Identification of PRKCB, NLRC4, and TNFSF10 as Key Regulators of the Lipid Metabolism‐Autophagy Network in Atherosclerosis

ABSTRACT

Background

The occurrence and development of atherosclerosis (AS) are closely related to disorders of lipid metabolism and dysfunction of autophagy. However, the key regulatory genes and immune microenvironment characteristics require further exploration.

Methods

This study integrated the bulk transcriptome and single‐cell RNA sequencing data, and utilized bioinformatics methods to screen the differentially expressed genes (DEGs) of AS. The weighted gene co‐expression network analysis (WGCNA) was used to identify the key gene modules related to lipid metabolism and autophagy. The core regulatory genes were further screened based on the machine learning algorithm (LASSO regression). The enrichment scores of 28 immune cell subtypes were calculated by ssGSEA, and the correlation between the immune infiltration characteristics and the expression levels of the three core genes (PRKCB, NLRC4, and TNFSF10) was analyzed by Spearman correlation analysis. The single‐cell transcriptome data analysis was performed to determine the cell subtype distribution characteristics of the core genes and to compare the expression differences in lipid metabolism‐autophagy‐related gene expression across distinct cell subtypes. In addition, we screened potential drugs targeting the key genes from the public drug database, and simulated the binding mode of approved drugs with the active site of PRKCB by molecular docking technology. Ultimately, the expression of target factors in mouse aortic tissues was determined by quantitative real‐time PCR (RT‐qPCR), Western blotting, and immunohistochemical (IHC) analysis. Concurrently, the levels of TNFSF10 in mouse plasma were confirmed through enzyme‐linked immunosorbent assay (ELISA).

Result

Three lipid metabolism‐autophagy core regulatory genes, PRKCB, NLRC4, and TNFSF10, were identified. These genes were significantly associated with the immune microenvironment of AS plaques and exhibited different cell subpopulation distribution patterns. The macrophage subpopulations showed higher lipid metabolism‐autophagy regulatory activity. Molecular docking revealed that the anti‐atherosclerotic drugs quercetin and atenolol can stably bind to the active site of PRKCB (with docking energies of ‐8.3 kcal/mol and ‐6.2 kcal/mol, respectively), a finding that identifies PRKCB as a candidate target worthy of further attention. Experimental evidence confirmed that the expression of these three key factors was significantly elevated in atherosclerotic plaque tissue, and the levels of TNFSF10 in the plasma of atherosclerotic mice were also significantly higher than those in the control group.

Conclusion

This study identifies PRKCB, NLRC4, and TNFSF10 as key hub genes that link lipid metabolism and autophagy in AS, and highlights their potential as diagnostic biomarkers and therapeutic targets.

Summary

This study integrates lipid metabolism and autophagy pathways to elucidate atherosclerosis pathogenesis. By combining bioinformatic analyses with experimental validation, we identified three hub genes (PRKCB, NLRC4, TNFSF10) as key regulators in disease progression. These findings provide molecular evidence for elucidating the “metabolism–autophagy–immune” interplay in AS and establish a theoretical foundation for the development of biomarkers for disease diagnosis as well as the screening of potential therapeutic targets.

Abbreviations

ARGs

autophagy‐regulated genes

AS

atherosclerosis

ASCVD

atherosclerotic cardiovascular diseases

BP

biological process

CC

cellular component

CMGs

cyan module genes

DEGs

differentially expressed genes

DGLdb

Drug Gene Interaction Database

ELISA

enzyme‐linked immunosorbent assay

GO

gene ontology

GS

gene significance

HADb

Human Autophagy Database

HMG

High Metabolic Groups

IHC

immunohistochemical

KEGG

Kyoto Encyclopedia of Genes and Genomes

LDL

low‐density lipoprotein

LMAGs

lipid metabolism‐autophagy genes

LMG

Low Metabolic Groups

LMGs

lipid metabolism‐related genes

MF

molecular function

MM

module membership

PCA

principal component analysis

RT‐qPCR

quantitative real‐time PCR

SPF

specific pathogen‐free

ssGSEA

single‐sample gene set enrichment analysis

TTD

Therapeutic Target Database

WGCNA

weighted gene co‐expression network analysis

1. Introduction

Atherosclerosis (AS) is a chronic inflammatory vascular disease induced by multiple risk factors, characterized by vascular endothelial dysfunction, abnormal lipid metabolism, and persistent inflammatory responses [1]. During disease progression, abnormal deposition of lipids (particularly oxidized low‐density lipoprotein) beneath the vascular intima triggers monocyte infiltration and differentiation into macrophages. These macrophages then extensively uptake modified lipoproteins via scavenger receptors, forming foam cells, which ultimately lead to the formation of atherosclerotic plaques and vascular remodeling in the arterial wall [2]. As the core pathological mechanism of atherosclerotic cardiovascular diseases (ASCVD), such as coronary heart disease, ischemic stroke, and peripheral artery disease, AS‐related disorders have become a major global public health challenge in the 21st century [3]. According to the latest Global Burden of Disease Study (GBD 2019), atherosclerotic cardiovascular diseases cause approximately 18 million deaths annually, accounting for 31.5% of total global deaths, posing a continuous threat to human health and imposing a substantial disease burden [4].

Autophagy is a self‐protective cellular catabolic pathway that relies on lysosomes and maintains intracellular homeostasis by selectively degrading damaged organelles, misfolded proteins, and lipids, etc [5]. As an important mechanism of cellular quality control, autophagy plays a crucial role in the occurrence and development of AS. Studies have shown that during the progression of AS plaques, various pathological factors can dynamically regulate autophagic activity, including pro‐inflammatory cytokines (such as TNF‐α, IL‐6), bioactive lipid metabolites (e.g., ox‐LDL), microbial‐associated molecular patterns (such as LPS), abnormal metabolic products (such as AGEs), and non‐coding RNAs (such as miR‐33, miR‐155) [6]. These factors participate in the regulation of key pathological processes of AS, such as foam cell formation, phenotypic transformation of vascular smooth muscle cells, and endothelial dysfunction, by affecting the integrity of the autophagic flux [7]. Notably, moderate activation of autophagy can exert an anti‐atherosclerotic effect by inhibiting inflammatory reactions, promoting cholesterol reverse transport, and enhancing plaque stability [8]. Conversely, excessive activation of autophagy can lead to “autophagic cell death” in smooth muscle cells and vascular endothelial cells, resulting in reduced collagen synthesis and destruction of the fibrous cap structure, ultimately increasing the risk of acute cardiovascular events associated with AS [9].

Despite the significant advancements in modern medical technology that have greatly improved the diagnosis and treatment of AS, especially through novel therapeutic strategies such as regulating autophagy and immunity, which offer targeted treatment possibilities for patients who do not respond well to traditional lipid‐lowering drugs, there are still deficiencies in our current understanding of the pathogenesis of AS. Disordered lipid metabolism, as a core pathological feature of AS, is closely associated with the autophagic process. However, the precise regulatory mechanisms of autophagy‐related genes in plaque formation, their impact on the local immune microenvironment, and their potential value as diagnostic biomarkers still require in‐depth exploration. Clarifying these key scientific issues will provide an important theoretical foundation for the development of more effective AS treatment strategies.

2. Materials and Methods

2.1. Data Acquisition and Processing

Both bulk RNA‐seq and single‐cell RNA‐seq data were obtained from the GEO database (https://www.ncbi.nlm.nih.gov/geo/). Four bulk transcriptomic datasets were analyzed: GSE43292 (32 controls vs. 32 atherosclerosis), GSE28829 (29 atherosclerosis), GSE57691 (10 controls vs. 9 atherosclerosis), and GSE100927 (35 controls vs. 69 atherosclerosis). The scRNA‐seq data were sourced from the GSE159677 dataset, which included 3 carotid atherosclerotic samples and 3 matched adjacent control samples. Detailed descriptions of the datasets are provided in Table 1. Batch effect correction and integration of the bulk RNA‐seq data were performed using the “sva” package in R. The scRNA‐seq data were analyzed using the R package Seurat (version 5.3.0). Cells were filtered based on the following criteria: nFeature_RNA (< 6,000 genes per cell), nCount_RNA (< 50,000 counts per cell), and percent_mito (< 15% mitochondrial gene ratio). Doublets were removed, and the data were normalized using the LogNormalize method. The top 2,000 highly variable genes were selected for principal component analysis (PCA) to reduce dimensionality. Cell clusters were identified through unsupervised clustering and visualized using UMAP. Cell subpopulations were manually annotated based on established marker genes.

Table 1.

Basic characteristics of the GEO datasets used in this study.

GEO Series	Data type	Organism	Tissue type	Sample		Platform
				Control	AS
GSE43292	RNA‐seq	Homo sapiens	Carotid artery	32	32	GPL6244
GSE28829	RNA‐seq	Homo sapiens	Carotid artery	0	29	GPL570
GSE57691	RNA‐seq	Homo sapiens	Abdominal aorta	10	9	GPL10558
GSE100927	RNA‐seq	Homo sapiens	Carotid, femoral, and popliteal arteries	35	69	GPL17077
GSE159677	scRNA‐seq	Homo sapiens	Carotid artery	3	3	GPL18573

2.2. Functional Enrichment Analysis

To elucidate the biological functions of differentially expressed genes during disease progression, we performed Gene Ontology (GO) enrichment analysis (including Biological Process [BP], Cellular Component [CC], and Molecular Function [MF]) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis using the “clusterProfiler” R package. The functional association networks of key candidate genes and their involved biological processes were predicted via the GeneMANIA database (http://genemania.org).

2.3. Screening of Key Lipid Metabolism‐Autophagy Genes

In this study, 17,095 lipid metabolism‐related genes (LMGs) were retrieved from the GeneCards database (https://www.genecards.org/) using the keyword “lipid metabolism” (Supporting File 1). Additionally, 371 autophagy‐regulated genes (ARGs) were obtained from the Human Autophagy Database (HADb, http://www.autophagy.lu/), with the complete list provided in Supporting File 2. The Venn Diagram package was used to identify 346 overlapping genes involved in both lipid metabolism and autophagy (Supporting File 3).

2.4. Co‐Expression Network Construction and Key Module Identification

Lipid metabolism‐autophagy scores were calculated for the integrated dataset (GSE43292/GSE28829) using the ssGSEA algorithm, and samples were stratified into low metabolism (LMG) and high metabolism (HMG) groups. The optimal soft threshold β was determined using the “pickSoftThreshold” function. Modules with highly correlated eigengenes were merged via the “mergeCloseModules” function. Heatmaps were used to visualize module‐trait relationships. The biological relevance of each module was assessed by examining module membership (MM) and gene significance (GS).

2.5. Immune Cell Infiltration Analysis

The single‐sample gene set enrichment analysis (ssGSEA) algorithm was employed to quantify immune cell infiltration. Based on a published gene set representing 28 immune cell types, the GSVA package was used to transform the gene expression matrix of the integrated dataset (GSE43292/GSE28829) into an immune cell infiltration score matrix [10]. Heatmaps and violin plots were generated to visualize intergroup differences. Spearman correlation analysis was performed to assess the associations between immune cells and hub genes. All statistical results were visualized using the “ggplot2” package.

2.6. Molecular Docking

The full‐length sequence of PRKCB was retrieved from the PDB database. The three‐dimensional structures of quercetin and atenolol were downloaded from the PubChem database. Docking analyses were performed using AutoDock Vina 1.5.7. Polar hydrogen atoms were added, and charges were assigned to both the protein and ligands. A docking grid was defined based on the active site of PRKCB. Subsequently, binding conformations of the candidate drugs with the protein were obtained, and the optimal binding energy was determined. To further investigate intermolecular interaction patterns, the protein–ligand complexes were visualized using PyMOL 2.5.

2.7. Animal Experiments

All animal procedures were approved by the Animal Experiment Ethics Committee of Shanghai Jiao Tong University and strictly adhered to its guidelines. All mice were housed in a specific pathogen‐free (SPF) environment at the Experimental Animal Center of Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, under a 12‐h light/dark cycle with ad libitum access to food and water. ApoE⁻/⁻ mice were purchased from Shanghai Model Organisms Center Inc. Atherosclerosis model mice were fed a standard rodent diet until 8 weeks of age, followed by a Western diet (40% fat + 1.25% cholesterol) for 16 weeks. After successful modeling, mice were anesthetized via intraperitoneal injection of 2.5% Avertin and euthanized, with blood collected by cardiac puncture.

2.8. Histological Analysis of Atherosclerotic Lesions

Aortas from the aortic arch to the iliac artery were dissected under a stereomicroscope. After fixation with 4% paraformaldehyde, atherosclerotic plaques were visualized by Oil Red O staining (G1261, Solarbio, Beijing, China) and quantified with Image Pro Plus software.

2.9. Quantitative Real‐Time PCR

Total RNA was isolated from mouse aortas using the Eastep Super RNA Extraction Kit (Promega, Shanghai, China). Subsequently, cDNA was synthesized via reverse transcription with the iScript cDNA Synthesis Kit (1708890, Bio‐Rad, Hercules, CA). RT‐qPCR was then performed using SYBR Fast qPCR Mix (Takara, Otsu, Japan) and the gene‐specific primers provided in Supporting File 4 Table 1.

2.10. Western Blot

Protein extracts from mouse aortas were prepared using RIPA lysis buffer (PC101, Yamei, Shanghai, China), and concentrations were measured by BCA protein assay kit (G2026‐200T, Servicebio, Wuhan, China). Equal protein loads were resolved on 8% or 10% SDS‐PAGE gels and immunoblotted onto PVDF membranes. After blocking with 5% non‐fat dry milk in TBST, membranes were incubated overnight at 4°C with primary antibodies specific for PRKCB (1:5000, 68620‐1‐1 g, Proteintech, Wuhan, China), NLRC4 (1:1000, ab201792, Abcam, Cambridge, UK), and TNFSF10 (1:5000, A25394, Wuhan, China), followed by HRP‐conjugated secondary antibodies. Protein bands were visualized by enhanced chemiluminescence and quantified.

2.11. Immunohistochemical Staining

Mouse aortic root sections were subjected to baking, dewaxing, rehydration, and high‐temperature antigen retrieval. Subsequent steps included incubation with 3% hydrogen peroxide for peroxidase quenching, TBST washing, blocking with 5% non‐fat milk, and overnight incubation with primary antibodies at 4°C. The sections were then exposed to HRP‐conjugated secondary antibodies for signal amplification and developed with DAB. Finally, they were counterstained with hematoxylin, dehydrated through an ethanol gradient, mounted with neutral balsam, and analyzed by light microscopy.

2.12. Enzyme‐Linked Immunosorbent Assay

Fresh plasma was immediately separated from the collected whole blood samples by centrifugation. The concentration of TNFSF10 was quantitatively measured using a commercial enzyme‐linked immunosorbent assay kit (ab253210, Abcam, Cambridge, UK).

2.13. Statistical Analysis

All statistical analyses were performed using GraphPad Prism or R software (version 4.4.2). The threshold for statistical significance was set at a p‐value < 0.05. Experimental data are presented as mean ± standard error of the mean (SEM) in the figures, with *p < 0.05, **p < 0.01, and ***p < 0.001.

3. Results

3.1. Data Analysis and Processing

This study employed the SVA algorithm to perform batch effect correction and standardized integration of the training datasets (GSE43292/GSE28829) and the test datasets (GSE57691/GSE100927). Principal component analysis (PCA) was used to visualize the distribution characteristics of different samples, confirming that batch effects were effectively removed and that data quality met analytical requirements (Figure 1A–D, Supporting File 4). The bioinformatics analysis pipeline revealed characteristic genomic alterations in atherosclerosis (Figure 1A).

Figure 1.

The overall analytical workflow of the study.

3.2. Identification of Differentially Expressed Genes and Functional Enrichment Analysis

Based on the GSE43292/GSE28829 datasets, differential expression analysis was conducted using the limma package (FDR < 0.05 and |log2FC | > 0.5). A total of 572 differentially expressed genes (DEGs) were identified in the comparison between the AS group and the control group, with 347 upregulated and 225 downregulated genes (Figure 2A,B). GO enrichment analysis revealed that the DEGs were primarily enriched in biological processes related to immune response, including cell migration, adhesion, proliferation, and phagocytosis. In terms of cellular components, these genes were mainly distributed in subcellular structures such as the cell membrane, secretory granules, and endocytic vesicles. For molecular functions, they were involved in transporter receptor activity, scavenger receptor activity, and immune receptor activity, among others (Figure 2C). These results suggest that there might be an enhanced immune activity in the atherosclerotic lesion areas. KEGG analysis showed that these genes were significantly enriched in pathways such as chemokine signaling, cell adhesion molecules, neutrophil extracellular trap formation, endocytosis, and phagosome‐related signaling pathways (Figure 2D). The enrichment of these pathways further confirmed the presence of active immune responses in the atherosclerotic lesion areas.

Figure 2.

Differential gene expression and functional enrichment analysis in the training dataset. A. Volcano plot of DEGs between the control and AS groups; B. Heatmap displaying the expression patterns of DEGs; C. GO analysis showing the main biological processes involving the DEGs; D. KEGG pathway enrichment analysis based on the DEGs.

3.3. Identification of Key Modules Related to Lipid Metabolism‐Autophagy

This study integrated the lipid metabolism and autophagy‐related intersecting genes from the Genecard database and the Human Autophagy Database (Figure 3A), and employed the ssGSEA algorithm to score samples in the training dataset (GSE24702/GSE24495), which were then categorized into Low Metabolic Groups (LMG) and High Metabolic Groups (HMG). To explore the potential regulators of lipid metabolism‐autophagy in the process of atherosclerosis, WGCNA analysis was conducted on the dataset GSE43292/GSE28829. Based on the scale‐free topology criterion, when the soft threshold β = 5, the network scale‐free fit index R² = 0.8, which met the requirements for constructing the co‐expression network (Figure 3B–E). The clustering result of the module eigengenes was presented in Figure 3F. Correlation analysis further revealed the relationship between module genes and clinical characteristics. Among them, the cyan module had the strongest association with lipid metabolism‐autophagy‐related AS, with a correlation coefficient of 0.75 and a P‐value of 3e‐18 (Figure 3G). Additionally, within this module, GS and MMF showed a good correlation, with a correlation coefficient of 0.77 and a P‐value less than 1e‐200 (Figure 3F). Therefore, we selected the cyan module as the key hub module for subsequent analysis.

Figure 3.

Identification of Key Genes Related to Lipid Metabolism‐Autophagy by WGCNA. A. Venn diagram showing the overlapping genes in lipid metabolism and autophagy. B and C. Determination of the optimal soft threshold (β) based on the scale‐free fit index and average connectivity analysis. D and E. Validation of network topology characteristics at a soft threshold β = 5. F. Hierarchical clustering analysis of the module eigengene expression profiles. G. Heatmap analysis of the correlation between module genes and AS. H. Correlation plot of MM‐GS within the cyan module.

3.4. Identification of Key Hub Genes Related to Lipid Metabolism‐Autophagy in AS

To identify the key hub genes involved in lipid metabolism‐autophagy in AS, a Venn diagram analysis was performed on the DEGs from the training dataset, the key cyan module genes (CMGs) identified by WGCNA, and the lipid metabolism‐autophagy genes (LMAGs). This analysis screened a total of 9 core genes (Figure 4A, Supporting File 3). GeneMANIA interaction network analysis revealed that these genes were significantly associated with pathways related to immune defense, inflammatory response, lysosomal function, and lipid localization regulation (Figure 4B), suggesting their potentially important roles in the “immune‐metabolic” interactive regulatory network of AS.

Figure 4.

Identification of hub genes related to lipid metabolism‐autophagy in AS. A. Venn diagram showing the intersection of DEGs, CMGs, and LMAGs. B. GeneMANIA database analysis demonstrates the functions of key genes and their cooperative regulatory networks. C and D. Using ten‐fold cross‐validation with lambda. min as the tuning parameter, coefficients were calculated for each gene, with a total of nine genes exhibiting coefficient shrinkage. E and F. Expression distribution of core differentially expressed genes in control versus AS groups within the training and test datasets, respectively. G. ROC curve analysis illustrates the diagnostic efficacy of 3 core lipid metabolism‐autophagy genes for AS.

Subsequently, the optimal regularization parameters were determined through cross‐validation to ensure the stability of the selection process (Figure 4C). A Lasso regression coefficient plot was provided, illustrating the relationship between the 3 selected key genes and their corresponding coefficients (Figure 4D). Next, to validate the screening results, expression profile analyses were conducted in both the training and validation datasets. Box plots demonstrated a consistent pattern of differential expression for the 3 genes in the AS group (Figure 4E,F). The predictive value of these key gene markers for atherosclerosis was evaluated using ROC curves, which showed AUC values of 0.628 for PRKCB, 0.873 for NLRC4, and 0.710 for TNFSF10 (Figure 4G). This evidence suggests that these three core genes may be involved in AS progression by regulating the “lipid metabolism‐autophagy” network.

3.5. Immune Cell Infiltration Analysis

Both differential gene expression enrichment analysis and GeneMANIA interaction network analysis indicated that the identified functional candidate genes were significantly associated with immune responses (Figures 2C,D, and 4B). Therefore, we investigated the immune cell infiltration profile associated with the key lipid metabolism‐autophagy genes in AS. The results revealed significant differences in immune cell characteristics between the AS and control groups (Figure 5A and B). Furthermore, correlation analysis demonstrated strong associations between the three key LMAGs and the majority of immune cell subsets (Figure 5C), suggesting that these genes may contribute to AS pathogenesis by modulating immune cell functions.

Figure 5.

Analysis of immune cell infiltration. A. Heatmap displaying the differential infiltration characteristics of 28 immune cell types. B. Violin plots showing the inter‐group differences in specific immune cell subsets. C. Correlation analysis of core genes and immune cell infiltration based on Spearman's rank correlation.

3.6. Single‐Cell Sequencing Analysis of Vascular Cell Distribution Patterns of LMAGs

To investigate the cell‐specific expression features of lipid metabolism‐autophagy key genes, we conducted an in‐depth analysis of atherosclerosis‐related single‐cell transcriptome dataset (GSE159677). After quality control, a total of 46,276 cells (35,263 in the AS group and 11,013 in the control group) were obtained for subsequent analysis (Supporting File 4: Figure 2A–D). Following data dimensionality reduction, cell clustering was optimized using multi‐resolution clustering analysis, identifying 9 subpopulations annotated as 8 cell types (Supporting File 4: Figure 2E, Figure 6A,B). UMAP dimensionality reduction analysis visualized the differences in cell population distribution between the control and AS disease groups (Supporting File 4: Figure 2F). Box plots showed differences in cell types between the two groups, with an increased proportion of T cells and macrophages in the AS group (Figure 6C), further revealing the importance of immune cells in the progression of AS. Subsequently, using the AddModuleScore function, a comprehensive analysis of cell subpopulations in the vasculature was performed, demonstrating that macrophages exhibited more active lipid metabolism‐autophagy regulatory activity compared to other cell types (Figure 6D,E). These findings are consistent with existing research and suggest that the synergistic effect of lipid metabolism disorder and autophagy dysregulation is an important mechanism in the development of cardiovascular diseases [7, 11]. To elucidate the molecular mechanisms underlying this cell‐specific heterogeneity, we focused on three key lipid metabolism‐autophagy regulators (PRKCB, NLRC4, and TNFSF10) that were previously identified from bulk RNA‐seq data. Their differential expression patterns and characteristics were then investigated through multi‐dimensional analyses. Dimensionality reduction clustering analysis showed that these three genes had spatial distribution differences between specific cell subpopulations in the control and AS groups (Figure 6F–H). NLRC4 expression levels were significantly higher in macrophage populations in the AS group. PRKCB and TNFSF10 were expressed in T cells, smooth muscle cells, macrophages, endothelial cells, and fibroblasts in the AS group, suggesting that they may participate in disease progression by regulating intercellular cooperative functions (Figure 6I–K). The above evidence confirms the close association between key LMAGs and the pathological process of AS at the single‐cell level.

Figure 6.

Results of single‐cell data analysis. A. Cell‐type‐specific markers for cluster annotation. B. Cell type identification and annotation. C. Differences in the proportion of cell subpopulations between the control and disease groups. D. Distribution of LMAG scores in different vascular cells. E. Differences in LMAG scores among cell subpopulations. F‐H. Distribution differences of lipid metabolism‐autophagy core genes in cell subpopulations between the disease group and the control group. I‐K. Distribution of the three key genes in cell subpopulations.

3.7. Drug Screening Targeting Key Lipid Metabolism‐Autophagy Factors

To explore candidate drugs targeting key regulators of lipid metabolism and autophagy in AS, we conducted a screening using the Drug Gene Interaction Database (DGLdb) and the Therapeutic Target Database (TTD). The results indicated that no available drugs targeting NLRC4 were identified; 16 drugs targeting PRKCB were found, 5 of which have been approved by regulatory agencies; additionally, 8 candidate drugs targeting TNFSF10 were screened, though none have been approved (Figure 7A). Among these drugs, aside from quercetin and atenolol, most are primarily used for treating cancer and immune‐related diseases. Quercetin has been demonstrated to exert anti‐atherosclerotic effects by modulating macrophage M1/M2 phenotypic polarization and autophagy activity [12, 13]. Atenolol reduces cardiac output and blood pressure by blocking β₁‐adrenergic receptors, thereby alleviating mechanical stress on the arterial wall and slowing atherosclerosis progression [14]. To evaluate the potential interaction of the two approved drugs with PRKCB, a key target in atherosclerosis lipid metabolism and autophagy, we simulated the binding modes of quercetin and atenolol to the PRKCB protein using molecular docking. The binding energy of quercetin to the PRKCB active site (−8.3 kcal/mol) was significantly higher than that of atenolol (−6.2 kcal/mol), suggesting that quercetin may possess stronger targeted binding capability (Figure 7B).

Figure 7.

Drug screening targeting key factors of lipid metabolism‐autophagy. A. Screening for potential drugs targeting PRKCB and TNFSF10. B and C. Molecular docking binding modes of quercetin and atenolol with PRKCB, respectively.

3.8. Verification of Key Targets of Lipid Metabolism‐Autophagy

To verify the diagnostic value of the three core factors of lipid metabolism‐autophagy in AS, we compared the expression of these factors in aortic tissues of ApoE^‐/‐ mice fed a high‐fat diet for 16 weeks with those of mice fed a normal diet. The results showed that, compared with mice fed a normal diet, mice fed a high‐fat diet not only exhibited more severe atherosclerotic lesions (Figure 8A–D), but also had significantly upregulated mRNA and protein expression levels of PRKCB, NLRC4, and TNFSF10 in their aortic tissues (Figure 8E–I). Immunohistochemical analysis further confirmed the high expression of these three factors in the AS plaque area (Figure 8J,K). Additionally, ELISA assays showed that the plasma levels of TNFSF10 were also significantly elevated in mice fed a high‐fat diet (Figure 8L). These results indicate that PRKCB, NLRC4, and TNFSF10 may participate in the progression of AS by regulating the lipid metabolism‐autophagy pathway.

Figure 8.

Animal experiments validate the expression of key lipid metabolism‐autophagy targets in atherosclerosis. A and B. Representative images and quantification of Oil Red O‐stained en face aorta. C and D. Representative images of Oil Red O staining of the aortic roots, and the related quantitative analysis. E‐G. The mRNA expression levels of three key lipid metabolism‐autophagy genes in aortic tissue. H and I. Representative immunoblots and expression quantification of the three key lipid metabolism‐autophagy proteins. J and K. IHC staining of three key proteins in the aortic root from control and AS groups. L. TNFSF10 levels in the serum of control and AS group mice. Comparisons between groups were performed using an unpaired t‐test or the Mann‐Whitney test.

4. Discussion

Dyslipidemia is considered a key driver in the occurrence and development of AS. When lipid metabolism is abnormal, excessive low‐density lipoprotein (LDL) enters the vascular wall and oxidizes to form ox‐LDL, which is engulfed by macrophages and transformed into foam cells, forming the core of AS plaques. Autophagy, as an important metabolic regulatory mechanism of cells, reduces lipid accumulation by selectively degrading intracellular lipid droplets and promoting cholesterol excretion. Studies have shown that impaired autophagy function leads to impaired clearance of lipid droplets in macrophages, exacerbating foam cell formation and inflammatory responses, and ultimately promoting the development of atherosclerotic plaques. However, the key genes regulating the lipid metabolism‐autophagy network and their specific mechanisms of action in AS remain unclear, especially how these genes affect critical pathological processes such as the immune microenvironment and plaque stability, which require further exploration. Here, we integrated transcriptome and single‐cell data for bioinformatics analysis. Through differential analysis, 572 DEGs between the AS group and the control group were identified. GO and KEGG enrichment analysis revealed that these DEGs were mainly involved in biological processes related to immune responses, such as cell migration, adhesion, phagocytosis, immune receptor activity, and chemokine signaling pathways. There is a complex network of regulatory relationships among lipid metabolism, autophagy, and immune responses [11, 15]. Lipid metabolites can act as signaling molecules to influence autophagy and the activity of immune cells [7]. Therefore, we analyzed the relationship between lipid metabolism‐autophagy gene sets and AS pathogenesis to explore the role of these genes in the AS immune microenvironment and their potential diagnostic value. Through a combination of WGCNA and machine learning approaches, followed by validation in an independent dataset, three core genes (PRKCB, NLRC4, and TNFSF10) were ultimately identified for subsequent functional studies. Animal experiments showed that the levels of the three key factors in the aortic tissue of high‐fat diet mice were significantly increased, and the level of TNFSF10 in the plasma was also significantly elevated.

PRKCB (Protein Kinase C beta) is an important member of the PKC (Protein Kinase C) family and belongs to the classical subtype (cPKC). Its activity is regulated by Ca²⁺ and diacylglycerol (DAG). In AS, PRKCB is activated in response to various pathological stimuli such as high glucose, phorbol esters, ischemia, and hypoxia [16, 17, 18]. PRKCB can mediate cholesterol accumulation by regulating the signals of macrophage surface receptors SR‐A and CD36 [19]. Although related studies have confirmed that inhibition of PRKCB can significantly reduce the area of AS plaques, the underlying mechanism remains unclear [20, 21]. Based on bioinformatics analysis techniques, we identified it as a potential candidate gene that affects the progression of AS by regulating the lipid metabolism‐autophagy network. Although there is no direct evidence that PRKCB regulates autophagy to affect AS at present, some studies have confirmed that PRKCB‐mediated lipotoxicity can damage the acidic environment of lysosomes and lead to impaired autophagic flux [22, 23, 24].

Interestingly, single‐cell RNA sequencing data revealed that PRKCB is predominantly distributed in macrophage, T cell, and B cell populations, consistent with the results of immune infiltration analysis, suggesting that PRKCB may participate in atherosclerosis progression by regulating immune cell function. Atherosclerosis involves intricate interactions between vascular cells and immune effector cells. PRKCB overexpression not only impairs endothelial function but also promotes macrophage infiltration [17]. Mechanistically, activated PRKCB upregulates the expression of CD11c, CCL2, and IL‐1β in macrophages through the ERK1/2‐Egr‐1 pathway [25]. Conversely, PRKCB knockdown reduces monocyte‐to‐macrophage differentiation, as well as lipid phagocytosis and inflammatory responses [26]. Additionally, in diabetic nephropathy, PRKCB similarly mediates macrophage infiltration and renal interstitial fibrosis by regulating osteopontin expression [27]. Meanwhile, PRKCB promotes macrophage uptake of modified LDL and foam cell formation by upregulating scavenger receptor A expression, thereby inducing the expression of inflammatory cytokines and chemokines, which further amplifies macrophage infiltration [19]. Collectively, these lines of evidence suggest that PRKCB plays a critical role in atherosclerosis, at least in part, by influencing macrophage infiltration and function. T cell deficiency significantly delays the progression of atherosclerosis [28]. Although there is no direct evidence linking PRKCB to lymphocyte infiltration in atherosclerosis‐related studies, existing research supports the role of PRKCB in regulating lymphocyte migration and activation. Phorbol ester‐stimulated human peripheral blood T lymphocytes rely on PRKCB to regulate RhoA/Rho‐associated kinase activity and microtubule stability, thereby mediating their directional migration [29]. The selective PRKCB agonist 12‐deoxyphorbol‐13‐O‐phenylacetate (dPPA) induces T cell and B cell proliferation as well as IL‐2 pathway activation, confirming that PRKCB is a key signaling molecule in lymphocyte activation [30]. In summary, PRKCB may play an important role in the progression of atherosclerosis by regulating immune cell infiltration, migration, and activation.

By using online screening of drug databases and molecular docking technology, we discovered two important compounds targeting PRKCB for the treatment of AS. Among them, quercetin has a higher binding energy with the active site of PRKCB and plays an important role in regulating autophagic activity, suggesting that targeting PRKCB may exert anti‐AS effects by improving autophagic flux [12, 31]. Quercetin is a natural flavonoid that has significant protective effects on the cardiovascular system [32]. A study investigating the effects of quercetin on the development of atherosclerosis showed that quercetin intervention reduced lesions and plaque formation [33]. Its mechanisms of action mainly include the following aspects: 1) Improvement of lipid profiles and inflammation. Sixty percent of the studies observed that quercetin reduced total cholesterol, triglycerides, and low‐density lipoprotein cholesterol; 80% of the studies analyzed inflammatory markers, of which 23.4% reported significant decreases in tumor necrosis factor‐α and interleukin‐6 [32]. In addition, quercetin increased adiponectin levels, decreased leptin levels, and ameliorated obesity‐related metabolic disorders [34]. 2) Inhibition of inflammatory cell infiltration and migration. Quercetin reduced macrophage infiltration, inhibited monocyte chemoattractant protein‐1‐induced cell migration, and improved endothelial function [32]. 3) Anti‐oxidative stress. Quercetin inhibited NADPH oxidase activity, upregulated heme oxygenase‐1 expression, thereby reducing vascular reactive oxygen species production and reversing superoxide anion levels [35]. 4) Inhibition of lipid deposition and promotion of plaque stabilization. Quercetin reduced macrophage lipid deposition and increased collagen fiber content, thereby stabilizing plaques [36]. Although there is currently no direct evidence that quercetin exerts anti‐atherosclerotic effects by targeting PRKCB, its aforementioned multi‐target mechanisms have demonstrated significant cardiovascular protective effects. Moreover, multiple studies have indicated that targeting PRKCB is an important strategy for improving atherosclerosis, suggesting that quercetin may exert protective effects by antagonizing PRKCB or reducing its protein stability [21, 25].

NLRC4, a key component of the inflammasome, recruits and activates caspase‐1 via its CARD domain, triggering the inflammasome assembly pathway. Compared with healthy individuals, AS patients exhibit upregulated NLRC4 expression accompanied by inflammasome activation and increased IL‐8 production, which exacerbates atherosclerotic lesions [37, 38]. As a major component of atherosclerotic plaques, cholesterol crystals can induce inflammasome activation and trigger a series of inflammatory responses both in vivo and in vitro, revealing the mechanistic link between lipid metabolism and inflammation. Moderate inflammatory responses activate autophagy, whereas excessive inflammasome activation leads to impaired or defective autophagy [39, 40]. NLRC4 modulates autophagic flux by interacting with Beclin1, a key regulator of autophagy [39, 41]. These findings suggest that targeting lipid metabolism to suppress inflammasome activation and restore autophagy function may be crucial for ameliorating AS.

TNFSF10 encodes tumor necrosis factor‐related apoptosis‐inducing ligand (TRAIL), whose role in AS remains controversial. Some studies have found that earlier and more extensive AS lesions are associated with TRAIL deficiency [42], and low levels of TRAIL are closely related to the severity of acute ischemic stroke [43]. In contrast, other research indicates that TRAIL plays a significant role in accelerating AS progression by mediating the formation of membrane raft‐associated redox signaling and activating the p38 MAPK signaling pathway, thereby upregulating SR‐AI and enhancing lipid uptake [44, 45]. These divergent observations may reflect distinct compensatory mechanisms during different stages of disease pathogenesis. The TRAIL signaling pathway is the most complex among TNF family members, and it is currently known to bind five different receptors [46]. Binding of TRAIL to the death domain‐containing receptors DR4 (TRAIL‐R1) and DR5 (TRAIL‐R2/APO2/Killer/TRICK2) induces apoptotic signaling, whereas binding to the decoy receptors DcR1 (TRAIL‐R3) and DcR2 (TRAIL‐R4) inhibits apoptosis [47, 48]. In addition, TRAIL can bind to the secreted receptor osteoprotegerin (OPG), thereby regulating osteoclastogenesis [49]. TRAIL and its receptors are expressed in normal arteries, atherosclerotic lesions, and Monckeberg's sclerosis, and are distributed across various cell populations, including vascular smooth muscle cells, macrophages, endothelial cells, and inflammatory cells within the vessel wall, which is consistent with our single‐cell analysis results [46, 50]. Endothelial cells possess anti‐inflammatory, antithrombotic, and anti‐atherosclerotic properties, and their dysfunction is an early manifestation of atherogenesis [51]. Studies have shown that TRAIL pretreatment reduces the production of multiple chemokines, thereby improving leukocyte‐endothelial cell adhesion [52]. However, TRAIL can also play a critical role in endothelial cell injury and neointima formation through the activation of distinct receptors [53]. These findings suggest that TRAIL plays a dual role in the early initiation and progression of atherosclerosis. Interestingly, levels of TRAIL in both monocytes and the circulation are reduced in patients with cardiovascular disease [54]. The Canakinumab Anti‐inflammatory Thrombosis Outcomes Study demonstrated that IL‐18 is associated with residual inflammatory risk [55]. A negative correlation exists between IL‐18 and TRAIL, while anti‐inflammatory treatment with colchicine increases plasma TRAIL levels, supporting the anti‐inflammatory role of TRAIL [56]. Indeed, macrophages lacking TRAIL exhibit enhanced inflammatory responses, reduced efferocytosis, impaired cholesterol handling, and diminished migratory capacity—features characteristic of dysfunctional macrophages in lesions, which may accelerate the progression of atherosclerosis [53]. Additionally, Michowitz et al. found that TRAIL protein expression was increased in cells derived from rupture‐prone plaques compared with those from stable plaques [57]. Vascular smooth muscle cells also express DR4 and DR5 and are susceptible to apoptosis induced by TRAIL‐expressing CD4⁺ plaque cells [58]. Both pharmacological inhibitors and targeted suppression of key autophagy factors (e.g., Beclin 1 or ATG7) have been shown to enhance TRAIL‐induced apoptosis, confirming that TRAIL counteracts its pro‐apoptotic effects by activating protective autophagy pathways [59, 60, 61]. This suggests that activation of TRAIL signaling under certain conditions may be detrimental, and blocking this signaling may have therapeutic potential. The role of TRAIL and its receptors in cardiovascular function warrants further investigation.

In summary, through bioinformatic analyses and murine experiments, we identified PRKCB, NLRC4, and TNFSF10 as key regulatory genes at the intersection of lipid metabolism and autophagy. The occurrence and development of atherosclerosis are the combined result of multiple risk factors. The assessment of either lipid metabolism or autophagy activity alone is insufficient to elucidate the complex process of AS occurrence. There exists a close co‐regulatory relationship between dyslipidemia and autophagy [62]. Accordingly, researchers have proposed the more precise concept of “lipophagy” to specifically describe a selective autophagic process triggered by disordered lipid metabolism, which degrades excess lipids via lysosome‐dependent pathways to maintain cellular homeostasis [11, 63]. Therefore, actively exploring the roles of key lipid metabolism‐autophagy genes in the progression of AS is of paramount importance. Additional functional intervention studies are warranted to further verify the regulatory functions of the aforementioned three core genes in atherosclerosis [64].

4.1. Limitations

This study has certain limitations. First, the specific mechanisms of the three core genes (PRKCB, NLRC4, and TNFSF10) in atherosclerosis still require further validation through molecular biology experiments. Second, the analysis of the association between the lipid metabolism‐autophagy gene set and the immune microenvironment is primarily based on bioinformatics analysis, lacking direct experimental evidence regarding the dynamic changes of immune cells in early and advanced plaques. Future studies should combine functional intervention experiments with multi‐time‐point dynamic observations to elucidate the causal relationships and cell type‐dependent roles of these genes in the initiation and progression of atherosclerosis, thereby providing a more solid theoretical foundation for their clinical translation. Despite certain limitations, this work identified key genes in the lipid metabolism‐autophagy interaction regulation in AS, providing insights into a deeper understanding of the pathogenesis of AS.

5. Conclusion

This study successfully identified three core genes (PRKCB, NLRC4, and TNFSF10) regulating the lipid metabolism–autophagy pathway through bioinformatic analysis and experimental validation. These genes not only participate in the pathological progression of AS by modulating immune cell infiltration but also demonstrate significant diagnostic value. Furthermore, continued investigation into their specific functions and interactions will deepen the understanding of AS pathogenesis, paving the way for developing targeted therapeutic strategies and their eventual clinical translation.

Author Contributions

All authors contributed to this study. Min Li conducted data collection and analysis, and wrote the initial draft. Yanjun Sun conceived the study and designed the experimental scheme. Yanyi Sun participated in the experimental design and performed the animal experiments. Zhenyue Chen was responsible for critical revision and final approval of the manuscript.

Ethics Statement

All animal procedures were approved by the Animal Experiment Ethics Committee of Shanghai Jiao Tong University and strictly adhered to its guidelines (No. RJ2023015).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1

Supporting File 2

Supporting File 3

Supporting File 4

Data Availability Statement

The data that supports the findings of this study are available in the supporting material of this article.