Background: Cancer-associated fibroblasts (CAFs) play critical roles in tumor growth, angiogenesis, metastasis, and therapy resistance. This study aimed to investigate the characteristics of CAFs in gastric cancer (GC) and develop a CAF-based risk signature for predicting the prognosis of GC patients.
Methods: Single-cell RNA sequencing (scRNA-seq) data were obtained from the GEO database. The data related to the survival prognosis of gastric cancer patients is derived from TCGA. The scRNA-seq data were analyzed using the Seurat R package, resulting in the identification of CAF clusters based on characteristic markers. Subsequent differential expression analysis was performed on the TCGA dataset to identify genes exhibiting differential expression (DEGs) between normal and tumor samples. Pearson correlation analysis was then utilized to uncover DEGs correlated with CAF clusters, followed by univariate Cox regression analysis to identify prognostic genes associated with CAFs. Lasso regression was applied to construct a risk signature based on these prognostic CAF-related genes. Lastly, a comprehensive scoring model, integrating the risk signature and clinicopathological characteristics, was meticulously developed.
Results: From the scRNA-seq data of gastric cancer (GC), six distinct clusters of cancer-associated fibroblasts (CAFs) were delineated, among which five clusters exhibited significant associations with GC prognosis. A total of 557 differentially expressed genes (DEGs) were identified that showed strong correlations with these CAF clusters. From this set, a refined risk signature comprising six key genes was derived. These pivotal genes were predominantly implicated in 39 biological pathways, encompassing crucial processes such as angiogenesis, apoptosis, and hypoxia. Notably, the risk signature demonstrated notable correlations with stromal and immune scores, as well as specific immune cell types. Furthermore, multivariate analysis underscored the independent prognostic role of the risk signature in GC, showcasing its potential for predicting outcomes of immunotherapy. The integration of the stage with the CAF-based risk signature yielded a novel scoring model with robust predictive performance and reliability in prognosticating GC outcomes.
Conclusion: The CAF-derived risk signature emerges as a potent prognostic tool for GC, offering valuable insights into the intricate landscape of CAFs within the tumor microenvironment. Such comprehensive profiling may hold promise in guiding personalized immunotherapeutic strategies and refining treatment modalities for GC.