Multi-omics-based construction of a palmitoylation-driven prognostic model reveals tumor immune phenotypes in osteosarcoma
摘要
Osteosarcoma (OS), the leading primary malignancy of bone in adolescents, is known for its aggressive metastatic behavior and poor responsiveness to conventional therapies. As a lipid-mediated post-translational modification, protein palmitoylation has gained attention for its pivotal role in modulating oncogenic signaling pathways and facilitating tumor immune escape. However, its prognostic value and functional role in OS remain unclear.
MethodsTranscriptomic and clinical data from the TARGET and GEO cohorts were used to identify palmitoylation-related prognostic genes. The palmitoylation-related prognostic signature (PPS) was constructed using univariate Cox and LASSO regression. The model was validated by GSE39058 and assessed via survival analysis, ROC curves, and nomogram construction. Functional enrichment (GO, KEGG, GSVA) and immune infiltration analyses were performed. Single-cell expression profiles were explored using the TISCH2 database, and the predictive value of PPS for immunotherapy response was evaluated in the IMvigor210 cohort.
ResultsA three-gene PPS (ZDHHC3, ZDHHC21, ZDHHC23) was identified and shown to independently predict survival in OS. Elevated PPS levels correlated with unfavorable clinical outcomes, diminished immune cell presence, and suppressed immune checkpoint molecule levels. Functional analysis revealed enrichment of oncogenic and immunosuppressive pathways in the high-PPS group. Single-cell analysis confirmed PPS gene expression in malignant and immune cells. In the IMvigor210 cohort, high PPS predicted worse response to anti-PD-L1 immunotherapy.
ConclusionsThis study establishes a novel palmitoylation-related prognostic signature in osteosarcoma, which reflects tumor aggressiveness and immune evasion. PPS holds promise as both a stratification indicator and an intervention point for osteosarcoma treatment.