Development and Validation of a Novel Prognostic Signature Based on M6a/m5c/m1a-Related Genes in Hepatocellular Carcinoma

Yu Xiao,Jinluan Li,Junxin Wu
DOI: https://doi.org/10.1186/s12920-023-01611-x
2023-01-01
BMC Medical Genomics
Abstract:Background RNA methylation modification plays an important role in cancers. This study sought to examine the association between m6A/m5C/m1A-related genes and hepatocellular carcinoma (HCC). Methods Gene expression and clinical data of HCC patients were obtained from the TCGA database. Unsupervised consensus clustering was performed according to the expression of m6A/m5C/m1A-related genes in HCC. The relationships among prognosis, clinicopathological features and molecular subtypes were analyzed. Least absolute shrinkage and selection operator (LASSO) regression analysis was used to establish the m6A/m5C/m1A-related gene prognostic signature. Furthermore, the prognostic signature was validated based on the ICGC dataset. RT‒qPCR was used to detect the expression of the model genes in HCC. Clinicopathological features, functional enrichment, gene mutations, immune cell infiltration, and immunotherapy response in different risk groups were analyzed. A nomogram based on risk score and stage was constructed to predict HCC patient prognosis. Results Two m6A/m5C/m1A-related molecular subtypes were identified in HCC, and the prognosis of cluster C1 was worse than that of cluster C2 ( p < 0.001). Highly expressed genes in cluster C1 are significantly correlated with G3-4, T3-4, stage III-IV ( p < 0.05). An m6A/m5C/m1A-related prognostic signature was established and validated. The RT‒qPCR results showed that the risk signature genes were significantly upregulated in liver cancer tissue ( p < 0.05). The prognosis of HCC patients in the high-risk group was worse than that of those in the low-risk group ( p < 0.05). Multivariate Cox analysis indicated that the risk score was an independent factor predicting prognosis in HCC patients. ssGSEA revealed that the risk score correlated with the tumor immune microenvironment in HCC. Gene mutation analysis showed that the tumor mutation burden of patients in the high-risk group was much higher ( p < 0.05), and the prognosis of HCC patients with high risk scores and high mutation burden was the worst ( p = 0.007). A nomogram combining risk scores with clinicopathological features showed performed well in predicting HCC prognosis. Conclusions The m6A/m5C/m1A-related genes could predict the prognosis and tumor microenvironment features of HCC and can be important biomarkers relevant to the immunotherapy response.
What problem does this paper attempt to address?