Establishment of a prognostic risk prediction model incorporating disulfidptosis-related lncRNA for patients with prostate cancer

Abstract

Purpose: Prostate cancer (PCa) is one of the major tumor diseases that threaten men's health globally, and biochemical recurrence significantly impacts its prognosis. Disulfidptosis, a recently discovered cell death mechanism triggered by intracellular disulfide accumulation leading to membrane rupture, is a new area of research in the context of PCa. Currently, its impact on PCa remains largely unexplored. This study aims to investigate the correlation between long non-coding RNAs (lncRNAs) associated with disulfidptosis and the prognosis of PCa, seeking potential connections between the two.

Methods: Transcriptomic data for a PCa cohort were obtained from the Cancer Genome Atlas database. Disulfidptosis-related lncRNAs (DDRLs) were identified through differential expression and Pearson correlation analysis. DDRLs associated with biochemical recurrence-free survival (BRFS) were precisely identified using univariate Cox and LASSO regression, resulting in the development of a risk score model. Clinical factors linked to BRFS were determined through both univariate and multivariate Cox analyses. A prognostic nomogram combined the risk score with key clinical variables. Model performance was assessed using Receiver Operating Characteristic (ROC) curves, Decision Curve Analysis (DCA), and calibration curves. The functional impact of a critical DDRL was substantiated through assays involving CCK8, invasion, migration, and cell cloning. Additionally, immunohistochemical (IHC) staining for the disulfidptosis-related protein SLC7A11 was conducted.

Results: The prognostic signature included AC026401.3, SNHG4, SNHG25, and U73166.1 as key components. The derived risk score from these signatures stood as one of the independent prognostic factor for PCa patients, correlating with poorer BRFS in the high-risk group. By combining the risk score with clinical variables, a practical nomogram was created, accurately predicting BRFS of PCa patients. Notably, silencing AC026401.3 significantly hindered PCa cell proliferation, invasion, migration, and colony formation. IHC staining revealed elevated expression of the dithiosulfatide-related protein SLC7A11 in tumor tissue.

Conclusions: A novel prognostic signature for PCa DDRLs, possessing commendable predictive power, has been constructed, simultaneously providing potential therapeutic targets associated with disulfidptosis, among which AC026401.3 has been validated in vitro and demonstrated inhibition of PCa tumorigenesis after its silencing.

Keywords: AC026401.3; Biochemical recurrence-free survival; Disulfidptosis; Prostate cancer; lncRNA.

Fig. 1

Identification of disulfidptosis-related lncRNAs in prostate cancer cohort. (A) Heatmap of the differentially expressed lncRNAs showing expression signature in patients with PCa. (B) Volcano plot of differentially expressed lncRNAs. (C) Sankey diagram of disulfidptosis-related lncRNAs.

Fig. 2

Identifying candidate disulfidptosis-related lncRNAs associated with the Biochemical Recurrence-Free Survival of prostate cancer patients. Biochemical Recurrence-Free Survival of AC008610.1, AC026401.3, AL023803.1, SNHG4, SNHG25 and U73166.1 in TCGA. (A) Univariate Cox regression analysis of disulfidptosis-related lncRNAs. (B, C) Determination of the best penalty value. LASSO regression of screening disulfidptosis-related lncRNAs with patients’ prognosis. (D) Biochemical Recurrence-Free Survival of AC008610.1, AC026401.3, AL023803.1, SNHG4, SNHG25 and U73166.1 in TCGA

Fig. 3

Validation of risk score through data derived from TCGA. The distribution of the four disulfidptosis-related lncRNAs risk score and survival status for each patient in training set (A) and validation set (B); Kaplan–Meier curve of survival time in high- and low-risk groups in training set (C) and validation set (D); ROC curve of 1-, 3- and 5-year BRFS in training set (E) and validation set (F)

Fig. 4

Prognostic Nomogram Establishment. Univariate (A) and multivariate (B) Cox regression analysis of clinical information (Age, T stage, N stage, Gleason score and risk score). (C) The prognostic nomogram that combines clinicopathological factors and risk scores predicts the 1-, 3-, and 5-year BRFS of PCa patients. Calibration curves, the Decision Curve Analysis (DCA) curves and ROC curves evaluating the prognostic efficiency of the very model in the training set (D) and validation set (E), respectively

Fig. 5

Cellular Functional Validation of AC026401.3. (A) Transfection efficiency of the control group and the siRNA group in both cells. (B, C) CCK-8 assay results indicate lower cell viability in the siRNA group compared to the control group in both cells. The p-value for the fifth day is noted. (D) Significantly reduced invasion and migration capabilities were observed in the siRNA group compared to the control group through invasion and migration assays in both cells. (E) Significantly decreasing of colony-forming ability was observed in the siRNA group compared to the control group in the cell cloning experiments. (*, p < 0.05; **, p < 0.01; ***, p < 0.001)

Fig. 6

Immunohistochemical analysis of SLC7A11 expression in prostate tissues. Representative micrographs of normal (upper panels) and tumor (lower panels) prostate tissues stained for SLC7A11 at 200x and 400x magnification (left and right in each tissue type, respectively). Insets show higher magnification of the selected areas. The bar graph on the right quantifies the relative areal density of SLC7A11 staining, with significant increase in tumor tissue compared to normal (**, p < 0.01)