Atherosclerosis (AS) is a chronic inflammatory disease of the arterial intima. It is typically manifested by lipid deposition, proliferation of smooth muscle cells (SMCs), and fibrous tissue hyperplasia, ultimately leading to the formation of arterial plaques, luminal narrowing, and impaired blood flow (Jebari-Benslaiman et al., 2022). As a common pathological basis for various cardiovascular diseases, AS can directly contribute to severe conditions such as hypertension, myocardial infarction, and ischemic stroke (Virani et al., 2020). According to global epidemiological statistics, cardiovascular diseases account for up to 18 million deaths annually, consistently ranking as the leading cause of human mortality (Di Cesare et al., 2024). Notably, alterations in lifestyle and diet, factors including elevated consumption of sugar, fat, protein, and cholesterol, alongside growing smoking and alcohol use, have thereby contributed to the rising global incidence of AS, with a noticeable trend toward earlier onset (Chen W. et al., 2023). Current clinical management of AS primarily involves pharmacological interventions and surgical treatments. Drug therapy regimens include lipid-lowering agents (such as statins, ezetimibe, and PCSK9 inhibitors), antiplatelet agents, microcirculation-improving drugs, and thrombolytics (Libby, 2021). However, even with standardized lipid-lowering therapy, the risk of major cardiovascular and cerebrovascular events is reduced by only approximately 30%, and a considerable proportion of patients still experience plaque progression. Antiplatelet therapy increases the risk of gastrointestinal bleeding, while interventional procedures such as angioplasty, stent implantation, and bypass grafting, although improving perfusion, are limited by complications including postoperative restenosis and thrombosis (Crea, 2023; Borovac et al., 2019; Ma et al., 2023). Hence, gaining a comprehensive insight into the pathogenesis of AS and establishing dependable prognostic indicators for vulnerable populations are paramount for enabling early intervention, timely diagnosis, and decelerating the progression of the disease.

The pathogenesis of AS involves multiple molecular and cellular mechanisms, including lipid infiltration and modification, inflammatory responses, oxidative stress, and cell death. Among these, metabolic dysregulation is not only a central feature but may also drive disease progression through interactions with other mechanisms. Under physiological conditions, cardiac energy metabolism primarily relies on mitochondrial oxidative phosphorylation of fatty acids, while other metabolites such as glucose, amino acids, and ketone bodies also contribute to energy supply. However, under pathological conditions, significant reprogramming of metabolic pathways and preferences occurs, manifesting as metabolic remodeling or reprogramming (Gandoy-Fieiras et al., 2020). The development and advancement of AS are intimately linked to metabolic reprogramming (MR) in macrophages, SMCs, and endothelial cells (Dai et al., 2024). Within the pathological microenvironment, infiltrated macrophages exhibit enhanced glycolysis to meet energy demands, accompanied by impaired mitochondrial oxidative phosphorylation. This metabolic shift not only promotes the release of pro-inflammatory cytokines (e.g., IL-1β, TNF-α) but also increases the generation of reactive oxygen species (ROS), which further oxidize low-density lipoprotein (LDL) to form more pathogenic oxidized LDL (oxLDL). Excessive oxLDL uptake causes defective cholesterol efflux in macrophages, promoting cholesterol ester deposition and irreversible foam cell formation—a key event in lipid core development and inflammation amplification (Vendrov et al., 2024; Bekkering et al., 2014). SMCs shift from oxidative phosphorylation to glycolytic dependency, supporting their proliferation, migration, and extracellular matrix synthesis. Some SMCs even transdifferentiate into macrophage-like phenotypes and similarly uptake lipids to form foam cells, further exacerbating intraplaque lipid accumulation and structural complexity (Bennett et al., 2016). Metabolic disturbances and oxidative stress in endothelial cells compromise nitric oxide bioavailability, leading to endothelial dysfunction. This dysfunction manifests as diminished vasodilation, heightened platelet aggregation, increased monocyte adhesion and infiltration, along with lipoprotein accumulation beneath the intima. These events collectively initiate early atherosclerotic lesions (Zhao et al., 2024). Notably, lipid raft microdomains on the cell membrane serve as signaling platforms and play a key role in integrating metabolic stress and inflammatory signals, thereby regulating endothelial NO synthesis and thrombotic propensity. Studies have shown that specific pathological stimuli (such as autoantibodies) can inhibit endothelial nitric oxide synthase (eNOS) phosphorylation via a lipid raft-dependent LRP8 receptor signaling pathway, exacerbating NO synthesis impairment and pro-thrombotic phenotypes (Riitano et al., 2023). These intertwined MR processes create a vicious cycle of “metabolic dysfunction–inflammatory response–lipid deposition–cellular impairment,” ultimately driving plaque growth, lipid core expansion, and increased instability. Given its central role in disease progression, systematically elucidating the key regulatory genes and networks involved in MR may facilitate the identification of reliable diagnostic biomarkers and inspire novel interventional strategies.

Nevertheless, although the critical role of MR in AS has been recognized, several key issues remain unresolved. First, the core gene regulatory network driving MR in AS has not been systematically elucidated. Second, most existing studies have focused on single pathways or cell types, lacking an analytical framework centered on MR that integrates multi-omics data to reveal its global regulatory mechanisms. Finally, whether molecular changes associated with MR can be translated into robust diagnostic biomarkers with clinical utility—specifically for early screening, patient risk stratification, or plaque vulnerability assessment—remains to be explored. Based on these challenges, we propose the central hypothesis that the progression of AS is governed by a core MR module composed of key genes and their upstream transcription factors. This module serves as a molecular hub linking immune–metabolic dysregulation with plaque progression and holds potential as an effective diagnostic biomarker. To test this hypothesis, we developed a multi-omics integrative analytical framework with MR as the focal point. By integrating transcriptomic, single-cell sequencing, and immune microenvironment data, we aimed to systematically map the overall landscape of metabolic regulatory networks in AS. Methodologically, we established a screening pipeline that combines weighted gene co-expression network analysis (WGCNA) with multiple machine-learning algorithms to enhance the robustness of key biomarker discovery. Using this approach, we not only identified a core gene signature consisting of LYN proto-oncogene, Src family tyrosine kinase (LYN), fatty acid binding protein 5 (FABP5), matrix metallopeptidase 9 (MMP9), and alanyl aminopeptidase, membrane (ANPEP) but also revealed the transcription factor early growth response 1 (EGR1) as a potential common upstream regulatory node for these genes, providing new insights into the transcriptional cascade governing MR. A diagnostic model constructed based on these four core genes suggested good discriminative performance in an independent validation cohort, offering potential experimental support for the development of novel molecular tools for early AS screening and risk stratification (Figure 1). In summary, this study adopted a research paradigm that integrates multi-omics data, computational screening, and experimental validation to systematically uncover key regulatory genes and networks involved in AS-MR, thereby providing novel insights for a deeper understanding of its pathological mechanisms and the development of innovative diagnostic strategies.

AS, as the primary pathological basis of cardiovascular diseases, has become a major driver of global mortality and disease burden. Despite continuous advancements in the prevention and treatment of cardiovascular diseases, patients with AS still face significant clinical challenges, particularly the difficulty in early diagnosis, due to the current lack of highly sensitive and specific biomarkers. MR is a critical biological process through which cells alter their metabolic patterns in response to environmental stresses to meet energy and biosynthetic demands. In recent years, disease-associated metabolic abnormalities have become a major research focus, with the metabolic features of AS attracting particular attention. Studies have revealed a correlation between the metabolic state of plaques and their clinical risk. In 2002, Rudd et al. reported that symptomatic unstable plaques exhibited significantly higher uptake of [¹⁸F]-fluorodeoxyglucose compared to asymptomatic plaques, while healthy carotid tissue showed no such uptake (Rudd et al., 2002). This finding was later expanded by Tomas et al., who demonstrated that highly vulnerable plaques display a characteristic metabolic profile: compared to low-risk plaques, they show reduced fatty acid oxidation (FAO) but significantly enhanced glycolysis and amino acid metabolism. Interestingly, these plaques showed much higher levels of mRNA expression for genes involved in the pentose phosphate system and glycolysis. Significantly higher levels of mRNA expression for genes linked to the pentose phosphate pathway and glycolysis were seen in these plaques. This MR signature was significantly associated with a pro-inflammatory state and poorer clinical outcomes, including an increased incidence of cardiovascular events during a 7-year follow-up (Tomas et al., 2018). These pieces of evidence suggest that systematically decoding the regulatory network of MR holds promise for providing new targets for disease diagnosis and treatment. Based on the aforementioned background, this study developed a multi-omics integration framework centered on MR, aiming to systematically identify key regulatory genes and their clinical translational potential in AS. We identified a core gene signature composed of four genes (LYN, FABP5, MMP9, ANPEP). This signature was consistently upregulated in AS tissues and showed excellent diagnostic performance in independent validation cohorts. Further analysis revealed that the transcription factor EGR1 may serve as a common upstream regulatory node coordinating the expression of these core genes. Single-cell transcriptomic and immune infiltration analyses indicated that the expression of these genes was significantly correlated with the enrichment of pro-inflammatory immune cells (e.g., M0 macrophages), suggesting their potential role in shaping the AS immune microenvironment. Finally, experimental validation in an Apoe ^−/− mouse model revealed the significant upregulation of these core genes and Egr1 in AS lesions.

The non-receptor tyrosine kinase encoded by LYN, a member of the Src family of kinases, has several regulatory functions in the development of AS. First, by stimulating macrophage scavenger receptors (including CD36 and SR-A), LYN encourages the uptake of oxLDL, resulting in intracellular lipid buildup and the generation of foam cells, which is a crucial stage in the development of atherosclerotic plaques (Rahaman et al., 2006). Additionally, LYN contributes to fibrillary amyloid-triggered CD36-dependent signaling cascades that increase ROS and tumor necrosis factor-alpha (TNF-α) production, aggravating macrophage inflammatory responses and promoting AS development (Medeiros et al., 2004). LYN can increase the production of monocyte chemoattractant protein-1 (MCP-1), which promotes monocyte recruitment and exacerbates vascular inflammation. It also plays a significant role in lipid metabolism (Miki et al., 2001). YOU et al. found that inhibiting LYN and its downstream factor Akt increased free fatty acid levels and enhanced mitochondrial fatty acid oxidation, driving macrophages toward an anti-inflammatory phenotype (You et al., 2020). Notably, the role of LYN is cell type- and time-dependent; regulated by the SNX10–Lyn–TFEB axis, it can also influence macrophage energy metabolism and phenotypic switching. These studies indicate that LYN not only promotes AS progression through inflammatory and lipid metabolic pathways but may also modulate macrophage polarization via the regulation of energy metabolism.

Intracellular lipid chaperones known as fatty acid-binding proteins (FABPs) are extensively expressed in macrophages and adipocytes and are essential for controlling inflammation and energy metabolism (Furuhashi and Hotamisligil, 2008; Furuhashi et al., 2011). FABP5 is one of those that has attracted a lot of interest because of its crucial roles in AS. Mounting evidence indicates that FABP5 is not only a potential biomarker for AS but also a promising therapeutic target (Yeung et al., 2008; Bagheri et al., 2010; Hong et al., 2011). Mechanistic studies have demonstrated that loss of FABP5 alleviates inflammation and modulates monocyte recruitment by activating PPARγ activity in macrophages, thereby effectively attenuating the progression of AS (Babaev et al., 2011). Moreover, FABP5 levels are closely associated with macrophage cholesterol efflux capacity (CEC), which in turn is inversely correlated with residual risk of atherosclerotic cardiovascular disease (ASCVD). Studies have shown that reduced FABP5 expression significantly enhances CEC, offering a novel interventional strategy for ASCVD prevention (Furuhashi et al., 2017).

MMP9 is a key member of the zinc-dependent endopeptidase family and participates in tissue remodeling through the degradation of the extracellular matrix (Nagase et al., 2006). According to clinical research, patients with plaque rupture have noticeably higher plasma MMP9 levels than healthy people (Kosowski et al., 2022). During the development and progression of AS, MMP9 contributes to pathological changes through multiple mechanisms: (a) regulation of lipid metabolism; (b) exacerbation of vascular inflammation; (c) induction of endothelial dysfunction; (d) promotion of smooth muscle cell proliferation and migration; and (e) acceleration of plaque calcification and destabilization (Brown et al., 2017; Myasoedova et al., 2018; Wang and Khalil, 2018). Existing pharmacological studies further underscore the therapeutic relevance of MMP9. Statins can effectively inhibit collagen degradation in advanced atherosclerotic plaques by downregulating MMP9 expression (Stasinopoulou et al., 2019). Meanwhile, metformin directly binds to MMP9 and facilitates its degradation, thereby significantly improving plaque stability (Chen X. et al., 2023). These results support the potential of MMP9 as a therapeutic target and suggest its involvement in AS.

ANPEP (CD13), a widely expressed ectoenzyme, regulates various biological processes—including peptide activity modulation, phagocytosis, and cholesterol metabolism—by cleaving N-terminal amino acids (Mina-Osorio, 2008). In AS, ANPEP exhibits complex and likely cell type-specific regulatory roles. On the one hand, research indicates that ANPEP regulates the absorption of cholesterol and could be a target of the medication bexarotene. By downregulating the expression of Anpep and Npc1l1, bexarotene significantly decreased intestinal cholesterol uptake and enhanced plasma cholesterol levels in an Apoe2-KI animal model fed a high-fat diet, thereby slowing the progression of AS (Lalloyer et al., 2006). On the other hand, some evidence indicates that deficiency of ANPEP may exacerbate AS. Devarakonda et al. reported that Anpep-deficient cells exhibit heightened sensitivity to oxidative stress inducers such as oxLDL, leading to aberrant increases in TFs levels and enhanced cell death, suggesting a protective role of Anpep against AS via oxidative stress resistance (Devarakonda et al., 2019). This apparent contradiction may reflect spatiotemporal specificity of ANPEP function, implying divergent roles across different cell types or disease stages.

Early Growth Response 1, or EGR1, is a zinc-finger TF that is quickly activated in response to a variety of extracellular stimuli, such as mechanical pressures, cellular stress, growth hormones, and inflammatory cytokines. In the development and course of AS, EGR1 is a crucial regulator. Research has shown that EGR1 is markedly elevated in both human atherosclerotic lesions and Ldlr ^−/− animal models (McCaffrey et al., 2000). Furthermore, its expression level within plaques shows a strong positive correlation with those of its downstream target genes—such as TGF-β1, TNF-α, ICAM-1, and M-CSF—which supports a potential role for EGR1 in transcriptional regulation during AS pathogenesis (McCaffrey et al., 2000).

The immune microenvironment of AS is a dynamically evolving inflammatory network, wherein the function and fate decisions of immune cells are tightly coupled with their metabolic states. Recent advances in single-cell and spatial transcriptomics have revealed the substantial heterogeneity of this network in terms of cell types, states, and spatial distributions. For example, studies have identified multiple macrophage subpopulations in human carotid plaques, including pro-inflammatory (e.g., IL1B or C1Q), efferocytotic, proliferative, and SMC-like subsets, alongside highly heterogeneous populations of T cells, SMCs, and fibroblasts (Bashore et al., 2024). A more recent study further employed single-cell RNA sequencing combined with spatial transcriptomics to systematically map senescent vascular cells in a mouse AS model. This work identified distinct senescent subpopulations in vascular SMCs, fibroblasts, and T cells, and found that these subpopulations were enriched in the plaque “cap” and “core” regions (Mazan-Mamczarz et al., 2025). This suggests that senescence, as a special cellular state, constitutes a significant dimension of cellular heterogeneity within the atherosclerotic microenvironment. Classic studies have shown that the metabolic state of immune cells is a determining factor for their phenotype and function. Pro-inflammatory M1 macrophages and effector T cells rely on glycolysis for rapid energy production, whereas regulatory or reparative M2 macrophages and regulatory T cells predominantly utilize oxidative phosphorylation and fatty acid oxidation (Liu et al., 2021). This MR is not only a consequence of cell activation but also serves as the engine driving their specific functions. Notably, single-cell analyses have further unveiled specific associations between this metabolic state and cellular subpopulations. For example, macrophages in symptomatic plaques exhibit higher senescence and glycolytic signatures (Bashore et al., 2024), while intraplaque T cells display a highly activated and dysfunctional state (Horstmann et al., 2024). Similarly, recent research has found that senescent vascular SMCs and fibroblasts highly express genes related to extracellular matrix remodeling, whereas senescent T cells are enriched in inflammatory pathways, indicating that senescent subpopulations of different cell types possess distinct functional and metabolic preferences (Mazan-Mamczarz et al., 2025). In this study, the core genes we identified (LYN, FABP5, MMP9, ANPEP) may deeply participate in shaping this imbalanced microenvironment by regulating the metabolic states of immune cells. Specifically, LYN and FABP5, as key nodes in lipid metabolism and inflammatory signaling, may directly regulate macrophage polarization toward a pro-inflammatory phenotype by influencing their lipid metabolic flux and energy utilization patterns. This could potentially explain why specific macrophage subpopulations (such as those associated with lipid metabolism) correlate with disease severity in single-cell atlases. The expression of MMP9 itself is driven by upstream pro-inflammatory metabolic pathways (e.g., HIF-1α signaling), and its mediation of extracellular matrix degradation can alter the mechanical and biochemical cues of the microenvironment, thereby affecting the metabolic adaptation and migratory behavior of immune cells (Yang et al., 2019). As an ectoenzyme, ANPEP may indirectly regulate the recruitment of immune cells and their surrounding metabolic stress environment by modifying local peptide signals (such as the chemokines CXCL8 and CCL2).

The findings of this study provide a novel perspective on understanding MR in AS, while the identified core biomarkers (LYN, FABP5, MMP9, ANPEP) also suggest potential for clinical translation. These genes not only exhibit high diagnostic performance but are also closely associated with the core pathological features of AS, including immune-inflammatory responses, extracellular matrix remodeling, and dysregulated lipid metabolism. Accordingly, they may play multiple roles in the comprehensive management of AS: in the early stage, these markers hold promise as non-invasive or minimally invasive detection indicators, aiding in the identification of pre-clinical or early-stage lesions, particularly enabling more precise risk stratification in populations with traditional risk factors (e.g., hyperlipidemia, diabetes); in disease progression assessment, the multi-gene model constructed in this study can provide molecular evidence for patient risk stratification, helping to distinguish between stable and progressive plaques, thereby guiding the adjustment of individualized treatment intensity; in plaque stability evaluation, the expression levels of core genes (especially MMP9) are significantly correlated with a pro-inflammatory immune microenvironment, suggesting their potential as supplementary biological markers for assessing plaque vulnerability and predicting the risk of acute cardiovascular events. Furthermore, these core genes and their regulatory pathways lay a molecular foundation for developing novel targeted therapies. Predictions based on the DGIdb database indicate that known inhibitors exist for several core genes, pointing the way for drug repurposing or the development of new agents. Of particular importance, the upstream key transcription factor EGR1 identified in this study, which acts as a “master regulatory switch” coordinating the expression of multiple core genes, provides a highly promising high-level intervention target for developing novel therapies aimed at reprogramming the plaque metabolic-immune microenvironment, with the potential to achieve deeper modulation of disease progression.

Although this study has revealed the critical roles of the core genes (LYN, FABP5, MMP9, and ANPEP) and the transcription factor EGR1 in MR during AS through multi-omics analysis and experimental validation, several limitations should be acknowledged. First, the primary mechanistic validation in this study relied on the Apoe ^−/− mouse model. While this model is a standard tool for AS research, inherent species differences exist between mice and humans in the pathophysiology of AS (e.g., plaque composition, immune responses, and metabolic regulatory networks). Therefore, the universality and translational value of the core gene regulatory network identified here, particularly the upstream dominant role of EGR1, require further confirmation using clinical samples or models that more closely resemble human pathology (e.g., organoids or humanized models). Second, research based on high-dimensional omics data and complex bioinformatics pipelines inherently carries a risk of false-positive findings. Although we employed multiple dataset integration, cross-validation with various machine learning algorithms, and independent animal experimental validation to maximize the robustness of our screening results, the association between the selected biomarkers and AS remains observational and predictive. The precise causal roles of these genes in disease pathogenesis and progression, as well as the specific molecular mechanisms by which EGR1 transcriptionally regulates them, ultimately require confirmation through orthogonal experimental strategies such as functional gain-/loss-of-function experiments (e.g., CRISPR-Cas9 gene editing) coupled with phenotypic rescue studies. Future research should combine multi-center clinical cohorts to validate the diagnostic potential of these biomarkers and employ cell type-specific gene-editing technologies to dissect their functions in depth. Simultaneously, targeted intervention strategies based on the core genes should be explored, integrating spatial transcriptomics, metabolomics, and epigenomics data to comprehensively unravel the regulatory network of MR in AS and provide novel strategies for precision therapy.

This study established a multi-omics and machine learning integrated analytical framework centered on MR to systematically identify and validate four core genes (LYN, FABP5, MMP9, and ANPEP) that play critical roles in AS. Our analysis suggested that the transcription factor EGR1 could be an upstream regulator of these genes. Furthermore, the altered expression of these core genes was validated in an animal model. By integrating single-cell sequencing and immune infiltration analysis, this study elucidated the close association between these core genes and the remodeling of a pro-inflammatory immune microenvironment. In summary, these findings not only deepen the understanding of the pathogenesis of AS from the novel perspective of “metabolism-immune” interactions but, more importantly, the identified gene signature and its regulatory network provide crucial molecular targets and a theoretical foundation for developing novel biomarkers applicable to the full-cycle management of AS, as well as for devising precision therapeutic strategies targeting MR.