MicroRNA Profiling of Human Acute Myeloid Leukemia and Normal Hematopoietic Stem/Progenitor Cells Reveals a Leukemia Stem Cell Signature.

While numerous studies have contributed important insights into the molecular origins of human acute myeloid leukemia (AML), many may not accurately reflect molecular pathways critical to AML development or maintenance because they ignore the inherent heterogeneity among AML blasts. One subset of blasts - leukemia stem cells (LSCs) - exhibits the unique ability to self-renew and to engraft disease in immunodeficient mouse hosts, suggesting that their elimination is critical to developing curative therapies. In addition, there is little information regarding the role of microRNAs (miRNAs) in regulating gene expression or biologic function in AML. In order to assess the potential contribution of miRNAs to AML LSC biology, we have evaluated the expression profile of 315 mature miRNAs in FACS-purified AML LSC and compared it to both non-LSC blasts as well as normal human bone marrow (BM) derived hematopoietic stem cells (HSC) and committed progenitors using a multiplexed TaqMan-based real-time PCR strategy. SAM analysis with stringent criteria (at least 25% samples with Ct <30, FDR <1%) reveals that AML LSC and non-LSC blasts are more similar to one another than to normal HSC or committed progenitors. Among the BM populations tested, AML LSC and non-LSC populations are most similar to the granulocyte-macrophage progenitor (GMP). A set of miRNAs distinguishes AML LSC and non-LSC from normal HSC and committed progenitors, including 35 miRNAs that are under-expressed and 33 miRNAs that are over-expressed in both AML fractions versus the normal populations; many of these differentially expressed miRNAs show a range of expression exceeding 3 orders of magnitude. Supervised clustering analysis of AML LSC and non-LSC blasts reveals an LSC signature composed of 89 miRNAs, with nearly all differentially expressed miRNAs (86/89) exhibiting lower expression levels in AML LSC than non-LSC blasts. Finally, supervised clustering identifies a “stem-cell” signature composed of 17 miRNAs that are over-expressed in AML LSC and HSC versus committed progenitors. This group of miRNAs does not include miRNAs previously described as being highly expressed in embryonic stem cells. Together, these studies represent the first direct comparison of miRNA expression in a human cancer stem cell to its normal counterpart, thereby identifying miRNAs that may regulate AML LSC and/or normal HSC/committed progenitor function. Initial functional studies in vivo using LNA knockdown strategies indicate that a subset of miRNAs highly expressed in HSC and LSC is important in regulating normal HSC function. We are currently expanding these studies to test the role of these miRNAs in maintaining engrafted AMLs in the xenotransplant setting.