Abstract
BACKGROUND: Disulfidptosis is a novel type of cell death that cannot be explained by the previous cell death approaches. Research on disulfidptosis may open the door to new therapeutic strategies for cancer. Long non-coding RNA (lncRNA) exerts a regulatory role in the cell death process. However, the potential value of disulfidptosis-associated lncRNAs in pancreatic adenocarcinoma (PAAD) has not yet been explored. Therefore, the aim of this study is to identify DRLncI related lncRNAs as a basis for establishing new predictive biomarkers in PAAD. METHODS: The RNA-sequencing matrices of PAAD were extracted from The Cancer Genome Atlas (TCGA) cohort. Co-expression algorithm, Cox and the least absolute shrinkage and selection operator (LASSO) regression were conducted to determine a disulfidptosis-related lncRNA index (DRLncI). Kaplan-Meier method, Cox regression, and receiver operating characteristic algorithms were applied to assess the predictive stability and effectiveness of the DRLncI. Gene ontology, Gene Set Variation Analysis (GSVA) and tumour mutation burden analysis were employed for index-based mechanistic exploration. Additionally, the Estimation of STromal and Immune cells in MAlignant Tumor tissues using Expression data (ESTIMATE), single-sample Gene Set Enrichment Analysis (ssGSEA), Tumor Immune Estimation Resource (TIMER) platform and drug sensitivity were utilised to assess the predictive value of DRLncI for tumour immune microenvironment (TIME) and drug efficacy. In addition, consensus clustering algorithm was applied to distinguish PAAD subgroups with different molecular characteristics. RESULTS: Based on disulfidptosis-related lncRNAs, we established a DRLncI consisting of seven lncRNAs. Multi-validation showed that DRLncI had better predictive stability and sensitivity than age and other clinical features. Additionally, DRLncI can well differentiate individuals with different TIME. Furthermore, DRLncI-based consensus clustering algorithm divided all individuals into two clusters. Systematic evaluation showed that the cluster 1 population not only had better prognosis, but also showed higher immune cell levels and immune checkpoints expression. Finally, DRLncI and consensus clustering analysis based on DRLncI can help determine the sensitivity of patients to different chemotherapeutic agents and targeted drugs, providing a reference for personalized treatment. CONCLUSIONS: The DRLncI and the DRLncI-based consensus clusters developed in the present research help to stratify the prognosis of individuals with PAAD, determine clinical outcomes and differentiate between patients with different TIME, providing a basis for personalized and precise oncology treatment.