Low Level BCR-ABL Mutations and Response to Treatment: A Substudy of the ENEST1st Trial

The presence of BCR-ABL kinase domain mutations below the detection limit of sanger sequencing (low level mutations, LLM) predicts the outcome of subsequent therapy in patients with imatinib resistance (Parker et al. JCO 2011, Blood 2012). We have evaluated LLM in the context of the ENEST1st trial, which addresses the frequency of deep molecular response (MR⁴) after 18 months on nilotinib 300mg BID (NI) in newly diagnosed patients with chronic myeloid leukemia (CML) in chronic phase (CP). We have investigated the incidence of detectable LLM in the CD34⁺ progenitor cell compartment and correlated the mutation status (LLM vs. no LLM) with the primary endpoint of the ENEST1st trial, MR⁴at 18 months. Additionally, we evaluated MMR at 18 months.

Seventy eight ENEST1st study patients (pts) provided 10ml of peripheral blood or 2ml bone marrow after written informed consent. CD34⁺ selection was carried out by MACS^® (Miltenyi Biotec) and the CD34⁺purity was subsequently determined by fluorescent activated cell sorting (FACS).

Aliquots of 10⁵ CD34⁺ were used for RNA extraction, cDNA synthesis and BCR-ABL amplification followed by Ligation PCR (L-PCR) for mutations T315I, Y253H, E255K/V, and F359V. This method has previously been shown to achieve a dynamic detection range of 100% to <0.1% mutant allele (3–3.5 log). No patient showed BCR-ABL kinase domain mutations detected by Sanger sequencing spanning ABL exons 4–9. Fifty one of 78 pts (65%) with 10⁵ CD34⁺ cells and a documented CD34⁺ purity of >50% were available for BCR-ABL amplification. Amplification was successful from 41 (52%) of these CD34⁺samples.

A total of 205 L-PCR assays of CD34⁺ cells identified 29 (14%) mutations (T315Ix12, Y253Hx7, E255Kx8/Vx1 and F359Vx1) from 21/41 pts (51%). The median quantitative level of mutant alleles in CD34⁺ cells was 0.13 (range 0.06–0.54) BCR-ABL^mutant/ BCR-ABL^unmutated.

One pt was taken off the study because of adverse events before the 3 months evaluation, 2 pts between 3 and 6 months and 1 after the 6 months evaluation (9.7%). Two pts were screened for the trial and did not fulfil the inclusion criteria (4.9%). One pt lost complete hematological response after 6 months and was switched to a different TKI and one pt had a persistent suboptimal response and switched TKI after 12 months on NI (4.9%). These two pts were considered as non-optimal responders at all time points after the change of treatment. All patients except one had typical transcripts (b2/a2 or b3/a2). One pt had an atypical transcript and was not available for MR4 assessment.

At 18 months, 31 of 41 patients were evaluable for BCR-ABL levels. Thirteen out of 31 (42%) pts had MR⁴ or higher at 18 months (6 pts in the LLM group). Twenty six (84%) pts had MMR (10 in the LLM group), while five pts did not reach MMR at 18 months (4 pts with LLM). No statistically significant differences in the proportion of patients achieving MR4 or MMR could be detected between the two groups.

In conclusion, T315I, Y253H, E255K or F359V mutations in CD34 selected cells below the detection limit of sanger sequencing do not predict the achievement of MR4 or higher in patients on nilotinib in the ENEST1st study. Therefore, screening for low level BCR-ABL mutations at diagnosis of CML is currently not recommended.