HEXIM1 Overexpression Alters GATA1 Activity and Promotes Fetal Globin Expression

Erythropoiesis is a highly organized process whereby multipotent HSCs become committed to the erythroid linage and mature into enucleate erythrocytes. This process is associated with significant changes in gene expression as non-erythroid genes are switched off while expression of erythroid genes is maintained. HEXIM1 is a key regulator of RNA Polymerase II (RNPII) activity that is essential for erythropoiesis. Following transcription initiation, RNPII transcribes 30-50 base pairs before pausing. Phosphorylation of the c-terminal domain of RNPII by the dimer pTEFb (positive transcription elongation factor beta) is necessary for RNPII to resume active transcription elongation. HEXIM1 associates with pTEFb in the 7SK complex and renders it inactive. Phosphorylation of HEXIM1 releases pTEFb from the 7SK complex, allowing it to act on RNPII. As 7SK can be targeted to specific sequences, HEXIM1/7SK mediated pTEFb delivery is a crucial mechanism to facilitate on-site pTEFb activation and RNPII pause release.

Recent work from our lab demonstrated that HEXIM1 is essential in the regulation of erythroid proliferation, viability and gene expression, and also promotes the expression of GATA1 target genes (Murphy Blood 2021). New analyses of HEXIM1 overexpression (OE) in HUDEP2 cells revealed a significant increase of gamma globin production, with a higher proportion of cells expressing fetal hemoglobin. These changes were accompanied by downregulation of beta globin and multiple key repressors of fetal hemoglobin, including BCL11A, MYB, and SOX6. These phenotypes were dependent on pTEFb activity, as the pTEFb (CDK9) inhibitor NVP-2 rescued the proliferative advantage granted by HEXIM1 OE and greatly decreased both gamma globin levels and the portion of fetal hemoglobin expressing cells. Further suggesting the effects of HEXIM1 OE are pTEFb-dependent, overexpression of HEXIM1 with a tyrosine to alanine mutation (Y271A) that prevents phosphorylation of HEXIM1 and subsequent release of pTEFb cannot enhance proliferation or gamma globin production.

To gain insights into the mechanisms underlying the increased fetal globin expression, we performed CUT&RUN assays for Ser5 phosphorylated and total RNPII in HEXIM1 OE and control cells. Surprisingly, OE of HEXIM1 had only modest impacts on RNPII pausing, but instead lead to changes in the genes occupied by RNPII. Notably, the gene that lost the most RNPII was HBB and the gene that gained the most RNPII was HBG. Of note, GATA1 motifs were highly enriched in regions that gained RNPII occupancy. CUT&RUN identified peaks of GATA1 in the HEXIM1 OE cells at regions of increased RNPII occupancy that were not present in the EV or Y271A controls. These novel GATA1 sites occurred in regions of previously closed chromatin, suggesting that HEXIM1 OE enhanced the pioneer ability of GATA1. Similar to RNPII, GATA1 occupancy shifted from the beta globin gene in cells transduced with EV or HEXIM Y271A to the gamma globin genes in cells transduced with WT HEXIM1. Novel RNPII and GATA1 peaks were also identified in other genes associated with a fetal erythroid phenotype, most notably ARID3A and LIN28B. Together, these results support a model where elevated HEXIM1 levels promote the pioneer activity of GATA1, and facilitate the expression of a fetal transcriptional program in adult cells in a pTEFb-dependent manner.

No relevant conflicts of interest to declare.