Abstract
BACKGROUND: Long non-coding ribonucleic acids (lncRNAs) regulate messenger RNA (mRNA) expression and influence cancer development and progression. Cuproptosis, a newly discovered form of cell death, plays an important role in cancer. Nonetheless, additional research investigating the association between cuproptosis-related lncRNAs and prostate cancer (PCa) prognosis is required. METHODS: Sequencing data and copy number variant data were obtained from 492 patients with PCa from The Cancer Genome Atlas (TCGA) Program. Prognostic models of PCa based on cuproptosis-related lncRNAs were constructed using a multi-level attention graph neural network (MLA-GNN) deep learning algorithm. Immune escape scoring was performed using Tumor Immune Dysfunction and Exclusion. Cellular experiments were conducted to explore the correlation between key lncRNAs and cuproptosis. RESULTS: Data from 492 patients with PCa were randomized into two groups at a 1:1 ratio. Prognostic modeling was successfully established using MLA-GNN. Survival analysis suggested that patients could be divided into high- and low-risk groups according to model scores and that there was a significant difference in disease-free survival (DFS) (P < 0.01). The area under the receiver operating characteristic (ROC) curve (AUC) indicated a strong predictive performance for the model, with AUCs of 0.913, 0.847, and 0.863 for the training group and 0.815, 0.907, and 0.866 for the test group at 12, 36, and 60 months, respectively. The immune escape score and immune microenvironment analysis suggested that the high-risk group corresponded to a stronger immune escape and a poorer immune microenvironment (P < 0.05). Cellular experiments revealed that the expression of all six key lncRNAs was upregulated in the presence of copper ion carriers (P < 0.05). CONCLUSIONS: This study identified cuproptosis-related lncRNAs that were strongly associated with PCa prognosis. Key lncRNAs could affect copper metabolism and may serve as new therapeutic targets.