Leukemia: A model for drug development

Early attempts at preclinical model development for cancer drug development relied heavily on mouse leukemias and lymphomas to detect agents with antitumor activity. These models were applied clinically, and the concepts of combination chemotherapy, remission induction, and maintenance treatment all developed in leukemia. Subsequently, the predominant impact of cytogenetics on probability of response to treatment and survival was first illustrated in leukemia. The power of a single drug to change the natural history of a disease was noted in acute myelogenous leukemia, in which a previously incurable disease was rendered potentially curable with 1-beta-D-arabinofuranosylcytosine. Additional studies illustrated the exquisite relationship between karyotype and response to specific agents. The ability to achieve a high proportion of complete remissions and to control the complication of intravascular coagulation, acute promyelocytic was noted with all-trans retinoic acid. The concept that new drug activity would only be demonstrated in patients with minimal prior therapy has been challenged by the curative potential of a number of agents in far-advanced hairy cell leukemia. In addition, fludarabine monophosphate (Fludara) was sufficiently active in advanced refractory patients that approval for this agent in chronic lymphocytic leukemia was granted by the Food and Drug Administration without comparative clinical trials. Fludara was initially a drug with limited therapeutic range, active only in indolent lymphoproliferative disorders. However, understanding of the multiple biochemical actions of this agent has led to its use in combinations with 1-beta-D-arabinofuranosylcytosine in acute myelogenous leukemia and myelodysplastic syndrome and with DNA active agents such as novantrone and cyclophosphamide in other lymphoproliferative disorders. The understanding of the various actions of this drug gives rise to a wide range of possibilities for biochemical modulation with agents active in solid tumors. The evolution of this understanding of the new role of Fludara has occurred over a period of 10 years. A drug with similar potential in the next decade is compound 506U78, an analogue of arabinosyl guanosine. This agent has potent activity in acute T-cell leukemia. Because it shares many of the activities of Fludara in interfering with enzyme systems important in DNA and RNA synthesis and DNA repair, it is likely that this agent will also have a wider scope than is presently obvious. The unique accessibility of leukemia cells for study has allowed hematologists to understand more fully the range of activities of new agents and has led to important new concepts in the area of drug development.