(P 267) NEW MOUSE MODEL OF SYNOVIAL SARCOMA BASED ON CONDITIONAL EXPRESSION IN MESENCHYMAL PROGENITORS

Objective: Conditional mouse models are an important strategy to uncover cancer biology and develop new treatments. However, despite recent advancements in our understanding of synovial sarcoma biology and its SS18SSX fusion oncogene driver, the cell of origin (in which conditional expression of SS18SSX would best model the disease) remains undefined. We hypothesize that Hic1+ mesenchymal progenitor cells may be relevant cells of origin for synovial sarcoma.

Methods: We previously generated a specific Hic1CreERT2 system that has enabled us to identify mesenchymal progenitors, characterize and define their lineage trajectories through embryonic development into adulthood. This mesenchymal progenitor-specific Cre line has been used to generate a unique mouse model of synovial sarcoma where the human SS18-SSX fusion oncoprotein (with an EGFP marker) is conditionally expressed in Hic1 (Hypermethylated in cancer 1)-expressing mesenchymal progenitor cells. To generate this model, a genetically engineered mouse model was created by crossing the Hic1CreERT2 driver with a floxed SS18-SSX2 allele. Oncoprotein expression is induced in 8-week old mice by administration of tamoxifen (which enables CreERT2 to enter the nucleus).

Results: This model shows robust formation of sarcomas with 100% penetrance after a mean latency of 15.5 weeks. Histologic analyses of tumors reveal characteristic cellular morphology reflective of the three major sub-types of synovial sarcoma, expressing key markers including human SS18SSX and mouse TLE1 and BCL2 proteins. Tumors primarily form in two sites: the stifle joints and the tongue. The orthotopic sarcomas in the stifle are most commonly monophasic and can be found in ~ 90% of mice at endpoint. Tumors in the tongue are faster growing, present with the histology of poorly differentiated biphasic synovial sarcomas, and are found in 100% of mice. These tongue tumors allow for robust, accessible, and replicable monitoring of tumor development. Microscopic precursor lesions are detectable in the tongue by 4 weeks post-tamoxifen induction, are grossly visible at 8 weeks allowing for ready monitoring of tumor growth and drug response, and result in endpoint on average at 16 weeks post-induction. Single-cell RNA-seq gene expression profiles of these tumors recapitulate orthologous genes expressed in human synovial sarcoma, highlighting the genetic programs involved in sarcomagenesis.

Conclusion: This new mouse model of synovial sarcoma closely recapitulates human disease as assessed by tumor location, gene expression, diagnostic markers, and histomorphology, identifying a subset of Hic1+ mesenchymal progenitor cells as permissive cells of origin for this sarcoma. The robustness and predictability of the tumor formation in this model opens a unique opportunity to test novel compounds for their therapeutic potential in synovial sarcoma.