Kinetics of BCR-ABL Mutant Clones Determines Resistance in CML on Second Generation TKI Treatment.

BCR-ABL kinase domain (KD) mutations are the major mechanism of acquired imatinib resistance in patients (pts) with chronic myeloid leukemia (CML). Second generation tyrosine kinase inhibitors (TKI) are effective against most imatinib resistance mutations but treatment can be hampered by the emergence of secondary drug-resistant clones over time. Furthermore, individual KD mutations differ according to their in vitro transforming potency and tyrosine kinase activity which in vivo upon presence of multiple mutations may result in competition between different clones. Using a novel sensitive and quantitative monitoring approach we systematically investigated the kinetics of drug-resistant mutants on second generation TKI in order to identify patterns of dynamics and to understand mechanisms of polyclonal drug resistance. Fourty CML pts (24 m; median age 64 years, range 39–74) with resistance to imatinib in chronic phase (n=31), accelerated phase (n=7), or blast crisis (n=2) were treated with dasatinib (D, n=20) or nilotinib (N, n=20). Peripheral blood samples taken at baseline and after 3, 6, 9, and 12 months on second generation TKI therapy were subjected to standard genotyping performed by D-HPLC/sequencing and two high-sensitive allele-specific approaches (ligation PCR [L-PCR] and amplification refractory mutation system PCR [ARMS-PCR]) for a panel of 13 key mutations: G250E, Q252H, Y253F/H, E255K/V, V299L, T315I, F311I, F317L, M351T, E355G, F359V. All mutational findings obtained by at least two methods were subjected to quantitative monitoring of BCR-ABL^mutant/GUS by ARMS-PCR allowing (i) a dynamical detection range of mutant BCR-ABL over 3 to 4 orders of magnitude and (ii) quantification of the mutant cell subset towards BCR-ABL/GUS ratio. We identified a total of 53 mutated clones in 28 imatinib resistant subjects, of which 46 were assessed quantitatively over time. The following patterns of kinetics were observed: I. A parallel decrease (>2 orders of magnitude) of BCR-ABL^mutant/GUS and BCR-ABL/GUS ratio without further mutations emerging (monoclonal resistance, good molecular response, n=11). II. Persistence of a single mutated clone (change <0.5 orders of magnitude) upon treatment combined with a lack of molecular response indicating inadequate treatment (n=6). III. A divergent dynamics - stable, high BCR-ABL/GUS ratios paralleled by decrease of a primary mutation (n=11) - which in the further course is replaced by a secondary (n=11) or even a tertiary (n=4) clone (suboptimal response in all cases, polyclonal resistance). Identification of group III subjects was already possible after three months of treatment (in average 5 months before mutations appeared by sequencing). In four group III subjects, rapidly evolving, secondary clones [F317L (D)/T315I (N)] were replaced by a tertiary clone [T315I (n=2), V299L (n=1) or E255K (n=1)] in the further treatment course. Such a deselection of T315I upon emergence of other drug-resistant clones does not accommodate data derived from in vitro characterization of mutated clones where T315I always heads the list of resistant mutations towards TKI. Our data indicate that co-existing, drug-resistant clones may interfere according to mechanisms which are still poorly understood. Using systematic - in vivo - quantitative analysis of drug-resistant cell subsets we identified clonal interference as a co-factor influencing selective patterns of drug resistance. Understanding the mechanism behind such a phenomenon could eventually improve therapeutic options used for relapsing patients harboring multiple drug-resistant clones inclusive pan-resistant variants like the T315I.