ABL kinase domain (AKD) mutations are found in 20%–40% of pts with CML who fail therapy with the tyrosine kinase inhibitor (TKI) imatinib. Except for T315I, resistance conferred by most mutations can be overcome by the 2nd generation TKI dasatinib. In pts failing TKI therapy with no detectable AKD mutations by conventional direct sequencing (DS) of the AKD, other mechanisms of resistance (e.g., BCR-ABL1 amplification, SRC overexpression) have been proposed. However, these mechanisms occur at low frequencies in vivo. To ascertain whether imatinib-resistant pts may harbor mutations not detected by standard DS, we evaluated 61 pts with CML after imatinib failure by DNA expansion of specific clones (DESC) followed by DNA sequencing of at least 10 clones. At this point and prior to start of dasatinib therapy, 26 pts were in CP, 14 in AP, and 21 in BP. A total of 118 distinct AKD mutations at 112 amino acid positions were detected (77 previously unreported) in 58/61 (95%) pts. As previously reported, most mutants mapped to 4 AKD regions: P-loop (16%), catalytic domain (17%), 315–317 region (13%), and activation loop (9%). In addition, mutations were also found to cluster at high frequency to 4 novel AKD regions. One of them spans the residues flanked by amino acid positions 295 and 312 and contained 18% of AKD mutations. In addition to 5 pts who developed the highly dasatinib-resistant mutation V299L, 16/61 (25%) other pts harbored mutations within 295–312 prior to dasatinib start. Eleven (69%) of them never achieved any cytogenetic response on dasatinib and this was associated with significantly worse overall survival than that of patients expressing any other AKD mutation (p=0.02), except for T315I (Figure 1). Structurally, 295–312 mutations can potentially interfere with the N-lobe:helix αC interface. Although the exact energetic consequences of mutations at these residues are difficult to predict, structural analysis appears to indicate that these may hinder the outward torsion of the adjacent helix αC, potentially hampering the transition to the inactive conformation (“αC-Glu In” conformation) to which imatinib binds. Alternatively, the 295–312 region may serve as the structural scaffold for amino acid position 299, a direct dasatinib contact residue. Mutations mapping to 295–312 might distort the topography surrounding 299, thus altering the 299/dasatinib interface. Notably, 5/16 (32%) of these pts carried clones expressing more than 1 mutation within 295–312. All but 1 (80%) of these pts are dead. In summary, the use of techniques with higher sensitivity than conventional DS reveals that AKD mutations are highly prevalent (95%) in pts failing imatinib therapy, which could explain TKI resistance in pts not found to carry resistant mutations by conventional DS. We present evidence supporting the deleterious effect of mutations mapping to the novel 295–312 region. Experiments designed to prove these hypotheses are ongoing.
Disclosure:Research Funding: Dr. Kantarjian, Dr. Cortes, and Dr. Talpaz receive research funding from Bristol Myers-Squibb.