxicity, selectivity and stability may be relatively poorly predictable.

The present invention is directed to an improved method of selecting a candidate drug compound with receptor binding ability in which a novel combination of rational and combinatorial drug design techniques is used.

Viewed from this aspect the invention provides a method for selecting a candidate drug compound, said method comprising: i) identifying a starting oligopeptide having a binding affinity for a cell surface or other receptor site; ii) generating a plurality of homologous oligopeptides having single site amino acid modifications relative to said starting oligopeptide and identifying modification sites which produce oligopeptides having an increase (or decrease) in the ratio of binding affinity between oligopeptide and said receptor site to biological response of binding between oligopeptide and said receptor site; iii) identifying components of said starting oligopeptide responsible for the secondary structure thereof ; iv) for modification sites identified in step (ii) selecting a set of amino acids, e.g. a set having a functional equivalence to the amino acid at that modification site in the starting oligopeptide and/or in the homologous oligopeptides having said increase (or decrease) ; v) for components identified in step (iii) selecting at least one non-natural amino acid which mimics the secondary structure of the component; vi) generating a first combinatorial library of oligopeptides having at said identified modification sites amino acid residues corresponding to said selected sets of amino acids and preferably having at the sites of said components amino acid residues corresponding to said non-natural amino acids; vii) identifying members of said first library having a binding affinity value above a predetermined minimum and optionally having a relatively high (or low) ratio of binding affinity to biological response; optionally viii) changing, by expanding and/or contracting, the sets of amino acids for said modification sites or non-natural amino acids for said components and generating a further combinatorial library of oligopeptides having members not present in said first library, and identifying members of said further library having improved binding affinity values relative to the identified members of said first library and/or increased (or decreased) ratios of binding affinity to biological response and/or other improved in vivo parameters such as stability, toxicity, metabolism, etc.; and, optionally after repeating step (viii), and with the proviso that at least one of said first and further libraries contains oligopeptides containing residues of said non-natural amino acid or acids ix) selecting from the identified members of said library or libraries a drug candidate.

In the selection of the amino acids for the sets for the generation of the combinatorial libraries and indeed for the selection of the potential modification sites, it will generally be preferable to identify amino acids and oligoamino acid motifs which are likely to result in poor in vivo stability or short blood pool lifetime and to seek to exclude these from the libraries to be tested. Thus for example the following amino acids or amino acid sequences will generally be ones to avoid or replace:

Tyr-Gly; peptides with free α-amines; Pro-x (where x is a natural amino acid) ; x-Phe-x; peptides with free α- carboxyls; x-Lys; Lys-Lys; Gln-Gln; x-Arg; Arg-Arg; Gly- Gly; Ala-Ala; and x-Tyr.

For the selection of non-natural amino acid mimetics of secondary structures, many possibilities are known in the art . Thus for example the following substitutions may be made:

β-turn mimetics

(where

(where R is an optional substituent, e.g. optionally hydroxylated C_λ.₁₀ alkyl, aralkyl, aryl or aralkyl)

γ-turn mimetics

β-sheet mimetics

Disulphide bridge mimetics Carba-sulphide analogs

Thus for example in E.Coli STa, an N-terminal Asn is identifiable as a helical end cap (capable of hydrogen-bonding to a nearby amide carbonyl , here of

Tyr⁴) which can stabilize a helix, here the 3₁₀ helix.

Thus were mutations which favour helicity to enhance enterotoxigenicity, it would be expected that an ST receptor agonist would require the functional groups at the N terminus to adopt a similar three-dimensional configuration .

Likewise the sequence Asn-Pro-XX (where XX is Ser,

Thr or Ala) is usually associated with a Type I beta turn. Such a sequence occurs at positions 11-13 of E.Coli STa. Other mutations for Asa or Ala generally give rise to other forms of β-turn or to γ-turns. Mutations disrupting the Type Iβ turn in E.Coli STa give rise to agonists (e.g. Alal3 Gly, Asnll Val or Asnll Gly) or antagonists (e.g. Prol2 Gly) . (By XX-n-YY is meant that XX" is changed to YY") .

The C-terminus of the E.Coli STa is also associated with secondary structure, namely a Type II β turn. The amino acid combinations responsible for secondary structure can in general either be predicted from known structures or can be determined by X-ray crystallography. (See "Protein folds: a distance-based approach", Eds. Bohr and Brunak, CRC Press, NY (1995) and "Protein Folding" Eds. Gierasch and King, AAAS, Washington DC, (1990)).

For the generation of the first library in step vi) , it is possible to use restricted sets of amino acids by panning one or more restricted libraries produced by modification at a restricted set of modification sites or secondary structure component sites, e.g. by panning a "sub- library" in which one, more or all of the secondary structure component sites are left unchanged or one in which one or more of the least promising (or indeed the most promising) of the modification sites are left unchanged.

In the generation of the further library, in optional steps (viii) , the sets of amino acids for the modification/component sites will typically be changed to avoid amino acids not found to be effective in the first library and to substitute amino acids which are functionally similar to the successful amino acids, e.g. amino acids omitted from the sets used in generation of the first library either to keep down the overall library size or because they are relatively expensive to purchase or complex to fabricate . Thus for example in the first library one might select a single member of a homologous group of amino acids, e.g. using a chloro compound as a representative of a group of halogenated compounds, and if this is a successful amino acid then other members of the homologous group would be added in for the generation of the further library.

It should be stressed that the sets of amino acids, for the modification sites may include natural (L) amino acids, as well as non-natural amino acids, e.g. D-acids, α-chain length homologues, α-chain substitution homologues, etc. It should also be stressed that the sets of amino acids used in the generation of the first library may, and in general will, also include the amino acid present in the particular modification/component site in the starting oligopeptide. The starting oligopeptide used in the method of the invention may be a native (i.e. naturally produced) oligopeptide (e.g. E.Coli STa) known to have a binding affinity for a cell surface or other receptor site and known to produce a particular biological response. Alternatively it may be a composite of a set of known oligopeptides (e.g. the set of ST enterotoxins listed in Table 1 above) in which case the reference point for binding affinity and biological response in step (ii) will either be the values for one member of the set, an average for the members of the set, or the value determined for a particular composite of the members of the set. Yet further, the starting oligopeptide may be a non-native oligopeptide, e.g. one known from the literature or one found using library screening techniques, e.g. one identified by the use of a phage display library.

In general, the method of the invention will be used to identify a candidate receptor antagonist; however in certain circumstances, e.g. where a drug which provokes biological response is desirable, the method of the invention will be used to identify a candidate drug compound which is a receptor agonist. Generation, panning and if required deconvolution of the combinatorial libraries used in the method of the invention may be performed using known techniques. Generally, however, oligopeptide synthesis will be carried out on a solid phase, and binding affinity will preferably be assessed m the solution phase, e.g. following release of library members from the solid phase. Where this is done, the solid phase bound library can either be released to produce a solution containing the library members or solutions containing individual library members. In the latter case deconvolution is not required, in the former it is.

Deconvolution techniques have been widely described in the literature m recent years but where deconvolution is required according to the present invention this will preferably involve orthogonal scanning, i.e. the use of sub-libraries to work back and identify the relevant oligopeptide structures. In this regard, the libraries will preferably be created on polymer beads using the split-and-mix technique.

If the avoidance of deconvolution is desired, the libraries may be generated by the multi-pin method, or they may be generated on an extensive (e.g. sheet-like) substrate using the spot technique or using masks and photo-deprotection. In these latter two techniques, the identity of the library member is determined by its location on the substrate. A combined solution phase library can be created by simultaneous release of all substrate bound oligopeptides - otherwise individual library members may be released into separate containers, e.g. by punching out pieces of substrate from the oligopeptide growth sites for the different oligopeptides or by pressing the substrate against an array of tubes and releasing the different oligopeptides into their adjacent tubes.

If desired, a selection for binding affinity may be carried out before biological response is determined, thus reducing the number of compounds for which biological response is determined. This will be particularly preferable where testing for biological response requires a tissue sample or a live animal. (Binding affinity will generally be determinable using only cell cultures or tissue samples) . Thus before step (ix) , if a selection for a biological response has not already been carried out in steps (vii) or (viii) , such a selection will generally be effected. For binding affinity, standard assays may be used, e.g. turbidification, competitive binding against labelled receptor binding agents, etc. For biological response, one may use an assay appropriate to the particular response as well as to the animal model if the response is being tested in vivo. Design and performance of such assays are well within the normal skill of those skilled in the technical field.

Following drug candidate identification using the method of the invention, the drug candidate may be manufactured, optionally conjugated to a therapeutically or diagnostically effective moiety as discussed above. Such a process, as well as the products thereof, pharmaceutical compositions containing them optionally together with at least one pharmaceutically acceptable carrier or recipient (e.g. water for injections, physiological saline, Ringers solution, tabeletting aids, sweeteners, pH adjusters, viscosity modifiers, propellants, bulking agents, etc), and methods of treatment or diagnosis using them are all deemed to form further aspects of the invention.

Following identification of an oligopeptide candidate drug compound using the method of the invention, it may be possible to produce an analog in which the peptide backbone structure is wholly or partially replaced by a non-peptide backbone, e.g. replacing NH-CH-CO-NH-CH-CO-

R_αl R_α2

(where R_αn are peptide side chains) backbone components by equivalently R_αn side chain substituted backbone components in which some or all of the peptide bonds are replaced, e.g. by backbone structures such as

-NH-CH-CH₂-NH-CH-CH₂-

I I

-CH₂-CH-CO-CH₂-CH-CO-

-CH₂-CH-CH₂-CH₂-CH-CH₂-

=CH-CH-CH=CH-CH-CH=

-NH-CH-CH=CH-CH-CO-

R «,ι R«2

R_αl R ^■,al -0-CH-CH₂-0-CH-CH₂-

R-αl R,l

-NH-CH-CH,-CHOH-CH-CO-

10

- CHOH - CH - CH₂ - CHOH - CH - CH₂ -

15

-NH-CH-CHOH-CHo-CH-CO-

20

■ CH₂ - CH - CHOH - CH₂ - CH - CHOH -

25

-NH-CH-CH-S-CH-CH₂-

30

-NH-CH-CH,-S-CH-CO-

35 -S-CH-CH₂-S-CH-CH₂-

X

-NH-CH-C(H)_m-L-CH-CO-

-X ^x^ ^x-

-L-CH-C(H)_m-L-CH-C(H)_m-

(where X = =N-N=N- , L = N, and m is 0 or

X = -S-CR-,,= , L = =N-C-, and m is 1^*

-NH-CH-CS-NH-CH-CO-

-NH-CH-CS-NH-CH-CS- , etc I I

(eg, wherein amide oxo groups are optionally replaced by CH₂, CS or CHOH groups and amide NH groups are optionally replaced by -0-, -S-, CH₂ or CHOH groups or wherein an amide group is optionally replaced by a 5- or 6- membered homo or most preferably ring nitrogen- containing hetero-cyclic group) or wherein the peptide structure is replaced in part or full by a peptoid structure , i . e . where

-NH-CH- CO-NH-CH- CO-

becomes

- CH₂-N-CO-CH₂-N-CO-

Such non-peptide, peptoid or partially peptide analogs, their manufacture and their use form further aspects of the present invention.

While not restricted to the identification of the ST receptor binding agents, the invention will now be described further using such agents as an appropriate means of illustration. The native ST enterotoxin scaffold has undesirable physicochemical characteristics which reduce its attractiveness for use as a vector for therapeutic or diagnostic moieties, eg the multiple disulphides, the diarrhoeagenic activity and its short blood pool residence time. A number of mutations have been made with retention, to a greater or lesser degree, of the binding affinity and biological response.

Cys and Cys mutations

This disulphide pair is able to tolerate the loss of the amino function at Cys⁵ or the reversal of its stereochemistry (ie D-Cys replacement) with no detectable loss in biological response. The stereochemistry of Cys¹⁰ however is important since its replacement with D-Cys results in a 5000 fold loss of biological response as compared with E.Coli STa. However, if the Cys⁵ Cys¹⁰ disulphide is replaced with a carbo-sulphide analog (resulting in a simpler bicyclic oligopeptide) only a 100 fold loss in biological response is observed.

Cys⁶ and Cys¹⁴ mutations

This disulphide cannot tolerate substitution of either Cys by a D-Cys without a major loss (>770 fold) in biological response.

Cys⁹ and Cys¹⁷ mutations

This disulphide is not apparently necessary for pharmacophore formation. Substitution of Cys¹⁴ with D- Cys results in a 3 -fold loss of biological response (measured as a drop in enterotoxigenicity) while substitution of Cys⁹ by D-Cys results in a 0.5-fold loss and the simultaneous mutation of each to an alanine only yields a 200-fold loss in activity, eg as determined in a suckling mouse assay.

Glu⁷ mutations

Mutations which vary the hydrogen bonding capacity (eg Asp or Gin substitutions) or the length of the side chain in position 7 appear to have relatively little effect on biological response. However mutations which change the stereochemistry of the Glu⁷ residue (eg D-Glu substitution) result in significant loss of biological response. There is minimal difference observed with mutations replacing Glu' by helix disrupting Pro or helix promoting Ala indicating that the 3₁₀ helix observed in the X-ray structure of native ST is not a required secondary structure. It likewise appears that Glu⁷ may be replaced by cyclic analogs of aspartic acid, eg Pro (a reduced pyroglutamic acid) . Thus acceptable Glu⁷ replacements include Asp, Gin, Ala and Pro.

Leu⁸ mutations

Mutations such as Val, Ala, and Lys, but not D-Leu are acceptable at this position indicating that creating a hydrophobic interaction on a specific region of the oligopeptide surface is important.

Asn¹¹ mutations

Neither change of stereochemistry nor positive charges appear to be acceptable at this position. However, changes in hydrophobicity, hydrogen bonding potential and steric hindrance appear to be tolerated. Overall this position seems to be fairly tolerant to mutation. Thus replacement by Val, lie, Ala, Tyr, Glu, Asp, His, Gly and Leu and less preferably Arg and Lys are tolerated. Indeed replacement by any β-branched, non-basic amino acid or analog is tolerated.

Pro¹² mutation

This position appears to require rigidity rather than any particular side chain identity. Thus replacement by D-Pro and Ala is tolerated but replacement by Gly is not.

Ala mutations

This position appears to be the least tolerant to changes in stereochemistry, hydrogen bonding potential, steric hindrance or conformational flexibility. Replacement by Ser or Gly may be tolerated, but replacement by Phe, Lys, Leu, Arg or Val does not appear to be tolerated. Ala¹⁵ mutations

This position appears to allow for a wide range of substitutions without significant change in biological response, eg substitution by Thr or D-Ala.

Gly¹⁶ mutations

This position appears to allow for a wide range of substitutions without significant change in biological response, eg substitution by Ala.

The major trigger for enterotoxigenicity appears to be the Type Iβ turn of the component Asnⁿ-Pro¹²-Ala¹³. Mutation here can destroy both binding affinity and biological response. However, it is feasible to find mutations that, while dropping the binding affinity cause an increase m the binding affinity to biological response ratio. Such mutations include for example D-

Cys ,9³, Gln,y-1¹2 a_„ndJ rD,-C-y-As^**"

Accordingly for an orthogonal scanning of enterotoxin E.Coli STa analogs, the following sets of amino acids/mutation sites are proposed for the construction of a first library:

Positions 7, 8, 9, 11, 12 and 13:

G, A, N, P, Y, E, R, L, dA, dN, dP, dY, dE, dR, dL or equal mixtures of all 15 amino acids.

Positions 6, 10, and 15: Cys

Position 5: Bromo butyryl acid (precursor to carbo- sulphide 5:10 bridge)

Positions 1 - 4 and 15 to 17: omitted. As an extra modification, Asn¹¹ - Pro¹² -Ala¹³ may be replaced by the β-turn mimetic

Other secondary structure components which may be varied include the structures set out above under the headings β-turn mimetics, γ-turn mimetics, β-sheet mimetics and disulphide bridge mimetics.

Thus ST analogs that may be produced include

(CH₂) _p-NH-CO- (CH₂) _p ^

H₂N-CH-CO-Cys-Glu-Leu-Ala-NH-CH-CO-Asn-Pro-Ala-Cys-COOH

(where p is 1 to 4 and ... represent a Cys:Cys disulphide bridge)

(ιCuHn₂₂) _p —ι

H_?N-CH-CO-NH-CH-CO-Glu-Leu-Ala-NH-CH-CO-Asn-Pro-Ala-NH-CH-COOH (CH₂)₂ S CH₂

H₂N-Cys-Cys-Glu-Leu-Ala-Cys

BTM

' Cys-^ HOOC

(where BTM is a β turn mimetic]

H₂N-Cys-Cys-Glu-Leu-Ala-Cys-NH- -Ala-Cys-COOH

and

H₂N-CH-CO-Cys -Glu-Leu-Ala-NH-CH-CO-Asn- Pro-Ala-Cys -COOH

(CH₂) ₂ — SCH₂

As mentioned above, the combinatorial libraries may be synthesized using solid state peptide synthesis, eg on beads or sheets of polymers such as polystyrenes, polystyrene copolymers (eg PEG-polystyrene polymers), PEGs, celluloses (eg paper, cotton, etc) , carbohydrates (eg dextrins) , or polyacrylamides, or controlled pore glass. Advantageously, the substrate will be a material which is swellable in aqueous and organic solvents, eg a PEG-polystyrene copolymer. Since the oligopeptide products may contain multiple disulphide bridges it is particularly desirable to use acetamidomethyl (Acm) protected resins. This use is novel and forms a further aspect of the invention. Viewed from this aspect the invention provides a process for the solid state synthesis of an oligopeptide on a resin substrate, optionally followed by release of the oligopeptide from the substrate, characterised in that as said substrate is used an acetamidomethyl-protected polymer. The oligopeptide can be cleaved from such a resin using a reagent such as iodine.

In this way, C-terminal cysteine residues may conveniently be attached to the resin surface.

Thus for example an ST enterotoxin analog may be prepared as follows:

DCC Resin-NH₂ + HOOC (CH₂ ) ₂CONH₂ Resin-NH-CO- (CH₂ ) ₂CONH₂

( I )

KOH ( I ) -^ Resin-NHCO- (CH₂ ) ₂ - CONHCH₂OH

HCHO ( I I ) L-Cysteine-OCH₃ NH₂

(II) Resin-NHCO(CH₂)₂CONH-CH₂-S-CH₂CHOCH₃

TFMSA/TF (1:9)

(III) ("Cys (Acm) Resin")

(III) + Stepwise peptide growth

-> Bromo-Butyryl-Cys (Acm) -Glu (tBu) Leu Ala Cys (StBu)Asn (Trt) Pro-Ala-Cys (Acm) Resin

(IV)

90% TFA (tBu)₃P H₂0/pH8 (IV) ^ s. ^

, S-CH₂

(CH₂) ₃CO . Cys (Acm) .Glu . Leu.Ala.NHCHCO .Asn. Pro .Ala. Cys (Acm) COOH

(V)

(V) . CH₂CO-Cys-Glu-Leu-Ala-NHCHCO-Asn-Pro-Ala-Cys-COOH

(CH₂) ₂ SCH₂

(VI)

The invention will now be described further with reference to the follow non-limiting Examples

Example 1

Synthesis of a library of

(Acm) -Resin A) Synthesis of the FMOC-Cys-COOCH₃

i) Linker synthesis (cf. F. Albericio, et al . , Synthesis, p 271 et seq (1987))

(I)

(Nb: The CH₂ group can alternatively be CR'X, where X = H or R' and where R' can be any alkyl, aryl, heteroalkyl or heteroaryl group, or an amino acid, for example) .

0.17 moles succinamic acid (19.9g) and 0.018 mole KOH (l.Og) are dissolved in a solution of formaldehyde:water (7:3 v/v; 12 ml, 0.012 moles). The mixture is stirred at 70°C for 5 minutes, then overnight at room temperature and the acidified with 6N HC1 to pH 7.0. The solution is rotovapped to a syrup and redissolved in ethylacetate, then dried over MgS0₄. Crude product is obtained by filtration and concentration of the filtrate. ii) Cysteine-linker synthesis

29 moles (4.32g) of N-hydroxymethyl succinamic acid and 24 moles (4.12g) of L-cysteine, methyl ester, hydrochloride are dissolved in 5.6 ml H₂0. The mixture is cooled in an ice bath and TFMSA/TFA (37 ml of a 1:19 mix) is added. After stirring for 90 minutes under argon, the solution is rotovapped to an oil and then again after additions (5 x 20 ml) of diethyl ether to yield compound II.

iii) FMOC-Cysteine-linker synthesis

Compound II oil from step A(ii) is redissolved in 50 ml 10% Na₂C0₃ (aq) and adjusted to pH 10 with solid Na₂C0₃. To this solution (cooled on ice) is added 21 mmoles FMOC-OSu (7.08g) in 50 ml dioxane . After stirring for two hours, the product is purified by acidifying (HC1 addition) and extraction into ethyl acetate. The organic layers are combined, dried over MgSO„ and recrystallized from chloroform/hexane to yield compound III in pure form.

iv) FMOC-Cysteine- (Sum) -Resin synthesis

(Sum)-Resin

To 1.0 mmole of compound III was added 1.0 mmoles DCC in a total volume of 5 ml CHC1₃. After one minute of stirring, this slurry was added dropwise to 1.0 mmoles of aminomethylPEG/polystyrene graft copolymer. After 10 minutes, 1 mmole of hydroxybenzotriazole in 50 ml DMF was added and the slurry was stirred for 2 hours. An additional 1 mmole of DCC in 5 ml CHCA was added and the reaction allowed to shake overnight at 25°C.

After 16-24 hours, excess amino groups were capped with acetic anhydride and the resin was thoroughly washed and dried. Substitution level was determined on a weighed amount of resin spectrophotometrically, based on cleaved FMOC group and a quantitative ninhydrin assay on the same resin aliquot. B. Synthesis of an ST analog library based on β and L amino acids

Into each of 15 reaction vessels (each vessel has a sintered frit and stopcock for filtration) is added 150 mmoles of FMOC-Cys (Sum) -resin (compound IV). The FMOC group is removed and the resin washed according to normal protocols (see G.B. Fields et al . "Principles and practice of solid phase peptide synthesis" in "Synthetic Peptides: a user's guide" W.H. Freeman & Co., New York, Editor Gregory A. Grant) for each reaction. Each reaction vessel is then individually reacted with an excess of one of the representative protected amino acids (Gly, D and L versions of Ala, Asn, Pro, Tyr, Glu, Arg, Leu) . One-sixth of each of these resins is set aside and the remaining five-sixths (by dry weight) is mixed to homogeneity with the other resins . At this point, there are 16 reaction vessels containing resin, as represented in Figure 1 of the accompanying drawings .

Each of the individual ^#1 dipeptide resins is split equally into 15 individual reaction vessels, deprotected, fully coupled with each of the representative amino acids and then recombined. This cycle is repeated for each amino acid position to be explored, as shown in Figure 2 of the accompanying drawings. Evaluation of the bioactivity of each of these peptide sets will enable determination of which amino acid is most preferable in the first incorporated position.

Similarly, if the homogeneous XC-{P) mixture (*2 in Figure 1) is split equally into 15 individual reaction vessels, deprotected, fully coupled with each of the representative amino acids, the set shown in Figure 3 of the accompanying drawings is created in order to fully explore the second combinatorial position. The strategy can be carried through the peptide to determine the best sequence (s) classes for bioactivity. If it is desired, sub-libraries can be similarly made based on which class of side chain is identified as optimal. For example, Gly, Ala and Asn are representatives of the polar amino acids and their identification as preferred amino acids would necessitate synthesis of the Ser, Thr and Gin analogs to fully explore the class. The class designations and their representatives are set out below.

Representative amino acids and subsets

The underlined amino acids have been chosen to represent their respective subsets.

Conformational (Gly. Pro) Aromatic (Phe. Tyr. Tip)

Charged (Asp. Glu. Lys. Arg)

Aliphatic (Ala. Val. Leu. Ile_^ Met]

Claims

Claims :

1. A method for selecting a candidate drug compound, said method comprising: i) identifying a starting oligopeptide having a binding affinity for a cell surface or other receptor site; ii) generating a plurality of homologous oligopeptides having single site amino acid modifications relative to said starting oligopeptide and identifying modification sites which produce oligopeptides having an increase (or decrease) in the ratio of binding affinity between oligopeptide and said receptor site to biological response of binding between oligopeptide and said receptor site; iii) identifying components of said starting oligopeptide responsible for the secondary structure thereof ; iv) for modification sites identified in step (ii) selecting a set of amino acids; v) for components identified in step (iii) selecting at least one non-natural amino acid which mimics the secondary structure of the component ; vi) generating a first combinatorial library of oligopeptides having at said identified modification sites amino acid residues corresponding to said selected sets of amino acids; vii) identifying members of said first library having a binding affinity value above a predetermined minimum and optionally having a relatively high (or low) ratio of binding affinity to biological response; optionally viii) changing, by expanding and/or contracting, the sets of amino acids for said modification sites or non-natural amino acids for said components and generating a further combinatorial library of oligopeptides having members not present in said first library, and identifying members of said further library having improved binding affinity values relative to the identified members of said first library and/or increased (or decreased) ratios of binding affinity to biological response; and, optionally after repeating step (viii) , and with the proviso that at least one of said first and further libraries contains oligopeptides containing residues of said non-natural amino acid or acids; and ix) selecting from the identified members of said library or libraries a drug candidate.

2. A method as claimed in claim 1 wherein the starting oligopeptide is an up to 19-mer.

3. A method as claimed in either of claims 1 or 2 wherein in step (iii) said set of amino acids is a set having polar, aromatic, charged, aliphatic, hydrophilic, conformationally constrained or lipophilic side chains where the amino acid at the identified modification site in the starting oligopeptide itself or in the homologous oligopeptide has a polar, aromatic, charged, aliphatic, hydrophilic, conformationally constrained or lipophilic side chain respectively.

4. A method as claimed in any one of claims 1 to 3 wherein in step (iv) said at least one non-natural amino acid is selected from ╬▓-turn mimetics, ╬│-turn mimetics, ╬▓-sheet mimetics and disulphide bridge mimetics.

5. A method as claimed in claim 4 wherein said ╬▓-turn mimetics are selected from

(where X =

(where R is an optional substituent, e.g. optionally hydroxylated C^_o alkyl, aralkyl, aryl or aralkyl) .

6. A method as claimed in either of claims 4 and 5 wherein said ╬│-turn mimetics are selected from (4-(l- amino-alk-1-yl) -1 , 4-diaza-5 , 7-dioxo-cycloheptyl) - ethanoic acids.

7. A method as claimed in any one of claims 4 to 6 wherein said ╬▓-sheet mimetics are selected from compounds of formula

(where R is an optional substituent, e.g. optionally hydroxylated C_x_₁₀ alkyl, aralkyl, aryl or aralkyl).

8. A method as claimed in any one of claims 1 to 7 wherein said first combinatorial library contains oligopeptides having at modification sites of said components amino acid residues corresponding to non- natural amino acids.

9. A process for the manufacture of a drug compound comprising performing a method as claimed in any one of claims 1 to 8 and manufacturing said drug candidate, optionally conjugated to a therapeutic or diagnostic moiety, and optionally with replacement of the peptidic backbone or a part thereof with a non-peptidic bacbone, as said drug compound.

10. A process as claimed in claim 9 wherein at least one amide oxo group is replaced by a CH₂, CS or CHOH group.

11. A process as claimed in either of claims 9 and 10 wherein at least one peptidic backbone amide NH group is replaced by a 5 or 6 membered homo or heterocyclic group .

12. A process as claimed in any one of claims 9 to 11 wherein at least one peptidic backbone amide group -CONH- is replaced by a peptoid backbone amide group -NH-CO-.

13. A process for the solid state synthesis of an oligopeptide on a resin substrate, optionally followed by release of the oligopeptide from the substrate, characterised in that as said substrate is used an acetamidomethyl-protected polymer.