Elucidating the Functions of CDK8 and CDK19 in Acute Myeloid Leukemia

Abstract

Transcriptional dysregulation is an important cause of acute myeloid leukemia (AML). The majority of mutations in AML reside in genes that encode transcriptional regulators including transcription factors, chromatin modifiers, DNA methylation proteins, and cohesin. Inhibition of these and other transcriptional regulators has emerged as a promising therapeutic strategy in AML. This dissertation describes our development of marine natural product cortistatin A (CA) as a potent and selective inhibitor of CDK8 and CDK19 (CDK8/19)—two kinases that reversibly associate with a general transcriptional co-activator complex called Mediator. We established that CDK8/19 are therapeutic targets in AML, as their inhibition suppresses AML growth both in cell culture and in animal models. Using CA, we also discovered previously unknown functions of CDK8/19 in restricting the expression of super-enhancer-associated genes in AML. CDK8/19 inhibition leads to the upregulation of these genes, which are in part responsible for the antiproliferative activity of CA. The effect of CA on super-enhancer activity is opposite to that of BRD4 or CDK7 inhibitors (which downregulate super-enhancer-associated genes), yet all of these inhibitors have equivalent efficacy in AML. This observation suggests that AML exerts tight control over the expression of super-enhancer-associated genes, and perturbations to the expression of these genes in either direction are sufficient to suppress AML growth. To further understand the antiproliferative mechanism of CDK8/19 inhibition in AML cells, we performed a genome-wide CRISPR-Cas9 suppressor screen with CA in an AML cell line, and we found that Notch signaling is essential for CA’s activity. CA inhibits CDK8/19-mediated degradation of the Notch1 and Notch2 intracellular domains (NICD1/2), thereby activating Notch signaling which suppresses AML growth. We also found that expression of Notch ligands Dll1 and Jagged1 is an important pre-requisite for the cell’s response to CA along the Notch axis. Lastly, we identified >60 previously unknown substrates of CDK8/19 through a phosphoproteomics experiment in combination with CA treatment. Most of these substrates associate either directly or indirectly with the chromatin—an observation that is consistent with the transcriptional role of CDK8/19—and span many functional annotations including DNA repair, transcription factors, and RNA polymerase II. Our findings have not only expanded the field’s understanding of CDK8/19 and their roles in transcriptional regulation, cell biology, and cancer, but also opened up new avenues for therapeutic developments of CDK8/19 inhibitors for the treatment of AML and other diseases.