Abstract 3816: Functional atlas of prostate mesenchyme

Abstract

Abstract Prostate cancer has a heterogeneous prognosis, and genetic alterations alone do not fully explain clinical behavior. We previously characterized the stroma of localized human prostates by Laser Capture Microdissection, and found that stroma was substantially different in prostates with and without tumor. Furthermore, a stromal gene signature reflecting bone remodeling was upregulated in high compared to low Gleason grade cases. To determine how stromal cells contribute to carcinogenesis and progression we study whether specific genetic alterations in the epithelium induce unique stromal changes. To do this, we utilized Genetically Engineered Mouse Models (GEMMs) representing common prostate cancer mutations and compared these to their wild-type conterparts: the Tmprss2-ERG fusion knock-in murine model induces histological alterations in the stroma in the absence of an epithelial phenotype; the Pten deletion mouse model (PtenKO) results in prostate intraepithelial neoplasia (PIN) but not invasive cancer; the Hi-Myc GEMM, leads to PIN and subsequently invasion; and the Pb4-Cre +/-;Pten f/f; LSL-MYCN +/+; Rb1 f/f (MNRPDKO) mouse model that leads to neuroendocrine prostate cancer (NEPC). We generated a comprehensive single-cell transcriptomic atlas of the mouse prostate cancer mesenchyme in these models. Using deep generative modeling followed by graph-based clustering and gene regulatory network inference, six (6) distinct subsets of fibroblasts and two (2) subsets of smooth muscle cells (myofibroblasts and pericytes) were identified. Notably, some subsets were common across all GEMMs and WT mice, while others aligned with specific genotypes. Moreover, we found a variable pattern of positive and negative Ar expressing cells between genotypes. Analysis by CellphoneDB of mesenchymal-epithelial communications revealed the complex cross-talk between mutated epithelial cells and the tumor microenvironment. Multiplex immunofluorescence phenotyping of mesenchymal cell confirmed the cluster subtypes by both expression and spatial location. Finally, stromal transcripts defining mesenchymal cluster subtypes associated with Tmprss2-ERG were conserved between mouse and human genotypes.This study lays the groundwork for understanding and ultimately targeting stromal-epithelial interactions in prostate cancer. Citation Format: Hubert Pakula, Ryan Carelli, Nicolo Fanelli, Madhavi Jere, Caitlin Unkenholz, Mohamed Omar, Caroline Ribeiro- Fidalgo, Filippo Pederzoli, Cory Abate-Shen, David S. Rickman, Brian Robinson, Luigi Marchionni, Massimo Loda. Functional atlas of prostate mesenchyme [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 3816.