Exploring Programmed Cell Death-Related Biomarkers and Disease Therapy Strategy in Nasopharyngeal Carcinoma Using Transcriptomics
Jiangyu Yan,Linrong Wu,Mengmeng Zheng,Fangfang Pan
DOI: https://doi.org/10.31083/j.fbl2907240
2024-06-27
Abstract:Background: Uncontrolled cellular proliferation may result in the progression of diseases such as cancer that promote organism death. Programmed cell death (PCD) is an important mechanism that ensures the quality and quantity of cells, which could be developed as a potential biomarker for disease diagnosis and treatment. Methods: RNA-seq data and clinical information of nasopharyngeal carcinoma (NPC) patients were downloaded from the Gene Expression Omnibus (GEO), and 1548 PCD-related genes were collected. We used the "limma" package to analyze differentially expressed genes (DEGs). The STRING database was used for protein interaction analysis, and the least absolute shrinkage and selection operator (Lasso) and support vector machines (SVMs) regression analyses were used to identify biomarkers. Then, the timeROC package was used for classifier efficiency assessment, and the "CIBERSORT" package was used for immune infiltration analysis. Wound healing and transwell migration assay were performed to evaluate migration and invasion. Results: We identified 800 DEGs between our control and NPC patient groups, in which 59 genes appeared to be PCD-related DEGs, with their function closely associated with NPC progression, including activation of the PI3K-Akt, TGF-β, and IL-17 signaling pathways. Furthermore, based on the STRING database, Cytoscape and six algorithms were employed to screen 16 important genes (GAPDH, FN1, IFNG, PTGS2, CXCL1, MYC, MUC1, LTF, S100A8, CAV1, CDK4, EZH2, AURKA, IL33, S100A9, and MIF). Subsequently, two reliably characterized biomarkers, FN1 and MUC1, were obtained from the Lasso and SVM analyses. The Receiver operating characteristic (ROC) curves showed that both biomarkers had area under the curve (AUC) values higher than 0.9. Meanwhile, the enrichment analysis showed that in NPC patients, the FN1 and MUC1 expression levels correlated with programmed cell death-related pathways. The enrichment analysis and cellular experimental results indicated that FN1 and MUC1 were overexpressed in NPC cells and associated with programmed cell death-related pathways. Importantly, FN1 and MUC1 severely affected the ability of NPC cells to migrate, invade, and undergo apoptosis. Finally, medroxyprogesterone acetate and 8-Bromo-cAMP acted as drug molecules for the docking of FN1 and MUC1 molecules, respectively, and had binding capacities of -9.17 and -7.27 kcal/mol, respectively. Conclusion: We examined the PCD-related phenotypes and screened FN1 and MUC1 as reliable biomarkers of NPC; our findings may promote the development of NPC treatment strategy.