Identification of autophagy-related long non-coding RNAs in endometrial cancer via comprehensive bioinformatics analysis

BACKGROUND: Endometrial cancer is a common gynaecological malignancy with an increasing incidence. It is of great importance and value to uncover its effective and accurate prognostic indicators of disease outcomes. METHODS: The sequencing data and clinical information of endometrial cancer patients in the TCGA database were downloaded, and autophagy-related genes in the human autophagy database were downloaded. R software was used to perform a Pearson correlation analysis on autophagy-related genes and long non-coding RNAs (lncRNAs) to screen autophagy-related lncRNAs. Next, univariate and multivariate Cox regression analyses were performed to select autophagy-related lncRNAs and construct the prognostic model. Finally, the accuracy of the prognostic prediction of the model was evaluated, the lncRNA–mRNA network was constructed and visualized by Cytoscape, and the gene expression profile of endometrial cancer patients was analysed by GSEA. RESULTS: A total of 10 autophagy-related lncRNAs were screened to construct the prognostic model. The risk factors were AC084117.1, SOS1-IT1, AC019080.5, FIRRE and MCCC1-AS, and the protective factors were AC034236.2, POC1B-AS1, AC137630.1, AC083799.1 and AL133243.2. This prognostic model could independently predict the prognosis of endometrial cancer patients and had better predictive performance than that of using age and tumour grade. In addition, after classifying patients as high-risk or low-risk based on the prognostic model, we found that the enrichment of the JAK-STAT and MAPK pathways was significantly higher in the high-risk group than that in the low-risk group. CONCLUSIONS: The 10 autophagy-related lncRNAs are potential prognostic biomarkers. Compared with using age and tumour grade, this prognostic model is more predictive for the prognosis of endometrial cancer patients.