Identification and validation of a novel prognosis model based on m5C-related long non-coding RNAs in colorectal cancer.

Background: The prognosis of tumor patients can be assessed by measuring the levels of lncRNAs (long non-coding RNAs), which play a role in controlling the methylation of the RNA. Prognosis in individuals with colorectal adenocarcinoma (CRC) is strongly linked to lncRNA expression, making it imperative to find lncRNAs that are associated with RNA methylation with strong prognostic value.

Methods: In this study, by analyzing TCGA dataset, we were able to develop a risk model for lncRNAs that are associated with m5C with prognostic significance by employing LASSO regression and univariate Cox proportional analysis. There were a number of methods employed to ensure the model was accurate, including multivariate and univariate Cox regression analysis, Kaplan analysis, and receiver operating characteristic curve analysis. The principal component analysis, GSEA and GSVA analysis were used for risk model analysis. The CIBERSORT instrument and the TIMER database were used to evaluate the link between the immune cells that infiltrate tumors and the risk model. In vitro experiments were also performed to validate the predicted m5C-related significant lncRNAs.

Results: The m5c regulators were differentially expressed in colorectal cancer and normal tissue. Based on the screening criteria and LASSO regression, 11 m5c-related lncRNAs were identified for developing the prognostic risk model. Multivariate and univariate Cox regression analysis showed the risk score is a crucial prognostic factor in CRC patients. The 1-year, 3-year, and 5-year AUC curves showed the risk score was higher than those identified for other clinicopathological characteristics. A nomogram using the risk score as a quantitative tool was developed for predicting patients' outcomes in clinical settings. In addition, the risk profile of m5C-associated lncRNAs can discriminate between tumor immune cells' characteristics in CRC. Mutation patterns and chemotherapy were analyzed between high- and low- risk groups of CRC patients. Moreover, TNFRSF10A-AS1 was chosen for the in vitro verification of the m5C-connected lncRNA to demonstrate impressive effects on the proliferation, migration and invasion of CRC cells.

Conclusion: A risk model including the prognostic value of 11 m5C-associated lncRNAs proves to be a useful prognostic tool for CRC and improves the care of patients suffering from CRC based on these findings.