The BET Proteins Brd4 and CBP Contribute to Severity of Anemia in EKLF/KLF1 Mutants By Driving Selective Release of Paused RNA Polymerase II

EKLF/Klf1 is a master transcriptional activator of critical genes that specify erythroid fate and drive terminal erythroid maturation. EKLF binds to DNA using three Zn-fingers, and heterozygous point mutations in the middle zinc finger cause severe anemia in humans (Congenital Dyserythropoietic Anemia Type IV, CDA IV/CDAN4; KLF1-E325K) and mice (Neonatal Anemia, Nan; Klf1-E339D). CDA IV is a rare disease affecting individuals at birth, and patients present with severe anemia, bone marrow dyserythropoiesis, and other dysmorphic features rendering them transfusion dependent. Homozygous Nan/Nan mutants are embryonic lethal at day E10-11, and Nan/+ heterozygotes are severely anemic at all stages of life. Both CDA IV and Nan mutations are autosomal dominant due to neomorphic defects resulting from altered DNA binding properties of mutant EKLF, which dysregulates gene expression and impairs erythropoiesis. To investigate the mechanism of gene dysregulation caused by the Nan and CDA IV mutations, we compared RNA-Seq data from differentiating EKLF-E325K mutant BEL-A proerythroblast cells, and from Nan/+ E13.5 fetal livers. Both the Nan and CDA IV mutants show altered gene expression with 376 genes ectopically expressed, and 167 genes downregulated in both CDA IV and Nan/+ revealing the similarities between the effects of the Nan and CDA IV mutations on gene expression. The ectopic genes common to CDA IV and Nan are mostly involved in immune response, inflammation, and stress responses. Thus, this suggests that ectopic expression of inflammatory genes likely contribute to the severity of anemia in CDA IV and Nan. To uncover the basis of the dominant anemia and neomorphic expression, we have used the Nan fetal liver to examine embryonic erythropoiesis. First, we found that Nan-EKLF is unable to bind to a subset of WT EKLF binding sites leading to loss of expression of important EKLF targets involved in cell cycle regulation and membrane functions. Second, we found that neomorphic defects are due to ectopic DNA binding of Nan-EKLF at an altered, less stringent binding site at intronic and intergenic locations. This leads to increased chromatin accessibility and recruitment of the BET domain protein acetylase CBP at these ectopic sites, which acetylates H3K27 and activates cryptic enhancers. This correlates with the formation of a transcription elongation-competent RNA Polymerase II complex at the transcription start sites of ectopic genes that ultimately drives untimely expression of their mRNAs. Third, we found that at some ectopic loci, RNA Pol II remains in a paused state near the TSS during normal WT erythropoiesis, and the fortuitous binding of Nan-EKLF at a nearby cryptic enhancer causes selective release of paused Pol II and ectopic mRNA expression. These high pausing ectopic sites have CBP deposition at enhancers in Nan/+ but CBP or H3K27ac levels at the TSS remains unchanged in Nan/+ compared to WT. This indicates that selective Pol II pause-release from the TSS is being driven by additional mechanisms or factors. To further clarify this, we examined the occupancy of the BET protein Brd4, which binds to acetylated histones and recruits transcription elongation factors such as P-TEFb. We find that Brd4 occupancy at both the TSS and enhancers correlates highly with gene expression and mirrors the altered gene expression in Nan/+ where Brd4 is reduced at downregulated sites and increased at ectopic sites. Strikingly, at the subset of high pausing genes, Brd4 occupancy is high at both TSS and intergenic enhancers, correlating with elongation-competent RNA Pol II. This strongly suggests that the presence of Brd4 at both TSS and enhancers is a driver of selective release of paused RNA Pol II at ectopic sites in Nan. Thus, we conclude that BET proteins such as CBP and Brd4 are central to the gene dysregulation and neomorphic defects observed in Nan/+, and we hypothesize that a similar mechanism might be occurring in CDA IV. We are currently examining the involvement of the BET proteins Brd4 and CBP in altering gene expression in CDA IV BEL-A cells by selective release of RNA Pol II, with the long-term translational goal of testing whether BET inhibitors can be used as a therapeutic measure against CDA IV.