Abstract 5749: Core transcriptional regulatory circuitry in AML

Abstract

Abstract Cell states are established by small sets of lineage-restricted transcription factors (TFs). A generally accepted model posits that core regulatory TFs positively regulate their own and each other’s genes, forming a network of interconnected feed-forward loops termed core regulatory circuitry (CRC). Here, we sought to test the CRC concept by defining the direct gene-regulatory programs of the critical oncogenic TFs in acute myeloid leukemia (AML). We employed a targeted protein degradation (dTag) strategy with a homozygous knock-in of the FKBP12F36V-coding DNA sequence into the endogenous TF-coding loci of 7 core myeloid TFs and 1 cofactor (MYB, PU.1/SPI1, GFI1, RUNX1, RUNX2, MEF2D, IRF8, IRF2BP2), all of which represent selective AML dependencies and display the canonical pattern of feed-forward CRC binding. Following complete degradation of TFs for 1-2 hours, we measured the genome-wide rates of nascent mRNA synthesis by SLAM-seq. Genes whose transcription rates change significantly and immediately after a TF’s degradation represent its direct transcriptional targets. Strikingly, each TF directly regulates only between 100-450 target genes. Furthermore, rather than forming a core regulatory circuit with fully interconnected feed-forward loops, the core TFs form a very sparsely interconnected hierarchy with a cascading regulatory structure. Each evaluated TF is capable of both direct gene activation and repression. For instance, MYB directly inhibits 135 genes, most of which are associated with myeloid differentiation and inflammation. Feed-forward loops appear to be much less common than previously anticipated. Indeed, we detected only 336 feed-forward loops formed by any combination of the 8 TFs genome-wide. GFI1 and IRF2BP2 negatively regulate their own expression, and the rest of the TFs show no evidence of direct self-regulation. The entire regulatory network converges on MYC and immune signaling. A time-course SLAM-seq experiment after MYB degradation highlighted the secondary nature of eventual global transcriptional collapse and revealed bi-phasic kinetic behavior of feed-forward loops that MYB forms with GFI1 and IRF2BP2. A pseudo-steady state of cellular transcription, ensuing 8-12 hours after MYB degradation, accurately predicts MYB-dependent gene expression patterns in human AML samples. Indeed, a much higher expression of inflammatory mediators and cytokines is observed in patients with low MYB expression, consistent with the observation that MYB directly regulates inflammatory pathways and potentially explaining the known survival benefit of high MYB expression in human AML. Accordingly, MYB degradation results in immediate release of the pro-inflammatory cytokines TNF and CCL2. Our data represent the first example of systematic elucidation of direct regulatory functions of multiple core regulatory TFs in a single cellular context and support a new model of core regulatory circuitry organization in human cells. Citation Format: Taku Harada, Monika Perez, Jeremie Kalfon, Kenneth Eagle, Flora Dievenich Braes, Rashad Batley, Kimberly Stegmaier, Stuart Orkin, Maxim Pimkin. Core transcriptional regulatory circuitry in AML. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5749.