Molecular analysis of matched PIN, invasive prostate cancer, and adjacent normal prostate tissue samples reveal distinct transcriptional signatures and clonal relationships

Abstract Background: Prostate cancer (PCa) is highly heterogeneous, and while some tumors remain indolent throughout a patient's lifetime, others are highly aggressive, leading to significant morbidity and death. The current methods for PCa diagnosis and prognosis are highly suboptimal, resulting in over-diagnosis and overtreatment of many clinically insignificant tumors. To understand the nature of PCa heterogeneity, we conducted comprehensive genomic analyses on matched samples of adjacent normal prostate tissue, prostate intraepithelial neoplasia (PIN), and invasive PCa. Methods: To investigate the changes cells undergo during PCa development and progression, we analyzed genomic and transcriptomic alterations in 92 adjacent normal, 80 PIN, and 99 PCa samples from 34 patients. For each sample type, both epithelial and stromal components were analyzed separately using laser capture microdissection of formalin fixed paraffin embedded tissue. Transcriptomic analysis was performed using smart3-SEQ and copy number analysis was performed by low-pass whole genome sequencing (WGS). The samples were analyzed as three separate data sets. Results: We studied the spectrum of molecular changes present in both epithelial and stromal samples. A negative-binomial regression model was used to identify differentially expressed genes between adjacent normal prostate tissue and invasive PCa in epithelium and stroma separately using the discovery data set. The resulting gene lists were used to perform non-negative matrix factorization clustering of all samples in the three datasets separately. This analysis identified two epithelial and two stromal clusters with distinct RNA expression profiles in all three data sets. For invasive PCa samples, all except three epithelial samples and five stromal samples clustered in the ‘invasive-like’ clusters. Similarly for adjacent normal samples, all except three of the epithelial and six of the stromal samples clustered as ‘normal-like’ across the three datasets. For PIN samples, 54% of epithelial and 62% of stromal samples classified as ‘normal-like’. We performed phylogenetic analysis of WGS data and identified PIN lesions with and without clear clonal relation to invasive PCa lesions. Recurrent copy number variations were identified in both PIN and PCa lesions. Conclusion. We identified two distinct epithelial and two distinct stromal expression clusters with ‘invasive-like’ and ‘normal-like’ signatures. WGS analysis revealed recurrent copy number alterations in PIN and PCa lesions, and evidence of clonal relationships between PIN and invasive PCa. These studies provide new insight into PIN biology and the relationship between PIN and invasive PCa. Citation Format: Siri H. Strand, Okyaz Eminaga, Sujay Vennam, Chunfang Zhu, Jason Wang, Sushama Varma, Rosalie Nolley, Christian Kunder, Jonathan Pollack, Andreas Roeder, Karina D. Sorensen, James D. Brooks, Robert B. West. Molecular analysis of matched PIN, invasive prostate cancer, and adjacent normal prostate tissue samples reveal distinct transcriptional signatures and clonal relationships [abstract]. In: Proceedings of the AACR Special Conference: Advances in Prostate Cancer Research; 2023 Mar 15-18; Denver, Colorado. Philadelphia (PA): AACR; Cancer Res 2023;83(11 Suppl):Abstract nr A017.