Control of Fetal Hemoglobin Levels By NFI Transcription Factors

Sickle cell disease and some forms of b-thalassemia are disorders, which can be improved by therapies that elevate HbF, the fetal form of hemoglobin. It is currently thought that in adult type erythroid cells, silencing of the fetal type b-globin-like genes HBG1 and HBG2 is accomplished predominantly by two major transcription factors: BCL11A and ZBTB7A. However, it is unknown whether there are additional transcription factors that contribute to the repression of the HBG1/2 genes.

Here, using a DNA-binding domain focused CRISPR-Cas9 screening approach, we identified NFIA as a novel regulator of HbF expression. NFIA belongs to the NFI transcription factor family, which is composed of NFIA, NFIB, NFIC and NFIX. NFIA and NFIX are the predominantly expressed forms in human erythroid cells, and their expression is higher in adult erythroblast cells when compared to fetal erythroid cells. CRISPR-Cas9 mediated disruption of NFIA in an immortalized human umbilical cord blood-derived erythroid progenitor cell line 2 (HUDEP2) or CD34+ hematopoietic stem and progenitor cell-derived primary erythroblast cells led to a modest reactivation of HBG1/2 mRNA expression, whereas disruption of NFIX had little effect. However, combined NFIA and NFIX disruption produced a substantial increase in HBG1/2 expression, suggesting that these factors function in a partially redundant manner. ChIP-seq and RNA-seq studies showed that NFIA and NFIX have comparable chromatin binding and activity profiles in human erythroid cells. ChIP-seq failed to detect NFI protein occupancy at or near the HBG1/2 genes. However, given the known difficulty in detecting repressor molecules by ChIP at the silent HBG1/2 genes [Martyn et al., Nature Genetics 2018; Liu et al., Cell 2018], we tested a direct involvement of NFI proteins by disrupting putative NFI binding sites near the HBG1/2 genes. Perturbation of one such NFI motif residing upstream of the transcription start site in the HBG1/2 promoters both in HUDEP2 and primary human erythroid cells markedly increased HBG1/2 mRNA levels, comparable to those achieved by combined disruption of NFIA and NFIX. Mutation of another putative NFI motif within intron1 in the HBG1/2 gene also significantly raised HbF levels. While these results implicate NFI proteins in the direct silencing of the HBG1/2 genes, the identity of the bound factors at the NFI motifs remains to be established.

Studies are currently ongoing that use alternative approaches such as Cleavage Under Target & Release Using Nuclease (CUT & RUN) to map the chromatin occupancy of NFI factors at the HBG1/2 genomic region. We will also discuss results from ongoing studies of NFI factors in NBSGW mice models.

In sum, we uncovered NFI transcription factors as a novel HbF regulator suggesting that the silencing of HbF involves a transcription factor network that is more complex than previously appreciated.