Molecular characterization and prognostic modeling of liquid-liquid phase separation-related genes in osteosarcoma based on single-cell sequencing and weighted gene co-expression network analysis.

Background: In recent years, liquid-liquid phase separation (LLPS) has garnered increasing attention in the field of oncology. However, its role in osteosarcoma remains largely unexplored. We aimed to construct a prognostic risk model associated with LLPS and to investigate the impact of LLPS-related genes on osteosarcoma briefly. Methods: Based on the single-cell dataset GSE162454, LLPS gene expression scores were calculated to stratify osteosarcoma samples into high and low LLPS expression cohorts, followed by differential analysis between the two groups. Using bulk transcriptomic data from GSE21257, LLPS scores were computed for each sample, and weighted gene co-expression network analysis (WGCNA) was performed to identify modules most strongly related to LLPS. Cox and least absolute shrinkage and selection operator (LASSO) regression analyses were performed on the genes that overlap between the two datasets to identify LLPS-related prognostic genes. A prognostic model was subsequently constructed, a nomogram was developed for clinical application, and its independent prognostic value was evaluated. Additionally, selected results were validated experimentally. Results: The LLPS-related prognostic model comprising seven genes, CCT6A, EXOSC8, ACO2, SEPHS1, SMARCB1, AASDH, and C15orf40, was successfully established and externally validated. Stratification of samples based on this gene signature revealed immunological differences between subgroups. The model was also shown to have independent prognostic value. Reverse-transcription quantitative polymerase chain reaction (RT-qPCR) and immunohistochemistry (IHC) confirmed that EXOSC8 was upregulated in osteosarcoma cell lines and tissues. Cell Counting Kit-8 (CCK-8), EdU incorporation, and clonogenic formation assays demonstrated the effect of EXOSC8 on proliferation. Conclusions: This study offers a novel framework for risk stratification in osteosarcoma, identifies seven novel LLPS-associated prognostic biomarkers, and presents valuable perspectives on the pathogenesis and potential therapeutic targets of osteosarcoma.