Despite the dramatic efficacy of imatinib and second-generation BCR-ABL kinase inhibitors (dasatinib and nilotinib) in chronic myelogenous leukemia (CML), these drugs fail to provide durable therapeutic benefit in patients with Philadelphia-chromosome positive (Ph+) BCR-ABL-induced acute lymphoblastic leukemias (ALL). In contrast to chronic-phase CML, deletion of the CDKN2A (INK4A-ARF) tumor suppressor locus occurs in diagnostic blast samples from 65% of Ph+ ALL patients. Introduction of cultured Arf−/− p185BCR-ABL-expressing (p185+) pre-B cells, but not their Arf+/+ counterparts, into healthy syngeneic recipient mice induces fulminant B cell leukemia. Virtually every p185+, Arf−/− cell has leukemia-initiating cell capacity. We have now monitored disease progression and dasatinib-responsiveness in vivo by following the fate of leukemogenic p185+ Arf−/− donor cells engineered to express luciferase. When dasatinib therapy is initiated in recipients with low leukemic burdens, most animals remain in remission for many weeks, but some ultimately relapse on therapy or soon after drug discontinuation. Most resistant leukemias sustained kinase domain (KD) mutations, most frequently in the ATP-binding (‘P’) loop, that only modestly impact drug sensitivity. In contrast, in animals bearing high leukemic burdens (simulating the human clinical condition at diagnosis), dasatinib therapy induced dramatic reductions in luminescent signals, but all animals harbored persistent, measurable deposits of drug-resistant cells. Residual drug-refractory ALL cells recovered from these otherwise healthy animals contained rare clones expressing the T315I mutation known to confer drug resistance to imatinib, dasatinib and nilotinib, and most frequently associated with clinical resistance in Ph+ ALL. However, the vast majority of BCR-ABL alleles were mutation-free, implying that KD mutations could not be solely responsible for drug resistance. Following continued therapy, these mice underwent clinical relapse preceded by dramatic increases in luminescent signals, both in the hematopoietic compartment and central nervous system. Many of these drug-resistant leukemias now harbored the T315I KD mutation. Animals that had been maintained in remission with 4 weeks of continuous dasatinib therapy quickly relapsed upon therapy discontinuation, almost always without evidence of BCR-ABL KD mutations. Together, these studies indicate that in this clinically-relevant Ph+ ALL model, several factors including Arf lossof- function, disease burden, intensity of therapy, and length of drug exposure interact to determine therapeutic outcome and to trigger confluent mechanisms of drug resistance. Critically, while continuous dasatinib therapy efficiently selects for and maintains cells harboring drug-resistant KD mutations that mediate eventual clinical relapse, most drugrefractory leukemic cells survive in the absence of KD mutations during maintenance therapy, implying that mutation-independent factors facilitate persistence of residual leukemic disease. We propose that Arf inactivation in Ph+ ALL (but not in CML) enhances the biological ‘fitness’ of leukemic cells and diminishes the efficiency with which targeted therapy can successfully eradicate drug-refractory disease. This facilitates the subsequent emergence of cell-intrinsic drug resistance, most frequently manifested as BCR-ABL KD mutations.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author