22.9 N&V Feat Fishman MH

Drug discovery is a tough business scientifically, the traditional steps taking nearly a decade. In this article we are concerned only with the very beginning of the process, the identification of a target for a new drug. It is this step, the choice of a gene product of clinical relevance, that is the greatest impediment to expanding the pharmaceutical arsenal. In the United States, only 20–30 new chemical entities are approved as drugs each year, and the picture is much the same in Europe. Of these, only a quarter act on targets not already hit by an existing drug. Why has the deciphering of the 25,000 or so genes in the human genome not swelled the ranks of new targets? The major reason, of course, is that targets are of value for drug discovery only if they can be convincingly related to disease. Such validation really takes not the one year shown on standard pipeline charts (Fig. 1), but decades. For example, the discovery of statins as agents that lower cholesterol levels in the bloodstream rested on investigations that began 40 years earlier with a study showing the relationship between cholesterol level and vascular disease. The recent discovery of Gleevec, which has revolutionized the therapy of chronic myeloid leukaemia, certain gastrointestinal tumours and other cancers, was based on work that began in the 1950s (ref. 2). Thus, it has been argued that well-validated targets, the lowhanging fruit as it were, are simply exhausted. Validation is generally not a one-step process, but is rather an edifice built from studies in epidemiology and disease physiology, and from the results of research with animal models. The one exception is mendelian disease. In these disorders, such as cystic fibrosis and sickle-cell anaemia, the inheritance of a mutation in a single gene can be incontrovertibly linked to a physical characteristic, or phenotype — in this case the disease. Of the 6,000 or so illnesses with a mendelian pattern of inheritance, the gene responsible has been identified in approximately 1,200 (refs 5, 6). Historically, pharmaceutical companies have not concentrated on these diseases, in some cases because the affected protein is not tractable to pharmaceutical approaches, in others, perhaps, because the number of people affected is small. But the powerful role of a single gene in mendelian disease can provide insight into complex diseases where the same gene accounts for part of the phenotype. Statin therapy, for example, was initially directed to patients with a genetic predisposition to excessive levels of blood cholesterol. But after the drug’s efficacy and safety had been tested, the treatment was extended to a wider population of patients who had the same condition but due to many causes. In seeking new targets, we first need to think about how we describe disease. Medical textbooks are organized by organ system, and diseases are classified by their pathology or physiology. If we hope to come to grips with the heterogeneity of common disease and its consequences for the choice of targets for drug discovery, clinical manifestations of disease must be causally related to a molecular definition. A simple concatenation of all the molecular changes in a disease, for example a compilation of profiles of which genes are being transcribed, and when and where, would be chaotic. A further step of integration and interpretation is needed. Ideally, this would provide a ‘grammar’ that would apply right through from the discovery of a drug target, to its testing in animal models and finally to the treatment of patients. In this context, a grammar means a set of rules, not for organizing words into sentences but for translating gene products into medicines. Just as the grammar we use in our speech is embedded in language, a grammar of drug discovery should be embedded in biological systems.