Deciphering biomarker related to 12 programmed cell death to predict prognosis and response to immunotherapy in lung squamous cell carcinoma

Abstract

Abstract Background Different programmed cell death (PCD) plays different roles in lung squamous cell carcinoma (LUSC). We integrated twelve programmed cell death patterns, investigated the expression patterns of PCD-related genes to identify promising PCD-related biomarkers. Methods Twelve PCD patterns (apoptosis, pyroptosis, necroptosis, cuproptosis, entotic cell death, autophagy-dependent cell death, netotic cell death, parthanatos, ferroptosis, lysosome-dependent cell death, oxeiptosis and alkaliptosis) were analyzed for model construction, resulting in 1388 PCD-related genes. We explored the expression changes of PCD-related genes in LUSC patients from TCGA database, and constructed a combined prognostic signature by Cox regression analysis and LASSO Cox regression analysis. The independent prognostic performance of the gene signature was evaluate based on consensus clustering, univariate and multivariate Cox regression and Kaplan–Meier survival. The GEO dataset was used for validation. Finally, we investigated the role of the immune microenvironment in different prognosis groups. Results We constructed a network of seven PCD-related genes (FGA, CHEK2, PTGIS, CSF2, STXBP1, NACC2, TFR2). Utilized these 7-gene network to establish a cell death index (CDI) and grouped patients using the median of CDI. We found that LUSC patients with low CDI had a better prognosis. More importantly, CDI was associated with tumor microenvironment components according to integrated analysis, and the response to immunotherapy in the low CDI group was better than that in the high CDI group. Conclusion Our study identified 7-gene network based on PCD to establish a new model of CDI to predict the clinical prognosis of LUSC patients. We proposed that CDI may serve as a new biomarker to predict the prognosis and immunotherapy efficacy in LUSC.