Dysfunction or death of pancreatic β cells underlies both types of diabetes. This functional decline begins with β cell stress and de-differentiation. Current drugs for type 2 diabetes (T2D) lower blood glucose levels but they do not directly alleviate β cell stress nor prevent, let alone reverse, β cell de-differentiation. We show here that Urocortin 3 (Ucn3), a marker for mature β cells, is down-regulated in the early stages of T2D in mice and when β cells are stressed in vitro. Using an insulin expression-coupled lineage tracer, with Ucn3 as a reporter for the mature β cell state, we screen for factors that reverse β cell de-differentiation. We find that a small molecule inhibitor of TGFβ receptor I (Alk5) protects cells from the loss of key β cell transcription factors and restores a mature β cell identity even after exposure to prolonged and severe diabetes.

AML1-ETO is one of the most common chromosomal translocation products associated with acute myelogenous leukemia (AML). Patients carrying the AML1-ETO fusion gene exhibit an accumulation of granulocyte precursors in the bone marrow and the blood. Here, we describe a transgenic zebrafish line that enables inducible expression of the human AML1-ETO oncogene. Induced AML1-ETO expression in embryonic zebrafish causes a phenotype that recapitulates some aspects of human AML. Using this highly tractable model, we show that AML1-ETO redirects myeloerythroid progenitor cells that are developmentally programmed to adopt the erythroid cell fate into the granulocytic cell fate. This fate change is characterized by a loss of gata1 expression and an increase in pu.1 expression in myeloerythroid progenitor cells. Moreover, we identify scl as an early and essential mediator of the effect of AML1-ETO on hematopoietic cell fate. AML1-ETO quickly shuts off scl expression, and restoration of scl expression rescues the effects of AML1-ETO on myeloerythroid progenitor cell fate. These results demonstrate that scl is an important mediator of the ability of AML1-ETO to reprogram hematopoietic cell fate decisions, suggesting that scl may be an important contributor to AML1-ETO-associated leukemia. In addition, treatment of AML1-ETO transgenic zebrafish embryos with a histone deacetylase inhibitor, Trichostatin A, restores scl and gata1 expression, and ameliorates the accumulation of granulocytic cells caused by AML1-ETO. Thus, this zebrafish model facilitates in vivo dissection of AML1-ETO-mediated signaling, and will enable large-scale chemical screens to identify suppressors of the in vivo effects of AML1-ETO.