Abstract
Background: Due to the highly variable prognosis of low-grade gliomas (LGGs), it is important to find robust biomarkers for predicting clinical outcomes. Aging cancer-associated fibroblasts (CAFs) within the senescent stroma of a tumor microenvironment (TME) have been recently reported to play a key role in tumor development. However, there are few studies focusing on this topic in gliomas. Methods and Results: Based on the transcriptome data from TCGA and CGGA databases, we identified aging CAF-related genes (ACAFRGs) in LGGs by the weighted gene co-expression network analysis (WGCNA) method, followed by which LGG samples were classified into two aging CAF-related gene clusters with distinct prognosis and characteristics of the TME. Machine learning algorithms were used to screen out eight featured ACAFRGs to characterize two aging CAF-related gene clusters, and a nomogram model was constructed to predict the probability of gene cluster A for each LGG sample. Then, a powerful aging CAF scoring system was developed to predict the prognosis and response to immune checkpoint blockage therapy. Finally, the ACAFRGs were verified in two glioma-related external datasets. The performance of the aging CAF score in predicting the immunotherapy response was further validated in two independent cohorts. We also confirmed the expression of ACAFRGs at the protein level in glioma tissues through the Human Protein Atlas website and Western blotting analysis. Conclusion: We developed a robust aging CAF scoring system to predict the prognosis and immunotherapy response in LGGs. Our findings may provide new targets for therapeutics and contribute to the exploration focusing on aging CAFs.