Acute myeloid leukemia (AML) is a heterogeneous disease with diverse gene mutations and chromosomal abnormalities. Core binding factor (CBF) leukemias, those with translocations or inversions that affect transcription factor genes RUNX1 or CBFB, account for approximately 24% of adult acute myeloid leukemia (AML) and 25% of pediatric acute lymphocytic leukemia. The encoded proteins, RUNX1 and CBFbeta, form a heterodimer to regulate gene expression, and they are both required for hematopoiesis in vertebrate animals from zebrafish to man. Extensive clinical studies have demonstrated that CBFB-MYH11 and RUNX1-ETO, the two common fusion genes in CBF leukemia, are the best biomarkers for diagnosis, prognosis, and residual disease monitoring of CBF leukemia patients. Over the years we have used mouse models and a variety of research tools to characterize the CBFB-MYH11 fusion gene, determine the effect of the encoded protein, CBFbeta-SMMHC, on normal hematopoiesis, and understand the leukemogenesis process associated with the fusion gene. We have generated both conventional and conditional knock-in mouse models to study CBFB-MYH11. Using these models we showed that CBFB-MYH11 is necessary but not sufficient for leukemia, and we were able to identify cooperating genetic events in the mouse models. We have generated knock-in mouse models expressing truncated CBFB-MYH11 to determine the importance of functional domains of CBFbeta-SMMHC. Overall our lab has been recognized in the field as the major contributor to the understanding of CBFB-MYH11 leukemia. The C-terminus of CBFbeta-SMMHC contains domains for self-multimerization and transcriptional repression, both of which have been proposed to be important for leukemogenesis by CBFbeta-SMMHC. To test the role of the fusion proteins C-terminus in vivo, we generated knock-in mice expressing a C-terminally truncated CBFbeta-SMMHC (CBFbeta-SMMHCdC95). Embryos with a single copy of CBFbeta-SMMHCdC95 were viable and showed no defects in hematopoiesis, whereas embryos homozygous for the CBFbeta-SMMHCdC95 allele had hematopoietic defects and died in mid-gestation, similar to embryos with a single-copy of the full-length CBFbeta-SMMHC. Importantly, unlike mice expressing full-length CBFbeta-SMMHC, none of the mice expressing CBFbeta-SMMHCdC95 developed leukemia, even after treatment with ENU, although some of the older mice developed a non-transplantable myeloproliferative disease. Our data indicate that the CBFbeta-SMMHCs C-terminus is essential to induce embryonic hematopoietic defects and leukemogenesis (Kamikubo et al. Blood 121:638, 2013). In a more recent study, we generated a new Cbfb-MYH11 knock-in mouse model to dissect the role of the multimerization domain at the C terminus of CBFbeta-SMMHC. Specifically, we mutated six amino acids in the helices D and E (mDE) of the assembly competent domain, which is important for SMMHC multimerization. We found that the embryos with the mDE mutation did not develop hematopoietic defects seen in embryos with full-length CBF-SMMHC. More importantly leukemia development was abolished in the adult mice with CBFbeta-SMMHCdC95 even after mutagenesis treatment. In addition, the gene expression profile of the hematopoietic cells from the CBFbeta-SMMHCdC95 mice was more similar to that of wildtype mice than the CBFbeta-SMMHC mice. Our data suggest that the C terminal multimerization domain is required for the defects in primitive and definitive hematopoiesis caused by CBFbeta-SMMHC, and it is also essential for leukemogenesis caused by CBFbeta-SMMHC (Zhao et al., Leukemia in press). Studies of the mouse models have also demonstrated that CBFB-MYH11 dominantly inhibits RUNX1 and CBFB function during definitive hematopoiesis, resulting in total loss of definitive hematopoiesis in the heterozygous Cbfb-MYH11 knockin mouse embryos. The data suggest that CBFbeta-SMMHC is a dominant negative repressor of RUNX1. If this hypothesis is correct, reducing RUNX1 activity should facilitate leukemogenesis by CBFB-MYH11. In fact, loss of function mutations of RUNX1 are common in human AML. However, we previously demonstrated that CBFB-MYH11 has RUNX1-repression independent functions (Hyde et al., Blood 115:1433, 2010). Moreover, we recently showed that a dominant negative allele of Runx1, Runx1-lz, delayed leukemogenesis by CBFB-MYH11 in a mouse model (Hyde et al., Leukemia 29:1771, 2015). These findings challenged the RUNX1-repression model for CBFbeta-SMMHC mediated leukemogenesis. However, our previous findings are not conclusive since the Runx1-lz mice used in the previous study still retain some Runx1 function. To definitively address this question, we crossed Cre-based conditional Runx1 knockout mice with Cre-based conditional Cbfb-MYH11 knockin mice to generate mice that express CBFbeta-SMMHC but not Runx1 after pIpC (poly I:C) treatment to induce Cre expression. Our results showed that, as expected, conditional Cbfb-MYH11 knockin mice developed leukemia with an average survival of 4 months. On the other hand, none of the conditional Runx1 knockout/conditional Cbfb-MYH11 knockin mice developed leukemia up to one year after pIpC treatment. These results suggest that Runx1 is absolutely required for Cbfb-MYH11 induced leukemogenesis. To further study the mechanism of leukemogenesis, we performed RNA-Seq on C-kit+ bone marrow cells isolated from mice two weeks after pIpC treatment, to explore the global gene expression changes caused by Runx1 knockout on Cbfb-MYH11 expressing mice. Our preliminary data analysis showed that 1688 genes were differential expressed between conditional Cbfb-MYH11 knockin mice and conditional Runx1 knockout/conditional Cbfb-MYH11 knockin mice. Interestingly, many of these genes (48%) are Runx1 target genes. The above results suggest that CBFbeta-SMMHC induces leukemia through mis-regulating the expression of Runx1 target genes (Zhen et al., ASH abstract 2016).
Showing the most recent 10 out of 40 publications
