SIX1 Transcription Factor Enhances Human Erythropoiesis Via GATA1

Erythropoiesis is orchestrated by the coordinated action of multiple transcription factors. The master erythropoietic regulator GATA1 is itself modulated via interactions with multiple co-regulatory factors, such as FOG1, KLF1, and LMO2. Though the PAX-SIX-EYA-DACH network (PSEDN) of conserved transcription factors has been well characterized in the formation of eyes, kidney, branchial structures, and skeletal muscles, a role for PSEDN members in hematopoietic systems has only recently been recognized (Liu et al., Nature 2019 PMID:30894749). Here, we studied the PSEDN member SIX1 and discovered its ability to drive erythroid differentiation of human hematopoietic cells.

Enforced overexpression (OE) of SIX1 in human TF1 erythroleukemia cells or primary CD34⁺ hematopoietic stem-progenitor cells (HSPCs) stimulated the generation of erythroid cells, as determined by increased numbers of cells expressing erythroid-selective surface markers (CD235a^hiCD71^hiCD34^-) and hemoglobin (HBB). Conversely, SIX1 knockout in TF1 cells or primary HSPCs reduced erythroid cell generation in response to erythropoietin (EPO). SIX1 OE could also stimulate TF1 cell erythroid differentiation in the absence of EPO. Further analysis of SIX1 OE in TF1 cells revealed that SIX1 stimulated the expression of multiple functionally important erythroid molecules including ALAS2, SLC4a1, EPOR, SPTA1, KLF1 and ANK1. By gene set enrichment analysis (GSEA) of global RNA-seq data, SIX1 OE stimulated heme metabolism genes as well as many genes known to be regulated by GATA1, including FOG1-dependent and -independent genes. SIX1 OE reduced GATA2 and increased GATA1 protein and RNA expression, resembling GATA switching downstream of EPO signaling. To determine whether GATA1 was necessary for SIX1 to stimulate erythropoiesis, we generated GATA1 knockout cells using CRISPR/Cas9 technology. In contrast to control cells, SIX1 OE in GATA1 knockout cells failed to stimulate erythropoiesis, indicating that SIX1 stimulation of erythropoiesis requires GATA1.

To gain further insight into the mechanism by which SIX1 stimulates erythropoiesis, the promiscuous biotin ligase, BirA (Choi-Rhee et al., Protein Sci. 2004 PMID:15459338), was fused in-frame to SIX1 to determine the SIX1 proximal interactome. Streptavidin-enrichment of biotinylated proteins in SIX1-BirA OE lysates revealed GATA1 and FOG1 as proximal interactors of SIX1-BirA, but not of BirA alone. When co-expressed in HEK293T cells GATA1 and SIX1 were found to co-immunoprecipitate, suggesting the two proteins can physically interact in a complex. We demonstrated the functional consequence of the SIX1 interaction with GATA1 using a GATA1-dependent luciferase reporter gene harboring three copies of GATA binding sites. Cells in which SIX1 and GATA1 were co-expressed exhibited significantly higher levels of luciferase expression compared to cells expressing only GATA1, suggesting SIX1 could stimulate GATA1-dependent transcription. Introduction of mutations in SIX known to cause Branchio-Oto-Renal (BOR) (Ruf et al., PNAS 2004 PMID:15141091) syndrome did not inhibit the ability of SIX to bind GATA1 nor its ability to drive erythropoiesis. Taken together our results suggest that SIX1 can stimulate erythropoiesis via multiple mechanisms, including increased GATA1 expression and function. Our findings provide the first demonstration of a role for the PSEDN in erythropoiesis and reveal unknown physical and functional interactions between two central developmental transcriptional networks (GATA:FOG network and PSEDN).