Background Osteosarcoma is a primary malignant tumor, characterized by its high incidence and recurrence rate in children and adolescents. Ferroptosis, an iron-dependent form of regulated cell death, has recently been recognized as a potential therapeutic vulnerability in cancer treatment. However, its prognostic significance and underlying regulatory mechanisms in osteosarcoma remain largely unexplored. Materials and methods We constructed a prognostic model based on 12 ferroptosis-related genes using LASSO regression and validated across independent GEO cohorts (GSE21257 and GSE39055). We identified hub genes via machine learning algorithms (SVM, RF, XGBoost, BORUTA) and single-cell RNA sequencing. The exosomal transfer of COX4I2 protein from CAFs to 143B osteosarcoma cells was evaluated by Western blot, confocal microscopy, and transmission electron microscopy. Ferroptosis indicators, including Fe 2+ , MDA, ACSL4, and ROS levels, were assessed in vitro . We performed tumorigenicity assays in vivo in nude mice to validate biological function. Results The ferroptosis-based risk model exhibited robust prognostic performance. We identified COX4I2 as a stromal hub gene, highly enriched in cancer-associated fibroblasts (CAFs). Functional experiments demonstrated that exosome-mediated delivery of COX4I2 suppressed ferroptosis in osteosarcoma cells and enhancd cell proliferation and mitochondrial integrity. Studies in vivo further revealed that overexpression of exosomal COX4I2 markedly promoted tumor growth while inhibiting ferroptosis. Conclusion These findings underscore the potential of exosomal COX4I2 as a biomarker and therapeutic target for ferroptosis-based interventions in osteosarcoma.
Niu et al. (Thu,) studied this question.