Genetics and Epigenetics of the PU.1 Upstream Regulatory Element: AML1 Binding Sites Are Critical for Leuekmia Induced by the AML/ETO9a Fusion Oncogene

The level of expression of the transcription factor PU.1 is a critical determinant of lineage commitment in normal hematopoiesis, and dysregulation of PU.1 leads to development of leukemia. In mice with targeted disruption of the PU.1 upstream regulatory element (URE), expression of PU.1 is decreased to 20% of wild type levels and results in development of acute myeloid leukemia (AML). These data suggests that tightly regulated PU.1 expression is important to maintain normal hematopoiesis and prevent leukemogenesis. Previously, we reported that AML1 (RUNX1) regulated PU.1 expression. Here we demonstrate that AMLl regulates PU.1 through 3 AML1 binding sites in the URE. Mice with targeted mutations in the 3 AML1 binding sites have decreased PU.1 expression in multiple hematopoietic lineages at multiple different developmental stages. Conditional targeting of AML1 in transgenic mice in which the URE homology region 2 (H2, containing all 3 AML1 binding sites) is used to drive expression of a reporter decreased reporter gene expression, suggesting that AML1 regulates PU.1 through these 3 sites in URE homology region 2. Using a second mouse model with a targeted mutation in the PU.1 binding site in the PU.1 URE (which is flanked by the 3 AML1 sites), we demonstrated that PU.1 indeed autoregulates itself through the URE. These results demonstrated that AML1 regulates PU.1 through the 3 AML1 sites in the URE. However, while low levels of PU.1 lead to leukemia, we have not observed frank leukemia development in AML1 conditional knockouts or in mice with targeted disruption of the 3 AML1 sites in the PU.1 URE. We hypothesized that this might be the case because disruption of AML1 or the AML1 sites reduces PU.1 levels to about 40% of wild type, but not as great as that found in PU.1 URE knockouts, which do progress to AML (20% of wild type). We hypothesized that downregulation of PU.1 as a result of binding of AML1/ETO fusion proteins to the URE might result in further reductions of PU.1 expression, and contribute to leukemogenesis. Therefore, we predicted that development of leukemia might be delayed in mice with mutations in the PU.1 URE AML1 DNA binding sites, and this was indeed the case in a modle using a retrovirus expressing the AML1/ETO9a form. We further explored the effect of AML1 and PU.1 binding on chromatin strucutre using chromatin immunoprecipitation (Chip) in the AML1 and PU.1 site URE knockin models, and found that AML1 is involved in H3K4me3 and H3/H4 acetylation of histone tails in the PU.1 URE, while PU.1 is involved in H3/H4 acetylation but not H3K4me3; H3K4 methylation and H3 acetylation decreased in AML1 sites mutant knockin mice and H3 acetylation decreased in PU.1 site mutant knockin mice. Mutation of the AML1 site in mice not only altered the chromatin structure of the URE region, but also interefered with the physical interaction between the URE and PU.1 promoter, as assessed by chromosome capture configuration (3C) assays. Interestingly, the AML1/ETO9a fusion oncogene has a unique role on the epigenetic status of the PU.1 URE in addition to its dominant effect on the 3 AML1 sites. AML1-ETO9 blocks the autoregulation of PU.1 through the PU.1 site in the URE. In summary, our data suggests that AML1 regulates PU.1 expression through 3 AML1 binding sites in the PU.1 URE by modifying chromatin structure in the URE region. In addition, PU.1 can autoregulate itself by facilitating similar epigenetic changes. Dysregulation of the epigenetic status by chromosome translocation products such as AML1-ETO might play an important role in leukemogenesis.

First two authors contribute equally to this work.