Transient abnormal myelopoiesis (TAM) is a temporal preleukemic status specific to neonates of Down syndrome (DS). TAM is characterized by leucocytosis and/or thrombocytopenia with the trisomy of chromosome 21 and somatic GATA1 mutation. Morphologically, most blasts found in TAM are myeloblasts, and some other types of blasts are similar to those typically found in DS-related acute megakaryoblastic leukemia (DS-AMKL). Although TAM patients usually experience spontaneous remission within 6 months, 20 to 30 % of develop DS-AMKL. The cell population that causes TAM phenotypes in vitro or in vivo mouse and human models has not yet been determined.

Recently, the hematopoietic differentiation systems from human pluripotent stem cells (PSCs) have revealed the pathogenesis in various diseases including TAM. Previous TAM models have succeeded in recapitulating the phenotypes, such as arrest and/or skewness of differentiation from immature hematopoietic progenitor cells (HPCs). However, identification of the responsible cell population for TAM at a precise subpopulation level based on functional evaluations has not yet been accomplished. Therefore, we applied our 2D-hematopoietic differentiation system for identifying the responsible cell population through the progenitor assay. A comprehensive understanding of TAM pathogenesis will enable us to predict and eventually prevent the disease progression from TAM into life-threatening DS-AMKL.

First, we differentiated artificial trisomy 21 embryonic stem cells (Ts21-ESCs) generated by transfer of human artificial chromosome (HAC) 21 in our 2D-differentiation system. Five subpopulations of HPCs were harvested at day6 and day9, and cells were subjected to expression analysis of hematopoiesis-related genes. Principal component analysis (PCA) successfully mapped each HPC subpopulation corresponding to the hematopoietic developmental pathways, proving that our system can be used for isolating functional HPC subpopulations.

Next, we used genome edition to establish strictly controlled isogenic ES/induced pluripotent stem cell (iPSC) pairs derived from TAM, non-TAM/DS patients and the Ts21-ESCs. Then we differentiated ES/iPSCs into each hematopoietic lineage to identify the abnormal phenotypes in vitro . GATA1 -mutated clones (G1s-clones) showed obviously different phenotypes, compared to GATA1-WT clones (WT-clones). In particular, completely restricted erythroid lineage cells (CD42b-CD71+CD235a+) and increased myeloid lineage cells (CD34-CD235a-CD41-CD42b-CD43+CD45+) were observed in G1s-clones. Moreover, in megakaryocytic lineage, the frequency of immature (CD235a-CD41+CD42bdim) and mature (CD235a-CD41+CD42bbright) cells differed between G1s- and WT-clones. Notably, the proportion of immature megakaryoblasts increased in G1s-clones. These data indicated that we have succeeded in recapitulation the TAM phenotype with our in vitro differentiation system.

Then, we performed progenitor assay based on in vitro TAM phenotypes. We differentiated 5 isolated HPCs subpopulation into erythrocyte, megakaryocyte and myeloid cells with a suspension culture and colony forming assay. As a result, we found strong correlations between the increased or decreased specific cell fractions and abnormal phenotypes. Correlation pattern is slightly different between clones.

In conclusion, we have identified putative responsible cell populations that contribute to the pathological phenotype in TAM. These populations were located on a series of spectra in the differentiation stage. Ubiquitous expression of GATA1 during the differentiation process and wide heterogeneity of the differentiation systems in general may be the reason why precise identification of the single responsible population was not possible. However, our model of TAM using iPSCs enabled us to narrow down the candidate cell populations which have the potential to be a source of DS-AMKL. The identification of responsible progenitor cells is an important first step to predict the transformation from TAM into DS-AMKL and to prevent its progression. Moreover, the method used in this study, such as 2D-differentiation system and progenitor assay, is a promising way to identify the responsible cell population of various types of leukemia other than TAM.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.