The role of non-canonical Hippo signaling in PAX3-FOXO1 fusion-positive rhabdomyosarcoma

Childhood cancer persists to be the leading cause of death by disease in children in the United States. While advances in survival have been made in recent decades for certain subtypes, others have seen little to no improvement in survival outcomes. One such example is cancer of the soft tissues, called soft tissue sarcoma. As compared to adults, the incidence of sarcoma is considerably higher in children. Soft tissue sarcomas are diagnosed and treated based on a variety of histological and molecular features, such as the presence or absence of a fusion oncoprotein, or mutations in oncogenic signaling pathways. Understanding the relationship between molecular signatures of soft tissue sarcoma and clinical outcomes such as metastasis and refractory disease is necessary to develop new and effective treatments. Fusion-positive rhabdomyosarcoma (FP-RMS) is an example of a subtype of soft tissue sarcoma in children and young adults for which advances in standard of care have been limited, and outcomes remain poor. Despite decades of work aimed at understanding the significance of the chromosomal translocation that produces the pathognomonic fusion oncoprotein PAX3-FOXO1 (PF), efforts to directly target and inhibit the fusion protein have failed, in part due to the nature of the PAX3-FOXO1 as a fusion of two essential transcription factors. Therefore, identification of downstream targets of PF may reveal more actionable therapeutic opportunities. In this study, we identify an axis in which PAX3-FOXO1 transcriptionally upregulates RASSF4, which in turn binds to and inhibits MST kinase. Using a characterized genetically engineered mouse model of FP-RMS, we crossed in animals with conditional loss of MST1 and MST2 and monitored for tumor development. Animals with additive loss of MST1 and MST2 developed tumors with increased penetrance and decreased latency. Using tumor-derived cell lines, we found that loss of MST increased invasiveness and decreased the cells’ ability to differentiate down the myogenic lineage. Investigation of signaling downstream of the MST kinase in the Hippo pathway revealed that phosphorylation of LATS is unchanged as a result of this MST loss, and YAP1 and WWTR1 remain active. Instead, we identified MOB1 as a candidate effector of this MST-mediated phenotype. In preliminary studies, we confirmed that loss of phospho-MOB1 is observed in xenografts of human patient-derived FP-RMS cells. In a cell-based model of FP-RMS, in which intrinsic stemness characteristics of these cells are supported, we discuss how aberrant Notch- driven developmental signaling contributes to chemoresistance and increased tumorigenic potential. In additional studies, we developed a model of STS metastasis and in an unbiased manner, identified genes that contribute to metastatic clonal outgrowth. Understanding the common drivers of metastasis in human patients is essential to developing new interventions that slow or prevent refractory and metastatic disease. Together, these studies provide rationale for using specific model systems to investigate the features of STS that are most difficult to treat in the clinical setting. Finally, in a fourth system, we show that inhibition of developmental pathways at a defined node is an effective approach to slow growth of STS. In these fusion-negative RMS (FN-RMS) studies, we show that genetic and pharmacologic inhibition of HES1, downstream of both Hippo and Notch, impedes tumor growth and encourages myogenic differentiation. In summary, aberrant non-canonical Hippo signaling is an effector of FP-RMS driven by the PAX3-FOXO1 fusion oncoprotein. Hippo signaling, like Notch and other developmental pathways, is required for both RMS tumorigenesis and normal human development. Using sophisticated modeling systems, we can identify nodes at which these critical pathways can be pharmacologically inhibited to develop therapeutics that are effective against refractory and metastatic disease.