Sequencing of Subcloned PCR Products Facilitates Earlier Detection of BCR-ABL1^T315I Mutants Compared to Direct Sequencing of the ABL1 Kinase Domain.

ABL kinase domain mutations represent the most frequent cause of resistance to tyrosine kinase inhibitors (TKIs). The BCR-ABL1^T315I mutation affects a highly conserved “gatekeeper” threonine near the ABL catalytic domain, thus causing steric hindrance that precludes ABL TKIs binding. BCR-ABL1^T315I retains kinase activity even in the presence of micromolar concentrations of imatinib or dasatinib. Thus, early detection of BCR-ABL1^T315I has important prognostic and therapeutic implications. We evaluated the sensitivity of detection of BCR-ABL1^T315I in 62 CML pts after failure (n=51) or intolerance (n=11) of imatinib enrolled in a phase I study of dasatinib by direct sequencing (DS) of nested PCR-amplified BCR-ABL1 products as well as by DNA expansion of specific clones (DESC) followed by DNA sequencing of at least 10 clones. Ten (15%) pts were found to carry BCR-ABL1^T315I, 4 prior to dasatinib start and 6 during dasatinib therapy. Four pts never responded to imatinib whereas 5 had achieved a complete hematologic response (CHR) and 1 a complete cytogenetic response (CCyR). Imatinib was stopped due to rash (n=1), hematologic resistance (n=4), and progression to accelerated (AP; n=1) or blastic (BP; n=4) phase. Of the 4 pts in whom BCR-ABL1^T315I was detected by DESC prior to dasatinib start, 2 are dead and 2 are alive. DS performed on the same samples detected BCR-ABL1^T315I only in 2 of them. In pts in whom direct sequencing failed to detect BCR-ABL1^T315I, the percentage of clones carrying BCR-ABL1^T315I was 10% and 100%, respectively. In one of them, DS detected BCR-ABL1^T315I 12 months later, whereas in the other case, DS failed to detect BCR-ABL1^T315I in 3 separate occasions. These 2 pts are still alive (1 on hydrea, 1 on bosutinib) but never achieved any cytogenetic response. BCR-ABL1^T315I was detected in 6 additional pts after a median time of 5 months (range, 1–6) on dasatinib (dosing ranging from 70 to 140 mg/d): 5 of 5 analyzed by DESC and 1 of 2 determined by DS. Of them, 4 had no response to dasatinib and 2 had transient cytogenetic responses (1 minor, 1 partial) and only one is still alive. In 1 that had no response, for whom paired samples were available, BCR-ABL1^T315I was present in 10% of clones but was not detected by DS. Overall, the median number of clones harboring BCR-ABL1^T315I was 90% (range, 10%–100%) and the median number of mutants co-expressed with BCR-ABL1^T315I was 3 (range, 1–9). DS failed to identify 40 non-BCR-ABL1^T315I mutants (including F317L in 1 patient), regardless of the percentage of clones in which they were expressed, except for E355G in 1 patient. Eight pts received dasatinib for more than 3 weeks (median, 5 months; range, 2–13) and were evaluable for response. Six failed to achieve any cytogenetic response and 2 had transient cytogenetic responses (1 minor and 1 partial). Seven (70%) pts died and 3 are alive with no cytogenetic response. In conclusion, DS has a poor sensitivity to detect ABL kinase mutations, particularly when the proportion of mutated clones is low. With the advent of novel T315I inhibitors, prompt detection of this highly-resistant mutation must be prioritized and included in therapeutic algorithms. To maximize the sensitivity of T315I detection, sequencing of subcloned PCR products might be preferable to DS.