Somatic Mutations Are Frequently Detected in Chronic Myeloid Leukemia in Chronic Phase and Do Not Affect Response to Tyrosine-Kinase Inhibitors

The development of chronic myeloid leukemia (CML) is driven by a unique genetic aberration, i.e. the BCR-ABL rearrangement. Somatic mutations typical of myeloid neoplasms have been described in patients with CML, mainly associated with blastic phase (BP) or with additional chromosomal aberrations (ACA). Recently, somatic mutations have been identified in a significant proportion of healthy subjects without a hematological phenotype, resulting in a significantly higher risk of developing a hematologic cancer. However, the clinical relevance of these mutations in patients with chronic phase (CP)-CML needs to be clarified. In this study, we performed a comprehensive mutation analysis of genes implicated in myeloid malignancies in a large cohort of CP-CML patients with the aim to estimate their prevalence and the impact on response to TKI and clinical outcomes.

Seventy-four patients (pts) with CP-CML were enrolled, after providing written informed consent (M/F: 44/30; median age 54 years, range 18-82; low or intermediate Sokal risk score: 85%; no ACA at diagnosis or during follow-up).

Samples were collected in 53 pts (group1, median time from TKI start: 66 months, range 3-177). In addition, in order to better clarify the clonal architecture, 21 consecutive pts were analyzed at diagnosis and sequentially during follow-up (group 2, median follow-up 16 months, range 3-47, median number (no.) of samples per pt 3, range 2-5). Targeted sequencing of 54 genes was performed using an Illumina HiSeq platform (Illumina, San Diego, CA) in bone marrow or peripheral blood granulocytes, whereas CD3+T lymphocytes were used as control tissue.

At least one somatic mutations was identified in 14/53 (26.4%) patients in the group1; out of these, 5 pts showed 2 mutations, and 1 pt showed 3 concomitant mutations. The most frequently mutated genes were DNMT3A (11%), TET2 (6%), ASXL1 (6%), as reported in age-related clonal hematopoiesis. In addition, somatic mutations in CUX1 and KDM6A were observed (4% each). No mutations in splicing factors were identified. Median variant allele frequency (VAF) was 0.07 (range 0.02-0.49). Most of mutated pts (11/14, 78.6%) were in MR3 at the time of sampling (median IS 0.014%, range 0-0.09%), thus suggesting that additional clones were independent from the BCR-ABL-positive one.

In pts from group 2 somatic mutations were identified in 12/21 (57.1%) cases: 2 pts harboured mutations at diagnosis that remained stable despite of clearance of BCR-ABL (JAK2 and MPL, in pts with associated Ph-negative myeloproliferative neoplasm). The other 10 pts showed transient small clones (median VAF 0.06, range 0.02-0.1), not detectable at diagnosis and fluctuating during follow-up. Only one of these pts had a STAG2 mutation at diagnosis, which disappeared during follow-up along with clearance of BCR-ABL.

The mutations most frequently identified involve genes demonstrated to be recurrently mutated in healthy subjects and associated with increasing age. In our cohort, no significant association was found between age and either presence (median age of wild-type pts 60 years vs 61 in mutated pts, P=0.35) and number (rho=0.13, P=0.26) of somatic mutations. However, the incidence of mutations was significantly higher in CP-CML pts in comparison with the age-matched general population (Jaiswal et al, NEJM, 2014), especially in young patients (P<0.001).

After a median follow-up in the whole cohort of 5.3 years (range 0.4-38.4), no progression to accelerated phase/BP was observed. In multivariate analysis, mutational status did not affect the probability of achieving MR3 or better under TKI therapy, neither when evaluating the presence of somatic mutations (HR=4.1, 95%CI: 0.7-23.3, P=0.111) nor when we considered the no. of mutations per pt (HR=1.9, 95%CI: 0.8-4.4, P=0.148).

In conclusion, this study showed that somatic mutations in genes frequently mutated in myeloid neoplasms and associated with age-related clonal hematopoiesis are detectable in CP-CML pts with a significantly higher incidence compared with the general population. These mutations are likely driving small transient clones, independent from the Ph-positive one, and become detectable when BCR-ABL is suppressed by TKI therapy. The presence of these mutations do not affect the probability of achieving an optimal molecular response to TKIs, although this observation needs to be confirmed in larger studies.