Leukemic stem cells (LSCs) share many of the same properties of normal hematopoietic stem cells (HSCs) including their highly quiescent state, capacity to self-renew, low levels of reactive oxygen species (ROS) and enhanced DNA repair program. These properties make the efficient and specific eradication of these cells challenging. Foxo3 and p53 are two transcription factors essential for the modulation of HSC quiescence and self-renewal. While Foxo3 is inhibited by signaling from several oncoproteins but crucial for the maintenance of the LSCs in both chronic and acute myeloid leukemia (CML and AML respectively), mutations of p53 although rare, are associated with poor prognosis in advanced stages of these diseases. In vivo ROS-mediated activation of p53 is known to lead to loss of quiescence, alterations of cell cycle and exhaustion of the Foxo3-/- HSC pool. Seeking to understand the contribution of p53 to Foxo3-/- HSC cycling defects, we crossed p53+/- and Foxo3+/- mice. To our surprise we found the bone marrow (BM) frequency of both p53+/-Foxo3-/- and p53-/-Foxo3-/- LSK (Lin-Sca1+cKit+) and long-term-HSC (LT-HSC, LSK Flk2-CD34-) populations greatly increased as compared to their Foxo3-/- counterparts (n=5 mice per genotype; p<0.05). Using Ki67 and DAPI staining we found that loss of one or both alleles of p53 gradually rescued the cell cycle defect of Foxo3-/- HSC and increased the frequency of LSK cells in Go by 2-fold. Loss of p53 also rescued the impaired capacity of Foxo3-/- LSK cells to competitively repopulate multilineage blood over 16 weeks, as shown by the higher frequency of p53+/-Foxo3-/- and p53-/-Foxo3-/- donor-derived cells in the peripheral blood of recipient animals (∼47% recipients of double-mutant cells versus 20% in Foxo3-/- recipients, n=5 per group). Furthermore, loss of p53 significantly improved the compromised self-renewal of Foxo3 mutant HSC in serial BM transplantations. In our quest to identify mechanisms whereby p53 depletion improves Foxo3-/- HSC function, we noticed that the DNA damage accumulated in Foxo3-/- HSC at the steady-state was remarkably ameliorated by removal of one or both alleles of p53 from Foxo3-/- HSCs, as measured by flow cytometry levels of phospho-H2AX (gamma-H2AX) and DNA breaks by comet assay (n=3, p<0.05). Unexpectedly, ROS levels were also significantly reduced by 30% in p53+/-Foxo3-/- in comparison to Foxo3-/- LSK cells, while ROS levels in p53+/- LSK cells were similar to that in WT cells. Consistent with these results, the expression of several anti-oxidant enzymes including Sod1, Sod2, Catalase, Gpx1, Sesn1 and Sesn2 (n≥2), was highly upregulated while a number of genes implicated in mitochondrial generation of ROS were significantly deregulated as a result of loss of one or both alleles of p53. These combined findings suggest that a switch from anti-oxidant to pro-oxidant activity of p53 contributes to Foxo3-/- HSC defects. Despite their apparent normal stem cell function, p53+/-Foxo3-/- HSC were highly altered in their gene expression profile. Interestingly, Gene Set Enrichment Analysis (GSEA) of the microarray analysis (Illumina bead chip mouse-Ref8) of WT, p53+/-, Foxo3-/-, and p53+/-Foxo3-/- LSK cells showed that a cluster of genes associated with fatty acid metabolism was highly enriched in p53+/-Foxo3-/- HSCs (ES=0.746; p<0.01). In addition, from 3976 genes exclusively deregulated in p53+/-Foxo3-/- LSK cells, 201 (out of 1051) overlapped with genes downregulated, while 9 (out of 14) overlapped with genes exclusively upregulated in a LSC-gene signature. To evaluate whether this pre-leukemic profile was associated with increased susceptibility to malignancy, we compared the potential and timeline of BCR-ABL-transformed p53+/-Foxo3-/- HSC as compared to controls in establishing CML in mice. We found a shorter time to the onset of the disease and decreased survival of the recipients of p53+/-Foxo3-/- transformed HSCs (n=4 per group, p<0.05) as compared to WT and Foxo3-/- controls. We propose that the p53+/-Foxo3-/- double-mutant HSCs are enriched for preleukemic stem cells based on their quiescence and self-renewal capacity, low ROS, robust DNA repair, susceptibility to transformation and aberrant gene expression profile. These findings raise the possibility that the coordinated Foxo3 and p53 regulation of ROS wires together the stem cell program.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.