Cytokine‐induced megakaryocytic differentiation is regulated by genome‐wide loss of a uSTAT transcriptional program

Abstract

Metazoan development is regulated by transcriptional networks, which must respond to extracellular cues including cytokines. The JAK/STAT pathway is a highly conserved regulatory module, activated by many cytokines, in which tyrosine-phosphorylated STATs (pSTATs) function as transcription factors. However, the mechanisms by which STAT activation modulates lineage-affiliated transcriptional programs are unclear. We demonstrate that in the absence of thrombopoietin (TPO), tyrosine-unphosphorylated STAT5 (uSTAT5) is present in the nucleus where it colocalizes with CTCF and represses a megakaryocytic transcriptional program. TPO-mediated phosphorylation of STAT5 triggers its genome-wide relocation to STAT consensus sites with two distinct transcriptional consequences, loss of a uSTAT5 program that restrains megakaryocytic differentiation and activation of a canonical pSTAT5-driven program which includes regulators of apoptosis and proliferation. Transcriptional repression by uSTAT5 reflects restricted access of the megakaryocytic transcription factor ERG to target genes. These results identify a previously unrecognized mechanism of cytokinemediated differentiation.

Thrombopoietin signaling to chromatin elicits rapid and pervasive epigenome remodeling within poised chromatin architectures

“…knockdown not only upregulated megakaryocytic-affiliated genes, but it also repressed lineage-inappropriate genes including lymphoidand pregranulocyte/monocyte-affiliated genes (Park et al 2015). This result, along with our feature importance analysis (Fig.…”

“…However, megakaryocytic-affiliated genes were not induced within 30 minutes of TPO (Supplemental Fig. S1A-D and S2), consistent with a slower induction kinetic (Park et al 2015). In contrast, genes involved in mitosis, chromosome maintenance and DNA repair were rapidly repressed by TPO ( Fig.…”

“…To this end, we considered activated and repressed cis-regulatory elements and trained least absolute shrinkage and selection operator (lasso) logistic regression models (Tibshirani 1996) using different sets of features (Fig. 6D) and after (n = 3) 30 minutes TPO stimulation (generated as part of this study or previously published (Wilson et al 2010(Wilson et al , 2016Park et al 2015), see Supplemental Table S1), and normalized Hi-C signals as a proxy for interaction frequencies. Model performances were evaluated on a test set consisting of differentially acetylated cisregulatory elements that were not used for learning model parameters (Fig.…”

“…We previously demonstrated that in the absence of TPO, chromatin-bound uSTAT5 restrains a megakaryocytic transcriptional program in hematopoietic stem/progenitor cells, and that TPO-mediated phosphorylation of STAT5 triggers a genome-wide relocation of STAT5 to canonical GAS motifs (Park et al 2015). Intriguingly, uSTAT5…”

“…To test this hypothesis, we took advantage of a mutant (Y699F) STAT5B that ablates TPO-induced tyrosine phosphorylation of STAT5 and does not act as dominant negative (Park et al 2015). We expressed wild-type or mutant STAT5B in HPC-7 cells, and measured H3K27ac levels at uSTAT5-bound enhancers within the Hlf, Sox4 and Cxcr4 loci (Fig.…”