(P 428) COMPREHENSIVE GENOMIC PROFILING OF CIC-DUX4 SARCOMA (CDS) CELL LINES AND MODELS

Objective: Ewing-like sarcomas constitute a rare and highly aggressive subtype of soft tissue sarcoma which histologically resemble Ewing but lack the Ewing-defining EWS-ETS gene fusion. In 2020, the W.H.O. began recognizing Ewing-like sarcomas as distinct disease entities which are now defined by the specific gene rearrangement or fusion these cancers possess. Although the incidence of these new small round cell sarcomas is not known, the most common appears to be driven by a translocation at t(4;19)(q35;q13) leading to a fusion between the DNA binding domain of capicua transcriptional repressor (CIC) with the transcriptional activation domain of double homeobox 4 (DUX4). CIC-DUX4 fusion-positive sarcomas (CDS) have an unfavorable clinical course and dismal 5-year overall survival attributable to its propensity for rapid and distant metastasis and poor response to conventional chemotherapies. For this reason, novel therapeutic strategies and treatments are desperately needed.

Methods: To elucidate a mechanism of disease initiation/maintenance and identify CDS-specific vulnerabilities, we comprehensively profiled human and mouse CDS cell lines. Review of the literature and public databases identified eight cell lines harboring a CIC-DUX4 gene fusion including four primary cell lines (ECD1, NCC-CDS2, Kitra-SRS, and TE441.T) and four mouse xenograft-derived cell lines (NCC-CDS-X1C1, NCC-CDS-X3C1, PDX-CDS1, PDX-CDS3). Additionally, two primary cell lines derived from spontaneous tumors in a genetically engineered mouse model of HA-CIC-DUX4 sarcoma were included (CDS-001, CDS-108). RNA-sequencing, Assay for Transposase-Accessible Chromatin (ATAC)-sequencing, and Chromatin Immunoprecipitation (ChIP)-sequencing were performed. Immunoprecipitation reactions were prepared using antibodies against the DUX4 CTD (abcam: ab124699), HA-epitope tag (abcam: ab9110), and H3K27ac (Active Motif 39134). 150bp paired-end libraires were prepared and sequenced on an Illumina HiSeq platform. Raw reads were aligned using STAR. Differential gene expression was called using DESeq2 relative to other small round cell sarcoma cell lines including Ewing sarcoma (ES) [RD-ES, A-673] desmoplastic small round cell sarcoma (DSRCT) [BER-DSRCT, JN-DSRCT], synovial sarcoma (SS) [Yamoto-SS, Aska-SS, SYO1], fusion-negative rhabdomyosarcoma (FN-RMS) [RD, SMS-CTR], and fusion-positive rhabdomyosarcoma (FP-RMS) [Rh4, Rh28, Rh30] all of which were sequenced in parallel. ATAC-seq peaks and ChiP-seq peaks were called using macs3 and then integrated to determine the core regulatory circuity using the ROSE and Coltron computational pipelines.

Results: Principal Component Analysis (PCA) separated the 20 small round cell sarcoma cell lines into two primary clusters based on global RNA expression profiles. Cluster 1 was composed of the eight human CDS cell lines and cluster 2 was composed of all other small round cell sarcomas (‘other’) which further segregated according to type along principal component 2 (figure 1). Differential expression analysis comparing CDS to ‘other’ revealed 2,054 upregulated genes (log2FC> 1, padj < 0.01) and 2,760 downregulated genes (log2FC < 1, padj < 0.01). CIC-DUX4 ChIP-seq using an anti-DUX4 antibody and anti-HA antibody identified 1,386 and 1,043 direct downstream targets (+/- 3kb of TSS) in human and mouse respectively. 72 direct downstream targets were conserved between the two species including: ETV1, ETV4, ETV5, FOS, SHC3, SCH4, DUSP4, DUSP6, and VGF. Regions bound by CIC-DUX4 were also enriched for H3K27ac and open chromatin as evaluated by ATAC-seq consistent with the notion that CIC-DUX4 is behaving as a neomorphic transcriptional activator in CDS (figure 2). Interestingly, CIC-DUX4 bound regions in human and mouse cell lines were also strongly enriched for a de novo motif matching the consensus ‘ETS’ transcription factor binding site [GGAA]. Further analysis of Super-Enhancer associated transcription factor binding suggests that ETV1/4/5 are key mediators of the core regulatory circuitry (CRC) in CDS.

Conclusion: Identifying therapeutic vulnerabilities and novel treatment strategies necessitates a firm understanding of disease mechanisms. Because of its rare incidence, human tissues are not readily available however numerous cell lines containing the CIC-DUX4 gene fusion have been established; in some cases, unknowingly. Here we obtained these cell lines and used Next Generation Sequencing based approaches to comprehensively profile the transcriptional and chromatin landscape of CDS in human and mouse. These datasets suggest that CDS is distinct from other small round cell sarcomas including ES, DSRCT, SS, and RMS. This is underscored by its unique transcriptional profile and enrichment at gene promoters. We also found that CIC-DUX4 occupancy strongly correlates with H23K27ac, open chromatin, and transcriptional upregulation in human cell lines which is recapitulated at notable targets in tumors derived from mice genetically engineered to express the CIC-DUX4 fusion gene. Taken together, these data strongly support the notion that CIC-DUX4 represents a neomorphic transcriptional activator. It also points to an unexpected co-operation between CIC-DUX4 and the Pea3-subfamily of ETS transcription factors in regulating the CDS transcriptional program.