Whole Exome Sequencing Detecting Kinesin Family Gene Defects In Myeloid Neoplasm

Clinical and pathomorphologic diversity in MDS is a reflection of heterogeneity of molecular lesions. Somatic mutations and chromosomal deletions/amplifications affect various pathways in a convergent and divergent fashion, generate phenocopy and can occur in a variety of combinations. Recent technological advances, including high density arrays and the new generation sequencing (NGS) led to the discovery of novel pathway mutations or gene families affected by somatic defects, e.g., cohesin or spliceosomal mutations.

We have performed whole exome NGS of paired (tumor/germ line) samples in 222 patients with myeloid neoplasms from the Cleveland Clinic and University of Tokyo. Clinical parameters were studied including age, gender, overall survival (OS), bone marrow blast count, and metaphase cytogenetics. Additionally, we also used in our analysis data sets from 197 AML included in the Cancer Genome Atlas (TCGA). We found 1.4% (6/419) of non-canonical somatic mutations of KIF2Bwhich is a member of kinesin13 family located on the long arm of chromosome 17; 3 cases from our cohort (p.V32M (c.G94A), p.T113M (c.C338T), p.R163C (c.C487T)) and 3 cases from TCGA database (p.T47M (c.C140T), p.T310M (c.C929T), p.H551N (c.C1651A)).

By analyzing clonal architecture and intra-tumor heterogeneity in 2 cases (RCMD and RAEB) by targeted deep sequencing, allelic frequencies of KIF2B mutations were more than 45% and larger than for any other concomitant mutations, suggesting that KIF2B mutations might consequently constitute ancestral events followed by subclonal acquisitions of the other mutations. Of note is that 6 non-sense mutations were also reported in lung cancer. Based on SNP-array mapping of chromosomal abnormalities, deletions of 17q involving the KIF2B locus (17q22) was present about 3% (6/215) of myeloid neoplasm. KIF2B defects were frequently detected in higher-risk MDS and AML phenotypes (9%). KIF2B performed an important role in regulation of kinetochore-microtubule attachment. Previous studies showed that the velocity of chromosomes’ movement in KIF2B-deficient cells is reduced 80% comparing to control and fail to perform cytokinesis. In our series, 56% of myeloid neoplasms with KI2B defects had complex cytogenetics and 67% cases of them were also UPD, suggesting that KIF2B defects might lead to inducing abnormal chromosomal movements and segregations. We then, expanded our study to the whole kinesin gene family: 17 somatic mutations and 57 deletions were identified in KIF1A (n=6), KIF23 (n=1), KIF26A (n=1), KIF27 (n=7), KIF1C (n=9), KIF21B (n=2), KIF13A (n=10), KIF14 (n=2), KIF17 (n=15), KIF25 (n=1), KIF3C (n=8), KIF6 (n=2) and CENPE (n=10). All mutations were heterozygous and mutually exclusive. By survival analysis of such mutated cases, a tendency towards worse prognosis was observed (HR; 1.72, 95%CI 0.86-3.37). Analysis of concomitant mutations associated with whole kinesin family mutations or deletions showed that most frequently affected genes are TET2 (n=14), DNMT3A (n=8), IDH1/2 (n=8) and MLL (n=5), all involved in epigenetic regulation.

In conclusion, somatic mutations in kinesin family genes are found in myeloid malignancies and might be responsible for another pathogenesis of the disease. KIF2B is most frequently found in myeloid malignancies and associated with aggressive type of MDS. Since knockout mice of multiple kinesin family genes (KIF5A, KIF16B and EG5) were lethal in embryo and all the mutations occur in a heterozygous configuration, it is likely synthetic lethal approach might create therapeutic window between defective malignant cells and healthy controls. Kinesin family of motor proteins may be an emerging novel therapeutic target. In fact some kinesins have been already successfully targeted in solid tumors.