an inhibitor. Although side-effects occur, these are acceptable due to the life-threatening nature of the disease.


Once the target enzyme has been identified then usually a 'lead' inhibitor has previously been reported or can be predicted from studies with related enzymes. The design process is then initially concerned with optimising the potency and selectivity of action of the inhibitor to the target enzyme using in vitro biochemical tests. Candidate drugs are then examined by in vivo animal studies for oral absorption, stability to the body's metabolising enzymes and toxic side effects. Since many candidates may fall at this stage further design is necessary to maintain desirable features and design out undesirable features from the in vivo profile. Since an in vivo profile in animal studies is not directly translatable to the human situation studies with human volunteers are also required before a drug enters clinical trials.

Computerised molecular modelling is nowadays an essential part of the design process but its relative importance in this process is determined by the state of knowledge concerning the target enzyme (see Chapter 3).

Ideally a high resolution crystal structure of the target enzyme with the active site identified by co-crystallisation with an inhibitor provides a knowledge of binding sites on the inhibitor and enzyme and their relative disposition. Furthermore an additional binding site may be identified so that a modified inhibitor using this additional site may be more potent or selective towards its target. Once the enzyme crystal structure is known the mode of binding of inhibitors fortuitously discovered later can be clarified (hindsight) to explain structural features responsible for their mechanism of action.

Usually for a newly discovered target the enzyme crystal structure is not known and the 3D-structure of the protein has to be less satisfactorily predicted from either NMR studies of by homology modelling from a related protein of known 3D-structure. For homology modelling the sequence similarity between the two proteins should be at least 30%. Either of these techniques can lead to the identification of prospective binding sites at the enzyme active site and on a lead inhibitor by "docking" the inhibitor at the active site. Observations can lead to further structural modification of the inhibitor to either improve fit or improve potency by taking advantage of additional binding areas such as hydrogen bonding groups or hydrophobic residues on the enzyme.

The relative positions of potential binding areas at the active site can provide a pharmacophoric pattern which can be used for de novo inhibitor design. Also, searching of 3D structural data bases can provide novel structures, designed for another purpose, with binding groups held in the correct 3D pattern through an appropriate carbon skeleton.

If a model of the enzyme active site does not exist, as is usual for a new target enzyme, then design is based on a knowledge of the substrate, a lead inhibitor (perhaps from a related enzyme) and of the mechanism of the catalytic reaction. Molecular modelling may enter into the design process at a later stage. A few selected examples are now given to illustrate this approach.

The anti-hypertensive drug captopril (8.48), an inhibitor or angiotensin I-converting enzyme (ACE), was designed from a knowledge of the substrate specificity and a known lead inhibitor of its target enzyme, together with a guess that the mechanism of action of ACE was similar to that of the zinc metalloprotease carboxypeptidase A about which much was known (see Section Further structural modification gave the related enalaprilat and, from molecular modelling using inhibitor superimposition, cilazaprilat (8.53).

Many mechanism-based inactivators of pyridoxal phosphate-dependent enzymes are known, some of which were designed from a knowledge of the mechanism of action of their respective target enzymes. Inhibitors of AADC, histamine decarboxylase, ornithine decarboxylase and GABA- transaminase designed in this way have proved to be useful drugs (see Section 8.6.4).

Aspartate proteases, such as renin and HIV-protease catalyse the hydrolysis of their substrates by aspartate ion-catalysed activation of the weak nucleophile water effectively to the strong nucleophile, hydroxyl ion. The hydroxyl ion attacks the carbonyl of the scissile amide bond in the substrate to give a tetrahedral intermediate which collapses to the products of the reaction (Equation [8.33]).

HIV-protease is an aspartate protease which cleaves polyproteins formed in viral reproduction to the correct length for viral maturation. Inhibitors of HIV-protease have been designed based on the amino acid sequence around a scissile bond of the polyprotein substrate and the structure of the tetrahedral intermediate. Using the substrate sequence 165-9 (Leu-Asp-Phe-Pro-ILeu) for a particular polyprotein a stable tripeptide analogue possessing a hydroxyethylamine moiety (-CH(OH)-CH2) to resemble the tetrahedral intermediate (-C(OH)2-) has been developed (see Section This compound, Saquinavir (8.79), has IC50=0.4 nM and is now in clinical trials as an agent to prevent the spread of viral infection. Stable amino- and carboxyl terminal blocking groups are present and the hydrophobicity of the proline in the substrate has been increased in the perhydro isoquinoline residue.

Other HIV protease inhibitors have been developed for other scissile bonds in the polyprotein substrate using a variety of functions (see Section which simulate the tetrahedral intermediate formed during catalysis. The crystal structure of the protease is now available leading to further designed inhibitors.

Modelling with a series of inhibitors by superimposition (matching) of key functional groups, similar areas of electrostatic potential, and common volumes may identify areas i.e. the pharmacophore, with similar physical and electronic properties in the more active members of a series. Whereas this approach is suitable for rigid structures it is less applicable to flexible molecules since the conformation in solution may be different to that required to efficiently bind to the enzyme active site.

Alternatively the common conformational space available to a range of active inhibitors can be used to distinguish this from the space available to less active or inactive analogues which may lead to a defined model for the pharmacophore.

More sophisticated methods have more recently been used to correlate a wide range of physicochemical properties with enzyme inhibitory activity and whereas some of these methods merely rationalise structure-activity relationships others may lead to new inhibitor design (see Chapters 3 and 5)).



The development of an inhibitor from its inception through to clinical trials is thwart with difficulties. After satisfactory in vitro screening of a potent inhibitor for selectivity towards its target enzyme (i.e. little effect on related enzymes) in vivo studies in animals are undertaken to establish that the candidate drug is well absorbed when administered orally, has a low rate of metabolism (long biological half life, and is free from toxic side effects. The in vivo studies present a formidable barrier to the development process and many candidate drugs can fall at this stage as has been described in Chapter 6 for the development of an aromatase inhibitor.

8.4.1 Oral absorption

Oral absorption of a drug may be improved by chemical manipulation to a biologically inert but more absorbable form of a drug which after absorption is converted by the body's enzymes to the active parent drug i.e. prodrug, (see Chapter 7). This approach has proved particularly useful for drugs possessing a carboxylic acid group which being in the ionised form at pH7 may not be well absorbed in the small intestine. Examples are ampicillin where well absorbed esters in the form of pivampicillin, becampicillin, talampicillin release ampicillin in the plasma by initial hydrolysis by esterases to an intermediate which degrades in the aqueous media (see Section 7.4.1).

The ACE inhibitor enalaprilat (see is well absorbed as its ethyl ester, enalapril, and the enkephalinase A (MEP) inhibitor SCH 32615 (8.3), a dicarboxylic acid, is well absorbed as the acetonide of the glycerol ester, SCH 34826 (8.4).

The potent enkephalinase inhibitor thiorphan (8.5) is not active parenterally but the protected prodrug, acetorphan (8.6) is absorbed through the blood-brain barrier and

/"N^ CH2-CH — NH-CH — CONH(CH3)rCOO" coo-

0 0

Post a comment