Long Noncoding RNAs in Erythropoiesis and Megakaryopoiesis.

Abstract 2331

Lnc (long noncoding) RNAs are RNA transcripts greater than 200nt that regulate gene expression independent of protein coding potential. It is estimated that thousands of lncRNAs play vital roles in diverse cellular processes and are involved in numerous diseases, including cancer. We hypothesize that multiple lncRNAs regulate erythrocyte and megakaryocyte formation by modulating gene expression. To identify lncRNAs in erythro-megakaryopoiesis, we purified two biological replicates each of murine Ter119⁺ erythroblasts, CD41⁺ megakaryocytes and bipotential megakaryocyte-erythroid progenitors (MEPs) [Lin⁻ Kit⁺, Sca1⁻, CD16/32⁻, CD34⁻]. We performed strand-specific, paired-end, 200nt-read-length deep sequencing (RNA-Seq) to a depth of ∼200 million reads per sample using the Illumina GAII platform. We used the Tophat and Cufflinks suite of bioinformatic tools to assemble and compare de-novo transcriptomes from these three cell types, producing a high-confidence set of 69,488 transcripts. We confirmed that the RNA-seq assemblies accurately reflect gene expression predicted from prior studies. For example, Ter119⁺ cells were highly enriched for key erythroid transcripts encoding globins, heme synthetic enzymes and specialized membrane proteins. Megakaryocytes expressed high levels of gene encoding lineage-specific integrins and platelet markers. MEPs expressed numerous progenitor genes including Gata2, Kit and Myc. Thus, the RNA-seq data are of high-quality and sufficient complexity to accurately represent erythroid, megakaryocytic and MEP transcriptomes. We used a series of Unix-based bioinformatic filtering tools to identify lncRNAs that are expressed in these transcriptomes. We identified 605 “stringent” lncRNAs, and 813 “potential noncoding” transcripts. 47% of the lncRNAs are novel unannotated transcripts, validating the use of de-novo RNA-Seq in unique cell populations for lncRNA discovery. Among the 605 “stringent” lncRNAs, 103 are erythroid-restricted, 133 are meg-restricted and 280 are MEP-restricted, consistent with reports that lncRNAs exhibit exquisitely cell-type specific expression. Current efforts are aimed at generating a more comprehensive map of lncRNA expression at specific stages of erythroid and megakaryocyte/platelet development, and performing high throughput functional screens to analyze currently identified lncRNAs. Our studies are beginning to define new layers of gene regulation in normal erythro-megakaryopoiesis and are relevant to the pathophysiology of related disorders including various anemias, myeloproliferative and myelodysplastic syndromes and leukemias.