Leukemogenesis by Truncated CBFβ-SMMHC Defective in Repressing RUNX1.

Chromosome 16 inversion, inv(16)(p13q22), is one of the most frequent chromosome abnormalities in human acute myeloid leukemia (AML), comprising almost 100% of subtype M4Eo and up to 15% of all AML. This inv(16) leads to an in-frame fusion of CBFB and MYH11 genes. CBFB-MYH11 encodes a fusion protein between CBFβ, which is an obligate partner of RUNX1 or AML1, and smooth muscle myosin heavy chain (SMMHC). Using knock-in mouse models we have previously demonstrated that Cbfb-MYH11 dominantly blocks Runx1/Cbfb function in hematopoiesis and predisposes mice to AML (requiring a second hit from ENU or retroviral mutagenesis). However, the molecular mechanism underlying these findings remains unclear. Current hypotheses, which are based on previous in vitro studies, focus on the ability of CBFβ-SMMHC to dominantly inhibit RUNX1/CBFβ and include

1. CBFβ-SMMHC binding RUNX1 with higher affinity than CBFβ;

2. CBFβ-SMMHC sequestration of RUNX1 in the cytoplasm;

3. CBFβ-SMMHC stabilization of RUNX1 by decreasing RUNX1 ubiquitination; and

4. CBFβ-SMMHC repression of RUNX1 transactivation, which is dependent on SMMHC multimerization and repressor recruitment.

To test these hypotheses in vivo, we generated knock in chimeric and F1 heterozygous mice expressing CBFβ-SMMHC with C-terminal and internal deletions. One of the deletions removes only the domain responsible for high affinity binding of RUNX1 (HABD). The CBFβ-SMMHC protein with HABD deletion did not bind RUNX1 with higher affinity than CBFβ and was hypothesized to be unable to dominantly repress RUNX1. Consistent with this hypothesis, the HABD-deleted protein was less efficient in sequestering RUNX1 and caused less severe hematopoietic defects in F1 embryos than full length CBFβ-SMMHC. In contrast to mice expressing full length CBFβ-SMMHC, which develop AML only after ENU or retroviral mutagenesis, most HABD-deleted chimeric and all HABD-deleted F1 mice developed AML spontaneously shortly after birth. A larger deletion removed both the HABD and the RUNX1 stabilization domain (RSD). CBFβ-SMMHC with the HABD-RSD double deletion did not bind RUNX1 with high affinity and could not sequester RUNX1 in the cytoplasm. Mice expressing HABD-RSD double deleted CBFβ-SMMHC had normal hematopoiesis, did not develop leukemia spontaneously and developed T cell (not myeloid) malignancies after ENU treatment. These data suggest that HABD and RSD are important for CBFβ-SMMHC to block RUNX1 function and to dominantly impair hematopoiesis. The accelerated leukemogenesis associated with HABD deletion and induction of T cell malignancies in mice expressing HABD-RSD double deleted CBFβ-SMMHC strongly support the hypothesis that CBFβ-SMMHC can induce leukemia through RUNX1-inhibition-independent pathways.