Aberrant transcriptional programs play a critical role in the development of acute myeloid leukemias (AMLs). Although persistent over-expression of MEIS1 and HOXA9 has been shown to be essential for the initiation and maintenance of MLL-associated leukemia, it is still poorly understood what additional transcriptional regulators, independent of the MLL fusion-driven MEIS/HOX pathway, dictate the development of MLL leukemia. Considering that AMLs with MLL translocation are typically associated with the monocytic lineage (FAB M4 and M5), we explored the potential role of the monocytic lineage-specific transcriptional program in MLL leukemia.

Using 97 genome-wide expression profiles of human MLL leukemias, we constructed an MLL distinctive transcriptional regulatory network. In addition to well-known transcriptional factors in leukemia development such as MEIS1 and HOXA family genes, we identified a highly active monocyte-specific gene signature that includes transcription factor PU.1. In our effort to determine the functional role of PU.1 in MLL leukemia, we found that lower PU.1 expression significantly delayed the onset of MLL- AF9 induced leukemia in primary bone marrow transplantation assay. MLL leukemia failed to maintain in vivo upon induced deletion of the PU.1 gene. To examine the clinical relevance of the PU.1 in AML patients, we further performed multivariate Cox proportional-hazards regression analysis in four published datasets of patients with AML, for whom gene expression and time-to-event data were available. We found that a PU.1-regulated 40-gene signature showed profound concordance with prognosis in segregating high-risk and low-risk AML patients. When specific subgroups of AMLs were examined, the PU.1 expression signature could predict patient outcome for MLL patients, but not in other major AMLs, such as t(8;21), t(15;17) and inv(16).

We further explored the molecular mechanisms underlying the critical role of the PU.1 program in MLL leukemia. Functional annotation of this PU.1 expression signature identified the MEIS/HOX pathway (MEIS1, FLT3, KIT), as well as key genes in the inflammatory response (AIF1, NF-KB1 and CD180). We showed that PU.1 is required to maintain high expression of Meis1 and Pbx3 and also important downstream genes in the MEIS/HOX pathway that includes known MEIS/HOX targets c-Kit and Flt3. Using ChIP-sequencing, we demonstrated that PU.1 interacts with the MEIS/HOX regulatory program through co-binding with MEIS1 at the target genomic regions in a MLL-ENL cell line. In our effort to determine the role of PU.1-controlled inflammatory response genes, we found that the growth inhibition in PU.1 knockdown MLL leukemic cells was partially rescued by addition of the monocytic inflammatory cytokine AIF1. AIF1 provides an anti-apoptotic effect through activation of the NF-ƒÛB pathway and additional known apoptosis regulators. Interestingly, AML patients with higher expression of both AIF1 and MEIS1 had a significantly shorter overall survival time than those with lower expression of both genes. Patients with high expression of either MEIS1 or AIF1 had medium survival possibilities. Notably, the prognostic value of AIF1 and MEIS1 remained in those with monocytic AMLs (P=0.00079), but not in the non- monocytic group of patients (p=0.105). Collectively, these results strongly suggest that the monocyte-specific inflammatory cytokine AIF1 is an MEIS/HOX independent essential regulator in monocytic AMLs such as MLL leukemia.

Loss of function PU.1 is leukemogenic in mouse models. Suppression of PU.1 activity is also required for the development of human myelocytic M2/M3 leukemia. Here we reveal a converse role for PU.1 as an essential positive regulator in the development of MLL myeloid leukemia, mostly M4/M5 monocytic AMLs. Our study demonstrats that the monocyte-specific PU.1-driven transcriptional program independently contributes to the development of myeloid MLL leukemia, in parallel with the MLL fusion pathway. PU.1 and downstream macrophage specific inflammatory cytokine AIF1 have important prognostic value and may serve as novel therapeutic targets for MLL leukemias.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.