Prognosis and Immune Correlation Analysis of m1A/m5C/m6A/m7G Regulated Genes in Gastric Cancer

Abstract

Abstract Background In gastric cancer (GC), biomarkers that reliably predict prognosis and patient response to immune checkpoint blockade (ICB) are lacking. Accumulating evidence indicate that RNA modification of m1A/m5C/m6A/m7G has a close relationship with the initiation and progression of cancer, particularly in GC. Here, our objective is to identify a significant signature based on m1A/m5C/m6A/m7G-regulated genes for prognosis prediction and immune correlation analysis in GC. Methods Firstly, The Cancer Genome Atlas (TCGA)-GC dataset was sifted for m1A/m5C/m6A/m7G-regulated genes that were significantly differentially expressed in normal and GC samples. By combining clinical survival prognostic information of the samples, the most optimal gene combination that was significantly associated with GC prognosis was then systematically sifted. Following that, a novel prognostic risk score (RS) model was constructed. The GSE62254 dataset was used for the RS model validation, with own RT-qPCR conducted for biological validation. Furthermore, a nomogram was founded to better predict the overall survival (OS) of GC. Finally, the RS model and its relevance to immune infiltration, drug sensitivity and pathway enrichment were investigated. Results On the basis of the m1A/m5C/m6A/m7G-regulated genes, we developed a prognostic RS model that classified GC patients as high or low risk. The predicted capability of the RS model was well validated in both TCGA-GC training and GSE62254 validation sets. After identifying the RS model as an independent prognostic factor via univariate and multivariate analyses, we built a nomogram with high accuracy to enhance the RS model's clinical suitability. When compared to low-risk patients, high-risk patients had a shorter OS and more activated oncogenic pathways. More importantly, the high-risk group exhibited higher ESTIMATE, immune, and stromal scores, as well as higher expression of immune checkpoint-related genes and human leukocyte antigen (HLA)-related genes. Lastly, we observed that the majority of commonly used GC chemotherapeutic agents had lower IC50 values in high-risk patients. Conclusion We created a reliable prognostic RS model based on m1A/m5C/m6A/m7G regulated genes that can predict GC prognosis and guide individualized treatment decisions-making.