PSC-Derived Hematopoietic System to Elucidate the Cooperation Between Gene Alterations and Cell Lineages in Leukemogenesis

Onset of acute myeloid leukemia (AML) has been accounted for by accumulated genetic mutations including chromosomal abnormalities. For example, MLL fusion genes, which have been proven to impair cell differentiation, proliferation and epigenetic regulations, are among the common causes of pediatric AML. However, although those alterations are thought to occur in very immature stages, leukemia cells often show phenotypes similar to specific later-stage progenitors such as myeloid and monocytic cells. Until today, it remains unclear how those lineage specifications and mutations are cooperatively involved in disease pathogenesis. To address this issue in reproducible manner, we generated the lines of human pluripotent stem cells (hPSCs) harbouring doxycycline (Dox)-inducible leukemic gene cassettes, applied them for hematopoietic differentiations in a step-wise manner, and sought to identify the mechanisms behind the phenomena.

Using our system, we first evaluated the in vitro and in vivo phenotypes compatible with AML symptoms. When MLL-AF9 (MF9) transgene expression was induced in combination with cKIT (822K) or FLT3-ITD mutations, PSC-derived hematopoietic cells showed reinforced growth in liquid culture and prolonged colony forming efficacies in serial replanting assays in methylcellulose-containing semisolid media. In addition to in vitro assays, in vivo transplantations also showed the prolonged detection of graft cells as long as eight months, while those sets of "leukemia-like" phenotypes were not observed under the condition of any single transgene induction. In order to find what lineages were most responsible for those phenomena, we next induced the combination of MF9 and FLT3-ITD in various sorted subpopulations of cells. As a result, CD34^+/-CD43⁺CD13⁺ myeloid precursors showed the strongest tendencies to emerge highly proliferative clones followed by CD34⁺CD43⁺CD13^- immature progenitors. In the contrast, CD34^-CD71⁺CD41⁺ erythro-megakaryocytic cells hardly emerged those kinds of long-term expanding clones. Those results together indicated the bias of cell lineages and stages in our disease model, and encouraged us to explore a key pathway. Interestingly, we found that a set of NFkB pathway-associating genes were significantly activated when the transgenes were induced not in erythroid but in myeloid cells, which indicated the myeloid specific mechanisms forming a bridge between this pathway and leukemic gene alterations.

In conclusion, we succeeded in establishing the way to dissect the leukemogenesis from the view of relationships between cell stages and gene alterations using PSCs. We believe that our model will enable us to better understand the pathogenesis of leukemia.