Disulfidptosis classification of pancreatic carcinoma reveals correlation with clinical prognosis and immune profile

Abstract

Background: Disulfidptosis, a novel form of metabolism-related regulated cell death, is a promising intervention for cancer therapeutic intervention. Although aberrant expression of long-chain noncoding RNAs (lncRNAs) expression has been associated with pancreatic carcinoma (PC) development, the biological properties and prognostic potential of disulfidptosis-related lncRNAs (DRLs) remain unclear.

Methods: We obtained RNA-seq data, clinical data, and genomic mutations of PC from the TCGA database, and then determined DRLs. We developed a risk score model and analyzed the role of risk score in the predictive ability, immune cell infiltration, immunotherapy response, and drug sensitivity.

Results: We finally established a prognostic model including three DRLs (AP005233.2, FAM83A-AS1, and TRAF3IP2-AS1). According to Kaplan-Meier curve analysis, the survival time of patients in the low-risk group was significantly longer than that in the high-risk group. Based on enrichment analysis, significant associations between metabolic processes and differentially expressed genes were assessed in two risk groups. In addition, we observed significant differences in the tumor immune microenvironment landscape. Tumor Immune Dysfunction and Rejection (TIDE) analysis showed no statistically significant likelihood of immune evasion in both risk groups. Patients exhibiting both high risk and high tumor mutation burden (TMB) had the poorest survival times, while those falling into the low risk and low TMB categories showed the best prognosis. Moreover, the risk group identified by the 3-DRLs profile showed significant drug sensitivity.

Conclusions: Our proposed 3-DRLs-based feature could serve as a promising tool for predicting the prognosis, immune landscape, and treatment response of PC patients, thus facilitating optimal clinical decision-making.

Keywords: Disulfidptosis; Drug sensitivity; Immune microenvironment; Long noncoding RNA (lncRNA); Pancreatic carcinoma; Prognostic signature.

Fig. 1

The Sankey relation between DRGs and DRLs

Fig. 2

Identification of the prognostic features of pancreatic carcinoma (PC) linked to DRLs. A Univariate cox forest map showing the top 21 prognostic DRLs. B LASSO coefficient profiles of the expression of DRLs. C Selection of the penalty parameter in the LASSO model via tenfold cross-validation. D The relationships between the three DRLs and DRGs. *, p < 0.05; **, p < 0.01; ***, p < 0.001

Fig. 3

Evaluation and validation of the independent prognostic ability of 3-DRLs signature model in training, testing, and all sets. A The distribution of patient with increasing risk scores. B The survival time of patients and risk scores. C The K-M survival analysis of survival status and overall survival (OS) of PC patients between two risk groups (The red line represents high-risk groups, and the blue line represents low-risk groups). D A univariate Cox regression analysis and multivariate Cox regression analysis of clinical variables and risk score

Fig. 4

Validation of the predictive model and construction of a nomogram combining clinical characteristics. A The ROC curves show the predictive accuracy in 1-, 3-, and 5-year of the predictive risk model. B The ROC curves show the predictive accuracy of the predictive risk model and clinicopathological characteristics. C ROC curves of the nomogram and clinical features demonstrating superior prediction of prognosis

Fig. 5

The K—M survival analysis of low- and high-risk patients with different clinical variables. A Age (> 65, ≤ 65); B Gender (Male, Female); C Stage (Stage I-II, Stage III-IV); D T stage (T1-2, T3-4); E N stage (N0, N1); F M stage (M0, M1); G Grade (G1-2, G3-4). The red line represents high-risk groups, and the blue line represents low-risk groups

Fig. 6

Biological functional and pathway enrichment analysis of the DRLs prognostic signature. A PCA about DRGs of patients in two risk groups. B PCA about DRLs of patients in two risk groups. C PCA about the three DRLs used in the predictive model of patients in two risk groups. D GO analysis reveals the diversity of BP, CC, and MF. E KEGG analysis shows the significantly enriched pathways. F GSEA analysis demonstrates the enriched pathways in two risk groups

Fig. 7

Analysis of immune cell infiltration in PC patients. A Differential expression of tumor microenvironment scores (Stromal Score, Immune Score, and ESTIMATE Score) between two risk groups. B Infiltration abundance of tumor immune cells in two risk groups. C Differential expression of immune functions scores between two risk groups. *, p < 0.05; **, p < 0.01; ***, p < 0.001

Fig. 8

Mutation profile and drug sensitivity analysis of the high- and low-risk groups. A-B Waterfall plots of somatic mutations in tumors in two risk groups. C Analysis of the difference for TMB between two risk groups. D The K—M survival analysis of PC patients between high- and low-TMB groups. E The K-M survival analysis of PC patients regarding TMB combined with risk score. F The violin plot of TIDE analysis for two risk groups

Fig. 9

The relationship between risk groups and drug sensitivity