The Role of Mediator Kinases in Blood Development and Hematological Malignancies

Abstract

Throughout the past decade, our understanding of the shared molecular features of development and cancer has evolved. Both processes rely on epigenetic alterations, in concert with transcriptional machinery, to support gene expression programs that drive cellular differentiation or oncogenesis. As such, transcriptional regulators are emerging therapeutic targets for cancer, especially acute myeloid leukemia (AML). My dissertation research has involved using the natural product cortistatin A (CA) as a chemical probe to elucidate the biology of transcriptional kinase CDK8 and its paralogue CDK19 in hematopoiesis and its therapeutic potential in hematological malignancies. Super-enhancers (SEs) are large clusters of enhancers with markedly high densities of transcriptional cofactors such as Mediator that drive the expression of essential cell identity and disease genes. We found that Mediator-associated kinases CDK8/19 restrain key SE-associated genes from further activation in AML cells. We characterized CA as a potent and selective inhibitor of CDK8/19 and determined that it had antileukemic activity in human cell lines and mouse models of AML. In sensitive AML cells, CA disproportionately upregulated SE-associated genes with tumor suppressor and cell identity functions. BRD4 and CDK7 inhibitors (currently in clinical trials or preclinical development for AML, respectively) have opposite effects on SEs. Thus, we found that AML cells depend on precise dosage of key cell identity genes for survival. Overall, we showed that Mediator kinase inhibition represents the first instance of SE-associated gene activation as a mechanism for therapeutic targeting of AML. Next, we characterized one of the downstream substrates of CDK8 that contributes to its anti-proliferative effect in leukemia. In AML cells driven by over-activation of the JAK-STAT pathway, we found that phosphorylation of STAT1 S727 by CDK8 maintains growth and attenuation of differentiation. Inhibition of serine phosphorylation by CA triggered upregulation of formerly STAT1-pS727-bound SE-associated cell identity genes, promoting growth arrest and differentiation. These phenotypic and transcriptional effects were distinct from inhibition of STAT tyrosine phosphorylation by ruxolitinib, an FDA-approved JAK1/2 inhibitor. CA suppressed the growth of both naïve and ruxolitinib-resistant JAK2-driven AML cell lines as well as primary patient samples, and exhibited synergy with ruxolitinib. These results suggest that Mediator kinase inhibition represents a novel therapeutic strategy to regulate JAK-STAT signaling in JAK2-driven AML. To further explore how CDK8 regulates the transcriptional activity of STAT1, we performed targeted proteomics studies to assess how the chromatin interactome of STAT1 changes in response to serine phosphorylation. We observed that RUNX1, a master myeloid transcription factor, co-localized with STAT1 more strongly when it was not serine-phosphorylated, and that STAT1 and RUNX1 co-regulated a subset of SE-associated target genes. AML and acute lymphoid leukemia cell lines with dysregulated RUNX/GATA programs were particularly sensitive to Mediator kinase inhibition. We found that CDK8 functionally restrains the RUNX/GATA program, partly by extruding RUNX1 from key genomic loci that regulate cell identity. At the mechanistic level, we found that CA treatment stimulated RUNX1 target gene expression, contributing to growth arrest. Furthermore, in a model of RUNX/GATA- dependent megakaryocytic differentiation, CA stimulated gene expression programs related to development of adult megakaryocytes. These results suggest that a significant percentage of leukemias with underlying RUNX/GATA dysregulation could be particularly amenable to therapeutic intervention with Mediator kinase inhibitors.