Screening and verification of long noncoding RNA promoter methylation sites in hepatocellular carcinoma

BACKGROUND: Long noncoding ribonucleic acid (lncRNA) promoter methylation is closely related to the occurrence and development of hepatocellular carcinoma (HCC). Thus, we aim to screen and verify the lncRNA promoter methylation sites associated with overall survival (OS), vascular invasion, pathological grade, and clinical stage in HCC. METHODS: Methylation-related data including clinical characteristic, transcriptome, methylation, and messenger RNA (mRNA) expression were taken from the Cancer Genome Atlas (TCGA) database. The OS, vascular invasion, pathological grade, and clinical stage-related lncRNA promoter methylation models were developed by the least absolute shrinkage and selection operator (LASSO) algorithm based on the lncRNA promoter methylation sites screened via R software. The Kaplan–Meier analysis, the area under the receiver operating characteristic (ROC) curve (AUC), the calibration curve (C-index) were performed to evaluate the performance of these models. Finally, the methylation-specific polymerase chain reaction (MS-PCR) was performed to verify the accuracy of these models based on 146 HCC tissues from our hospital. RESULTS: A total of 10 methylation sites were included in the OS-related lncRNA promoter methylation model that could effectively divide HCC patients into high-risk and low-risk groups (P < 0.0001) via survival analysis. COX univariable and multivariable regression analysis found that the OS-related model (P < 0.001, 95% CI 1.378–2.942) and T stage (P < 0.001, 95% CI 1.490–3.418) were independent risk factors affecting OS in HCC patients. The vascular invasion-related model contained 8 methylation sites with its AUC value of 0.657; the pathological grade-related model contained 22 methylation sites with its AUC value of 0.797; the clinical stage-related model contained 13 methylation sites with its AUC of 0.724. Target genes corresponded to vascular invasion-related lncRNA promoter methylation sites were involved in many kinds of biological processes in HCC such as PI3K-Akt signaling pathway. The accuracy of the vascular invasion-related model was consistent with our bioinformatics conclusion after being verified via MS-PCR. CONCLUSION: The lncRNA promoter methylation sites are closely correlated with the process of HCC and can be utilized to improve the therapy and prognosis of HCC.