Abstract
RUNX1-ETO (also known as AML1-ETO and AML1-MTG8) is a fusion gene generated from t(8;21), which is a common chromosome translocation in acute myeloid leukemia (AML). It has been shown that t(8;21) requires additional aberrations to induce leukemia. Interestingly, 32-59% of t(8;21) patients also display loss of a sex chromosome (LOS) in their leukemia cells. Therefore, loss of the genes located on the sex chromosomes, especially in the pseudoautosomal regions (PARs) that are shared between the X and Y chromosomes, may contribute to RUNX1-ETO leukemia development. One gene of interest in the PARs is CSF2RA, which encodes the alpha subunit of the granulocyte-macrophage colony-stimulating factor (GM-CSF) receptor. When the GM-CSF receptor is bound to its ligand, downstream signaling events promote various functional outcomes including proliferation, differentiation, self-renewal, and survival of myeloid cells. Thus, GM-CSF signaling has the potential to regulate both normal and malignant hematopoiesis. We previously reported that mice expressing RUNX1-ETO in GM-CSF deficient hematopoietic cells displayed higher incidence of leukemia (Matsuura S et al. 2012 Blood 119:3155). This result suggests that GM-CSF signaling is inhibitory to RUNX1-ETO dependent leukemogenesis. Furthermore, GM-CSF treatment reduces the self-renewal potential of RUNX1-ETO expressing cells and promotes myeloid differentiation in replating assays. We therefore hypothesize that the negative effect of GM-CSF on RUNX1-ETO induced leukemia development is due to the activation of selected GM-CSF downstream signaling pathway(s) that diminish self-renewal capacity and promote myeloid differentiation.
To understand the molecular mechanism of the negative effect of GM-CSF on t(8;21) leukemogenesis, in the current report, we conducted a gene expression profiling assay to examine the effect of GM-CSF on RUNX1-ETO cells. MigR1 vector control or MigR1-RUNX1-ETO retrovirus transduced lineage negative/c-Kit positive (Lin-/c-Kit+) murine hematopoietic stem/progenitor cells (HSPCs) were cultured with or without GM-CSF for 24 hours. Then, Lin-/c-Kit+/GFP+ HSPCs were isolated for the profiling study. We observed little response to GM-CSF in control HSPCs, with only 4 genes being differentially expressed after a 2-fold cutoff. Conversely, 122 genes were differentially expressed in RUNX1-ETO cells treated with GM-CSF. These results clearly indicate that RUNX1-ETO specifically enhances GM-CSF responsiveness in HSPCs. Gene Set Enrichment Analysis (GSEA) of the differentially expressed genes in RUNX1-ETO cells reveals that this response resembles that of GM-CSF-induced myeloid differentiation. Furthermore, pathway analysis of these differentially expressed genes predicts MEK1/2 and ERK1/2 to be activated after GM-CSF treatment in RE cells. We previously reported that ERK1/2, downstream targets of MEK1/2, are hyper-phosphorylated after GM-CSF treatment of RUNX1-ETO cells, and MEK-ERK activation has been shown to regulate cell proliferation and myelopoiesis. Other GM-CSF induced genes are predicted targets of MYD88. MYD88 is upregulated during myeloid differentiation. Its in vivo knockout has been reported to result in an increase of hematopoietic stem cells (HSCs) and reduction of mature granulocytes. Most interestingly, a subset of genes upregulated in GM-CSF treated RUNX1-ETO cells are predicted to be activated by CEBPβ. CEBPβ can heterodimerize with CEBPα and is induced during myelopoiesis, critical for macrophage differentiation, capable of promoting granulopoiesis, and involved in regulating granulopoiesis in vivo.
In conclusion, our data suggest that RUNX1-ETO expression results in hyper-responsiveness to GM-CSF. Such enhanced GM-CSF signaling activates the expression of a specific group of genes and results in the reduced self-renewal capacity and increased myeloid differentiation of HSPCs. These GM-CSF effects are likely involved in reducing the leukemogenic potential of RUNX1-ETO and may be considered for specific therapeutic interventions.
Disclosures:
No relevant conflicts of interest to declare.
The effect of vascular diseases on bioimpedance measurements: mathematical modeling
