Kinase Profiling as a Tool in Molecular Dissection of Imatinib Resistant Chronic Myeloid Leukemia (CML).

Imatinib (Gleevec) resistance (IR) in CML is a multifactorial phenomenon related to ineffective blockage of the bcr-abl kinase through point mutation and/or gene amplification, or alternatively kinase bypass through clonal evolution or presently unknown mechanisms. The newer generation of bcr-abl small molecule inhibitors (dasatinib, nilotinib) may be useful in overcoming ineffective bcr-abl blockage in IR-CML but may not be effective in other mechanisms of resistance. As a tool in classifying the pattern of disease resistance, we profiled kinase levels in IR-CML following switch to new kinase inhibitors or with imatinib dose escalation. In Ficoll-purified blood and bone marrow samples, we examined levels of BCR-ABL autophosphorylation (Tyr245), and phosphorylation of two bcr-abl targets STAT5 (Tyr694/699) and CRKL (Tyr207) by Western blot. Kinase profiling was compared to hematologic responses, cytogenetic analysis, FISH for the bcr-abl locus using dual fusion probes and quantitative(q)PCR for bcr-abl transcript. We profiled 26 patients with IR-CML undergoing therapy shifts as well as serial samples from 10 patients receiving frontline dasatinib therapy for newly diagnosed CML. In all but 1 of 68 analyzed samples, detection of phospho(p)-CRKL in unsorted white blood cells was associated with bcr-abl/abl qPCR ratios above 1%. Among all 68 samples, pSTAT5 and pBCR-ABL levels declined in parallel to pCRKL but were absent in more samples than pCRKL. Persistence of detectable pCRKL despite switch to nilotinib, dasatinib or high dose imatinib was seen in 5 patients with bcr-abl kinase domain (KD) mutations including 2 with G250E and 1 each with L248V, F311L and T315I, 2 patients with bcr-abl gene amplification and 3 patients without KD mutations including 2 with cytogenetic clonal evolution. Among these 10 patients with IR-CML and persistent pCRKL, complete hematologic responses were noted in 4 and complete cytogenetic response in 3. Response to therapy switch and absence of pCRKL was seen in tumors with E355G, L387M and E459K KD mutations and in 12 patients with no mutation or genomic amplification. Among the patients receiving frontline therapy with dasatinib, effective hematologic responses correlated well with loss of detectable pCRKL and pSTAT5. However, detectable pCRKL did not disappear until at least 8 weeks following initiation of therapy, and lagged hematologic response. The 2 patients treated with frontline dasatinib therapy who had persistent pCRKL after 20 weeks of treatment also showed amplification of the bcr-abl locus in tumor cells by FISH. Kinase profiling, particularly for pCRKL, is useful for monitoring molecular response in the therapy switch period in IR-CML and correlates well with the predicted pattern of response in tumors with KD mutations. The value of monitoring kinase levels frequently following initiation of frontline kinase inhibitor therapy in newly diagnosed CML remains to be established.