Abstract PR16: Targeting the chromatin architecture established by PAX3-FOXO1 in rhabdomyosarcoma

Abstract

Master transcription factors establish enhancers to regulate cell identity genes by recruiting epigenetic machinery, and are sequentially exchanged during changes in cell identity (ie, differentiation). Commonly, the fusion of transcription factors profoundly alters proper progression of cell identity, serving as the signature oncogenic event in many malignancies. The most common soft tissue cancer of childhood, rhabdomyosarcoma (RMS), is characterized by an inability to exit the proliferative myoblast-like state, presumably by blocking myogenic transcription factors from advancing the active enhancer landscape. This is achieved by either chromosomal translocation resulting in the oncogenic fusion transcription factor PAX3/7-FOXO1 (Fusion-Positive alveolar subtype, FP-RMS) or mutations in the tyrosine kinase/RAS/PIK3C axis (Fusion-Negative embryonal subtype, FN-RMS). Patients who harbor a PAX-fusion typically relapse despite aggressive therapy and have very poor survival. Here we hypothesized that the PAX3-FOXO1 fusion gene causes epigenetic reprogramming resulting in increased proliferation and a failure to terminally differentiate. Furthermore we hypothesized that disrupting the epigenetic machinery recruited by this fusion gene would provide a tractable target for therapy. We mapped the landscape of epigenetic alterations caused by the PAX3-FOXO1 fusion gene using a combination of RNA-seq, DNase hypersensitivity, and ChIP-seq against histone marks and transcription factors in cell lines and models of FP-RMS. We found high expression of several master transcription factors (including MYOD1, MYOG, MYCN, and SOX8) in FP-RMS primary tumors and cell lines, resembling human skeletal muscle myoblasts. ChIP-seq revealed that PAX3-FOXO1 is exclusively bound to distal, active enhancers and the histone modification most enriched surrounding PAX3-FOXO1 was acetylated H3K27. Furthermore we found that the introduction of the fusion gene into fibroblast cells opened up the chromatin at these same sites, completely rewiring the active enhancer landscape, recapitulating a transcriptome locked in a myoblast-like state. Genome-wide profiling of MYOD1, MYOG and MYCN reveals that all three master regulators collaborative bind at nearly every PAX3-FOXO1 driven super enhancer (SE), while typical enhancers (TEs) rarely have more than two of these four. PAX3-FOXO1 has a 7-fold preference for SEs over TEs. We also find that PAX3-FOXO1 bound, myogenic enhancers are decommissioned throughout normal skeletal muscle differentiation. To identify small molecules that would inhibit the PAX3-FOXO1 induced epigenetic machinery we treated a panel of FP-RMS cell lines with 1912 targeted agents and chemical probes at multiple concentrations and measured cell viability. Classes of molecules selectively potent for PAX3-FOXO1 driven cells (as compared to normal fibroblasts) hit connected biologically relevant targets including SE controlled receptor tyrosine kinases (including FGFR4, IGF1R, ALK), and transcriptional cofactors involved in SE complexes (including HDACs and BRD). In an expanded panel of RMS cell lines we confirmed that FP-RMS is selectively sensitive to the BET bromodomain inhibitors with the most potent being JQ1. These inhibitors selectively suppress PAX3-FOXO1 dependent transcription as measured by reporter assays and RNA-seq analysis. Indeed, coactivators of looped chromatin p300, MED1 and BRD4 excessively co-localize with PAX3-FOXO1 genome wide. In vivo, JQ1 selectively ablated PAX3-FOXO1 dependent SE driven transcription, and significantly delayed tumor progression in xenografts of PAX3-FOXO1 driven cell lines. In conclusion we found that PAX3-FOXO1 establishes myogenic super enhancers that are sensitive to BET bromodomain inhibition which constitutes a novel therapeutic strategy for children with PAX-fusion driven rhabdomyosarcoma. This abstract is also presented as Poster A16. Citation Format: Berkley E. Gryder, Marielle E. Yohe, Jack Shern, Hsien-Chao Chou, Young Song, Rajesh Patidar, Sam Li, Sivasish Sindiri, Abigail Cleveland, Hongling Liao, Xinyu Wen, Xiaohu Zhang, Lesley Mathews-Griner, Rajarshi Guha, Paul Shinn, Marc Ferrer, Scott Martin, Madhu Lal, Craig Thomas, Javed Khan. Targeting the chromatin architecture established by PAX3-FOXO1 in rhabdomyosarcoma. [abstract]. In: Proceedings of the AACR Special Conference on Advances in Pediatric Cancer Research: From Mechanisms and Models to Treatment and Survivorship; 2015 Nov 9-12; Fort Lauderdale, FL. Philadelphia (PA): AACR; Cancer Res 2016;76(5 Suppl):Abstract nr PR16.

The PAXFOXO1s trigger fast trans-differentiation of chick embryonic neural cells into alveolar rhabdomyosarcoma with tissue invasive properties limited by S phase entry inhibition

“…To refine the list of genes that define this identity [9], we combined and re-analysed microarray-based tumour transcriptomes obtained from 133 PAXFOXO1 positive ARMS patients and 59 patients affected by other RMS subtypes (Fig 2A, S1 Table, S1A Fig) [42][43][44][45][46]. We identified 1194 genes enriched in ARMS biopsies; 40% of which were in the vicinity of previously identified PAX3FOXO1 bound cis-regulatory modules (CRM) [11,12] (Fig 2B). This list of genes largely comprised previously identified PAX3FOXO1 dependent ARMS markers, such as ALK, ARHGAP25, or FGFR4 [9,47].…”

“…Genes upregulated in ARMS versus ERMS are squared in red and named ARMS signature genes. (B) Percentage of genes present nearby at least one known PAX3FOXO1 bound CRM [12] out of those present in our complete transcriptomic data set (all genes), non-differentially regulated between ARMS and other RMS (non DEG), the differentially expressed genes between ARMS and other RMS (DEG), downregulated in ARMS compared to other RMS (DOWN in ARMS), upregulated in ARMS compared to other RMS (UP in ARMS). (C) Gene ontology enrichment for biological processes terms linked to tissue specification applied to genes enriched in ARMS biopsies.…”

“…A large body of work, mainly focused on PAX3FOXO1 and aimed at identifying and functionally characterizing PAXFOXO1's target genes, argues the cell fate change characteristic of ARMS is driven by PAXFOXO1s [8,9]. This is hypothesised to arise from PAXFOXO1's strong transcriptional transactivation potential, which surpasses that of normal PAX3 and PAX7 [10][11][12][13]. PAX3FOXO1 binds to non-coding cis-regulatory genomic modules (CRMs), remodelling chromatin and enhancing transcriptional activity [11,12].…”

“…This is hypothesised to arise from PAXFOXO1's strong transcriptional transactivation potential, which surpasses that of normal PAX3 and PAX7 [10][11][12][13]. PAX3FOXO1 binds to non-coding cis-regulatory genomic modules (CRMs), remodelling chromatin and enhancing transcriptional activity [11,12]. These CRMs regulate the expression of genes associated with at least 3 traits deleterious for patients' health [8,9,11,12,14].…”

“…PAX3FOXO1 binds to non-coding cis-regulatory genomic modules (CRMs), remodelling chromatin and enhancing transcriptional activity [11,12]. These CRMs regulate the expression of genes associated with at least 3 traits deleterious for patients' health [8,9,11,12,14]. First, several of the target genes encode cell surface proteins required for ARMS cells migration [15][16][17][18][19].…”