Myeloid transcription factors: effect on stem cells, hematopoiesis, and Leukemia

Acute Myelogenous Leukemia (AML) is the most common form of acute leukemia in adults. It is characterized by a block in myeloid differentiation. Myeloid transcription factors, including PU.1 and CCAAT/Enhancer Binding Proteinalpha (C/EBP alpha), play critical roles in myeloid lineage differentiation. Disruption of PU.1 function in non-conditional knockouts in mice leads to an early multi-lineage block, and C/EBP alpha knockout mice demonstrate a block at the earliest stage of granulocyte differentiation and have myeloid blasts in the blood. Introduction of PU.1 or C/EBPalpha into multipotential precursor cells can restore lineage specific differentiation. Consistent with these phenotypes, mutations and other disruptions of C/EBP alpha and PU.1 have been found in human AML. However, the embryonic and/or perinatal mortality of both PU.1 and C/EBP alpha non-conditional knockouts has limited our ability to utilize them to model human AML. In order to further characterize the stage at which myelopoiesis is blocked, and to develop animal models of human AML, we have developed conditional knockouts of PU.1 and C/EBP alpha. Analyses of non-conditional fetal liver and conditional bone marrow PU.1 knockout hematopoietic cells demonstrate an early block at the transition from the hematopoietic stem cell (HSC) to common myeloid progenitor (CMP) and common lymphoid progenitor (CLP) stages. 3 weeks following induction of recombination in newborn PU.1 mice, the bone marrow of the animals has a large number of blasts, similar to what is observed in AML. At this time,　the recombination of the PU.1 gene is >90%. However, after 2 months the bone marrow appears normal, and no recombination of the PU.1 gene can be　observed. The conditional PU.1 mice demonstrate a block in the transition from long-term to short term phenotypic HSC, as well as a block from the HSC compartment to the CMP and CLP. However, the long-term PU.1 / HSC appear to have a qualitative defect in self-renewal, and eventually the non-recombined HSC have a selective advantage over HSC in which PU.1 has been disrupted. In contrast, non-conditional (fetal liver) and conditional (adult bone marrow) C/EBP alpha cells are blocked at a later stage, the CMP to GMP (granulocyte/macrophage progenitor) transition. Granulocytic differentiation is not observed in the peripheral blood, spleen, and bone marrow of the conditional C/EBP alpha knockout mice. However, erythroid and megakaryocyte development are normal. While CMP cells from C/EBP alpha / mice can differentiate to all different lineages in vitro, cells in GM colonies were mainly immature myeloid cells. Loss of C/EBP alpha function selectively blocks myeloid cell differentiation, but does not inhibit the development of other hematopoietic cell types. In contrast to what was observed for PU.1, C/EBP alpha / blasts persist in the bone marrow and peripheral blood for months. However, the animals do not develop a malignant leukemia. It is possible that additional abnormalities in addition to loss of C/EBP alpha function, such as FLT3 mutations and/or upregulation of anti-apoptotic signals, may be required to induce leukemia in the mouse, and this can be tested using our conditional knockout lines. We acknowledge the collaborative efforts of Julie Lekstrom-Himes, Chamorro Somoza, and Rich Murray in these studies.