Evidence of ETNK1 Somatic Variants in Atypical Chronic Myeloid Leukemia

Atypical Chronic Myeloid Leukemia (aCML) is a clonal disorder belonging to the Myeloproliferative/Myelodysplastic (MPN/MDS) group. The molecular lesions responsible for the onset of aCML remained unknown until 2013 when recurrent somatic mutations of SETBP1 were identified. However, the frequency of SETBP1 mutations in aCML does not exceed 25-30%, which suggests that other lesions may play a role in the remaining cases.

To gain further insight into the somatic variants responsible for the onset of aCML, we generated whole-exome and transcriptome sequencing data on 15 matched case/control aCML samples. A total of 151 non-synonymous and 42 synonymous single-nucleotide somatic variants were identified. Of these, 140 were transitions and 53 transversions. Of the non-synonymous mutations, 141 were missense and 10 nonsense mutations. In 2/15 (13.3%) samples we identified the presence of missense, single-nucleotide somatic variants occurring in the ETNK1 gene affecting two contiguous residues: H243Y and N244S. Sanger sequencing confirmed the presence and the somatic nature of the variants. Targeted resequencing of 383 clonal hematological disorders showed evidence of mutated ETNK1 in 7/70 aCML (10.0%, 95% C.I. 4.6-19.5%) and in 2/77 chronic myelomonocytic leukemia samples (CMML; 2.6%, 95% C.I. 0.2-9.5%) %), while no ETNK1 mutations were identified in the remaining hematological disorders. All the variants were heterozygous and clustered in the same, highly conserved region within the kinase domain (1/9 H243Y and 8/9 N244S).

Somatic, heterozygous ETNK1 variants have been also recently reported in 10% of Systemic Mastocytosis (SM) cases and in 22% of SM with associated hypereosinophilia (Lasho T et al., Abstract 4062, EHA2014); strikingly, there is a large overlap between the variants that we identified in aCML and CMML and those described for SM (3 N244S and 2 G245A), which suggests that the common hotspot region may play a critical and yet unknown functional role.

The hitherto described data suggest that ETNK1 variants are restricted to a limited subset of hematological disorders. This is further supported by the lack of somatic ETNK1 mutations in 60 paired whole-genome and over 600 exomes, comprising 276 paired tumor/germline primary samples and 344 cancer cell lines (http://cancer.sanger.ac.uk/cancergenome/projects/cell_lines/).

In 2/6 ETNK1 mutated aCML cases (33%, 95% C.I. 9%-70%), we detected the presence of a coexisting somatic SETBP1 variant. The fraction of SETBP1 mutations identified in this group is perfectly in line with the overall frequency of SETBP1 mutations in aCML, suggesting that mutations occurring in ETNK1 and SETBP1 are not mutually exclusive. To discriminate if ETNK1 and SETBP1 mutations occur in different or in the same clone, we performed colony assay experiments, revealing the coexistence of the two somatic mutations within the same clone.

Liquid Chromatography – Mass Spectrometry experiments revealed that in ETNK1 mutated cells the intracellular levels of phosphoethanolamine are over 5-fold lower than in the wild-type counterpart (p < 0.05), suggesting that ETNK1 mutations may impair the physiological catalytic activity of the kinase.

Taken globally these data identify ETNK1 somatic mutations as a new oncogenic lesion in aCML and CMML, two overlapping MDS/MPN neoplasms. They also show that ETNK1 variants apparently cause a loss-of-function effect, leading to a decrease in the intracellular levels of phosphoethanolamine.