Acute Myeloid Leukemia (AML)

Chromosomal translocations target master regulatory genes that affect growth, differentiation, and apoptosis and are closely associated with the development of acute leukemia and a growing number of solid cancers. In Acute Myeloid Leukemia (AML), the eight-twenty-one transcriptional co-repressor (ETO, also known as MTG8 or RUNX1T1) and its paralog, ETO2/MTG16, are fused to the RUNT DNA-binding domain of AML1 (RUNX1) by the t(8;21) and t(16;21). While the t(8;21) is generally regarded as a “good” prognostic factor, 1/3 of these patients will relapse after intensive chemotherapy. Additionally, the inv(16)(p13.3q24.3), observed in 27% of “non-Downs” AMKL cases, fuses the majority of MTG16 to GLIS2 and yields poor outcomes. Moreover, The Cancer Genome Atlas (TCGA) showed that members of this gene family were altered in some of the most frequent tumor types including squamous cell lung cancer, colorectal carcinoma, and invasive breast cancer (cBIO portal). Consistent with genetic inactivation in human tumors, our mouse deletion models showed that the loss of Mtg16 (CBFA2T3) or Mtgr1 (CBFA2T2) triggered intestinal tumor formation in the context of inflammatory carcinogenesis or mutant APC. Thus, understanding the mechanism of how these factors control gene expression networks will open up new therapeutic opportunities in a variety of tumor types.

While it is generally agreed that ETO is a transcriptional co-repressor and that the t(8;21) fuses the DNA binding domain of AML1 to this repressor, a deep mechanistic understanding of how AML1, ETO, and AML1-ETO regulate transcription is lacking. For many years, transcription has been studied using siRNA approaches or genetic deletion that take days to establish knockdown or knockout, which allows cells to compensate and alter chromatin structure. These genetic experiments coupled with ChIPseq have been used to suggest that transcription factors regulate hundreds to thousands of genes, but this could be due to the cells changing phenotype, rather than just the genes controlled by a single factor. This is especially true of master regulatory genes such as AML/RUNX1 and AML1-ETO.  In fact, we know that AML1 and AML1-ETO control the expression of other transcription factors (e.g., CEBPA and PU.1) to control myeloid differentiation, so shRNA or CRISPR knockout experiments done 2-7 days after infection are comparing undifferentiated cells to differentiating cells.

To overcome these technical challenges and address these critical questions, we have used CRISPR-directed homology-directed repair to add the FKBP12-F36V  degron tag (dTAG) into the endogenous allele of AML1-ETO in Kasumi-1 cells. The dTAG system provides the opportunity to use a drug to endogenous AML1-ETO to rapidly inactivate it and fully define the molecular mechanism of AML1-ETO action throughout the genome. At the same time, we can test how effective a drug to AML1-ETO would be, and we are already defining the therapeutic benefits using cord blood CD34+ human hematopoietic stem cell cultures. Most importantly, this system allows us to avoid compensation in transcription networks that can occur in just a few hours. By “dTAGging” AE and degrading the fusion protein within 1-2 hr and examining nascent transcription using precision nuclear run-on transcription sequencing (PROseq), we have defined a core transcriptional circuit regulated by AE and we are now poised to define the mechanism(s) underpinning this regulation. By defining how AML1-ETO controls a small core gene network, we will define how it triggers AML, which is a big step toward curing this disease.