Identification of a cancer-associated fibroblast classifier for predicting prognosis and therapeutic response in lung squamous cell carcinoma.

Reliable prognostic gene signatures for cancer-associated fibroblasts (CAFs) in lung squamous cell carcinoma (LUSC) are still lacking, and the underlying genetic principles remain unclear. Therefore, the 2 main aims of our study were to establish a reliable CAFs prognostic gene signature that can be used to stratify patients with LUSC and to identify promising potential targets for more effective and individualized therapies. Clinical information and mRNA expression were accessed of the cancer genome atlas-LUSC cohort (n = 501) and GSE157011 cohort (n = 484). CAFs abundance were quantified by the multi-estimated algorithms. Stromal CAF-related genes were identified by weighted gene co-expression network analysis. The least absolute shrinkage and selection operator Cox regression method was utilized to identify the most relevant CAFs candidates for predicting prognosis. Chemotherapy sensitivity scores were calculated using the "pRRophetic" package in R software, and the tumor immune dysfunction and exclusion algorithm was employed to evaluate immunotherapy response. Gene set enrichment analysis and the Search Tool for Interaction of Chemicals database were applied to clarify the molecular mechanisms. In this study, we identified 288 hub CAF-related candidate genes by weighted gene co-expression network analysis. Next, 34 potential prognostic CAFs candidate genes were identified by univariate Cox regression in the cancer genome atlas-LUSC cohort. We prioritized the top 8 CAFs prognostic genes (DCBLD1, SLC24A3, ILK, SMAD7, SERPINE1, SNX9, PDGFA, and KLF10) by a least absolute shrinkage and selection operator Cox regression model, and these genes were used to identify low- and high-risk subgroups for unfavorable survival. In silico drug screening identified 6 effective compounds for high-risk CAFs-related LUSC: TAK-715, GW 441756, OSU-03012, MP470, FH535, and KIN001-266. Additionally, search tool for interaction of chemicals database highlighted PI3K-Akt signaling as a potential target pathway for high-risk CAFs-related LUSC. Overall, our findings provide a molecular classifier for high-risk CAFs-related LUSC and suggest that treatment with PI3K-Akt signaling inhibitors could benefit these patients.