Characteristics and outcome of chronic myeloid leukemia patients with F317L BCR-ABL kinase domain mutation after therapy with tyrosine kinase inhibitors

Point mutations of the BCR-ABL kinase domain (KD) are a frequent mechanism of resistance to tyrosine kinase inhibitors (TKIs) in chronic myeloid leukemia (CML).^1-11  Newer TKIs demonstrate activity against most mutations noted with imatinib.^11,12  The sensitivity of different mutations and their potential to develop during therapy with different TKIs varies.^3,13-15  In vitro drug sensitivity and emerging clinical data suggest that mutations at codon 317 may develop after dasatinib therapy.^3,16  The aromatic ring in the side chain of phenylalanine 317 interacts directly with the pyrimidine and thiazole rings of dasatinib, impairing dasatinib binding⁵ ; these mutations are associated with dasatinib failure.^3,4,16-18  The objectives of this study were to assess the incidence, prognosis, and response to therapy in patients with F317L mutations

Between June 2003 and March 2007, 192 patients (186 previously reported, including 14 with F317L¹⁹ ) with CML were evaluated by DNA sequencing for BCR-ABL KD mutations after failure of therapy with TKI. The criteria to trigger mutation analysis were based on evidence of treatment failure as defined by the European LeukemiaNet.²⁰  Definitions of CML phases and responses were as described.^20-22  All patients were treated on institutional review board–approved protocols in accordance with the Declaration of Helsinki.

For mutational analysis screening, the entire KD of the BCR-ABL fusion transcript was sequenced by the Sanger method from codons 221 to 500 using a nested PCR approach, with a sensitivity of 10% to 20% mutation-bearing BCR-ABL transcripts.³  On follow-up samples, the ratio of F317L mutated to unmutated transcripts was performed by pyrosequencing using an HSQ96 Pyrosequencer (Biotage, Uppsala, Sweden), following a similar nested PCR approach with a sensitivity of 1% to 5% mutation-bearing BCR-ABL transcripts.

Descriptive statistics were analyzed using the χ² test.²³  Survival was calculated by the Kaplan-Meier method.²⁴  Overall survival (OS) was calculated from time of detection of mutation to date of death or last follow-up.

Mutations were identified in 99/192 (51%) patients at the time of imatinib failure. In addition, 36 (19%) had a mutation first detected after a second-generation TKI given after imatinib failure. F317L was detected in 20 patients: 12/99 (12%; 95% confidence interval [CI], 6%-20%) patients with mutations after imatinib failure, and 8/16 (50%; 95% CI 28%-72%) with new mutations after dasatinib failure (P = .001). All 8 had previously failed imatinib and one had also failed nilotinib (0/17 new mutations after nilotinib, 0/2 after bosutinib, and 0/1 after INN0-406). The median time from start of therapy to detection of the F317L mutations was 23 months (range, 2-69 months) for imatinib and 10 months (range, 2-22 months) for dasatinib. Two patients had concomitant mutations (with M351T and G250E, respectively, in addition to F317L; the later eventually also acquired T315I).

At the time mutation was detected, 8 patients were in chronic phase (CP), 6 in accelerated phase (AP; 3 of them with clonal evolution) and 6 in blast phase (BP; 4 myeloid and 2 lymphoid). Table 1 summarizes patients with F317L-mutated and those with other mutations or no mutations after TKI therapy. There was no difference in characteristics between patients with F317L and those with other or no mutations except for lack of response to second generation TKIs for patients with T315I (P = .003).

Among 12 patients with F317L postimatinib failure (one patient with primary refractory disease), 11 received salvage therapy. Three of them received dasatinib; their best response was partial hematologic response (PHR) in one and complete hematologic response (CHR) in 2 (lost after 10 and 12 months, respectively). The latter 2 patients subsequently received, respectively, nilotinib with no response and bosutinib that induced a transient CHR of 9 months' duration. None had a change in the ratio of the mutated to unmutated clones. An additional 4 patients received nilotinib and 3 responded: one achieved CHR lasting 15 months and subsequently again achieved CHR with bosutinib lasting 6 months (the mutated clone remained fully dominant), one achieved MMR (with eradication of the mutant clone) that was lost to partial cytogenetic response after 22 months, and one achieved ongoing CMR for 30 months. Salvage therapy in the other 4 included imatinib dose escalation in 2; both achieved sustained CCyR for more than 24 and more than 26 months, one of them with eradication of the mutant clone; stem cell transplantation in one (sustained CMR 20 months after transplantation), and combination of imatinib and farnesyl-transferase inhibitor in one without response.

Among 8 patients who developed F317L only after dasatinib, 4 were primary refractory (1 AP, 3 BP) and 4 responded to dasatinib (2 CP, 1AP, 1 BP). One achieved CHR and was lost to follow-up 4 months later. The second achieved a CHR that lasted 10 months; the mutant clone remained stable, around 60%. The third achieved CCyR that was lost 3 months after acquisition of the mutation (25% mutated); he subsequently achieved CCyR with nilotinib that lasted 4 months, with the ratio of the mutated to unmutated clone increasing to 43%. In the fourth patient, F317L (70% mutated) was detected after the loss of a CCyR of 9 months' duration; he subsequently failed bosutinib, but had a decrease of the mutated clone from 70% to 3% (Table 2). Three of the 8 patients who acquired F317L on dasatinib had treatment interruptions for a median of 10 days, with one of them having his dose reduced from 70 mg twice daily to 40 mg twice daily. To determine whether F317L may have been present at the time of imatinib failure at levels below the detection limit for direct sequencing, we performed pyrosequencing in 5 patients with available samples. F317L was not detected in any at the time of imatinib failure.

The median survival with F317L from mutation detection was 19 months, similar to that of patients with other mutations (P = .45). After a median follow-up of 25 months from the detection of F317L, 8 of 20 patients (40%) died, including 1 of 8 (13%) patients in CP at the time of mutation detection, 3 of 6 (50%) in AP, and 4 of 6 (67%) in BP. Two-year OS by CML phase from mutation's detection were 75%, 50%, and 20%, respectively (P = .04).

The high incidence of the F317L mutation after dasatinib therapy in this series (50% of all mutations detected, compared with 12% after imatinib failure) parallels the findings of in vitro studies that identified the potential emergence of mutations, mostly limited to contact points (including F317) that lead to dasatinib resistance.^5,13,15 

Patients with F317L have a similar survival compared with patients with other mutations, with outcome dictated mostly by the CML phase.¹⁹  Another factor influencing outcome of patients with F317L is response to salvage therapy. In vitro studies suggest a high sensitivity of F317L to nilotinib (IC₅₀ of 50 nM)¹⁸  and an intermediate sensitivity to imatinib (IC₅₀ of 1050 nM).¹⁷  In this report, 4 of 6 patients treated with nilotinib as a second- (n = 4) or third- (n = 2) line (after dasatinib failure) TKI responded: 1 CHR, 1 CCyR, 1 MMR, and 1 CMR. Imatinib dose escalation induced durable CCyR in 2 patients. Thus, in vitro sensitivity may guide therapy for patients with this mutation.²⁵ 

In conclusion, F317L occurs more often after dasatinib therapy, usually leading to resistance. The prognosis of patients with F317L is similar to that of patients with other mutations or no mutations and is dictated mostly by the CML phase. A switch to a TKI with better in vitro potency against this mutation may improve outcome.

Contribution: E.J. and J.C. performed research, analyzed data, and wrote the paper. D.J. performed research, analyzed data, and performed sequencing. N.R. performed sequencing. H.K., S.O., G.G.M., and J.B. performed research. All authors approved the manuscript.

Conflict-of-interest disclosure: H.K. and J.C. have received research grants from Novartis Oncology and Bristol-Myers Squibb. The remaining authors declare no competing financial interests.

Correspondence: Jorge Cortes, MD, Professor of Medicine, Department of Leukemia, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Box 428, Houston, TX 77030; e-mail: jcortes@mdanderson.org.