The formation of red blood cells begins with the differentiation of multipotent haematopoietic progenitors. Reconstructing the steps of this differentiation represents a general challenge in stem-cell biology. Here we used single-cell transcriptomics, fate assays and a theory that allows the prediction of cell fates from population snapshots to demonstrate that mouse haematopoietic progenitors differentiate through a continuous, hierarchical structure into seven blood lineages. We uncovered coupling between the erythroid and the basophil or mast cell fates, a global haematopoietic response to erythroid stress and novel growth factor receptors that regulate erythropoiesis. We defined a flow cytometry sorting strategy to purify early stages of erythroid differentiation, completely isolating classically defined burst-forming and colony-forming progenitors. We also found that the cell cycle is progressively remodelled during erythroid development and during a sharp transcriptional switch that ends the colony-forming progenitor stage and activates terminal differentiation. Our work showcases the utility of linking transcriptomic data to predictive fate models, and provides insights into lineage development in vivo.

Disruptor of telomeric silencing 1-like (Dot1l) is a histone 3 lysine 79 methyltransferase. Studies of constitutive Dot1l knockout mice show that Dot1l is essential for embryonic development and prenatal hematopoiesis. DOT1L also interacts with translocation partners of Mixed Lineage Leukemia (MLL) gene, which is commonly translocated in human leukemia. However, the requirement of Dot1l in postnatal hematopoiesis and leukemogenesis of MLL translocation proteins has not been conclusively shown. With a conditional Dot1l knockout mouse model, we examined the consequences of Dot1l loss in postnatal hematopoiesis and MLL translocation leukemia. Deletion of Dot1l led to pancytopenia and failure of hematopoietic homeostasis, and Dot1l-deficient cells minimally reconstituted recipient bone marrow in competitive transplantation experiments. In addition, MLL-AF9 cells required Dot1l for oncogenic transformation, whereas cells with other leukemic oncogenes, such as Hoxa9/Meis1 and E2A-HLF, did not. These findings illustrate a crucial role of Dot1l in normal hematopoiesis and leukemogenesis of specific oncogenes.

With the aim of finding small molecules that stimulate erythropoiesis earlier than erythropoietin and that enhance erythroid colony-forming unit (CFU-E) production, we studied the mechanism by which glucocorticoids increase CFU-E formation. Using erythroid burst-forming unit (BFU-E) and CFU-E progenitors purified by a new technique, we demonstrate that glucocorticoids stimulate the earliest (BFU-E) progenitors to undergo limited self-renewal, which increases formation of CFU-E cells textgreater 20-fold. Interestingly, glucocorticoids induce expression of genes in BFU-E cells that contain promoter regions highly enriched for hypoxia-induced factor 1α (HIF1α) binding sites. This suggests activation of HIF1α may enhance or replace the effect of glucocorticoids on BFU-E self-renewal. Indeed, HIF1α activation by a prolyl hydroxylase inhibitor (PHI) synergizes with glucocorticoids and enhances production of CFU-Es 170-fold. Because PHIs are able to increase erythroblast production at very low concentrations of glucocorticoids, PHI-induced stimulation of BFU-E progenitors thus represents a conceptually new therapeutic window for treating erythropoietin-resistant anemia.