snFLARE-seq and mrFRIGID for the transcriptomic and metabolomic landscape of prostate cancer with different anatomical origins

snFLARE-seq and mrFRIGID for the transcriptomic and metabolomic landscape of prostate cancer with different anatomical origins

Prostate cancer cells of different anatomical locations display remarkable heterogeneity. This poses a challenge to the clinical relevance of pre-clinical models and the efficacy of contemporary therapeutic approaches. Here we developed the snFLARE-seq and mrFRIGID methodologies to directly investigate the transcriptomic and metabolomic landscape of prostate cancer patients utilizing formalin-fixed paraffin-embedded (FFPE) specimens. A retrospective analysis revealed the clinical disparities of prostate cancer from peripheral zone (PZ), transition zone (TZ), and across PZ and TZ. The snFLARE-seq, refined for enhanced single-nucleus sequencing, unveiled distinct cell type distributions and signaling pathways between PZ and TZ samples. Hormone therapy substantially affected cancer cells and microenvironment, leading to a polarized feature of epithelial cells and a subverted immune microenvironment. With improvements on metabolite extraction, mrFRIGID revealed unique metabolic features of prostate cancer from different origins. The metabolomic results indicate that PZ cancer cells were in a metabolic-dormant status, which were probably awaken by hormone therapy. Integrative analysis of results from snFLARE-seq, mrFRIGID, and TCGA database uncovered four metabolic pathways and related genes associated with disease aggressiveness. Our work would accelerate investigations on disease heterogeneity and evolution in real-world clinical settings, stimulating patient-specific precision healthcare solutions.

This study utilized the Dynamic Network Biomarkers (DNB) model developed by ChenLab at the Chinese Academy of Sciences (CAS) to analyze single-cell data. For specific details, please visit https://github.com/Kaiyu-W/DNBr.

To install this package, use the following command in R: devtools::install_github("Kaiyu-W/DNBr")

The associated datasets are publicly available at Zenodo under the following persistent link: https://zenodo.org/records/15671856