Transcription factors RUNX1 and CBFbeta play key roles in leukemogenesis and normal hematopoiesis. Mutations or chromosome translocations affecting RUNX1 or CBFB (which encodes CBFbeta) are found in 20-30% of patients with acute leukemia. A chromosome 16 inversion inv(16) that generates a fusion gene between CBFB and MYH11 (which encodes the smooth muscle myosin heavy chain, SMMHC) is found in all patients with acute myeloid leukemia (AML) subtype M4Eo. My lab has studied RUNX1 and CBFbeta for their roles in leukemogenesis and normal hematopoiesis for the last 16 years. We have established a mouse model of human AML by targeted insertion of the fusion gene CBFB-MYH11 in mouse ES cells, which demonstrated the critical role of CBFB-MYH11 in leukemogenesis. Using transgenic mouse and zebrafish models we have demonstrated that RUNX1 and CBFbeta are required for multiple steps of normal hematopoiesis, starting from the hematopoietic stem cells. We have also conducted in vitro analysis to understand the molecular level mechanisms of CBFbeta-SMMHC function, which will be critical for designing new therapeutic approaches for AML. Dominant RUNX1 inhibition has been proposed as a common pathway for CBF-leukemia. Using knock-in mouse strategy, we previously demonstrated that Cbfb-MYH11 dominantly blocks Runx1/Cbfb function in hematopoiesis and induces AML when accompanied by cooperating mutations. CBFbeta-SMMHC inhibits RUNX1/CBFbeta via the functions of several domains in the fusion protein including (1) RUNX1 high affinity binding domain (HABD) and (2) multimerization and repression domain (MRD). However, it is not clear whether these domains are important for leukemogenesis in vivo, since the data came from experimental systems in vitro. To test the hypothesis that they are important for leukemogenesis, we generated knockin mice expressing CBFbeta-SMMHC with deletions of the proposed functional domains. Our results suggest critical roles for the MRD, but not the HABD. The HABD-deleted protein (CBFbeta-SMMHCd179-221) did not bind RUNX1 with high affinity, was less efficient in sequestering RUNX1 and caused less severe hematopoietic defects in mice than full-length CBFbeta-SMMHC. Surprisingly the Cbfb+/MYH11d179-221 (expressing the CBFbeta-SMMHCd179-221 protein) mice developed leukemia spontaneously without ENU treatment, much faster than those mice expressing full-length CBFbeta-SMMHC. A larger pool of leukemia initiating cells, increased expression of MN1, and retention of RUNX1 phosphorylation are potential mechanisms for accelerated leukemia development in these mice. Clinically one of the 10 CBFB-MYH11 fusions detected in AML patients also encodes a fusion protein that lacks HABD, which does not bind or repress RUNX1 efficiently. Our data suggest that RUNX1 dominant inhibition is not required for leukemogenesis by CBFbeta-SMMHC (Kamikubo et al., Cancer Cell, 2010). However, it is likely that RUNX1 interaction is still required, based on our data that hemizygous deletion of Runx1 led to slower leukemia development in the Cbfb-MYH11 knockin mice (Zhao et al., unpublished data). We also found that Cbfb-MYH11 caused defects in differentiation, proliferation, and apoptosis of primitive hematopoietic cells that were undetectable or less severe in Cbfb-/- and Runx1-/- embryos, indicating that Cbfb-MYH11 has activities independent of Cbfb/Runx repression. Gene expression profiling with RNA from primitive blood cells of Cbfb+/MYH11 and Cbfb-/- embryos was conducted and led to the identification of over 500 differentially expressed genes. Most of the genes (76%) upregulated in the Cbfb+/MYH11 embryos also showed high levels of expression in leukemic cells from the Cbfb+/MYH11 mice, as well as human AML cells expressing CBFβ-SMMHC, suggesting that these gene expression changes are involved in leukemogenesis (Hyde et al., Blood, 2010). We further characterized Csf2rb, Il1rl1, and Gata2, which were chosen because they were among genes with highest expression changes in the Cbfb+/MYH11 embryos, and because they are known to have multiple functional roles during hematopoiesis. In agreement with gene expression changes, Csf2rb, Il1rl1, and Gata2 proteins were expressed at high levels in leukemic cells from the Cbfb+/MYH11 mice. In mice with a conditional allele of Cbfb-MYH11, a transient population of myeloid progenitor-like cells expressing both Csf2rb and Il1rl1 was seen in the peripheral blood soon after induction of Cbfb-MYH11 expression. This implies that upregulation of Csf2rb and Il1rl1 are early events after Cbfb-MYH11 expression. Thus, in addition to its ability to repress CBFβ/RUNX, Cbfb-MYH11 has other activities that are likely important for leukemogenesis. This finding has important implications for future efforts to develop new treatments for AML (Hyde et al., Blood, 2010). Previous results from in vitro and in vivo studies have shown that the interaction between CBFbeta-SMMHC and RUNX1 is a key step in leukemogenesis, even though our recent studies are revealing RUNX1-repression independent functions by CBFbeta-SMMHC. Likewise, RUNX1-CBFbeta interaction might be critical for leukemia involving RUNX1 mutations, such as the AML1-ETO (also known as RUNX1-ETO) fusion gene generated by t(8;21) in AML. Thus, inhibitors of CBFbeta - RUNX1 interaction may have potential therapeutic applications for both (inv)16 and t(8;21) AML, which account for 20-30% of all AML cases. In collaboration with the NIH Chemical Genomics Center (NCGC), we developed a CBFbeta and RUNX1 bead-based proximity assay in Amplified Luminescence Proximity Homogenous Assay (ALPHA) Screen format and optimized it for high throughput screening. A total of 243,398 compounds were screened with this assay at NCGC, which led to the identification of 137 putative inhibitors by Structure-Activity Relationships and Curve Class. Confirmatory ALPHA and HTRF (homogeneous time resolved fluorescence) assays were performed and candidate compounds showing consistent results were further tested by Bioacore to characterize the kinetics and binding affinity of the compounds. These follow-up tests have so far identified 70 potential candidate compounds. Three related lead hits have been confirmed in tissue culture cells and in our zebrafish model. Administration of the lead compounds to CBFb-MYH11 leukemic mice results in statistically significant reduction of leukemic cell burden. Optimization in formulation and delivery as well as chemical modifications of these 3 compounds are underway to improve efficacy, which hopefully will lead to clinical trials in human patients in the near future.
Showing the most recent 10 out of 40 publications
