Development of a novel disulfidptosis-related lncRNA signature for prognostic and immune response prediction in clear cell renal cell carcinoma

Abstract

Disulfidptosis, a novel form of regulated cell death, occurs due to the aberrant accumulation of intracellular cystine and other disulfides. Moreover, targeting disulfidptosis could identify promising approaches for cancer treatment. Long non-coding RNAs (lncRNAs) are known to be critically implicated in clear cell renal cell carcinoma (ccRCC) development. Currently, the involvement of disulfidptosis-related lncRNAs in ccRCC is yet to be elucidated. This study primarily dealt with identifying and validating a disulfidptosis-related lncRNAs-based signature for predicting the prognosis and immune landscape of individuals with ccRCC. Clinical and RNA sequencing data of ccRCC samples were accessed from The Cancer Genome Atlas (TCGA) database. Pearson correlation analysis was conducted for the identification of the disulfidptosis-related lncRNAs. Additionally, univariate Cox regression analysis, Least Absolute Shrinkage and Selection Operator Cox regression, and stepwise multivariate Cox analysis were executed to develop a novel risk prognostic model. The prognosis-predictive capacity of the model was then assessed using an integrated method. Variation in biological function was noted using GO, KEGG, and GSEA. Additionally, immune cell infiltration, the tumor mutational burden (TMB), and tumor immune dysfunction and exclusion (TIDE) scores were calculated to investigate differences in the immune landscape. Finally, the expression of hub disulfidptosis-related lncRNAs was validated using qPCR. We established a novel signature comprised of eight lncRNAs that were associated with disulfidptosis (SPINT1-AS1, AL121944.1, AC131009.3, AC104088.3, AL035071.1, LINC00886, AL035587.2, and AC007743.1). Kaplan-Meier and receiver operating characteristic curves demonstrated the acceptable predictive potency of the model. The nomogram and C-index confirmed the strong correlation between the risk signature and clinical decision-making. Furthermore, immune cell infiltration analysis and ssGSEA revealed significantly different immune statuses among risk groups. TMB analysis revealed the link between the high-risk group and high TMB. It is worth noting that the cumulative effect of the patients belonging to the high-risk group and having elevated TMB led to decreased patient survival times. The high-risk group depicted greater TIDE scores in contrast with the low-risk group, indicating greater potential for immune escape. Finally, qPCR validated the hub disulfidptosis-related lncRNAs in cell lines. The established novel signature holds potential regarding the prognosis prediction of individuals with ccRCC as well as predicting their responses to immunotherapy.

Figure 1

Flow chart of the present study.

Figure 2

Identification of disulfidptosis-related lncRNAs and prognostic value preparation in ccRCC. (A) Sankey diagram for disulfidptosis genes and disulfidptosis-related lncRNAs. (B) LASSO coefficient profiles of 181 disulfidptosis-related lncRNAs. (C) Selection of tuning parameter lambda in the LASSO Cox regression model using ten-fold cross-validation. (D) Heatmap for the correlation between disulfidptosis genes and 8 disulfidptosis-related lncRNAs.

Figure 3

Prognostic value of the risk model in the train, test, and entire set. (A–C) Risk score distribution of patients in the low-risk and high-risk groups. (D–F) Survival status and time in the low-risk and high-risk groups. (G–I) Hierarchical clustering analysis of 8 disulfidptosis-related lncRNAs between the low-risk and high-risk groups. (J–L) Kaplan–Meier survival curves of OS between low-risk and high-risk groups. (M–O) Kaplan–Meier survival curves of progression-free survival between low-risk and high-risk groups.

Figure 4

Kaplan–Meier survival curves for low-risk and high-risk populations by different clinical variables. (A,B) Survival curve after grouping according to age (≤ 65, > 65). (C,D) Survival curve after grouping according to sex (male, female). (E,F) Survival curve after grouping according to clinical stages (stage1-2, stage3-4). (G,H) Survival curve after grouping according to T stage (T1-2, T3-4). (I,J) Survival curve after grouping according to N stage (N0, N1). (K,L) Survival curve after grouping according to M stage (M0, M1).

Figure 5

Independent prognostic analysis and further validation of the risk model. Forest plots of (A) univariate and (B) multivariate Cox regression analysis showed the effects between clinical characteristics (including the risk signature) and OS. (C) Time-dependent ROC curves of OS at 1-, 3- and 5-year. (D) Predictive accuracy of the risk model compared with clinicopathologic characteristics. (E) Concordance index of the risk model and other clinical information. (F) Nomogram combining the risk signature and clinical factors. (G) Calibration curves for the nomogram-predicted OS at 1, 3, and 5 years (C-index = 0.810, p < 0.001).

Figure 6

PCA between low-risk and high-risk groups. (A) PCA of all genes. (B) PCA of disulfidptosis genes. (C) PCA of disulfidptosis-related lncRNAs. (D) PCA of 8 risk lncRNAs.

Figure 7

Functional analysis of the risk model. (A,B) GO analysis demonstrated the richness of molecular biological processes (BP), cellular components (CC), and molecular functions (MF). (C,D) KEGG pathway analysis showed the significantly enriched pathways. (E,F) GSEA analysis based on KEGG pathway database of high-risk group and low-risk group. (G,H) GSEA analysis based on C7 gene set of high-risk group and low-risk group.

Figure 8

Differences in the tumor immune microenvironment between the low- and high-risk groups. (A) The abundance ratios of immune cells in the ccRCC samples. (B) Differentially expressed immune cells in the high- and low-risk score groups. (C) Immune function analysis in the high-risk and low-risk score groups. (D) Violin diagram comparing StromalScore, ImmuneScore and ESTIMATEScore between the low-risk and high-risk groups.

Figure 9

Relationship of model scores to TMB and TIDE. (A,B) Waterfall plots of somatic mutation characteristics in the two groups. (C) TMB between the low-risk and high-risk groups. (D) Kaplan–Meier analysis of the effect of TMB status on OS. (E) Kaplan–Meier analysis for OS of patients categorized by combing TMB status and risk score. (F) TIDE scores between the two groups.

Figure 10

Drug sensitivity. (A) Axitinib was more sensitive in the low-risk group. (B) Savolitinib was more effective in the high-risk group.

Figure 11

External validation of disulfidptosis-associated lncRNAs. (A,B) The expression of LINC00886 and SPINT1-AS1 in tumor tissues and paired normal tissues of TGCA database. (C,D) OS analysis of LINC00886 and SPINT1-AS1 in the Kaplan–Meier Plotter datasets.

Figure 12

Expression levels of hub risk disulfidptosis-associated lncRNAs in ccRCC cell lines and HK-2 cell. (A) The expression levels of LINC00886 measured by qPCR. (B) The expression levels of SPINT1-AS1 measured by qPCR. The significant differences from the HK-2 were indicated with star (*p < 0.05; **p < 0.01; ***p < 0.001, ****p < 0.0001).