Integrative analysis of programmed cell death pathways reveals prognostic biomarkers and immune infiltration signatures in coronary artery disease

Introduction

Coronary artery disease (CAD), also known as coronary heart disease (CHD), is a major cause of morbidity and mortality on a global scale (1). This disease is marked by the constriction or obstruction of coronary arteries caused by the accumulation of atherosclerotic plaques, which decreases blood flow to the heart muscle (2). This condition can result in chest pain (angina), heart attacks (myocardial infarctions), and other serious cardiovascular events (3). The pathogenesis of CAD is complex and multifactorial, involving genetic, environmental, and lifestyle factors (4-7). Both inflammation and immune responses are key contributors to the development and progression of atherosclerosis, the underlying pathology of CAD (8,9). Understanding the molecular mechanisms and cellular interactions involved in CAD is essential for developing effective diagnostic, prognostic, and therapeutic strategies.

Cell death is an essential process for the development and maintenance of multicellular organisms (10,11). It ensures the removal of damaged, infected, or unnecessary cells, thereby maintaining cellular homeostasis and tissue health. There are different types of cell death, including necrosis and programmed cell death (PCD). While necrosis is typically an uncontrolled process resulting from acute cellular injury, leading to inflammation and tissue damage (6), PCD is a regulated and orderly process that plays a vital role in normal development and disease prevention (12).

PCD encompasses several distinct pathways, including apoptosis, autophagy, necroptosis, pyroptosis, and others, each with distinct molecular mechanisms and physiological roles (13-16). Apoptosis, the most well-studied form, involves a highly regulated cascade of events leading to cell death without provoking an inflammatory response (17). Autophagy, on the other hand, is a process where cells degrade their own components through the lysosomal pathway, which can lead to cell survival or death depending on the context (18,19). Necroptosis and pyroptosis are forms of inflammatory cell death that are triggered by specific signaling pathways, contributing to the immune response and inflammation (20-22).

In the context of CAD, the regulation of PCD is particularly significant. Dysregulated PCD contributes to the pathology of CAD by promoting plaque instability, inflammation, and adverse vascular remodeling. Understanding the specific roles and regulation of different PCD pathways in CAD could provide offer new perspectives on disease mechanisms and uncover potential therapeutic targets for treatment and prevention. In this study, we carried out an extensive analysis of public databases to identify prognostic PCD-related genes. We then created a prognostic scoring system based on these genes to forecast patient outcomes and their response to therapies. Our findings were validated through a blend of advanced bioinformatics and machine learning methods. This multifaceted approach aimed to enhance our understanding of the role of PCD in CAD and to pinpoint potential biomarkers and therapeutic targets for improving patient outcomes. We present this article in accordance with the TRIPOD reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2024-2283/rc).

Methods

Data collection

To investigate the underlying mechanisms of CAD, we utilized both bulk RNA sequencing and single-cell RNA sequencing datasets. The bulk RNA datasets included GSE42148, comprising samples from 11 normal individuals and 13 CAD patients, and GSE66360, with samples from 22 normal individuals and 21 CAD patients. The single-cell RNA sequencing datasets included GSE114727 (breast cancer), GSE108989 [colorectal cancer (CRC)], GSE140228 [liver hepatocellular carcinoma (LIHC)], GSE176031 (prostate adenocarcinoma), and GSE139555 (uterine corpus endometrial carcinoma). These datasets facilitated the identification of differentially expressed genes (DEGs) and biomarkers associated with CAD, as well as to explore the cellular heterogeneity and molecular mechanisms underlying the disease. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

We also analyzed pan-cancer gene expression and clinical data obtained using the TCGAplot R package, which provided TPM gene expression data and clinical data with normal controls.

For the PCD gene set, we collected 18 types of PCD and key regulatory genes, including 580 genes related to apoptosis, 367 to autophagy, 7 alkaliptosis-related genes, and 338 anoikis-related genes. Additionally, we included 19 genes associated with cuproptosis, 15 with enteric cell death, 87 with ferroptosis, 34 with immunogenic cell death, 220 with lysosome-dependent cell death, and 101 with necroptosis. The dataset also encompassed 8 genes related to netotic cell death, 24 to NETosis, 5 to oxeiptosis, 52 to pyroptosis, 9 to parthanatos, and 66 to paraptosis. Moreover, the gene set comprised 8 methuosis-related genes and 23 entosis-related genes (23).

Analysis of key PCD forms of CAD and differential gene expression

Based on the relevant genes contained in each PCD, we used the Gene Set Variation Analysis (GSVA) R package to evaluate each sample for PCD model. The diagnostic effect of each PCD method on CAD was evaluated by receiver operating characteristic (ROC) curve. At the same time, the differential expression forms of PCD-related genes in CAD were analyzed by the limma algorithm. The thresholds for identifying DEGs were set at |log fold change (FC)| >0.58 and P value <0.05.

Core gene selection for candidate PCD

To further identify key PCD genes, three machine learning algorithms were employed: least absolute shrinkage and selection operator (LASSO) logistic regression, support vector machine (SVM), and random forest (RF). The LASSO logistic regression model was executed using the ‘glmnet’ package in R, with parameters set to ‘seed =3’ and ‘family = binomial’, iterated 1,000 times to prevent overfitting. SVM, a supervised machine learning technique, was run using the ‘caret’ package with ‘seed = 3’ and ‘method = svmLinear’. For the RF model, parameters were set to ‘seed =3’, ‘ntree =150’, and ‘Importance >0.3’ for DEGs, and it was executed using the ‘randomForest’ package. The intersection of genes identified by LASSO, SVM, and RF were selected as the core genes.

Construction of a risk model based on core genes

To evaluate the predictive ability of key PCD DEGs for CAD, a predictive model was established using LASSO regression analysis. The GSE42148 dataset served as the discovery cohort, and GSE66360 was used for validation. The risk score was calculated based on the expression of key PCD DEGs and the LASSO regression coefficients. To identify genes associated with the apoptotic phenotype, differential expression analysis was performed using the “limma” package in R, grouping high- and low-risk groups in the GSE42148 dataset.

Functional enrichment analysis

Gene Ontology (GO) analysis, encompassing biological processes (BP), molecular functions (MF), and cellular components (CC), was conducted using the “clusterProfiler” package for GO annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment of DEGs between high- and low-risk groups in the combined dataset. False discovery rate (FDR) <0.05 was considered significant, with selection criteria set at P_adj <0.05 and q<0.05, and the Benjamini-Hochberg (BH) method was used for P value correction. Immune-related genes, totaling 782 genes across 28 cell types, were sourced from a gene set database (24). We analyzed the infiltration levels of immune cells within the combined dataset utilizing the single-sample Gene Set Enrichment Analysis (ssGSEA) algorithm available in the GSVA package in R.

Single-cell analysis

Single-cell transcriptome data were processed in R using the Seurat pipeline. Quality control excluded cells with >20% ribosomal gene expression or >3% erythrocyte gene expression. Data normalization was performed via NormalizeData, followed by identification of highly variable genes using FindVariableFeatures. These genes were scaled via ScaleData, then subjected to principal component analysis (PCA) for dimensionality reduction. Cell clustering was achieved through FindNeighbors (k=20) and FindClusters (resolution =0.5). Pseudotime trajectories were reconstructed using Monocle3, with regulatory T cells (Treg) cell clusters defining the trajectory root. Significant trajectory-associated genes were identified via graph_test (q_value <0.01, Moran’s I >0.05).

Construction of a cell model of CHD

AC16 cell lines (Shanghai Zhong Qiao Xin Zhou Biotechnology Co., Ltd., Shanghai, China) were cultured under hypoxic conditions to simulate myocardial ischemia. Cells were placed in a hypoxia incubator with an oxygen concentration maintained at 1–2%. To enhance the hypoxic response, CoCl₂ was added to the culture medium at a final concentration of 400 µM. Hypoxia was maintained for 2–4 hours to induce cellular ischemia. Subsequently, cells were returned to normoxic conditions (21% O₂) for 1–2 hours to simulate reoxygenation.

Western blot analysis

Total protein was extracted from cells using RIPA lysis buffer supplemented with protease and phosphatase inhibitors. The lysates were incubated on ice for 30 minutes and centrifuged at 12,000 ×g for 15 minutes at 4 ℃. The supernatants were collected, and protein concentrations were determined using a BCA protein assay kit. Equal amounts of protein (30 µg) were separated by 10–12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to polyvinylidene fluoride (PVDF) membranes. Membranes were blocked with 5% non-fat dry milk in Tris Buffered Saline with Tween (TBST) for 1 hour at room temperature, then incubated overnight at 4 ℃ with the following primary antibodies: anti-GZMB (1:1,000, CST, Boston, USA), anti-SFN (1:1,000, CST), and anti-GAPDH (1:2,000, Abcam, Cambridge, UK) as a loading control. After washing with TBST three times for 5 minutes each, membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibody for 1 hour at room temperature. Protein bands were visualized using enhanced chemiluminescence (ECL) reagent and imaged with a chemiluminescence detection system. Band intensities were quantified using ImageJ software. Expression levels of GZMB and SFN were normalized to GAPDH. All experiments were performed in triplicate.

Statistical analysis

All statistical analyses were performed using R v4.2.1. To compare the results between different groups, one-way analysis of variance (ANOVA) or Student’s t-test was used. For survival analysis, Kaplan-Meier curves, log-rank tests, or Cox proportional hazards regression models were applied. A P value of less than 0.05 was considered statistically significant, with the following criteria: *P<0.05, **P<0.01, ***P<0.001; ns: not significant.

Results

Differential expression analysis of PCD genes

In the differential expression analysis of PCD genes associated with CAD, we discovered significant differential expression patterns. Using the discovery cohort GSE42148, we set thresholds for differential expression at |logFC| >0.7 and P value <0.05. This analysis identified 923 DEGs, with 392 downregulated and 571 upregulated (Figure 1A,1B). Furthermore, the number of differentially expressed genes for each PCD type is shown in a bar chart (Figure 1C), seven PCD types were found to have no differentially expressed genes associated with CAD and were excluded from further analysis. Additionally, PCD types with fewer than three differentially expressed genes were also excluded, highlighting apoptosis, anoikis, and autophagy as the most represented types. This analysis indicates significant differential expression in several PCD pathways, suggesting their potential involvement in CAD pathogenesis.

Figure 1 Differential expression analysis of PCD-related genes in CAD. (A) Volcano plot of differential expression analysis. (B) Heatmap of differentially expressed genes, with clustering indicating groups of genes with similar expression profiles. (C) The bar chart of the number of differentially expressed genes for each PCD type. CAD, coronary artery disease; DEGs, differentially expressed genes; FC, fold change; PCD, programmed cell death.

To explore the relationship between PCD types and CAD, we performed ssGSEA analysis based on the differentially expressed PCD genes to obtain the ssGSEA scores for each PCD type (Table S1). The heatmap (Figure 2A) indicates that NETosis, apoptosis, and ferroptosis exhibit notably higher expression scores in CAD samples compared to normal samples, while other PCD types like autophagy and lysosome also show higher expression in CAD but are less pronounced. This suggests a distinct role for these PCD types in the context of CAD. The correlation matrix (Figure 2B) reveals significant relationships between different PCD types based on their ssGSEA scores. Anoikis shows a strong positive correlation with apoptosis, autophagy and NETosis. Apoptosis is highly correlated with autophagy, lysosome and NETosis. Autophagy has significant positive correlations with lysosome, while NETosis is significantly correlated with pyroptosis. These correlations indicate that certain PCD pathways, such as anoikis and apoptosis, are closely related and may share common regulatory mechanisms or participate in similar BP, suggesting a coordinated role in the pathology of CAD. The diagnostic ability of each PCD type for CAD was assessed using ROC curves. The results (Figure 2C) demonstrate that, except for pyroptosis, all PCD types had area under the curve (AUC) values greater than 0.6, indicating good predictive ability. Therefore, only PCD types with AUC values greater than 0.6 were retained for further analysis.

Figure 2 ssGSEA analysis of PCD pathways in CAD. (A) Heatmap of ssGSEA scores for each PCD type. (B) Correlation matrix of ssGSEA scores for different PCD types. (C) ROC curves for different PCD types. *, P<0.05; **, P<0.01; ***, P<0.001. AUC, area under the curve; CAD, coronary artery disease; PCD, programmed cell death; ROC, receiver operating characteristic; ssGSEA, single-sample Gene Set Enrichment Analysis.

Identification of core apoptosis-related genes using multiple machine learning methods

To identify core apoptosis-related genes, we integrated the remaining PCD-related genes and employed three machine learning methods: LASSO regression, RF, and SVM (Figure 3A-3D). LASSO regression is a regularization technique used for variable selection and regularization to improve the prediction accuracy and interpretability of the model. This method helps in identifying the most influential genes while eliminating less significant ones, reducing the risk of overfitting. RF is an ensemble learning technique that constructs multiple decision trees during training and outputs the mean prediction of the individual trees. The RF model is robust to overfitting and provides insights into the relative importance of each gene. SVM is a supervised learning algorithm used for classification and regression tasks. Initially, the accuracy improves significantly with the addition of more variables, but it stabilizes after a certain point, indicating the optimal number of variables for the best model performance. The importance of each variable was evaluated, and genes with an importance score greater than 0.3 were considered significant (Figure 3E). By combining the results from these three methods, we identified the core genes that were consistently selected across all methods. The Venn diagram (Figure 3F) illustrates the overlap of significant genes identified by LASSO, RF, and SVM. The intersection of these methods resulted in five core genes: GZMB, CXCR4, ATP6V0A4, SFN and GSDMA.

Figure 3 Identification of core PCD genes using machine learning methods. (A) LASSO regression coefficient path plot. (B) MSE plot for LASSO regression. Each red dot indicates the MSE for a specific λ value, and the gray bars represent the standard error of the MSE. (C) RF error plot. The solid black line indicates the overall error rate, the dashed red line represents the error rate for one class, and the dotted green line represents the error rate for the other class. As the number of trees increases, the error rate stabilizes, indicating the optimal number of trees needed for the model to achieve a stable and accurate prediction. (D) Variable importance plot for the RF model. (E) The accuracy plot from a SVM model, showing cross-validation accuracy for different numbers of variables included in the model. (F) Venn diagram showing the intersection of significant genes identified by LASSO, RF, and SVM. LASSO, least absolute shrinkage and selection operator; MSE, mean squared error; PCD, programmed cell death; RF, random forest; SVM, Support Vector Machine.

Construction of a risk model based on key genes

To develop a risk model for CAD, we utilized the expression levels of the identified key genes as independent variables and the presence of CAD as the dependent variable. A binary classification LASSO regression was performed on the discovery cohort to construct the model. The risk score for each sample was calculated based on the regression coefficients and the expression levels of the key genes (Table 1). The samples were then divided into high and low-risk groups using the median risk score.

The risk score distribution plots (Figure 4A) illustrated the separation of samples into high and low-risk groups, clearly distinguishing between normal and CAD samples. The genes CXCR4 and SFN exhibit higher expression in the high-risk samples compared to the low-risk samples. Conversely, GZMB, ATP6V0A4 and GSDMA show lower expression in the high-risk samples. These findings indicate that elevated expression of CXCR4 and SFN in high-risk individuals may contribute to the pathology of CAD, while reduced expression of GZMB, ATP6V0A4 and GSDMA may suggest protective roles against CAD. The ROC curve (Figure 4B) further demonstrated the predictive accuracy of the risk model, achieving an AUC of 1.00, indicating perfect discrimination between CAD and non-CAD samples. Differential expression analysis revealed key genes associated with the risk score (Figure 4C). The heatmap (Figure S1) displayed the expression levels of these differentially expressed genes across all samples, grouped into high and low-risk categories, revealing distinct gene expression patterns associated with each risk group.

Figure 4 Construction and functional analysis of the PCD-related risk score model. (A) Distribution of risk scores among patients, with high and low-risk groups (top panel), and distribution among normal and CAD patients (middle panel). Heatmap shows the expression levels of key PCD-related genes (CXCR4, SFN, GZMB, ATP6V0A4, and GSDMA) across patients with increasing risk scores (bottom panel). (B) ROC curve for the risk score model demonstrating its high predictive accuracy (AUC =1.00). (C) Volcano plot illustrating the differentially expressed genes between high and low-risk groups. (D) GO enrichment analysis of differentially expressed genes. (E) KEGG pathway enrichment analysis of differentially expressed genes. AUC, area under the curve; CAD, coronary artery disease; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; PCD, programmed cell death; ROC, receiver operating characteristic.

To explore the biological significance of these differentially expressed genes, we conducted enrichment analyses. The GO enrichment analysis (Figure 4D) identified significant BP, such as gland morphogenesis and multi-organism reproductive process. KEGG pathway enrichment analysis (Figure 4E) highlighted pathways such as efferocytosis and primary bile acid biosynthesis, suggesting key BP involved in the pathology of CAD.

To validate the predictive ability of our risk model, we further evaluated the distribution of risk scores in the validation cohort, GSE66360. The risk score distribution plots (Figure 5A) show a clear separation of samples into high and low-risk groups, analogous to the discovery cohort. The correlation of risk scores with key gene expression was also examined in the validation cohort. Consistent with the discovery cohort, elevated expression of CXCR4 and SFN was observed in high-risk samples. These findings further confirm the association of these genes with CAD risk. The ROC curve analysis was performed on the validation cohort, demonstrating an AUC of 0.948 (Figure 5B). This high AUC value indicates that the risk model maintains strong predictive performance in an independent dataset, successfully distinguishing CAD patients from non-CAD individuals.

Figure 5 Validation of the PCD-related risk score model. (A) Distribution of risk scores among patients in the validation cohort, with high and low-risk groups (top panel), and normal vs. CAD patients (middle panel). The heatmap shows the expression levels of key PCD-related genes (ATP6V0A4, SFN, CXCR4, and GZMB) across patients with increasing risk scores (bottom panel). (B) ROC curve for the risk score model in the GSE66360 validation cohort, demonstrating its predictive accuracy (AUC =0.948). AUC, area under the curve; CAD, coronary artery disease; PCD, programmed cell death; ROC, receiver operating characteristic.

Immune infiltration analysis of key genes

Using ssGSEA, we conducted a comprehensive analysis of immune cell infiltration to explore the relationship between key gene expressions and immune cell populations. Significant differences in immune cell infiltration were observed between CAD patients and normal individuals, as illustrated in the box plot (Figure 6A). This analysis identified several immune cells with significant differences, including central memory CD4 T cells, memory B cells, eosinophils and mast cells. Subsequent correlation analysis of these differentially infiltrated immune cells with the expression levels of key genes (CXCR4, SFN, GZMB, ATP6V0A4, and GSDMA) revealed notable associations (Figure 6B, Table 2). The correlation matrix highlighted significant positive correlations between central memory CD4 T cells and memory B cells. Similarly, eosinophils showed positive correlations with mast cells. Further analysis revealed distinct correlation patterns between key genes and specific immune cells. For instance, CXCR4 and SFN were significantly correlated with increased infiltration of eosinophils and mast cells, while GSDMA and GZMB exhibited negative correlations with memory B cells (Figure 6C-6F). These findings suggest a complex interplay between gene expression and immune cell infiltration, highlighting the potential role of CXCR4 and SFN in promoting an inflammatory microenvironment in CAD.

Figure 6 Immune infiltration analysis of key genes. (A) Box plot of immune cell infiltration between CAD patients and normal individuals based on ssGSEA scores. (B) Correlation matrix of the relationships between differentially infiltrated immune cells and key gene expressions. (C-F) Correlation analysis between key genes and central memory CD4 T cells (C), eosinophils (D), mast cells (E) and memory B cells (F). *, P<0.05; **, P<0.01; ***, P<0.001. CAD, coronary artery disease; ns, not significant; ssGSEA, single-sample Gene Set Enrichment Analysis.

GZMB and SFN affect apoptosis of CAD

In order to verify the potential role of GZMB and SFN in promoting the inflammatory microenvironment of CAD, we constructed a CHD cell model and detected the expression of GZMB and SFN in CHD cells by protein immunoblotting. The results showed that the expression of GZMB and SFN genes was increased in CHD (Figure 7A). In addition, we constructed GZMB and SFN overexpression cell lines and detected the effects of GZMB and SFN on cell apoptosis by flow cytometry. The results showed that overexpression of GZMB and SFN promoted cell apoptosis (Figure 7B,7C).

Figure 7 Functional verification of GZMB and SFN. (A) Western blot detection of GZMB and SFN protein expression levels after establishing coronary heart disease model. (B,C) Flow cytometry was used to detect cell apoptosis after overexpression of GZMB and SFN. ****, P<0.0001. FITC, fluorescein isothiocyanate; PI, propidium iodide; OE, overexpression.

Pan-cancer analysis of key genes

Using TCGAplot, we conducted a comprehensive pan-cancer analysis to examine the differential expression and prognostic implications of key genes (CXCR4, SFN, GZMB, ATP6V0A4, and GSDMA) across various cancer types. This analysis provided insights into the roles of these genes beyond CAD and their potential involvement in cancer biology. The differential expression heatmap (Figure 8A) reveals significant variations in the expression levels of the key genes across multiple cancer types. Notably, CXCR4 was found to be differentially expressed in eight cancers, including over-expressed in glioblastoma multiforme (GBM), kidney renal papillary cell carcinoma (KIRP), kidney renal clear cell carcinoma (KIRC) and breast invasive carcinoma (BRCA). SFN showed significant differential expression in fifteen cancers such as LIHC and uterine corpus endometrial carcinoma (UCEC). GZMB was differentially over-expressed in eight cancers, including esophageal carcinoma (ESCA) and colon adenocarcinoma (COAD). ATP6V0A4 exhibited significant changes in expression in eight cancers, such as over-expressed in UCEC and lung adenocarcinoma (LUAD), down-expressed in KIRC and KIRP. GSDMA showed differential expression in five cancers including GBM and COAD. These findings underscore the diverse roles of these key genes in various oncogenic processes and their potential as biomarkers for specific cancer types (Table S1).

Figure 8 Pan-cancer analysis of key genes. (A) Heatmap depicting the differential expression of key genes (GZMB, CXCR4, SFN, ATP6V0A4, and GSDMA) across various cancer types. Red indicates upregulation, blue indicates downregulation, and white indicates no significant change. (B) Heatmap showing the prognostic significance of key genes across different cancer types. Red indicates poor prognosis, green indicates favorable prognosis, and white indicates no significant correlation with prognosis. DEGs, differentially expressed genes; ns, not significant.

Further, we analyzed the prognostic significance of these key genes in different cancers. The prognostic heatmap (Figure 8B) highlights the correlation between gene expression and patient outcomes: CXCR4 expression was associated with poorer prognosis in three cancers such as skin cutaneous melanoma (SKCM) and head and neck squamous cell carcinoma (HNSC), associated with better prognosis in KIRP and stomach adenocarcinoma (STAD). SFN expression correlated with worse outcomes in BRCA. GZMB showed a significant prognostic impact in LIHC and SKCM. ATP6V0A4 expression was linked to patient prognosis in KIRC, cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC). GSDMA exhibited negative prognostic relevance in cancers such as KIRP, READ and BRCA (Table S2).

Pan-cancer single-cell analysis

To further elucidate the role of key genes in different cancer types, we performed single-cell RNA sequencing analysis across five cancer datasets: BRCA, CRC, LIHC, prostate adenocarcinoma (PRAD) and UCEC. The analysis revealed both common and unique immune cell populations within the tumor microenvironments of these cancers.

Common immune cell types identified across all five cancer types included CD8 T cells [CD8 central memory T cells (Tcm), CD8 exhausted T cells (Tex), CD8 effector T cells (Teff)], and Treg. Additionally, CD4 T cells [CD4 naive T cells (Tn)] and proliferating T cells (Tprolif) were consistently found in BRCA, CRC, UCEC, and LIHC, indicating active immune responses and potential anti-tumor activity. Each cancer type also exhibited unique cell populations reflecting their distinct tumor microenvironments. For instance, PRAD had a significant presence of epithelial cells, fibroblasts, malignant cells, monocytes/macrophages, and plasma cells, highlighting the diverse cellular composition specific to prostate tumors. In UCEC, fibroblasts were notably present, suggesting a substantial component. LIHC showed unique populations of B cells, dendritic cells [conventional type 1 dendritic cells (cDC1)], innate lymphoid cells (ILC), and M1 macrophages, indicating a complex immune landscape in liver cancer (Figure 9).

Figure 9 Single-cell analysis of key genes in various cancer types: UMAP plot showing the distribution of cell types in BRCA (A), CRC (B), LIHC (C), PRAD (D), and UCEC (E). BRCA, breast invasive carcinoma; cDC1, conventional type 1 dendritic cells; CRC, colorectal cancer; ILC, innate lymphoid cells; LIHC, liver hepatocellular carcinoma; MAIT, mucosal-associated invariant T cells; NK, natural killer; PRAD, prostate adenocarcinoma; Tcm, central memory T cells; Teff, effector T cells; Tem, effector memory T cells; Tex, exhausted T cells; Tfh, follicular helper T cells; Th, T helper; Tn, naive T cells; Tprolif, proliferating T cells; Treg, regulatory T cells; UCEC, uterine corpus endometrial carcinoma; UMAP, Uniform Manifold Approximation and Projection.

The expression patterns of key genes (GZMB, CXCR4, GSDMA, SFN, ATP6V0A4) varied across these immune cell types (Figure 10). In BRCA, GZMB was highly expressed in CD8 Tex and Tprolif cells, CXCR4 in CD8 Tcm, Th17 and CD8 effector memory T cells (Tem), and GSDMA showed higher expression in CD8 Tem, Th17 and Treg. In CRC, ATP6V0A4 and SFN were highly expressed, with variations across Th17 and Treg cells, while GZMB had lower expression levels. In LIHC, SFN showed high expression in mast cells, GSDMA in M1 cells, and GZMB in ILC. In PRAD, malignant cells showed high SFN and ATP6V0A4 expression; Treg and CD8 T cells had significant GSDMA and CXCR4 expression. In UCEC, GSDMA and SFN were highly expressed in fibroblast cells, while GZMB and CXCR4 had varying expression levels across other immune cell types.

Figure 10 Heatmap of key gene expression in different cell types across cancer types: BRCA (A), CRC (B), LIHC (C), PRAD (D), and UCEC (E). The colors represent the relative expression levels, with red indicating high expression and blue indicating low expression. BRCA, breast invasive carcinoma; cDC1, conventional type 1 dendritic cells; CRC, colorectal cancer; ILC, innate lymphoid cells; LIHC, liver hepatocellular carcinoma; NK, natural killer; PRAD, prostate adenocarcinoma; Tcm, central memory T cells; Teff, effector T cells; Tem, effector memory T cells; Tex, exhausted T cells; Tfh, follicular helper T cells; Th, T helper; Tn, naive T cells; Tprolif, proliferating T cells; Treg, regulatory T cells; UCEC, uterine corpus endometrial carcinoma.

The trajectory and pseudotime analysis (Figure 11) across these cancers provide insights into the dynamic changes in key gene expression during cell differentiation. The pseudotime plots reveal the progression of different immune cell subtypes along the developmental trajectory, highlighting the temporal changes in gene expression that correspond to various stages of cell differentiation and tumor progression.

Figure 11 Pseudotime trajectories across cancer types: breast cancer (A), colorectal cancer (B), liver hepatocellular carcinoma (C), prostate adenocarcinoma (D), and uterine corpus endometrial carcinoma (E). cDC1, conventional type 1 dendritic cells; ILC, innate lymphoid cells; MAIT, mucosal-associated invariant T cells; NK, natural killer; Tcm, central memory T cells; Teff, effector T cells; Tem, effector memory T cells; Tex, exhausted T cells; Tfh, follicular helper T cells; Th, T helper; Tn, naive T cells; Tprolif, proliferating T cells; Treg, regulatory T cells; UMAP, Uniform Manifold Approximation and Projection.

Through the above analysis, we found that PCD plays an important role in tumors, to explore the overall prognosis ability of PCD in multiple cancers. After analysis, it was found that among the 7 PCD forms, apoptosis, pyroptosis and autophagy affected the tumor prognosis (Figure 12A). Based on these three death methods, we constructed a PCD pan-cancer prediction model. After analysis, it was found that the predictive model can predict the prognosis of pan-cancer tumors, and the prognosis with high scores is not effective (Figure 12B). To explore the relationship between PCD and immunotherapy, we further observed the relationship between PCD model and PCD specific form and immunotherapy disorders based on Tumor Immune Dysfunction and Exclusion (TIDE). After analysis, it was found that there was a negative correlation between PCD model and immunotherapy dysregulation, which further elucidated the potential mechanism of high scores and poor prognosis of PCD model (Figure 12C).

Figure 12 The effect of PCD on prognosis in pan-cancer. (A) Effects of each PCD form on the prognosis of pan-cancer. (B) Survival curve of PCD prognostic model on pan-cancer prognostic ability. (C) The relationship between PCD model and immunotherapy. *, P<0.05; **, P<0.01; ***, P<0.001. CI, confidence interval; HR, hazard ratio; MSI Expr Sig, microsatellite instability expression signature; OS, overall survival; PCD, programmed cell death; TIDE, Tumor Immune Dysfunction and Exclusion.

Discussion

The molecular mechanisms of regulated cell death significantly influence CAD progression through distinct pathways that modulate plaque stability and inflammatory responses. For example, apoptosis contributes to necrotic core formation by promoting macrophage and vascular smooth muscle cell death, with death receptor-mediated pathways and mitochondrial dysfunction amplifying endothelial injury. Necroptosis, driven by RIPK1/RIPK3/MLKL signaling, exacerbates plaque vulnerability through pro-inflammatory cytokine release and impaired efferocytosis, particularly under oxidative stress conditions. Since PCD includes many different forms, it is still unknown which PCD plays a more important role in CAD. In order to have a more comprehensive understanding of the role of PCD in CAD, we conducted the following research. In this study, we conducted a comprehensive analysis to understand the role of PCD in CAD. Our investigation included differential expression analysis of PCD-related genes, ssGSEA analysis to explore PCD pathways, identification of core PCD genes through machine learning methods, development of a prognostic risk model, analysis of immune infiltration, and pan-cancer analysis of key genes. These multifaceted approaches allowed us to elucidate the complex interplay between PCD and CAD, offering novel insights into the disease’s pathogenesis and potential therapeutic targets.

Emerging evidence highlights the complex interplay between PCD pathways and CAD pathogenesis, with NETosis, apoptosis, ferroptosis, and autophagy-lysosomal dysfunction exhibiting distinct mechanisms and therapeutic implications. NETosis, characterized by neutrophil extracellular trap release through histone citrullination and chromatin decondensation, is induced by inflammatory stimuli like angiotensin II and mechanical stretch, contributing to endothelial dysfunction, plaque instability, and thrombosis through neutrophil elastase-mediated matrix degradation and IL-1β activation. Apoptosis demonstrates stage-dependent roles, where early macrophage apoptosis may limit inflammation through efferocytosis, but advanced plaque smooth muscle cell apoptosis reduces collagen synthesis, weakening fibrous caps, while defective clearance of apoptotic bodies exacerbates necrotic core formation. Ferroptosis, driven by iron-dependent lipid peroxidation via GPX4 suppression, accelerates endothelial dysfunction through oxidized low-density lipoprotein (ox-LDL)-induced oxidative stress and promotes macrophage death, creating pro-inflammatory microenvironments. Autophagy exhibits dual roles: basal autophagy protects against oxidative damage by removing dysfunctional organelles, but excessive activation induces autophagic death of macrophages and vascular cells, while lysosomal dysfunction from ceroid accumulation impairs lipid clearance. Therapeutically, targeting PAD4-mediated NETosis, enhancing efferocytosis machinery, modulating GPX4 activity, and pharmacologically inducing lysosomal biogenesis with trehalose represent promising strategies to stabilize plaques by balancing cell survival/death dynamics. These interconnected pathways underscore the need for temporally and cellularly targeted interventions to address the multifaceted cell death mechanisms driving CAD progression.

The differential expression analysis identified several PCD-related genes that were significantly associated with CAD prognosis. Notably, genes such as GZMB, CXCR4, and SFN showed distinct expression patterns in CAD samples. GZMB is a serine protease primarily expressed in cytotoxic T cells and natural killer (NK) cells. It plays a crucial role in inducing apoptosis in target cells (25). In CAD, GZMB was found to be upregulated, suggesting its involvement in plaque instability and progression. The elevated levels of GZMB may contribute to the apoptotic death of vascular cells, promoting the development of vulnerable plaques prone to rupture, which is a critical event leading to acute coronary syndromes. CXCR4 is a chemokine receptor that regulates the migration and homing of immune cells (26). In CAD, differential expression of CXCR4 was observed, indicating its potential role in modulating the immune response (27-29). CXCR4 is known to mediate the recruitment of inflammatory cells to the sites of vascular injury and atherosclerotic plaques. Its altered expression in CAD suggests that it may influence the inflammatory milieu within the plaques, affecting plaque stability and the overall progression of the disease (30-32). ATP6V0A4 is a component of the vacuolar ATPase (V-ATPase) complex, involved in acidifying intracellular compartments. The differential expression of ATP6V0A4 in CAD points to its potential role in cellular metabolism and autophagy (33). Dysregulation of autophagy has been implicated in atherosclerosis, and ATP6V0A4 may contribute to the defective autophagic processes observed in CAD, thereby influencing plaque development and stability. GSDMA is a member of the gasdermin family, known to play a role in pyroptosis, a form of PCD associated with inflammation (34,35). In CAD, altered expression of GSDMA suggests its involvement in the inflammatory processes within atherosclerotic plaques (36). Pyroptosis mediated by GSDMA could exacerbate inflammation, promoting plaque instability and progression. SFN, also known as 14-3-3σ, is involved in regulating cell cycle (37) and apoptosis (38). The differential expression of SFN in CAD indicates its potential role in the disease’s pathogenesis. SFN can modulate various signaling pathways (39-41), including those involved in cell survival and death. Its expression in CAD suggests that it may influence the balance between cell proliferation and apoptosis within atherosclerotic plaques, contributing to their stability and progression.

The different forms of PCD—apoptosis, autophagy, necroptosis, and pyroptosis—are interconnected processes that collectively contribute to cellular homeostasis and disease progression (12,13,42,43). In the context of CAD, these PCD pathways interact in complex ways to influence plaque stability, inflammation, and overall disease outcomes. Apoptosis and necroptosis, for instance, both lead to cell death but with different downstream effects (44-46), apoptosis is non-inflammatory, while necroptosis triggers inflammation. Autophagy, which can promote cell survival or death, interacts with these pathways by degrading CC, potentially influencing apoptotic and necroptotic processes. Pyroptosis, another inflammatory form of cell death, further amplifies inflammatory responses within atherosclerotic plaques (47,48).

By integrating genes from these diverse PCD pathways—such as GZMB (apoptosis), CXCR4 (cell migration and immune response), SFN (cell cycle regulation and apoptosis), ATP6V0A4 (autophagy), and GSDMA (pyroptosis)—into a single prognostic model, we provide a comprehensive tool for assessing CAD prognosis. Our prognostic scoring system, the PCDscore, effectively stratifies patients into high and low-risk groups with distinct survival outcomes. This integration highlights the multifaceted role of PCD in CAD and underscores the potential for a more nuanced approach to predicting disease progression and therapeutic responses. The PCDscore’s ability to account for the crosstalk between different PCD pathways offers a robust prognostic marker and a promising guide for personalized treatment strategies in CAD.

Our prognostic scoring system, based on PCD-related genes, demonstrated high accuracy in predicting CAD prognosis. This model’s ability to stratify patients into distinct risk groups underscores its potential clinical utility. High-PCDscore patients exhibited shorter survival times and lower levels of CD4⁺ T cells, suggesting that PCD genes might be influencing immune cell dynamics and contributing to adverse outcomes. These findings align with previous studies that have highlighted the importance of immune responses in CAD progression (49-53). The immune infiltration analysis revealed a close association between the PCDscore and the levels of various immune cells. The lower sensitivity to chemotherapy and targeted therapies in high-PCDscore patients highlights the need for tailored therapeutic approaches based on the PCDscore.

In order to have a more comprehensive understanding of the role of PCD genes related to CAD, we further analyzed the role of five key genes in the immune microenvironment of pan-cancer and even pan-cancer. After analysis, we found that these genes play different roles in different tumors. This also indicates that this PCD may play a dual role in different tumors. Apoptosis typically suppresses tumorigenesis by eliminating damaged cells, but chronic apoptotic stress in the TME may paradoxically drive compensatory proliferation and immune tolerance through the release of anti-inflammatory cytokines. Conversely, lytic forms of PCD like necroptosis and pyroptosis promote immunogenic cell death by releasing damage-associated molecular patterns (DAMPs) and cytokines that enhance dendritic cell maturation and T-cell activation, though excessive inflammation may simultaneously fuel tumor angiogenesis and metastatic niche formation. The net oncogenic effect further depends on the immune contexture—tumors with pre-existing CD8⁺ T-cell infiltration and PD-L1 expression show better responses to PCD-induced immunogenicity, while those dominated by Tregs and M2 macrophages often exhibit PCD-driven immunosuppression. Genetic alterations in PCD-related genes additionally modulate antigen presentation and break immune tolerance, determining whether PCD triggers antitumor immunity or autoimmune pathology. This intricate balance underscores the need for precision modulation of PCD pathways based on tumor-specific molecular and immune profiles.

The immunological consequences of different cell death modalities significantly impact cancer therapy efficacy. While apoptosis is traditionally considered immunologically silent, often creating an immunosuppressive microenvironment that limits treatment efficacy, pyroptosis presents contrasting effects by releasing cellular antigens, DAMPs, and proinflammatory cytokines that convert “cold” tumors to “hot” tumors, enhancing immune checkpoint blockade therapy. Similarly, autophagy demonstrates dual roles—it can promote antigen presentation and T cell responses, yet may also suppress antitumor immunity through various mechanisms. Understanding these pathways’ interplay provides promising therapeutic opportunities through combination strategies targeting specific death mechanisms. Future cancer treatments may benefit from approaches that selectively induce immunogenic forms of cell death while preserving beneficial immune responses, potentially overcoming resistance and improving clinical outcomes across diverse malignancies.

Conclusions

Overall, our study highlights the critical role of PCD in the pathogenesis and progression of CAD. The identification of key PCD genes and pathways, along with the development of a prognostic risk model, offers new avenues for the diagnosis and treatment of CAD. Additionally, the insights gained from immune infiltration and pan-cancer analyses underscore the broader significance of PCD mechanisms in health and disease. Future studies should focus on the functional validation of these findings and the exploration of targeted therapies that modulate PCD pathways to improve patient outcomes in CAD and other related diseases.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2024-2283/rc

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2024-2283/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2024-2283/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Chen X, Fang M, Hong J, et al. JNK Pathway-Associated Phosphatase Deficiency Facilitates Atherosclerotic Progression by Inducing T-Helper 1 and 17 Polarization and Inflammation in an ERK- and NF-κB Pathway-Dependent Manner. J Atheroscler Thromb 2024;31:1460-78. [Crossref] [PubMed]
2. Virani SS, Newby LK, Arnold SV, et al. 2023 AHA/ACC/ACCP/ASPC/NLA/PCNA Guideline for the Management of Patients With Chronic Coronary Disease: A Report of the American Heart Association/American College of Cardiology Joint Committee on Clinical Practice Guidelines. Circulation 2023;148:e9-e119. [Crossref] [PubMed]
3. Kolhar M, Al Rajeh AM. Deep learning hybrid model ECG classification using AlexNet and parallel dual branch fusion network model. Sci Rep 2024;14:26919. [Crossref] [PubMed]
4. Zhang Y, Guo Z, Lai R, et al. Comprehensive Analysis Based on Genes Associated With Cuproptosis, Ferroptosis, and Pyroptosis for the Prediction of Diagnosis and Therapies in Coronary Artery Disease. Cardiovasc Ther 2025;2025:9106621. [Crossref] [PubMed]
5. Yang F, Xu F, Zhang H, et al. Proteomic insights into the associations between obesity, lifestyle factors, and coronary artery disease. BMC Med 2023;21:485. [Crossref] [PubMed]
6. Liu Y, Yu Z, Lu Y, et al. Progress in the study of the mechanism of ferroptosis in coronary heart disease and clinical intervention strategies. Front Cardiovasc Med 2025;12:1545231. [Crossref] [PubMed]
7. Khan K, Khan A, Rahman ZU, et al. Genetic Polymorphism in miRNA Genes and Their Association with susceptibility of Coronary Heart Disease: An Updated Review. Pathol Res Pract 2024;264:155675. [Crossref] [PubMed]
8. Yang YL, Wu CH, Hsu PF, et al. Systemic immune-inflammation index (SII) predicted clinical outcome in patients with coronary artery disease. Eur J Clin Invest 2020;50:e13230. [Crossref] [PubMed]
9. Zhao Z, Zhang X, Sun T, et al. Prognostic value of systemic immune-inflammation index in CAD patients: Systematic review and meta-analyses. Eur J Clin Invest 2024;54:e14100. [Crossref] [PubMed]
10. Newton K, Strasser A, Kayagaki N, et al. Cell death. Cell 2024;187:235-56. [Crossref] [PubMed]
11. Peng F, Liao M, Qin R, et al. Regulated cell death (RCD) in cancer: key pathways and targeted therapies. Signal Transduct Target Ther 2022;7:286. [Crossref] [PubMed]
12. Chen Y, Li X, Yang M, et al. Research progress on morphology and mechanism of programmed cell death. Cell Death Dis 2024;15:327. [Crossref] [PubMed]
13. Bedoui S, Herold MJ, Strasser A. Emerging connectivity of programmed cell death pathways and its physiological implications. Nat Rev Mol Cell Biol 2020;21:678-95. [Crossref] [PubMed]
14. D'Arcy MS. Cell death: a review of the major forms of apoptosis, necrosis and autophagy. Cell Biol Int 2019;43:582-92. [Crossref] [PubMed]
15. Kopeina GS, Zhivotovsky B. Programmed cell death: Past, present and future. Biochem Biophys Res Commun 2022;633:55-8. [Crossref] [PubMed]
16. Tower J. Programmed cell death in aging. Ageing Res Rev 2015;23:90-100. [Crossref] [PubMed]
17. Uti DE, Atangwho IJ, Alum EU, et al. Antioxidants in cancer therapy mitigating lipid peroxidation without compromising treatment through nanotechnology. Discov Nano 2025;20:70. [Crossref] [PubMed]
18. Kaplan O, Demircan G. Relationship of Autophagy and Apoptosis with Total Occlusion of Coronary Arteries. Med Sci Monit 2018;24:6984-8. [Crossref] [PubMed]
19. Klionsky DJ, Abdelmohsen K, Abe A, et al. Guidelines for the use and interpretation of assays for monitoring autophagy (3rd edition). Autophagy 2016;12:1-222.
20. Frank D, Vince JE. Pyroptosis versus necroptosis: similarities, differences, and crosstalk. Cell Death Differ 2019;26:99-114. [Crossref] [PubMed]
21. Wang Y, Kanneganti TD. From pyroptosis, apoptosis and necroptosis to PANoptosis: A mechanistic compendium of programmed cell death pathways. Comput Struct Biotechnol J 2021;19:4641-57. [Crossref] [PubMed]
22. Zhou P, Zhang S, Wang M, et al. The Induction Mechanism of Ferroptosis, Necroptosis, and Pyroptosis in Inflammatory Bowel Disease, Colorectal Cancer, and Intestinal Injury. Biomolecules 2023;13:820. [Crossref] [PubMed]
23. Qin H, Abulaiti A, Maimaiti A, et al. Integrated machine learning survival framework develops a prognostic model based on inter-crosstalk definition of mitochondrial function and cell death patterns in a large multicenter cohort for lower-grade glioma. J Transl Med 2023;21:588. [Crossref] [PubMed]
24. Charoentong P, Finotello F, Angelova M, et al. Pan-cancer Immunogenomic Analyses Reveal Genotype-Immunophenotype Relationships and Predictors of Response to Checkpoint Blockade. Cell Rep 2017;18:248-62. [Crossref] [PubMed]
25. Li Z, Ma R, Tang H, et al. Therapeutic application of human type 2 innate lymphoid cells via induction of granzyme B-mediated tumor cell death. Cell 2024;187:624-641.e23. [Crossref] [PubMed]
26. Bianchi ME, Mezzapelle R. The Chemokine Receptor CXCR4 in Cell Proliferation and Tissue Regeneration. Front Immunol 2020;11:2109. [Crossref] [PubMed]
27. Biasci D, Smoragiewicz M, Connell CM, et al. CXCR4 inhibition in human pancreatic and colorectal cancers induces an integrated immune response. Proc Natl Acad Sci U S A 2020;117:28960-70. [Crossref] [PubMed]
28. Wen F, Lu X, Huang W, et al. Characteristics of immunophenotypes and immunological in tumor microenvironment and analysis of immune implication of CXCR4 in gastric cancer. Sci Rep 2022;12:5720. [Crossref] [PubMed]
29. Werner Y, Mass E, Ashok Kumar P, et al. Cxcr4 distinguishes HSC-derived monocytes from microglia and reveals monocyte immune responses to experimental stroke. Nat Neurosci 2020;23:351-62. [Crossref] [PubMed]
30. Walter DH, Haendeler J, Reinhold J, et al. Impaired CXCR4 signaling contributes to the reduced neovascularization capacity of endothelial progenitor cells from patients with coronary artery disease. Circ Res 2005;97:1142-51. [Crossref] [PubMed]
31. Rath D, Chatterjee M, Borst O, et al. Platelet surface expression of stromal cell-derived factor-1 receptors CXCR4 and CXCR7 is associated with clinical outcomes in patients with coronary artery disease. J Thromb Haemost 2015;13:719-28. [Crossref] [PubMed]
32. Rath D, Chatterjee M, Borst O, et al. Expression of stromal cell-derived factor-1 receptors CXCR4 and CXCR7 on circulating platelets of patients with acute coronary syndrome and association with left ventricular functional recovery. Eur Heart J 2014;35:386-94. [Crossref] [PubMed]
33. Wang P, Du S, Guo C, et al. The presence of blastocyst within the uteri facilitates lumenal epithelium transformation for implantation via upregulating lysosome proteostasis activity. Autophagy 2024;20:58-75. [Crossref] [PubMed]
34. Deng W, Bai Y, Deng F, et al. Streptococcal pyrogenic exotoxin B cleaves GSDMA and triggers pyroptosis. Nature 2022;602:496-502. [Crossref] [PubMed]
35. LaRock DL, Johnson AF, Wilde S, et al. Group A Streptococcus induces GSDMA-dependent pyroptosis in keratinocytes. Nature 2022;605:527-31. [Crossref] [PubMed]
36. Yang H, Liu Y, Chen G, et al. Caspase-3/gasdermin-E axis facilitates the progression of coronary artery calcification by inducing the release of high mobility group box protein 1. Int Immunopharmacol 2024;127:111454. [Crossref] [PubMed]
37. Bryant CS, Kumar S, Chamala S, et al. Sulforaphane induces cell cycle arrest by protecting RB-E2F-1 complex in epithelial ovarian cancer cells. Mol Cancer 2010;9:47. [Crossref] [PubMed]
38. Wang H, Yang G, Tian Y, et al. Sulforaphane inhibits angiotensin II-induced cardiomyocyte apoptosis by acetylation modification of Nrf2. Aging (Albany NY) 2022;14:6740-55. [Crossref] [PubMed]
39. Wang Y, He X, Cheng N, et al. Unveiling the Nutritional Veil of Sulforaphane: With a Major Focus on Glucose Homeostasis Modulation. Nutrients 2024;16:1877. [Crossref] [PubMed]
40. Sailo BL, Liu L, Chauhan S, et al. Harnessing Sulforaphane Potential as a Chemosensitizing Agent: A Comprehensive Review. Cancers (Basel) 2024;16:244. [Crossref] [PubMed]
41. Pan J, Wang R, Pei Y, et al. Sulforaphane alleviated vascular remodeling in hypoxic pulmonary hypertension via inhibiting inflammation and oxidative stress. J Nutr Biochem 2023;111:109182. [Crossref] [PubMed]
42. Fuchs Y, Steller H. Programmed cell death in animal development and disease. Cell 2011;147:742-58. [Crossref] [PubMed]
43. Xu Y, Li P, Li K, et al. Pathological mechanisms and crosstalk among different forms of cell death in systemic lupus erythematosus. J Autoimmun 2022;132:102890. [Crossref] [PubMed]
44. Bertheloot D, Latz E, Franklin BS. Necroptosis, pyroptosis and apoptosis: an intricate game of cell death. Cell Mol Immunol 2021;18:1106-21. [Crossref] [PubMed]
45. Gielecińska A, Kciuk M, Yahya EB, et al. Apoptosis, necroptosis, and pyroptosis as alternative cell death pathways induced by chemotherapeutic agents? Biochim Biophys Acta Rev Cancer 2023;1878:189024. [Crossref] [PubMed]
46. Ketelut-Carneiro N, Fitzgerald KA. Apoptosis, Pyroptosis, and Necroptosis-Oh My! The Many Ways a Cell Can Die. J Mol Biol 2022;434:167378. [Crossref] [PubMed]
47. Lin Y, Liu M, Chen E, et al. Bone marrow-derived mesenchymal stem cells microvesicles stabilize atherosclerotic plaques by inhibiting NLRP3-mediated macrophage pyroptosis. Cell Biol Int 2021;45:820-30. [Crossref] [PubMed]
48. Yang Z, Shi J, Chen L, et al. Role of Pyroptosis and Ferroptosis in the Progression of Atherosclerotic Plaques. Front Cell Dev Biol 2022;10:811196. [Crossref] [PubMed]
49. Aravindhan V, Madhumitha H. Metainflammation in Diabetic Coronary Artery Disease: Emerging Role of Innate and Adaptive Immune Responses. J Diabetes Res 2016;2016:6264149. [Crossref] [PubMed]
50. Attiq A, Afzal S, Ahmad W, et al. Hegemony of inflammation in atherosclerosis and coronary artery disease. Eur J Pharmacol 2024;966:176338. [Crossref] [PubMed]
51. Depuydt MAC, Prange KHM, Slenders L, et al. Microanatomy of the Human Atherosclerotic Plaque by Single-Cell Transcriptomics. Circ Res 2020;127:1437-55. [Crossref] [PubMed]
52. Laera N, Malerba P, Vacanti G, et al. Impact of Immunity on Coronary Artery Disease: An Updated Pathogenic Interplay and Potential Therapeutic Strategies. Life (Basel) 2023;13:2128. [Crossref] [PubMed]
53. Zhao S, Wu Y, Wei Y, et al. Identification of Biomarkers Associated With CD8+ T Cells in Coronary Artery Disease and Their Pan-Cancer Analysis. Front Immunol 2022;13:876616. [Crossref] [PubMed]