Abstract
Hematopoietic proliferation, lineage commitment, and terminal differentiation are characterized by the emergence of a cell type-specific gene expression and transcriptional programs that determine the specific phenotype and function of cells in the erythroid lineage. Our objectives in this study were to identify unique gene expression patterns that characterize the transcriptional program of normal primary human erythroid precursors during terminal differentiation, and define the gene expression patterns seen in erythroblasts (EBL) of patients with polycythemia vera (PV). Homogenous populations of primary proEBL were generated from purified liquid cultures of CD34+ cells collected from healthy volunteers and PV patients. All patients with PV were diagnosed based on established criteria and had the JAK2-V617F mutation. Morphologic examination and surface expression of CD71 confirmed the purity of proEBL cell populations. ProEBL from normal individuals were induced to terminally differentiate generating orthochromatic EBL. RNA was extracted from normal proEBL, PV proEBL, and normal orthochromatic EBL. Affymetrix U133 Plus 2.0 arrays representing approximately 39,000 human genes were used for gene expression analysis. Four replicates from four independent primary cell cultures were analyzed for each comparison group (e.g. undifferentiated proEBL versus terminally differentiated orthochromatic EBL). Unsupervised hierarchical clustering showed distinct gene expression profiles in the proEBL and terminally differentiated EBL lineages. 1109 genes (2.0 fold change, P<0.01) were found to be differentially expressed. Numerous erythroid genes were found to be upregulated during terminal differentiation [e.g. globin genes, erythropoietin receptor, heme synthesis enzymes (ferrochelatase, ALAS2) erythrocyte membrane proteins (band 3, ankyrin, protein 4.1) and transcription factors (NFE2, Kruppel-like factors, myb, GATA2)]. As a proof of validation, the differential expression of 7 genes was verified by Northern blotting. To better understand the biologic role of the gene sets identified, using Ingenuity pathway analysis, individual genes were integrated into specific regulatory and signaling pathway networks. A total of 19 networks with significant scores (>23) were identified. Biological functions of the identified networks included RNA post-transcriptional regulation, cell cycle control, translational regulation, DNA replication and repair and cellular assembly/organization. In a proof of principle study, gene expression patterns in PV proEBL (n=6) were compared to normal proEBL (n=5). Unsupervised hierarchical clustering showed a distinct gene expression profile for PV. A binary regression predictive model was also developed to find gene expression patterns predictive for PV. Using this model a 150 gene predictor was found that could predict PV patients from control at 100% accuracy. Ingenuity pathways analysis of a subset of gene subsets demonstrated several biologically relevant networks that were distinct in patients with PV, including myc, CDC2, and JAK2. Deregulation of normal transcriptional mechanisms in hematopoietic cells is associated with the pathogenesis of PV. Further, our data shows that genomic studies provide new insights into transcriptional programs that govern erythroid differentiation, and identify biologically relevant deregulated pathways as potential targets for therapy in PV.
Hierarchical clustering of gene expression patterns in the Eomes + lineage of excitatory neurons during early neocortical development
