RUNX2 Inhibition Disrupts a PAX3::FOXO1-RUNX2 Feed-Forward Loop and Dismantles Oncogenic Gene Programs in Fusion-Positive Rhabdomyosarcoma

Abstract

Globally, hundreds of thousands of children will develop cancer each year, and in the United States, childhood cancer is the leading cause of death by disease. Despite decades of research and the development of cures in specific cancer subtypes, others such as soft tissue sarcomas rely on standard chemotherapy regimens that date back to the mid-twentieth century. Current World Health Organization criteria subclassify childhood sarcomas by histopathologic and molecular features to determine cancer grade and design treatment regimens. Some bone and soft tissue sarcomas are characterized by the presence of a signature transcription-factor based fusion oncogene which is not currently druggable. Alternatively, other non-fusion-driven sarcomas are characterized by the presence or absence of point mutations, many of which have led to the development of novel therapies for children. To identify novel druggable proteins in fusion-driven sarcomas, we must rigorously characterize patient tumors and cell lines to better understand the relationship between fusion drivers and the genomes that they govern. Rhabdomyosarcoma (RMS) is the most common soft tissue sarcoma of childhood and adolescence, characterized by the expression of skeletal muscle markers. RMS is subclassified into four types based on histopathologic features: embryonal RMS (ERMS), alveolar RMS (ARMS), spindle cell/sclerosing RMS (SCRMS), and pleomorphic RMS (PRMS). ERMS and ARMS are the most common, and now classified molecularly based on the absence (fusion-negative RMS; FN-RMS) or the presence (fusion-positive RMS; FP-RMS) of the PAX3/7::FOXO1 fusion oncogene. FP-RMS represents a disproportionally high percentage of high-risk RMS cases. Despite its role as an oncogenic driver, the PAX3::FOXO1 oncoprotein is refractory to therapeutic targeting due to its lack of catalytic activity, and its intrinsically disordered structure, which leaves few pockets to bind small molecules. However, the network of transcriptional programs driven by PAX3::FOXO1 may yield alternate therapeutic targets. In this study, we take a broad view of RMS and FP-RMS to nominate additional members of this network that may be more amenable to clinical targeting. We identify a PAX3::FOXO1 feed-forward loop whereby RUNX2 and PAX3::FOXO1 reciprocally regulate one another to drive transcription of downstream PAX3::FOXO1 targets, indicating that RUNX2 is a promising FP-RMS therapeutic target. Using messenger (mRNA) sequencing data of primary patient tumor tissue, we compared the four histopathologic types of RMS. With the DepMap resource we prioritized the transcription factors participating in this network. At the top of our list emerged RUNX2, a transcription factor previously identified as a PAX3::FOXO1 downstream target and part of the PAX3::FOXO1 interactome, but which has not been evaluated as a therapeutic target in FP-RMS. We investigated the phenotypic consequences of RUNX2 inhibition in FP-RMS, the mechanism by which RUNX2 promotes downstream PAX3::FOXO1 signaling, and evaluated the consequences of RUNX2 inhibition in vivo. Using both genetic and pharmacological inhibition of RUNX2 in vitro and in vivo, we found that RUNX2 is essential to maintaining a proliferative FP-RMS cell state, while preventing apoptosis and terminal myogenic differentiation. Using chromatin immunoprecipitation (ChIP) sequencing data whereby PAX3::FOXO1 was inhibited with RNAi, we found that PAX3::FOXO1 binds a RUNX2 enhancer to upregulate gene expression alongside MYOD1 and p300. Follow-up RNAi experiments inhibiting RUNX2 expression followed by mRNA sequencing and protein-based studies found that RUNX2 supports the expression of PAX3::FOXO1-associated enhancer networks, myogenic signatures, and PAX3::FOXO1 itself at the mRNA and protein level. In summary, our findings suggest that directly inhibiting RUNX2 phenocopies PAX3::FOXO1 suppression and highlight its role as a legitimate druggable driver of oncogenic phenotypes in FP-RMS. Future studies to degrade RUNX2 protein paired with HiBiT drug screening will identify additional compounds capable of eliminating this newly characterized FP-RMS driver.