Program: General Sessions
Session: Plenary Scientific Session
To assess the effects of LRF loss on the mouse erythroid transcriptome, we performed RNA-Seq analysis using splenic erythroblasts from control and LRF conditional knockout (Zbtb7aF/F Mx1-Cre+) mice. LRF-deficient adult erythroblasts showed significant induction of Hbb-bh1, but not Hbb-y. The results were validated at the protein levels via isoelectric focusing of peripheral blood (PB) hemolysates and MALDI-TOFMS analysis. LRF loss also reactivated human fetal globin expression in vivo in LRF conditional KO mice harboring the human β-globin gene cluster as a yeast artificial chromosome transgene (β-YAC).
To determine whether LRF loss could induce HbF in human erythroid cells, we employed human CD34+ hematopoietic stem and progenitor (HSPC)-derived primary erythroblasts and determined γ-globin expression levels upon shRNA-mediated LRF knockdown (KD). HbF levels in LRF KD cells (49-70%) were much greater than those seen in parental or scrambled-shRNA transduced cells. We next employed a novel human immortalized erythroid line (HUDEP-2). This line possesses an advantage over lines currently used for globin switching studies because it expresses predominantly adult hemoglobin (HbA: α2β2), with very low background HbF expression. Using CRISPR/cas9 gene modification, we knocked out ZBTB7A in HUDEP-2 cells and performed RNA-Seq analysis. As expected, γ-globin (HBG1 and HBG2) transcripts, but not those of embryonic ε-globin (HBE1), were markedly induced in ZBTB7A KO (ZBTB7AΔ/Δ) HUDEP-2 cells. ZBTB7AΔ/Δ cells exhibited HbF levels greater than 60%, while that of parental cells was less than 3%. Notably, the HbF reactivation occurred without changes in levels of transcripts encoding known HbF repressors, including BCL11A, the principal known switching factor.
We next performed chromatin-immunoprecipitation and sequencing (ChIP-Seq) with an anti-LRF antibody using HSPC-derived proerythroblasts and HUDEP-2 cells. The most enriched motif identified in either was concordant with that previously identified in vitro using CAST analysis (Maeda et. al. Nature 2005), confirming antibody specificity. Supporting a direct role of LRF at the β-globin cluster, we observed several significant enrichment of LRF-ChIP binding signals at adult (HBB), fetal (HBG1) globin loci and the upstream hypersensitivity (HS) sites within the locus control region (LCR). ATAC-Seq (for assay for transposase-accessible chromatin with high-throughput sequencing) analysis revealed strong chromatin accessibility at the γ-globin locus in ZBTB7AΔ/Δ cells. Strikingly, differential enrichment of ATAC-signals in ZBTB7AΔ/Δ cells was evident only at the γ-locus. Thus, while LRF binds to the HBB locus and HS sites as well as to the HBG1 locus, LRF depletion specifically opens chromatin at the γ-globin locus.
Finally, to determine whether LRF and BCL11A suppress γ-globin expression via distinct mechanisms, we established LRF/BCL11A double knockout HUDEP-2 cells. Strikingly, HUDEP-2 lines lacking both LRF and BCL11A exhibited almost a complete switch in expression from adult- to fetal-type globin, suggesting that these two factors cumulatively represent the near entirety of γ-globin repressive activity in adult erythroid cells. Our findings reveal a novel molecular mechanism regulating γ-globin silencing and may open a new window for therapeutic targeting in the treatment of hemoglobinopathies.
Disclosures: Bauer: Biogen: Research Funding ; Editas Medicine: Consultancy . Orkin: Editas Inc.: Consultancy .