Abstract 2584: <i>In situ</i> hybridization-based validation of transcriptionally-defined stromal clusters in the prostate cancer microenvironment

Abstract Introduction: Stromal cells in the prostate tumor microenvironment (TME) affect tumor progression and exhibit both a universal response to tumorigenesis and specific ones determined by the genotype that drives the tumor (Pakula et al., Nat Commun 2024). As a result, stromal clusters associated with murine and human prostate cancer identified by single cell RNA sequencing (scRNA-seq) fall into a) common subtypes: myofibroblasts/pericytes and innate immunity-associated; as well as b) two additional ones specific to either AR positive adenocarcinomas or neuroendocrine tumors (NEPC). Here, we defined their spatial localization leveraging spatial transcriptomics and multiplex in situ RNA hybridization. Methods &amp; Results: We computationally identify the most discriminative genes able to distinguish the four stromal groups. Specifically, we first identified highly discriminatory genes representative of each cluster. We then used integer programming-based optimization to obtain small sets of genes (n&amp;lt;4) that are differentially expressed in each group compared to the remaining three. For instance, group 0, encompassing myofibroblasts and pericytes, was characterized by the expression of Rgs5 and Myh11, while group 2, whose stromal cells are characterized by the expression of Ar and Wnt signaling genes, was characterized by the expression of Wif1 and Coch. A decision-tree analysis was able to discriminate all the different stromal groups. These were validated using 1) in situ RNA hybridization + multiplex immunofluorescence panels and 2) spatial transcriptomics. Myofibroblasts/pericytes were found to be distant to tumor cells, common clusters were associated with all tumors of diverse phenotype (Tmprss2-ERG, cMYC, Pten-/-, Pten-/-/ Rb1-/- NMYC+/+), while periostin high/AR low clusters were immediately adjacent to NEPC foci in mouse models. Conclusions: Our integrated computational and experimental approach establishes a robust pipeline for validating stromal markers identified through scRNA-seq, enabling spatial mapping of stromal heterogeneity in prostate cancer. Citation Format: Filippo Pederzoli, Hubert Pakula, Wikum Dinalankara, Jeanne Jiang, Mohamed Omar, Silvia Rodrigues, Pier Vitale Nuzzo, Fabio Socciarelli, Giuseppe Nicolò Fanelli, Luigi Marchionni, Massimo Loda. In situ hybridization-based validation of transcriptionally-defined stromal clusters in the prostate cancer microenvironment [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 2584.