(P 67) MULTIPLE MESENCHYMAL LINEAGES AS POTENTIAL CELL-OF-ORIGINS OF SMARCB1-DEFICIENT EPITHELIOID SARCOMA

Objective: Epithelioid sarcoma (EPS) is an aggressive soft-tissue sarcoma with multiple histological variations, including classical and proximal types. While SMARCB1 deficiency is recognized as a primary driving event, the mechanisms initiating EPS remain largely unknown. Progress in understanding EPS biology is limited by the small number of available patient samples and limited disease models to study sarcomagenesis. To address this, we performed a systematic study to develop mouse models of EPS.

Methods: We deleted Smarcb1 in a site- or lineage-specific manner using the Cre-loxP system. Genetically engineered mice (GEMs) expressing CreERT2 under lineage-specific promoters were used, including Pax7, Prrx1, Pdgfra, or Cdh5, to target myogenic, mesenchymal, fibro-adipogenic, or endothelial lineages, respectively. GEMs carrying floxed Smarcb1 (S), floxed p53 (P), or both alleles (SP) were used for Adeno-Cre and TAT-Cre injection to induce site-specific recombination or to generate CreERT2 mice carrying these alleles.

Results: Delivery of Adeno-Cre and TAT-Cre induced tumors in both S and SP mice, with higher incidence rates in SP mice. TAT-Cre induced earlier tumor onset, with a median latency of 68 days compared to 228 days with Adeno-Cre. Tumor formation was observed in all CreERT2 lines tested. Tumors in Pax7 and Pdgfra lines were restricted to the 4OHT-injection site, whereas Prrx1 and Cdh5 lines frequently developed tumors at other sites, such as those on the face (Prrx1) and genital tract (Cdh5). Histological analysis revealed tumor cells with characteristic epithelioid morphology, occasional rhabdoid cells, and loss of Smarcb1 staining. Principal component analysis (PCA) of bulk tumor RNA-seq revealed several distinct tumor clusters, where tumors with myogenic and fibro-adipogenic signatures constituted two major clusters.

Conclusion: In this study, we developed highly efficient conditional mouse models of EPS, providing a robust platform for preclinical testing of novel therapeutic approaches. Our results support a model where multiple muscle-resident cell types can give rise to EPS following SMARCB1 loss, potentially explaining the tumor heterogeneity of EPS across patients.