Yong Yean Kim, Robert Hawley, Berkley Gryder, Silvia Pomella, Josh Kowalczyk, Ranu Sinniah, Young Song, and Javed Khan

• Submitter: Yong Yean Kim (Postdoc or fellow)
• email: yong.kim@nih.gov

BACKGROUND: The PAX3/7-FOXO1 fusion gene is the major oncogenic driver in fusion positive rhabdomyosarcoma (FP-RMS), which is a highly aggressive soft tissue muscle sarcoma of childhood. The chimeric gene results from either t(2;13) or t(1;13) translocation and has been shown by our group to drive FP-RMS oncogenesis through transcription activation driven by super enhancers. Thus, this fusion gene represents a unique vulnerability in FP-RMS and direct or indirect targeting of the fusion gene by small molecules is actively investigated in our laboratory.

MATERIALS AND METHODS: Novel luciferase assays were developed for monitoring PAX3-FOXO1 super enhancer and general transcription activity. Using this methodology, a small molecule library of 62,643 compounds were screened to identify selective inhibitors of PAX3-FOXO1 activity. RNA-seq was performed on FP-RMS cell lines treated with the top hits and gene set enrichment analysis (GSEA) was performed to assess the molecular effects of the inhibitors. A top small molecule of unknown mechanism of action, PFI-63 and an analogue PFI-90, were selected for further characterization. Western analysis was performed for MYOG, PARP, and PAX3-FOXO1. Apoptosis was also assessed using Annexin V and Caspase Glow assays. In vitro direct enzyme inhibition analysis by lead compounds were performed using full length proteins. Ligand observed NMR techniques were used to determine biophysical binding of compound to KDM3B. ChIP-seq analysis was performed using antibodies against H3K4me3, H3K9me2, H3K27me3, H3K27ac, and PAX3-FOXO1. Mouse xenograft model of FP-RMS was used to determine in vivo efficacy of lead compound PFI-90.

RESULTS: Using a PAX3-FOXO1 selective cell-based reporter assay, 64 top hits were identified that inhibited PAX3-FOXO1 activity without a general inhibition of transcription or cell death at 24 hours. PFI-63 and a more water-soluble analogue, PFI-90, were identified for further characterization. RNA-seq and computational docking modeling indicated that the compounds had histone demethylase inhibitory activity. Analysis by RNA-seq showed activation of apoptosis and myogenic differentiation pathways while PAX3-FOXO1 gene sets were repressed. Activation of apoptosis and myogenic differentiation was validated by Western analysis for PARP cleavage and increase expression of MYOG respectively. In vitro enzymatic inhibition assay confirmed activity against multiple KDM family members with highest specificity to KDM3B. Western analysis for methylation of histone 3 lysines at position K27, K4, and K9 showed increase after treatment. Due to poor solubility of PFI-63, we performed a similarity search of PFI-63 and screened additional compounds leading to the identification of PFI-90. PFI-90 showed a superior inhibition of KDM family of proteins while maintaining highest inhibition of KDM3B by in vitro enzymatic assay. ChIP-seq analysis validated increase in methylation of histone 3 lysines at position K4 and K9 while acetylation at K27 and PAX3-FOXO1 was unchanged. In an in vivo xenograft model of FP-RMS, PFI-90 treatment group has smaller tumors compared to DMSO control.

CONCLUSION AND FUTURE DIRECTION: We have identified first in class KDM3B selective inhibitors which have activity against FP-RMS in vitro. Our initial pre-clinical validation via FP-RMS xenograft model showed PFI-90 was able to decrease tumor size. We are currently investigating the exact mechanism of how inhibition of KDMs results in disruption of the PAX3-FOXO1 downstream effect. Thus, we describe here novel inhibitors of KDM3B that results in inhibition of PAX3-FOXO1 activity and represents a potential new therapy for FP-RMS and other transcriptionally driven cancers.