Abstract
BACKGROUND: Long noncoding RNAs (lncRNAs) influence the onset of osteosarcoma. Cuproptosis is a novel cell death mechanism. We attempted to identify a cuproptosis-related lncRNA signature to predict the prognosis and immune landscape in osteosarcoma patients. METHODS: Transcriptional and clinical data of 85 osteosarcoma patients were derived from the TARGET database and randomly categorized into the training and validation cohorts. We implemented the univariate and multivariate Cox regression, along with LASSO regression analyses for developing a cuproptosis-related lncRNA risk model. Kaplan-Meier curves, C-index, ROC curves, univariate and multivariate Cox regression, and nomogram were used to assess the capacity of this risk model to predict the osteosarcoma prognosis. Gene ontology, KEGG, and Gene Set Enrichment (GSEA) analyses were conducted for determining the potential functional differences existing between the high-risk and low-risk patients. We further conducted the ESTIMATE, single-smaple GSEA, and CIBERSORT analyses for identifying the different immune microenvironments and immune cells infiltrating both the risk groups. RESULTS: We screened out four cuproptosis-related lncRNAs (AL033384.2, AL031775.1, AC110995.1, and LINC00565) to construct the risk model in the training cohort. This risk model displayed a good performance to predict the overall survival of osteosarcoma patients, which was confirmed by using the validation and the entire cohort. Further analyses showed that the low-risk patients have more immune activation and immune cells infiltrating as well as a good response to immunotherapy. CONCLUSIONS: We developed a novel cuproptosis-related lncRNA signature with high reliability and accuracy for predicting outcome and immunotherapy response in osteosarcoma patients, which provides new insights into the personalized treatment of osteosarcoma.