DNA Damage Repair Factors have a Tumor Promoting Role in MLL-fusion Leukemia

The congenital view of the DNA damage response postulates that DNA repair factors are activated upon oncogene expression and constitute an important barrier against tumor progression (tumor protective role). In this study, Santos et al. showed that, in the particular case of MLL-fusion leukemia, the DNA damage response has an opposite role: it promotes tumorigenesis by allowing leukemic cells to self renew (tumor promoting role).

The congenital view of the DNA damage response postulates that DNA repair factors are activated upon oncogene expression and constitute an important barrier against tumor progression (tumor protective role). In this study, Santos et al. showed that, in the particular case of MLL-fusion leukemia, the DNA damage response has an opposite role: it promotes tumorigenesis by allowing leukemic cells to self renew (tumor promoting role).

Cancers develop when cells accumulate DNA mutations that allow them to grow and divide inappropriately. Thus, proteins involved in repairing DNA damage are generally suppressors of cancer formation, and their expression is often lost in the early stages of cancer initiation. In contrast, cancer stem cells, like their normal counterparts, must retain their ability to self-renew, which necessitates maintenance of DNA integrity. In hematopoietic stem cells (HSC), for example, double strand breaks and oxidative damage exhaust their regenerative ability. André Nussenzweig, Ph.D., Chief of CCR’s Laboratory of Genome Integrity and his colleagues wondered whether leukemic stem cells might be similarly constrained by DNA damage.

The researchers decided to look at the histone methyltransferase MLL4 because MLL family translocations drive most infant acute myeloid leukemia (AML) and a significant fraction of adult AML. MLL4 is a recognized tumor suppressor, though its mechanism is unknown. They first knocked out the MLL4 gene in HSC and progenitor cells in a mouse model (MLL4ko). The total number of bone marrow cells was the same between wild type (WT) and MLL4ko mice, but loss of MLL4 significantly increased HSC and myeloid cells and reduced lymphoid progenitors and B cells. Bone marrow from MLL4ko mice also showed a reduced ability to repopulate the immune system of irradiated mice when transplanted in competition with WT bone marrow. The investigators observed that, after an initial division, the MLL4ko HSC were more likely to symmetrically produce two committed progenitors. These results suggested that loss of MLL4 led to increased HSC numbers during homeostasis but reduced stem cell capacity with stress.

To understand how MLL4 could regulate stem cell function, the scientists analyzed global gene expression in WT and MLL4ko HSC. They found that genes important for mediating resistance to oxidative stress were down-regulated in the absence of MLL4. In fact, MLL4ko HSC had increased reactive oxygen species (ROS) and DNA damage, implicating these pathways in the cells’ reduced stem cell activity.

FoxO transcription factor family members regulate similar types of genes and are required for maintaining leukemia initiating cells in a mouse model of AML containing a MLL1-AF9 gene fusion. To see whether MLL4 played a similar role in this system, the researchers introduced the fusion into WT and MLL4ko hematopoietic stem and progenitor cells using a retrovirus and injected the cells into irradiated mice. As expected, the WT fusion-containing cells caused lethal leukemia. In contrast, no leukemia was detected in mice given the MLL4ko MLL1-AF9 cells, suggesting that, in spite of its role as a tumor suppressor in some cancers, MLL4 is essential for MLL1-AF9-induced leukemia.

The MLL4ko MLL1-AF9 cells also grew poorly in culture but without a detectable change in cell death. Instead, in collaboration with Peter Aplan, Ph.D., of CCR’s Genetics Branch, the investigators observed reduced numbers of undifferentiated cells compared to WT MLL1-AF9 cells and morphological characteristics associated with myeloid differentiation on the majority of MLL4ko MLL1-AF9 cells. Disrupting MLL4 expression with a short hairpin RNA in WT MLL1-AF9 cells similarly enhanced their myeloid differentiation. Examining the gene expression of WT and MLL4ko fusion-containing cells, the scientists found that the majority of MLL1-AF9 target genes were not altered with the loss of MLL4. Rather, they saw down-regulation of antioxidant genes, including those regulated by the FoxO family, and higher levels of ROS and DNA damage.

To test whether elevated ROS might be responsible for the induction of differentiation in stem cells lacking MLL4, the researchers grew MLL4ko MLL1-AF9 cells with the antioxidants N-acetyl-L cysteine (NAC) or catalase. Antioxidant treatment increased the percentage of leukemic cells and correlated with reduced DNA damage. Similarly, mice that received MLL4ko MLL1-AF9 cells and consumed NAC daily had reduced survival compared to mice that received the same number of cells but not NAC. These results indicate that the role of MLL4 in MLL1-AF9-induced leukemia is to prevent differentiation by protecting leukemic cells from oxidative and DNA damage.

Since FoxO-regulated genes appear to lie downstream of MLL4, the investigators wondered whether expression of FOXO3 could rescue MLL4 loss. Over-expression of FOXO3 in MLL4ko MLL1-AF9 cells almost completely restored cell growth and antioxidant pathway transcription and prevented myeloid differentiation, supporting the idea that the FoxO pathway is an important target of MLL4-mediated ROS protection.

The scientists then asked whether oxidative stress was sufficient to induce leukemic stem cell differentiation. Indeed, they found that treating WT MLL1-AF9 cells with hydrogen peroxide increased ROS, DNA damage, and myeloid differentiation with an accompanying decrease in leukemic blast cells. Likewise, loss of other DNA protective proteins, including ATM and BRCA1, or treatment with ATM or ATR inhibitors caused reduced cell growth and increased differentiation in WT MLL1-AF9 cells. To see whether they could separate the effects of ROS and DNA damage, the researchers expressed endonucleases in WT MLL1-AF9 cells to generate DNA double strand breaks. While they observed no change in ROS, there was a significant reduction in the number of leukemic blasts. Thus, DNA strand breaks generated directly or via ROS can induce differentiation in MLL fusion cells.

Finally, the scientists asked whether the cell cycle checkpoint regulator p21 was involved in the DNA-damage-induced differentiation of MLL1-AF9 cells. Loss of p21 expression alone had no effect on cell growth or differentiation. After endonuclease or ATM inhibitor-induced DNA damage, however, the absence of p21 in MLL1-AF9 cells permitted continued proliferation without myeloid differentiation, suggesting that activation of p21 links cell cycle exit and cell differentiation in MLL1-AF9 cells with damaged DNA.

Together these studies have revealed the importance of MLL4 and similar protectors of DNA integrity to maintaining stem-ness in MLL fusion-initiated leukemia. They also suggest that DNA repair pathway inhibitors may be active therapies against these cancers by reducing proliferation and inducing differentiation of leukemic stem cells.

Summary Posted: 08/2014

Reference

Santos MA, Faryabi RB, Ergen AV, Day AM, Malhowski A, Canela A, Onozawa M, Lee JE, Callen E, Gutierrez-Martinez P, Chen HT, Wong N, Finkel N, Deshpande A, Sharrow S, Rossi DJ, Ito K, Ge K, Aplan PD, Armstrong SA, Nussenzweig A. DNA-damage-induced differentiation of leukaemic cells as an anti-cancer barrier. Nature. July 27, 2014 PubMed Link