Unbiased Analyses of Signaling Through Leukemia Associated MicroRNA

Abstract 2373

MicroRNA (miRNA or miR) expression profiling of primary human acute myelogenous leukemia (AML) biopsies and human AML cell lines has identified miRNA expression patterns associated with cytogenetic subtypes of AML. Elevated expression of miR-196b is associated with 11q23 translocations (MLL), whereas elevated expression of miR-126 is associated with AML1/CBFβ rearrangements in AML. Moreover, miR-196b blocks normal granulopoiesis and miR-126 inhibits erythropoiesis. To understand how these miRNAs function in normal and abnormal-malignant hematopoiesis, we sought to identify the target mRNAs they regulate. Currently miRNA target identification relies mostly on computational prediction algorithms with high error rates, or indirect methods such as microarray expression analyses of mRNAs altered upon overexpression or inhibition of the miRNA of interest. Here we applied a novel unbiased method to identify direct miR-196b and miR-126 target mRNAs by capture of miRNA-bound mRNA complexes from human fibroblasts and human AML cell lines. mRNAs that were bound to the miRNA of interest were applied to microarray analyses, and those that were present in at least 2-fold or greater quantities over those associated with a negative control C. elegans miRNA were identified as putative targets.

We identified 651 potential mRNA targets of miR-196b and 402 potential targets of miR-126. We validated this technique in several ways. First, we showed that captured mRNAs were significantly enriched for miRNA seed sequences by comparison with computer algorithm predicted targets (244/651 for miR-196b and 42/402 for miR-126), and then remaining mRNA 3'UTR sequences were examined using miRNA binding site prediction software. We also confirmed the microarray findings by qRT-PCR analyses of the captured mRNAs from multiple independent experiments. Bioinformatic gene network analyses revealed that the 651 miR-196b targets were significantly enriched for participation in cell cycle checkpoint control, kinetochore/centromere interaction during mitosis, and apoptosis. Additionally, some of the miR-196b candidate targets are know to be deregulated in bone marrow failure (myelodysplastic syndrome or MDS, hemolytic anemia, and thrombocytopenia) or have roles in granulocyte activation. Bioinformatic network analyses of the 402 miR-126 targets revealed significant enrichment for genes involved in mitotic spindle assembly, kinetochore/mitotic spindle interaction during mitosis, Rho-mediated cytoskeletal dynamics, and DNA damage repair. 199 targets were found in common between miR-196b and miR-126. Notably, cell cycle (G2/M checkpoint, and negative cell cycle regulators) was one of the most significantly enriched biological processes in the 199 shared target genes (p<0.001). Moreover, some of the miR-196b/miR-126 common targets participate in translational control complexes dysregulated in bone marrow failure (MDS), as well as transcriptional circuits deregulated in abnormal myeloid differentiation.

In sum, we optimized and implemented an innovative microRNA-mRNA target capture approach that resulted in identification of hundreds of previously unknown miRNA targets in human cells. These recent discoveries of putative miR-196b and miR-126 targets suggest novel mechanisms of miRNA gene regulation in normal and defective hematopoiesis and leukemogenesis. The identification of these miRNA targets may also help explain their distinct expression patterns in cytogenetic subtypes of AML.