Optimised dissociation and multimodal profiling of prostate cancer stroma reveal fibromuscular cell heterogeneity with clinical correlates

Elisabeth Damisch¹Elena Brunner¹Lukas Nommensen¹Lucy Neumann¹Georgios Fotakis²Zlatko Trajanoski²Sieghart Sopper³Georg Schäfer⁴Martin Puhr¹Isabel Heidegger¹Marianna Kruithof-de Julio^5,6Natalie Sampson^1*

• ¹Department of Urology, Innsbruck Medical University, Innsbruck, Austria
• ²Institute of Bioinformatics, Innsbruck Medical University, Innsbruck, Austria
• ³Univ. Clinic for Hematology and Oncology, Internal Medicine V, Innsbruck Medical University, Innsbruck, Austria
• ⁴Department of Pathology, Neuropathology and Molecular Pathology, Innsbruck Medical University, Innsbruck, Austria
• ⁵Universitat Bern Department for BioMedical Research, Bern, Switzerland
• ⁶Inselspital Universitatsspital Bern Universitatsklinik fur Urologie, Bern, Switzerland

Background: Dynamic remodelling of the tumour microenvironment (TME) plays a central role in prostate cancer (PCa) progression, immune evasion and therapy resistance. However, the co-existence of both tumour-promoting and tumour-restraining stromal elements necessitates extensive characterisation of the TME for effective targeting. Fibromuscular cell heterogeneity in PCa remains poorly characterised, in part due to challenges in isolating cells embedded within the desmoplastic stroma. This study therefore aimed to better characterise fibroblast and smooth muscle cell (SMC) populations as the major tissue-resident stromal cell subtypes within the PCa TME. Methods: A PCa single-cell RNA sequencing (scRNA-seq) dataset was re-analysed to define fibromuscular subtypes. Due to low fibroblast yields, an optimised tissue dissociation protocol was developed and benchmarked against two commercial kits via flow cytometry, immunostaining of clinical specimens and ex vivo culture. Dimensionality reduction and clustering were applied to the CD31-stromal fraction using a multiparameter surface marker panel. Annotation of the resulting clusters based on their surface marker profile was supported by integrating scRNA-seq and immuno-histological findings. Results: The optimised protocol yielded over twice the viable cells/mg tissue compared to two commercial kits, preserved surface marker integrity, enhanced successful cultivation of mesenchymal cells and recovered diverse stromal subpopulations from benign and malignant samples. Dimensionality reduction and clustering of flow cytometry counts identified 11 distinct CD31-stromal populations. Integration with transcriptomic data and immunofluorescence of clinical specimens identified spatially-and prognostically-distinct fibroblast subtypes, including inflammatory and myofibroblastic cancer-associated fibroblasts, pericytes linked to poor prognosis and a novel SMC subset associated with stromal activation.