Background The application of modern molecular genetics to the study of cancer biology has produced enormous advances in our understanding of the principles underlying mammalian ceil transformation. The implied promise of molecular oncology is that this knowledge can be exploited to create rational, mechanism-based inhibitors of specific biochemical processes that are essential to the malignant phenotype of cancer ceils. Recently, the detailed biochemical circuitry controlling cell replication has also begun to emerge. These insights have served as points of departure for defining molecular targets of drug discovery programs. Unfortunately, the translation of modern molecular biology concepts into practical cancer therapeutics has proven to be far more problematic than was first anticipated, and few true breakthrough agents have been found that significantly improve the survival of most cancer patients. A partial explanation for these difficulties lies in understanding the fundamental process of drug discovery and the nature of pharmaceutically useful molecular targets. it is not our intent here to review ail aspects of pharmaceutical research in oncology, but rather to focus on the principles of small molecule drug discovery that determine which molecular targets are likely to be most useful. We have selected specific examples from studies of oncogenes and signal transduction pathways that highlight many of the biochemical and pharmacological issues that should be considered in any pharmaceutical research program.

Drug Targets As the knowledge base in molecular oncology grows, so too does the number of potential targets increase. Therefore, some system of prioritization is required to select the targets most likely to yield effective therapeutics. One important question to ask is whether or not a target is genetically linked to the pathophysiology of human cancers. The most commonly altered gene products in human solid tumors include ErbB2lHER (a membrane-associated tyrosine kinase), Ras (a GTP-binding protein), and proteins that affect transcription, such as Myc, Rb, or~ 53 (Figure 1; Bishop, 1991). These genetically validated targets anchor our conceptual framework of how cancer ceils differ from normal cells and thus represent biologically sound targets for drug discovery programs. Moreover, the genetic evidence for involvement of a specific molecule in the pathophysiology of cancer provides strong presumptive evidence that inhibition of the abnormal activity of that gene product or restoration of its normal function would be an effective cancer therapy.