CELL INTELLIGENCE · 肿瘤细胞治疗研究

Osteosarcoma, the most common primary malignant bone tumour, presents significant treatment challenges due to its complex tumour microenvironment and the development of chemoresistance. This study employs single-cell transcriptomics to investigate chemotherapy-induced changes in osteosarcoma at both the cellular and molecular levels. Single-cell RNA sequencing data were analysed to identify cell subpopulations and their responses to chemotherapy. Differential gene expression and pathway enrichment analyses were performed to elucidate chemotherapy-induced changes. Additionally, we developed and validated a predictive model based on pyroptosis-related genes, named Pyroscore, using 101 different machine-learning algorithms. Chemotherapy led to an increased proportion of osteoclasts, endothelial cells, mesenchymal stem cells and pericytes, while decreasing T and NK cells, B cells, chondroblasts, monocytes and macrophages. Chemotherapy markedly upregulates the pyroptosis pathway in tumour cells, suggesting that chemotherapy induces programmed cell death in cancer cells through the activation of pyroptosis. Metabolic pathway analysis revealed significant inhibition of sulphur metabolism, starch and sucrose metabolism, pentose phosphate pathway, inositol phosphate metabolism, nitrogen metabolism and fatty acid metabolism. The Pyroscore model, which incorporates BAK1, CASP1, CASP5 and CASP6, demonstrated robust prognostic value across multiple data sets, with high scores correlating with improved survival outcomes. This study highlights the impact of chemotherapy on osteosarcoma cell subpopulations and the tumour microenvironment. The activation of the pyroptosis pathway and the development of the pyroscore prognostic model provide new insights into the mechanisms of chemotherapy response and potential therapeutic targets. These findings underscore the importance of personalized treatment strategies in improving outcomes for osteosarcoma patients.