Oncology drug discovery and clinical trial testing: who's listening?

Abstract

Considerable advances in the science of rational drug design have been realized since agents such as me4hotrexate, cyclophosphamide, and 5-ftuorouracil (5-FU) were introduced to the practice of cliical oncology. Although each of these compounds represents the product of a program of logical drug discovery, the accumulated clinical experience of approximately 90 years of testing these oncolytics has failed to provide the impact on oncology that had been widely anticipated. This disparity betwen realized and predicted clinical activity may simply represent the limitations of a scientific theory very much in its infancy in the 1950s; however, it may also reflect certain methodological flaws in the clinical trial process for evaluating new agents. Realization of the potential that now exists for the development of truly novel drugs based on a direct observation of the drug-receptor interaction (1,2) will r e q h a nearly seamless interface between the scientific laboratory and the clinical ward to fully exploit a therapeutic advantage. An analysis of the clinical development ofCB3717, a specific inhibitor of thymidylate synthase, suggests that this may not yet be fully realized. In 1981, Jones et al. announced the synthesis and preliminary biological activity of CB3717 (3). Interest in the development of a new thymidylate synthase inhibitor was driven by the meager clinical activity of 5-FU in gastrointestinal malignancies. This quinazoline-based antifolate was conceived through classical methods of medicinal chemistry: the parent antifolate nucleus, represented by methotrexate, was systematically varied until a compound with the desired biochemical properties was obtained. A folate-based nucleus for CB3717 was selected to circumvent the extensive metabolism required for activation of 5-FU to FdUMP. As a result of th is exercise, new insights regarding structural features essential for binding to the enzymes dihydrofolate reductase and thymidylate synthase were elucidated. The ternary inhibitory complex formed from CB3717 (thymidylate synthase/dUMP/CB37 17) was shown to be 30-fold more stable than the corresponding ternary complex with FdUMP. With the demonstration of encouraging preclinical activity in several cell lines and in animal models, CB3717 was recommended for initial Phase I testing in humans. In the Phase I clinical testing, an i.v. bolus dose administered every 3 weeks, and a 1 hour i.v. infusion administered every 3 to 4 weeks were evaluated (43). An MTD of 400 m g / d for the bolus schedule and 600 mg/m2 for the infusion schedule were defined. Doselimiting toxicities in each trial consisted of renal and hepatic toxicity as well as malaise. In the infusion trial, 7 of 76 patients had an objective response in ovarian,