1 Drug discovery process

Contemporary lead discovery is usually driven by high-throughput screening of large numbers (millions) of compounds in corporate collections, although this has its problems in the identification of viable leads. Corporate collections are notorious for their tendency to be populated by larger, more lipophilic molecules, which, whilst being a good means of generating active compounds, do not necessarily give the best leads. There is a trend for more focused screening (as compound-handling methods improve) to complement the mass-screening campaigns and also for the high concentration screening of smaller ‘fragment’ molecules using a range of biophysical techniques. Since the decoding of the human genome, it has also been possible to correlate disease states with previously unidentified gene products (proteins) and to devise potential ligands using computerised modelling.

The goal of all of this is to produce smaller, leaner (i.e. less lipophilic) leads. A ‘hit’ molecule would be described as a molecule or closely-related series of molecules, which has demonstrable desirable activity in a particular primary screen. The term ‘lead’ would describe the best of a series of molecules, with established Structure–Activity Relationships (SARs) and, perhaps, activity in secondary assays of some evidence of a desirable pharmacokinetic profile. From such work a ‘tool’ molecule may arise. This is a term sometimes used to describe a compound with demonstrable activity, which can be used to establish mechanism or prove the concept of a particular target being efficacious in animal models – but without the necessary overall profile to be considered good enough to be a candidate. To produce good drug candidates, compounds should not only possess potent action against the target protein, but should also have the appropriate physical properties (solubility, size, lipophilicity) that necessarily accompany effectiveness and selectivity in vivo. Much of the lead optimisation process, which involves the iterative synthesis of analogue structures to hone the activity of a molecule, balanced with appropriate physicochemical properties required to achieve good pharmacokinetic (what the body does to a drug) and pharmacodynamic (what the drug does to the body) profiles. Combinatorial and/or multiple parallel synthesis techniques can have an impact at many stages in the drug discovery process – although the balance may be shifting to more focused libraries with better properties in the earlier stages – and to compounds designed with the aid of computer graphics in lead optimisation.

Historically, adverse drug metabolism and pharmacokinetic properties have been a major reason for failure in clinical trials. Modern drug design tends to take these factors into account at a much earlier stage, thereby reducing this liability. The fact is that optimising purely for potency in an isolated protein assay has not necessarily resulted in the most effective drugs. Compounds with high lipophilicity can bind strongly to many protein active sites since most proteins have multiple regions with high lipophilicity. One effect is to reduce selectivity for the target; another is for this generally undesirable physical property to confer poor solubility in aqueous media, including body fluids, resulting in sub-optimal ADME outcomes. Thus, the compound with the highest intrinsic activity is almost invariably not the best drug candidate. Improvements in ADME ideally need to be part of the original drug design process, thereby offering improved selectivity and reducing the potential for undesirable side-effects.

(Note that the study and analysis of ADME factors is usually described as ‘drug metabolism and pharmacokinetics’ or DMPK studies.)

Taking the above factors fully into account improves the chances that any selected candidate will progress through to Phase 1 trials. Even then, at the present time, some 40% of candidates drop out of contention at the preclinical evaluation stage. Success in Phase 1 clinical trials, establishes the tolerability and safety of the compound, and is followed by Phase 2 and Phase 3 clinical trials in which efficacy against the target disease is sought. When these data have been collated, analysed and reviewed by regulatory authorities, the permission to market the drug as a new medicine, as it may now finally be termed, might be granted.

The key stages in the drug development process are:

1. Disease selection and target identification
   • Unmet medical needs.
   • Genomics approaches.
   • Reducing a concept to practice to enable a discovery programme.
2. Lead identification and target validation
   • Methods for discovering chemical leads to enable an optimisation programme to commence.
   • Investigative tools may be derived from active compounds.
3. Lead optimisation: the essence of drug discovery
   • Achieving the best balance between novelty, potency and acceptable pharmacokinetic and pharmacodynamic properties
   • The quality of drug candidates and their chance of making it to market depend on originality and good science at this stage.
   • Intellectual property rights are essential to protect new discoveries.
4. Candidate selection
   • Forming a short list of potential development candidates based on preliminary toxicology, ADME studies and efficacy in an animal disease model.
5. Preclinical evaluation
   • Selection and characterisation of preferred drug candidate.
   • Preparation for Phase 1 studies in humans.
6. Clinical evaluation
   • Assessment of the safety and dosing regimes of the candidate drug in humans.
   • Evaluation of the effectiveness of the drug in selected patients in a clinical context.
   • Compilation of a data package to enable product registration, prior to manufacturing and sale.