Abstract
Disulfidptosis is a newly discovered formation of programmed cell death. However, the significance of disulfidptosis in pancreatic adenocarcinoma remains unclear. Our investigation aims to elucidate the significance of disulfidptosis in pancreatic ductal adenocarcinoma by integrating diverse datasets, including bulk RNA sequencing data, microarray profiles, single-cell transcriptome profiles, spatial transcriptome data, and biospecimens. Utilizing various bioinformatics tools, we screened disulfidptosis-related genes based on single-cell RNA sequencing profiles, subsequently validating them through enrichment analysis. An 8-gene disulfidptosis-related prognostic signature was established by constructing massive LASSO-Cox regression models and validated by multiple external PDAC cohorts. Evaluation methods, such as Kaplan-Meier curves, ROC curves, time-dependent ROC curves, and decision curve analysis, were employed to assess the prognostic signature's reliability. High disulfidptosis-related scores were associated with a poorer prognosis and diminished sensitivity to immune checkpoint blockade. Further investigation uncovered that the potential components of elevated DPS involve malignant tumor hallmarks, extensive interactions between myCAFs and tumor cells, and the exclusion of immune cells. Cell-cell communication analysis highlighted myCAFs' role in signaling, potentially influencing tumor cells towards increased malignancy through collagen, laminin, and FN1 signaling networks. Spatial transcriptome analysis confirmed the crosstalk between myCAFs and tumor cells. Biospecimens including 20 pairs of PDAC samples and adjacent normal tissues further demonstrated the robustness of DPS and its correlation with CAF markers. In conclusion, our study introduces a novel disulfidptosis-related signature with high efficacy in patient risk stratification, which has the ability to predict the sensitivity to immune checkpoint blockade.