The molecular classification of cancer‐associated fibroblasts on a pan‐cancer single‐cell transcriptional atlas

Abstract

Abstract Background Cancer‐associated fibroblasts (CAFs), integral to the tumour microenvironment, are pivotal in cancer progression, exhibiting either pro‐tumourigenic or anti‐tumourigenic functions. Their inherent phenotypic and functional diversity allows for the subdivision of CAFs into various subpopulations. While several classification systems have been suggested for different cancer types, a unified molecular classification of CAFs on a single‐cell pan‐cancer scale has yet to be established. Methods We employed a comprehensive single‐cell transcriptomic atlas encompassing 12 solid tumour types. Our objective was to establish a novel molecular classification and to elucidate the evolutionary trajectories of CAFs. We investigated the functional profiles of each CAF subtype using Single‐Cell Regulatory Network Inference and Clustering and single‐cell gene set enrichment analysis. The clinical relevance of these subtypes was assessed through survival curve analysis. Concurrently, we employed multiplex immunofluorescence staining on tumour tissues to determine the dynamic changes of CAF subtypes across different tumour stages. Additionally, we identified the small molecule procyanidin C1 (PCC1) as a target for matrix‐producing CAF (matCAF) using molecular docking techniques and further validated these findings through in vitro and in vivo experiments. Results In our investigation of solid tumours, we identified four molecular clusters of CAFs: progenitor CAF (proCAF), inflammatory CAF (iCAF), myofibroblastic CAF (myCAF) and matCAF, each characterised by distinct molecular traits. This classification was consistently applicable across all nine studied solid tumour types. These CAF subtypes displayed unique evolutionary pathways, functional roles and clinical relevance in various solid tumours. Notably, the matCAF subtype was associated with poorer prognoses in several cancer types. The targeting of matCAF using the identified small molecule, PCC1, demonstrated promising antitumour activity. Conclusions Collectively, the various subtypes of CAFs, particularly matCAF, are crucial in the initiation and progression of cancer. Focusing therapeutic strategies on targeting matCAF in solid tumours holds significant potential for cancer treatment.

Keywords

• cancer‐associated fibroblast
• molecular classification
• pan‐cancer
• single‐cell transcriptional atlas