BCR-ABL Kinase Domain Mutations and Their Kinetics on Second Line Dasatinib or Nilotinib Therapy in CML Patients after Imatinib Failure: A Cooperative Evaluation of Four Different Detection Methods.

Major cause of acquired resistance to imatinib in patients (pts) with chronic myeloid leukemia (CML) are BCR-ABL kinase domain (KD) mutations. Various techniques to detect KD mutations have been employed, resulting in different frequencies of mutations and a heterogeneous pattern of individual mutations. Thus, we sought to compare direct sequencing (DS, Mannheim and Tromsø), denaturing high-performance liquid chromatography (DHPLC, Mannheim), and two different quantitative allele-specific PCR (A-PCR; ARMS assay, Tromsø and Ligation [L]-PCR technique, Leipzig) approaches for a panel of six key mutations within three different European laboratories. Blinded samples of 30 imatinib resistant CML pts were investigated before start of second line therapy with dasatinib (n=18) or nilotinib (n=12) and after three and six months on therapy. Pts (16 m, 14 f; median age 63 years, range 39–78) were in chronic phase (n=22), accelerated phase (n=6), or myeloid blast crisis (n=2). Analyzed sequences of DS and DHPLC overlapped at ABL1a amino acids 207 to 414 with an estimated detection sensitivity of 10% and 0.5–1% for DS and DHPLC, respectively. A-PCR was performed for a panel of five P-loop mutations (G250E, Y253F/H, E255K/V) and T315I according to the ARMS- and L-PCR technique with a sensitivity of 0.1–1% and 0.05–0.1%, respectively. In total 87/90 (97%) samples were comparable between DS and DHPLC. Fourty-five mutations affecting 15 different amino acids were detected by both techniques. DS identified one additional M244V, whereas DHPLC 12 additional mutations (Y253H, D276G, T277A, T315I, F317L (n=2), L324Q (n=2), L387F (n=3), H396R). All mutations detected by DHPLC within the A-PCR panel were confirmed by both A-PCR techniques and 17 additional mutations [G250E (n=2), Y253H (n=4), E255K (n=5), E255V, T315I (n=5)] were identified by both A-PCR techniques with a median proportion of mutant alleles of 1.4% (range 0.04–83%). Furthermore, 37 additional mutations [G250E (n=2), Y253H (n=5), E255K (n=3), E255V (n=2), T315I (n=25)] were detected by one A-PCR technique only with a median proportion of mutant alleles of 0.57% (range 0.095–25%). The quantitative values of both A-PCR techniques showed a significant correlation (r=0.67, p<0.001). On follow-up samples, the number of BCR-ABL T315I mutant clones increases significantly by A-PCR from 5/30 (17%) at baseline to 17 (57%) after 3 months of treatment (p=0.003). This was neither associated with lack of response (ratio BCR-ABL/ABL >1% at 6 months) nor with dasatinib or nilotinib treatment. We conclude:

1. DHPLC detects more mutations compared to DS.

2. A-PCR further increase the number of detected mutations and confirm by part mutations at low level.

3. Selection of low level T315I mutants on dasatinib and nilotinib without subsequent non response confirms the cross resistance of both inhibitors for T315I and supports also the varying transforming capacity of T315I mutants to a fully resistant phenotype.

4. Although the findings of low level T315I mutations are still investigational at the moment these selected pts could probably profit from a combination therapy including a T315I inhibitor in the future.