Much of the morbidity and mortality caused by Acute Myeloid Leukemia (AML) is secondary to failure of normal hematopoiesis. Several recent studies indicate that this is not due to "overcrowding" within the bone marrow, suggesting other mechanisms account for the loss of bone marrow regenerative function in the leukemic niche. We recently reported that AML patient leukemia cells produce exosomes – small, membrane-enclosed extracellular vesicles – that are enriched in microRNA (miRNA) content and alter gene expression, cytokine secretion, and homing upon entry into neighboring stromal and progenitor cells. In order to examine the consequences of exosome trafficking in vivo, we conducted a series of experiments in a xenograft model. Here, we show that introduction of either AML cells or isolated AML-derived exosomes into the marrow of immunodeficient mice causes suppression of marrow hematopoietic stem and progenitor cell (HSPC) clonogenicity and a dramatic increase in peripheral circulating HSPC that coincided with a marked decrease in expression of Scf and Cxcl12 in murine stromal cells. Hypothesizing that transferred miRNA mediates the changes observed in recipient cells, we evaluated the miRNA content of two AML cell lines and their exosomes using microarrays paired with qRT-PCR. These experiments revealed stark differences between the total miRNA produced by AML cells and the selective incorporation of miRNA in exosomes. Combining AML exosome-enriched miRNA identified in our screen with a review of miRNA previously identified as significant in AML biology, we selected a panel of highly enriched miRNA, including miR-155 (100-fold enriched in exosomes) for further study. In order to mechanistically identify the mRNA targets of these miRNA, we used the uniquely powerful RISC-Trap method (Cambronne et al, PNAS, 2012), beginning with miR-155. This assay revealed 131 likely targets of miR-155, a set which we compared with several other target detection methods previously attempted with this miRNA, finding both significant overlap and several promising new putative mRNA targets. Of these 131 RISC-Trap targets, 98 were predicted by one or more of six tested target prediction algorithms (miRWalk). We compared the miR-155 targets identified by RISC-Trap to predicted targets for several other miRNA in the panel selected from our microarray studies, identifying both miRNA potentially responsible for the expression changes seen in stroma as well as several potential targets common to multiple miRNA. Using bioinformatics analysis of direct targets of our miRNA of interest and their common interacting partners, we were able to identify a network of targets and cellular processes regulated by AML exosomes, including regulation of transcription (p53, SOX9, CEBPB) and apoptosis (ESR1, TRAF2, CHEK2). These networks may provide insight into the mechanisms by which AML impairs hematopoiesis, both directly, through effects on HSPC, and indirectly, through effects on stroma. Ongoing RISC-Trap studies of other miRNA provide a forward genetic screen to evaluate the paracrine targets of exosomal miRNA in a mechanistic way, in specific niche cell types. By identifying the wider networks in addition to the individual contributors of exosomal miRNA in the AML microenvironment, we hope to uncover new non-leukemia targeted “niche” therapies to relieve hematopoietic suppression in AML patients.


Cambronne:Clontech/Takara Bio: commercial licensing of the RISC-trap technology Patents & Royalties.

Author notes


Asterisk with author names denotes non-ASH members.