Acquired resistance to imatinib mesylate: selection for pre-existing mutant cells

Branford and colleagues have reported on the development of resistance to imatinib mesylate in patients with chronic myeloid leukemia (CML) and with Ph⁺ acute lymphoblastic leukemia (ALL) treated with this drug. Their report highlights the biologic interest of this phenomenon, which, at the same time, is cause for considerable concern from the clinical point of view.

These and other previous studies2-7 demonstrate that there are various mechanisms of resistance to imatinib. In a substantial proportion of patients, the basis for resistance is a genetic change in the BCR-ABL gene itself: particularly, point mutations within the protein tyrosine kinase (PTK) domain. This genetic mechanism has 2 interesting implications. First, it must be highly specific for imatinib; there is no reason to expect that there would be cross-resistance with any other commonly used chemotherapeutic agent. Everything leads one to believe that clinical resistance to imatinib, just like antibiotic resistance in bacteria, arises through a process whereby the drug itself selects for rare pre-existing mutant cells, which gradually outgrow drug-sensitive cells. Although this will require a very high sensitivity, it should be possible to find ways to detect such rare mutant cells in pretreatment samples.

Second, as we all share the excitement of imatinib being the herald of a new generation of antitumor agents,8-10 we may still have to learn some of the implications. “Conventional” cytotoxic agents target fundamental processes within the cell, such as DNA replication or the mitotic spindle. In principle, certain mutations in any of the genes involved in one such process could confer resistance toward a cytotoxic agent that is directed against that process. But since the genes involved in, say, DNA replication are indispensable in every cell, there may be enormous constraints for a mutation in any of them to yield a cell that is drug-resistant and viable at the same time. By contrast, since the inhibition of the normal Abl PTK by imatinib has relatively few side effects, it must mean that its function is dispensable in most cells, including normal granulocytes: therefore, there may be less stringent constraints for a mutation in the ABL gene to produce a PTK that is no longer inhibited by imatinib in a cell that is viable. The same reasoning may be extended to other genes that are involved in this or in any other of the multiple signal-transduction pathways known to exist in various types of cells and that, when mutated in tumors, are attractive targets for new drugs11 12 Thus a relatively high frequency of resistance mutations may be a price to pay for the target specificity and consequent reduced toxicity of new chemotherapeutic agents.

Is this going to be a major deterrent to the use of these drugs, or even to their development? Of course we hope not. Indeed, it is conceivable that new analogs can be synthesized that will break through the most common mutations conferring resistance to imatinib. But more in general, we can perhaps apply the principles learned from infectious diseases, where the use of a drug combination has been a time-honored approach aiming to minimize the risk of antibiotic resistance, since the statistical probability of a single bacterial cell having 2 rare mutations must be very low. There is every reason to assume that the same principles apply to mammalian somatic cells. One wonders whether, in a not too distant future, initial therapy of CML with imatinib alone will be frowned upon, just like single antibiotic therapy for tuberculosis would be frowned upon today. It seems not impossible that, just as “triple” or “quadruple” therapy is the standard of care today for infections by M tuberculosis, a 2-drug approach13 (perhaps cytosine-arabinoside or even the old busulphan together with imatinib) might become the standard of care for newly diagnosed CML.