Comprehensive Analysis of WGCNA - Derived Cancer Associated Fibroblasts Model For Prognosis, Immune Features, and Candidate Drug Development in LUSC

Abstract

Abstract Cancer-associated fibroblasts (CAFs) directly affect the behavior of surrounding cells and reshape extracellular matrix (ECM) in tumor microenvironment (TME) via cell-cell contact, releasing regulatory factors. This study aimed to explore stromal CAF - related genes for prognostic prediction and therapeutic response in LUSC. We downloaded mRNA expression and clinical information of 243 LUSC cases from Gene Expression Omnibus (GEO) and 504 cases from The Cancer Genome Atlas (TCGA) databases. weighted gene co-expression network analysis (WGCNA) was performed to identity the key gene module. The protein-protein interaction (PPI) network and machine learning methodology were used to construct a prognostic model. The risk score was involved in 5 genes (COL1A2, COL4A1 COL5A1 MMP2,FN1). In addition, a series of methods based on bioinformatics were used and the results indicated the cases in high risk group suffered less survival time, weaker immune response and higher likely to respond to chemotherapeutic agents. Subsequently, we characterized prognostic model by sing-cell sequencing and immunohistochemistry. This five - gene prognostic CAF signature may be a potential biomarker for guiding anti - CAFs therapy and a prognostic clue related to CAF for LUSC patients.