Chemotherapy resistance is the most important clinical challenge in treating patients with acute myeloid leukemia (AML). The majority of patients with AML initially respond to chemotherapy but over half of responders will relapse and eventually die of disease. Traditionally, chemotherapy resistance is hypothesized to represent selection for a genetically distinct clone that has properties of leukemic stem cells (LSC). Recent data and our preliminary results (below) indicate that neither part of this model is fully correct. Rather, we propose that chemotherapy resistance can and frequently is an acquired epigenetic response to chemotherapeutic challenge that may arise in any leukemic clone. In vitro, we have modeled chemotherapy resistance and demonstrate that both DNA methylation and histone methylation are altered rapidly (within days) after treatment with Ara-C. Specifically, H3K27me3 is increased and H3K4me2 is decreased. In order to initially study the contribution of epigenetic changes in primary cells, we have compared the genetic and DNA methylation signature of 140 pairs of AML cells at diagnosis and relapse. Importantly, preliminary evidence suggests that at least 20% of patients have no new detectable genetic abnormalities at relapse. In contrast, when comparing paired diagnosis vs. relapse AML specimens we observe extensive and partially convergent redistribution of cytosine methylation affecting specific pathways, regardless of genetic background. Taken together these studies suggest an epigenetic mechanism of chemotherapy resistance but do not address the role of leukemic stem cells. To address this, we have studied the effect of cytosine arabinoside (Ara-C) on AML cells xenografted in NSG mice. Importantly, we demonstrate for the first time that limiting dilution analysis of Ara-C treatd AML cells does NOT show an enrichment in SCID-leukemia initiating cells. Collectively these data lead us to hypothesize that chemotherapy resistance in AML is a biologically complex event that includes but is not fully explained by genetic lesions and is composed of both fixed and inducible epigenetic elements. To test and explore these ideas, we propose 3 Specific Aims:
Specific Aim 1) Identify and characterize candidate mechanisms of epigenetic response to Ara-C in AML cells in vitro.
Specific Aim 2) Identify epigenetic and genetic determinants of relapse in primary AML patient samples.
Specific Aim 3) Determine the dynamics and mechanism through which chemoresistant features emerge in leukemic cell populations.
Acute myeloid leukemia is a highly dangerous blood cancer with a high mortality rate. Most patients respond initially to chemotherapy but subsequently relapse with an incurable form of the illness. We propose a novel mechanism underlying chemotherapy resistance and describe approaches to reverse chemoresistance in order to develop better therapies for patients with AML.
