Differential Re-Programming of Myeloid Stem Cell Development by Activator Versus Repressor Isoforms of Human CCAAT/Enhancer Binding Protein Epsilon (C/EBPε).

Abstract

Human C/EBPε is expressed as four isoforms (32, 30, 27, 14kD) through alternative promoter usage, RNA splicing, and translational start sites. We are studying the C/EBPε isoforms in vivo in cord blood (CB) CD34+ progenitors to define their roles in granulocyte differentiation and hematopoietic lineage specification. The C/EBPε32/30 isoforms function as transcriptional activators. We reported that C/EBPε27 interacts with and is a repressor of GATA-1 transactivation of eosinophil-specific genes (Du J et al. JBC 2002; 277:43481). C/EBPε14, which contains DNA binding and bZIP domains but lacks a transactivation domain, may function as a dominant negative repressor of C/EBPε32,30 or other C/EBPs through heterodimerization or competition for C/EBP binding sites. We assessed mRNA expression of the C/EBPε isoforms during eosinophil differentiation of CB CD34+ progenitors to determine their temporal patterns of expression. CD34+ progenitors (>95% pure) induced to differentiate into eosinophils by SCF, IL-3, and IL-5 showed differentiation to >90% eosinophils by 17–19 days. Semi-quantitative RT-PCR using C/EBPε isoform-specific primers showed that CD34+ progenitors initially express only the repressor C/EBPε14 isoform, with induction of all isoforms by day 3 to different peak levels of expression by days 9–11 during the promyelocyte to myelocyte transition. These results define temporal changes in the expression ratios of the C/EBPε activator vs. repressor isoforms during eosinophil development that may differentially regulate gene transcription in this process. We subcloned the C/EBPε isoform cDNAs into an MSCV-based bicistronic retroviral vector (pGCDNsam IRES-EGFP) and ectopic expression was induced in CB CD34+ progenitors by retroviral transduction for 72hrs. The CD34+/GFP+ cells were sorted by FACS and plated in Collagen Cult™ media containing SCF, IL-3 and a lineage-specific cytokine (i.e. EPO, G-CSF, or IL-5), or in suspension culture containing SCF, IL-3 and IL-5 to drive eosinophil differentiation. Total and differential colony and cell counts were performed after 15–17 days based on colony morphology, histochemical and enzyme staining of the cells. The activator C/EBPε32 isoform significantly altered myeloid development, favoring eosinophil over neutrophil or erythroid development. Even in cultures containing EPO, cells transduced with C/EBPε32 failed to develop into erythroid colonies (BFU-E). C/EBPε27, a potent repressor of GATA-1 in vitro, inhibited erythroid colony growth by ~50%, and doubled the numbers of granulocyte-macrophage colonies compared to C/EBPε14 or empty vector control. C/EBPε32 strongly induced eosinophil colony formation at the expense of neutrophil other myeloid lineages regardless of the cytokines used, inducing ~90% eosinophil colonies. C/EBPε27 reduced eosinophil colonies by >50%. Likewise, >90% of cells transduced with C/EBPε32 and grown in suspension culture were eosinophils, whereas C/EBPε27 and C/EBPε14 inhibited eosinophil differentiation by ~50% and >98%, respectively. Thus, the C/EBPε isoforms: (1) are differentially expressed during eosinophil development, (2) have the capacity to reprogram stem cells to different myeloid lineages consistent with their predicted activator vs. repressor activities and interactions with hematopoietic transcription factors such as GATA-1 or other C/EBPs, and (3) may be useful in their ability to reprogram myeloid terminal differentiation for the development of novel approaches to treat myeloid or other leukemias.