Transient abnormal myelopoiesis (TAM) is a myeloid proliferation resembling acute megakaryoblastic leukemia (AMKL), mostly affecting perinatal infants with Down syndrome. Although self-limiting in a majority of cases, TAM may evolve as non-self-limiting AMKL after spontaneous remission (DS-AMKL). Pathogenesis of these Down syndrome-related myeloid disorders is poorly understood, except for GATA1 mutations found in most cases. Here we report genomic profiling of 41 TAM, 49 DS-AMKL and 19 non-DS-AMKL samples, including whole-genome and/or whole-exome sequencing of 15 TAM and 14 DS-AMKL samples. TAM appears to be caused by a single GATA1 mutation and constitutive trisomy 21. Subsequent AMKL evolves from a pre-existing TAM clone through the acquisition of additional mutations, with major mutational targets including multiple cohesin components (53%), CTCF (20%), and EZH2, KANSL1 and other epigenetic regulators (45%), as well as common signaling pathways, such as the JAK family kinases, MPL, SH2B3 (LNK) and multiple RAS pathway genes (47%).

Recent advances in high-throughput sequencing technologies have enabled a comprehensive dissection of the cancer genome clarifying a large number of somatic mutations in a wide variety of cancer types. A number of methods have been proposed for mutation calling based on a large amount of sequencing data, which is accomplished in most cases by statistically evaluating the difference in the observed allele frequencies of possible single nucleotide variants between tumours and paired normal samples. However, an accurate detection of mutations remains a challenge under low sequencing depths or tumour contents. To overcome this problem, we propose a novel method, Empirical Bayesian mutation Calling (https://github.com/friend1ws/EBCall), for detecting somatic mutations. Unlike previous methods, the proposed method discriminates somatic mutations from sequencing errors based on an empirical Bayesian framework, where the model parameters are estimated using sequencing data from multiple non-paired normal samples. Using 13 whole-exome sequencing data with 87.5–206.3 mean sequencing depths, we demonstrate that our method not only outperforms several existing methods in the calling of mutations with moderate allele frequencies but also enables accurate calling of mutations with low allele frequencies (≤10%) harboured within a minor tumour subpopulation, thus allowing for the deciphering of fine substructures within a tumour specimen.

Pediatric acute megakaryoblastic leukemia in non-Down syndrome (AMKL) is a unique subtype of acute myeloid leukemia (AML). Novel CBFA2T3-GLIS2 and NUP98-KDM5A fusions recurrently found in AMKL were recently reported as poor prognostic factors. However, their detailed clinical and molecular characteristics in patients treated with recent improved therapies remain uncertain. We analyzed molecular features of 44 AMKL patients treated on two recent Japanese AML protocols, the AML99 and AML-05 trials. We identified CBFA2T3-GLIS2, NUP98-KDM5A, RBM15-MKL1, and KMT2A rearrangements in 12 (27%), 4 (9%), 2 (5%), and 3 (7%) patients, respectively. Among 459 other AML patients, NUP98-KDM5A was identified in 3 patients, whereas CBFA2T3-GLIS2 and RBM15-MKL1 were only present in AMKL. GATA1 mutations were found in 5 patients (11%). Four-year overall survival (OS) and event-free survival (EFS) rates of CBFA2T3-GLIS2-positive patients in AMKL were 41.7% and 16.7%, respectively. Three-year cumulative incidence of relapse in CBFA2T3-GLIS2-positive patients was significantly higher than that of CBFA2T3-GLIS2-negative patients (75.0% vs. 35.7%, P = 0.024). In multivariate analyses, CBFA2T3-GLIS2 was an independent poor prognostic factor for OS (HR, 4.34; 95% CI, 1.31-14.38) and EFS (HR, 2.95; 95% CI, 1.20-7.23). Furthermore, seven (54%) of 13 infant AMKL patients were CBFA2T3-GLIS2-positive. Notably, out of 7 CBFA2T3-GLIS2-positive infants, six (86%) relapsed and five (71%) died. Moreover, all of CBFA2T3-GLIS2-positive patients who experienced induction failure (n = 3) were infants, indicating worse prognosis of CBFA2T3-GLIS2-positive infants. These findings indicated the significance of CBFA2T3-GLIS2 as a poor prognostic factor in AMKL patients, particularly in infants.