A N(7)-Methylguanine-Related Gene Signature Applicable for the Prognosis and Microenvironment of Prostate Cancer

BACKGROUND: Despite the constant iteration of small-molecule inhibitors and immune checkpoint inhibitors, PRAD (prostate adenocarcinoma) patients with distant metastases and biochemical recurrence maintain a poor survival outcome along with an increasing morbidity in recent years. N(7)-Methylguanine, a new-found type of RNA modification, has demonstrated an essential role in tumor progression but has hardly been studied for its effect on prostate carcinoma. The current study aimed to seek m(7)G (N7-methylguanosine) related prognostic biomarkers and potential targets for PRAD treatment. METHODS: 42 genes related to m(7)G were collected from former literatures and GSEA (Gene Set Enrichment Analysis) website. Then, RNA-seq (RNA sequencing) and clinical data from TCGA-PRAD (The Cancer Genome Atlas-Prostate) cohort were retrieved to screen the differentially expressed m(7)G genes to further construct a multivariate Cox prognostic model for PRAD. Next, GSE116918, a prostate cancer cohort acquired from GEO (Gene Expression Omnibus) database, was analyzed for the external validation group to assess the ability to predict BFFS (biochemical failure-free survival) of our m(7)G prognostic signature. Kaplan-Meier, ROC (receiver operator characteristic), AUC (areas under ROC curve), and calibration curves were adopted to display the performance of this prognostic signature. In addition, immune infiltration analysis was implemented to evaluate the effect of these m(7)G genes on immunoinfiltrating cells. Correlation with drug susceptibility of the m(7)G signature was also analyzed by matching drug information in CellMiner database. RESULTS: The m(7)G-related prognostic signature, including three genes (EIF3D, EIF4A1, LARP1) illustrated superior prognostic ability for PRAD in both training and validation cohorts. The 5-year AUC were 0.768 for TCGA-PRAD and 0.608 for GSE116918. It can well distinguish patients into different risk groups of biochemical recurrence (p =1e-04 for TCGA-PRAD and p =0.0186 for GSE116918). Immune infiltration analysis suggested potential regulation of m(7)G genes on neutrophils and dendritic cells in PRAD. CONCLUSIONS: A m(7)G-related prognostic signature was constructed and validated in the current study, giving new sights of m(7)G methylation in predicting the prognostic and improving the treatment of PRAD.