HIC2 Controls Developmental Hemoglobin Switching By Repressing BCL11A Transcription

One of the oldest and most deeply studied problems in developmental gene expression is the switch from fetal to adult type hemoglobin production in red blood cell precursors. Interest in this question has been fueled by its relevance to genetic blood disorders such as sickle cell disease (SCD) and thalassemia. BCL11A is a transcriptional repressor that is thought to directly silence the fetal β-type globin (HBG1/2) genes in adult erythroid cells. Transcriptome and RNA polymerase II profiling indicate that the BCL11A gene is transcribed considerably more highly in adult erythroblasts compared to fetal cells, accounting in large part for corresponding changes in BCL11A protein levels. Yet, the mechanism governing BCL11A developmental regulation is still unclear.

To identify novel regulators of the fetal-to-adult globin switch, we interrogated our recent CRISPR based genetic screens that employed single guide RNAs (sgRNAs) targeting transcription factors (Huang et al., Blood, 2020) and uncovered HIC2, a penta-dactyl zinc finger DNA binding protein bearing a BTB/POZ domain as a novel regulator of hemoglobin switching.

HIC2 is expressed more highly in fetal erythroblasts compared to adult cells, a pattern inverse to that of BCL11A. Overexpression (OE) of HIC2 in the adult type erythroid HUDEP2 cell line stimulated the expression of 322 genes while impairing that of 224 genes (FDR < 0.01 and fold change ≥ 2). The most highly upregulated genes (>150-fold) were HBG1/2. Upregulation was accompanied by gains in chromatin accessibility and histone H3K27acetylation of HBG1/2, and increased chromatin contacts between the distal globin gene enhancer (LCR) and the HBG1/2 genes. Overexpression of HIC2 in primary human erythroblasts also significantly increased HBG1/2 mRNA and protein levels, sufficient to reduce cell sickling in SCD patient-derived erythroid cells.

HIC2 OE lowered BCL11A mature and pre-mRNA production, indicating that HIC2 attenuates BCL11A transcription. Forced expression of BCL11A restored HBG1/2 silencing in HIC2 OE HUDEP2 cells, suggesting that BCL11A repression accounts for the effects of HIC2 on fetal globin genes. ChIP-seq revealed a strong HIC2 binding peak at the erythroid BCL11A +55 enhancer. HIC2 OE reduced chromatin accessibility and H3K27acetylation of the +55 enhancer, as well as the enhancer-promoter contacts, suggesting that HIC2 directly decommissions the enhancer to attenuate BCL11A transcription.

The BCL11A +55 enhancer contains two consensus HIC2 binding motifs under the HIC2 peak adjacent to GATA:E-box and GATA motifs. CRISPR-mediated mutagenesis of both HIC2 motifs raised BCL11A basal level transcription and diminished the ability of overexpressed HIC2 to repress BCL11A transcription. Notably, HIC2 OE impaired binding of transcription factor GATA1 at the +55 enhancer, suggesting that this enhancer is under developmental control. Indeed, GATA1 binding and chromatin accessibility of +55 enhancer were virtually undetectable in HUDEP1 cells, which represent a more fetal-like state. CRISPR-mediated depletion of HIC2 in HUDEP1 cells reversed this pattern with gains in GATA1 binding, chromatin accessibility, and BCL11A transcription.

In sum, HIC2 emerges as a critical regulator of hemoglobin switching that operates by imposing developmental stage-specific control onto a BCL11A transcriptional enhancer.