RUNX1-ETO-Dependent Transcriptional Repression of RASSF2 Contributes to t(8;21) Leukemia through Evasion of MST1-Driven Apoptosis Signaling

Abstract

The t(8;21) chromosomal translocation is among the most frequent recurring cytogenetic abnormalities associated with acute myeloid leukemia (AML), found in 8-12% of de novo AML patients. The t(8;21) results in the stable fusion of the RUNX1 and RUNX1T1 genes, and formation of the oncofusion protein RUNX1-ETO (AML1-ETO). RUNX1-ETO is composed of the N-terminal DNA-binding domain of RUNX1 and nearly the entire ETO protein. RUNX1-ETO promotes leukemia development via the recruitment of transcription factor/transcriptional repression complexes (including NCOR, HDACs, p300, etc.) to regulatory regions of RUNX1 target genes known to be critical for myeloid differentiation and function, such as CEBPA, SPI1 (PU.1), NFE2, and CSF1R. Despite this knowledge, additional RUNX1-ETO target genes remain poorly characterized, and the complete molecular mechanism through which RUNX1-ETO leads to leukemic transformation remains to be elucidated. We propose that a better understanding of additional RUNX1-ETO target genes will lead to the potential for development of novel therapeutics to treat these patients.One such gene that we initially identified as markedly downregulated in RUNX1-ETO leukemia cells using a mouse model of t(8;21) AML is RASSF2 (Lo et al, Blood, 2012). Assessment of publicly available gene expression data revealed that RASSF2 is specifically downregulated in the bone marrow of t(8;21) AML patients compared to patients of different cytogenetic subtypes or to non-t(8;21) FAB subtype M2 AML patients. Additionally, RT-qPCR analysis confirmed that RASSF2 transcript is downregulated 10-100-fold in the t(8;21) AML cell lines, Kasumi-1 and SKNO-1, compared to non-t(8;21) AML cell lines and normal CD34+ hematopoietic cells. Expression of RUNX1-ETO in a non-t(8;21) AML cell line led to a reduction in RASSF2 mRNA expression, while knockdown of RUNX1-ETO in Kasumi-1 cells resulted in a ~5-fold increase in RASSF2 expression. Assessment of published ChIP-seq data showed that RUNX1-ETO directly binds at two regulatory regions within the RASSF2 genomic locus in t(8;21) AML cell lines and patient samples.Re-expression of RASSF2 at physiological levels in t(8;21) AML cell lines resulted in a modest negative growth phenotype, and greatly sensitized these cells to apoptosis following stimulation with various pro-apoptotic agents. Re-expression of RASSF2 in RUNX1-ETO-transduced primary mouse bone marrow caused these cells to lose their long-term self-renewal ability after 3 weeks in a serial replating/colony formation assay. This loss of self-renewal ability in co-transduced cells was accompanied by a marked increase in apoptosis during each of the first three weeks of replating. Mechanistically, re-expression of full-length RASSF2, but not of a deletion mutant lacking the SARAH heterodimerization domain (RASSF2ΔSARAH), in t(8;21) AML cell lines resulted in increased protein amount of the pro-apoptotic kinase, MST1. This suggests that RASSF2 may be a critical regulator of MST1 protein stability in AML cells. Importantly, modest (2-3-fold) overexpression of MST1 in t(8;21) AML cell lines resulted in a significant increase in apoptosis and caused growth arrest. The effects of RASSF2 or MST1 expression in non-t(8;21) AML cell lines were greatly reduced, suggesting that the cellular context of RUNX-ETO-driven leukemias makes them highly susceptible to MST1-dependent apoptosis.Overall, we have identified the importance of a MST1-driven pro-apoptotic signaling axis in t(8;21) leukemia. RUNX1-ETO-dependent transcriptional repression of RASSF2 may be essential for evasion of this apoptosis signaling during leukemic transformation via reduction of MST1 protein stability. MST1, perhaps better known as the mammalian orthologue of the drosophila Hippo kinase, is a critical tumor suppressor in many solid tumor types; and we believe our studies warrant the continued investigation of this pathway in hematological malignancy. DisclosuresNo relevant conflicts of interest to declare.