(P 345) TRANSCRIPTOMIC AND PHENOTYPIC CHARACTERIZATION OF AN IMMUNOCOMPETENT LEIOMYOSARCOMA MOUSE MODEL.

Objective: Leiomyosarcomas (LMS) are rare, aggressive, malignant soft tissue tumors, burdened by dismal prognosis and limited therapeutic options in advanced stages. Novel treatment strategies, including immunotherapy, have been investigated both in clinical and research settings. However, these studies have been limited by the lack of reliable immunocompetent models. Main objective of this work was the thorough transcriptomic and phenotypic characterization of an immunocompetent LMS murine model, which could foster future immunotherapy research in LMS.

Methods: The immunocompetent Usp18-/- murine model was established starting from Chinyengetere F. et al. work (PMID: 26555296), showing for the first time that mice null for the deubiquitinase Usp18 spontaneously develop LMS. Starting from the mating of FVB.129(B6) Usp18+/- mice, we generated a progeny in which we identified by genotyping homozygous Usp18-/- mice (n=82). Usp18-/- homozygous mice developed palpable tumor masses around 210±60 days, with a penetrance of 25.6% (21/82). At tumor onset, the tumor was explanted and further processed in vitro. The tumor and derived cell lines were characterized, implanted subcutaneously into the right flank of FVB/N wild type mice, and explanted when target tumor volume was reached. Putative murine syngeneic LMS tumors were sent for centralized, blind pathology review with morphological and immunohistochemical (IHC) markers assessment. Total RNA was extracted from snap-frozen primary human LMS tumor tissue and putative murine LMS. RNA sequencing reads from snap frozen human LMS samples and from murine samples were compared. A collection of LMS gene signatures from the literature was obtained by mining five public LMS transcriptomic studies and data from TCGA and the Human Phenotype Ontology (HPO) (PMID: 29100075, 36672483, 32518213, 25896974, 29321523). ComplexHeatmap v2.6.2 was used to graphically represent the results of the hierarchical clustering analysis of the normalized gene levels in these samples.

Results: We explored the morphological, immunohistochemical, genomic and transcriptomic features of the Usp18-/- putative LMS model. The morphological and most relevant IHC features, concordant with LMS histotype, are shown in Figure 1. The syngeneic tumor 13.1 (SYN 13.1) shows spindle-shaped cells with human LMS-like morphology, while SYN14 has more prominent undifferentiated epithelioid morphology, however retaining desmin and focal caldesmon positivity. By means of total RNA sequencing, we evaluated the similarities between the putative murine LMS model and human LMS, taking advantage of LMS samples available at our Institution and of publicly available LMS datasets. The concordance between the identified LMS-related genes with LMS-specific expression patterns in the selected datasets and our human and murine samples is shown in Table 1. In Figure 2, intersections between the selected datasets are depicted. One-third (10/27) of the murine LMS samples clearly clustered with human LMS samples and TCGA LMS, as shown in Figure 3 (heatmap), while the other syngeneic murine LMS samples were less homogeneous. Of note, also a few human LMS samples showed similarly scattered expression profiles.

Conclusion: In this work, we provide a thorough morphological, phenotypic and molecular characterization of a putative LMS immunocompetent model, which could be instrumental to study and develop novel immunotherapeutic strategies in LMS.