We used a culture model

We used a culture model of human erythropoiesis that recapitulates all developmental stages [24] to screen a focused set of lncRNAs for potential contribution to this program. We found induction of lncRNA Fas-antisense 1 (Fas-AS1 or Saf) during RBC maturation and demonstrate that essential erythroid transcription factors GATA-1 and KLF1 drive Saf expression. Conversely, NF-κB, which is known to play important roles during hematopoiesis, suppresses Saf transcription. Saf was previously shown to participate in splicing of Fas pre-mRNA [25], and we recently demonstrated that Saf interacts with Fas pre-mRNA and human splicing factor 45 (SPF45) to facilitate splicing and production of a soluble Fas protein that protects qx 314 against Fas-mediated apoptosis [26]. Here, overexpression of Saf in erythroblasts derived from CD34+ hematopoietic stem/progenitor cells of healthy donors reduced surface levels of Fas receptor and protected cells from Fas-mediated cell death signals. Collectively, these results reveal a novel mechanism to modulate Fas-mediated apoptosis during human erythropoiesis by an lncRNA.
Materials and methods
Results
Discussion LncRNAs are becoming appreciated for their contribution to development [44], [45], [46]. A few studies have demonstrated that lncRNAs drive gene expression at the terminal stages of erythroid maturation in mice [21], [22], [47], but these sequences are not evolutionary conserved in humans. Here, we describe the first lncRNA (Fas-AS1 or Saf) to be induced and regulated during maturation of human RBCs through the activity of critical erythroid transcription factors and NF-κB. Furthermore, we show that Saf expression reduces surface levels of Fas and protects maturing erythroblasts from Fas-mediated cell death. These results support a functional role for this lncRNA in normal and pathological erythropoiesis. RBC development is a tightly controlled process that involves stage-specific regulatory factors to function properly (reviewed in [20], [41], [42], [43]). Efforts to characterize cytokine/hormone signaling pathways and lineage-specific transcription factors required for this critical developmental process have been particularly informative. However, these protein-coding sequences account for <2% of the human genome with the majority of transcriptional events resulting in noncoding RNAs. While evidence is emerging for a regulatory function of microRNAs in normal human erythropoiesis [48], [49], studies of lncRNAs are lacking because candidates are just beginning to be discovered [23]. Bioinformatic analyses showed enrichment for H3K27Ac marks, clustering of DNaseI hypersensitive sites, and sequence-specific binding sites to GATA-1, KLF1, and NF-κB upstream of the Saf TSS. ChIP assays confirmed binding of these transcription factors to their consensus sequences on the endogenous Saf genomic sequence. These findings are consistent with published ChIP-seq studies for GATA-1 and KLF1 that have shown chromatin co-occupancy between those two transcription factors that are critical for erythroid specific expression of target genes [50], [51]. The presence of two GATA-1 sites on Saf promoter is in accordance with the genome-wide analysis reporting the correlation between activated genes with the presence of multiple GATA-1 binding sites [50]. Mutation of GATA-1 and KLF1 binding sites in the Saf promoter confirmed their requirement for Saf transcriptional regulation. NF-κB is also important for erythropoiesis by regulating apoptosis of erythroid precursors following erythropoietin (EPO) treatment [52]. NF-κB was originally described as a transcriptional activator of pro-inflammatory genes (reviewed in [53], [54]), but recent evidence provides support for NF-κB suppression of target genes [55], [56], [57], [58]. Endogenous NF-κB activity decreased with erythroblast maturation which coincided with increasing Saf levels and induced NF-κB activity suppressed Saf transcription. The suppressive activity of NF-κB on erythroid-specific globin genes may be mediated by reduced levels of p45/NF-E2 [59]. Studies are underway to characterize the exact mechanism by which NF-κB suppresses Saf expression. Collectively, these results suggest a complex interplay of transcription factor expression and activity regulating Saf expression during erythropoiesis.