High expression levels of haem oxygenase-1 promote ferroptosis in macrophage-derived foam cells and exacerbate plaque instability

Abstract

Plaque rupture with consequent thrombosis is the leading cause of acute cardiovascular events, during which macrophage death is a hallmark. Ferroptosis is a pivotal intermediate link between early and advanced atherosclerosis. Existing evidence indicates the involvement of macrophage ferroptosis in plaque vulnerability; however, the exact mechanism remains elusive. The aim of this study was to explore key ferroptosis-related genes (FRGs) involved in plaque progression and the underlying molecular mechanisms involved. The expression landscape of FRGs was obtained from atherosclerosis-related GEO datasets. Molecular mechanism studies of ferroptosis were performed using bone marrow-derived macrophages (BMDMs) and macrophage-derived foam cells (MDFCs). Bioinformatics analysis and immunohistochemistry revealed that macrophage haem oxygenase-1 (HMOX1) is the key FRG involved in plaque destabilization. Hypoxic conditions induced a significant increase in Hmox1 expression in MDFCs but not in macrophages. In addition, the beneficial or deleterious effects of Hmox1 were dependent on the degree of Hmox1 induction. Hmox1 overexpression drove inflammatory responses and ferroptotic oxidative stress in MDFCs and aggravated the plaque burden in atherosclerotic model mice. Further mechanistic investigations demonstrated that hypoxia-mediated degradation of egl-9 family hypoxia-inducible factor 3 (Egln3) stabilized Hif1a, which subsequently promoted Hmox1 transcription. Our findings suggest that high Hmox1 expression under hypoxia is deleterious to MDFC viability and plaque stability, providing a reference for the management of acute cardiovascular events.

Graphical abstract

1. Introduction

Atherosclerosis (AS) is a progressive inflammatory disorder characterized by lipid accumulation and plaque formation [1]. As the leading cause of cardiovascular disease, AS can affect large and middle-sized arteries and increase cardiovascular risk [2]. AS is initiated with endothelial damage, followed by the migration of circulating monocytes into the subendothelial space and their subsequent differentiation into macrophages [3]. Subintimal macrophage internalization of modified lipids and intracellular lipid deposition are the most prominent features of early-stage AS [4]. Through scavenger receptors on the cell surface, macrophages engulf excessive amounts of low-density lipoprotein (ox-LDL) and transform into lipid-rich foam cells [3,5]. Macrophage-derived foam cells (MDFCs) are involved in all phases of atherosclerotic lesions and affect plaque status through cellular responses (i.e., cytokine secretion) [5]. The rupture of atherosclerotic plaques causes thrombosis, which is a major contributor to acute cardiovascular events [6,7]. Rupture-prone plaques are characterized by thin fibrous caps, large necrotic cores, and extensive inflammatory infiltration [[6], [7], [8]].

Since cell death is a hallmark of advanced plaque lesions, macrophage death has garnered increased attention [9]. Accumulating evidence has revealed that necrotic core formation and plaque rupture are largely attributed to macrophage death [10]. Plaque macrophages may experience different manners of death, ranging from canonical modes (i.e., apoptosis) to noncanonical modes (i.e., ferroptosis) [8,[10], [11], [12]]. Ferroptosis is an iron-catalysed manner of regulated cell death [13]. The interaction between intracellular free iron and hydrogen peroxide contributes to excessive ROS accumulation and lipid peroxidation, which promotes necrotic core formation [14]. Recent studies have also revealed significantly elevated iron levels in plaque regions, especially in macrophages [[15], [16], [17], [18]]. Given that intraplaque haemorrhage (IPH), iron overload, and redox imbalance are prominent features of advanced plaques, it is reasonable to postulate that macrophage ferroptosis is involved in atherosclerotic plaque vulnerability [10]. This study aimed to identify key ferroptosis-related genes (FRGs) involved in plaque progression and explore the underlying mechanisms involved.

2. Methods and materials

Materials and methods are available in the Supplementary data.

3. Results and discussion

3.1. Identification of macrophage HMOX1 as the key FRG for plaque destabilization

Differential expression analysis was conducted on the gene expression landscape of 69 atherosclerotic lesions and 35 control samples (GSE100927). A total of 500 DEGs were screened (Fig. S1A). Gene Ontology (GO) analysis revealed that the DEGs were enriched mainly in biological processes, including leukocyte activation (Fig. S1A). The GSVA results revealed ferroptosis pathway activation (Fig. S1A). Twelve differentially expressed FRGs (DE-FRGs, Figs. S1B–C) were obtained by mapping the identified DEGs to the ferroptosis gene set (Table S1). Further cluster analysis was performed with the NMF algorithm for the stratification of plaques on the basis of the expression landscape of DE-FRGs. The optimal factorization rank was ultimately determined, and all atherosclerotic samples were classified into 2 clusters (Fig. S1D), namely, the C1 cluster (n = 15) and the C2 cluster (n = 54). A significant difference was observed between the C1 and C2 clusters according to principal component analysis (Fig. S1E). The expression patterns of the DE-FRGs between the 2 clusters are shown in Fig. S1F.

GSVA enrichment analysis was performed on the hallmark gene set (h.all.v7.2.symbols.gmt). The results revealed that the hypoxia pathway was activated in the C2 cluster (Fig. S2A), with a greater abundance of infiltrating leukocytes (Fig. S2B). As a key element in plaque progression, hypoxia within plaques is caused mainly by a decreased oxygen supply and increased oxygen consumption [19]. DE-FRG-based cluster analysis revealed a potential correlation between hypoxia and ferroptosis, which may accelerate the growth of atherosclerotic lesions [20,21].

To explore potential mechanisms of plaque progression, differential biological characteristics between stable and unstable plaques were analysed via the GSE163154 (non-IPH and IPH) and GSE28829 datasets (stable and unstable plaques). The DEGs identified from GSE163154 were involved mainly in the inflammatory response (Fig. 1A), with activation of the ferroptosis pathway (Fig. 1B). Consistently, greater immune infiltration was observed in vulnerable plaques (Fig. 1C, Fig. S2C). To further identify key FRGs, multiple machine learning methods were used to analyse the expression profiles of 12 DE-FRGs (GSE163154). We subsequently overlapped the three sets of genes (Lasso, SVM-RFE, and random forest) with the hypoxia-related gene set (Table S2) and ultimately obtained haem oxygenase-1 (HMOX1; Fig. S2D). HMOX1 expression was significantly upregulated in vulnerable plaques (Fig. S2D) and closely related to immune infiltration (Fig. S2E). Further single-cell sequencing analysis revealed that HMOX1 was expressed mainly in macrophages (Fig. 1D, Fig. S2F). Immunohistochemical staining also confirmed the increase in HMOX1 expression in the macrophage-rich areas of unstable plaques (Fig. 1E), which was accompanied by plaque rupture, increased lipid deposition, and increased collagen volume. Previous studies have reported that hypoxic regions are rich mainly in macrophages and foam cells [19,22]. Hypoxia has been reported to regulate iron-related proteins (i.e., HMOX1), which enhance iron uptake and affect ferroptosis sensitivity [23]. On the basis of our findings, we hypothesized that hypoxic stress can induce the expression of proferroptotic HMOX1 and subsequently activate macrophage ferroptosis.

Fig. 1.

Identification of macrophage Hmox1 as the key FRG involved in plaque destabilization

A. GO enrichment analysis of DEGs (GSE163154) in terms of BP, CC, and MF. B. Upregulated and downregulated signalling pathways in non-IPH and IPH (GSE163154) samples based on GSEA enrichment analysis. C. Infiltration of 28 types of immune cells in non-IPH and IPH (GSE163154) patients. D. Assessment of Hmox1 expression in macrophages via scRNA-seq datasets. E. Histological analysis of stable and unstable plaques (HE, Masson, and oil red O staining) and immunohistochemistry analysis of Hmox1 expression. GSEA, gene set enrichment analysis; DEGs, differentially expressed genes; BP, biological process; CC, cellular component; MF, molecular function; IPH, intraplaque haemorrhage. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

3.2. Hypoxia upregulates Hmox1 expression and activates ferroptosis in MDFCs

We further investigated the effect of hypoxia on macrophages via the GSE16099 dataset (Table S3). DEGs were identified between the normoxia and hypoxia groups, with significant upregulation of HMOX1 expression observed in hypoxia-treated macrophages (Fig. 2A). The results of the GO and KEGG analyses are shown in Fig. 2A.

Fig. 2.

Hypoxia promotes Hmox1 expression in MDFCs and drives ferroptotic cell death.

A. Volcano plot of DEGs in GSE16099 and enrichment analysis (GO enrichment analysis and KEGG pathway analysis). B. Schematic representation of BMDM isolation and foam cell formation. C. Western blot analysis of macrophages and ox-LDL-treated macrophages (MDFCs) in the presence or absence of hypoxia. D. CCK-8 assays were used to assess the effects of different inhibitors on MDFC activity under hypoxia. E. The relative GSH/GSSG ratios were assayed. F. Iron and lipid peroxidation levels of macrophages and MDFCs in response to hypoxia. Scale bar, 50 μm. MDFC, macrophage-derived foam cell; DEGs, differentially expressed genes; BP, biological process; CC, cellular component; MF, molecular function; ns P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001.

BMDMs were treated with ox-LDL to induce MDFCs (Fig. 2B). We failed to observe differences in the expression of Hmox1 in macrophages exposed to normoxia or hypoxia (Fig. 2C). However, the expression levels of Hmox1 substantially increased in MDFCs under hypoxia. In addition, hypoxia inhibited the viability of MDFCs but had no significant effect on macrophages (Fig. 2D). To verify whether ferroptosis plays an essential role in hypoxia-mediated cell death, MDFCs were treated with various cell death inhibitors. Our results revealed that apoptosis and necrosis inhibitors failed to exert a protective effect, whereas the inhibition of ferroptosis protected against hypoxia-mediated cell inactivation (Fig. 2D). As shown in Fig. 2E, hypoxia contributed to a substantial reduction in the GSH/GSSG ratio in MDFCs, accompanied by markedly increased iron content and lipid peroxidation levels (Fig. 2F), which were not observed in macrophages. Our findings suggest that ferroptosis plays a key role in the cytotoxicity of MDFCs and is mediated by hypoxic stress.

As a rate-limiting enzyme in the haem oxygenase pathway, HMOX1 catalyses the degradation of haem to carbon monoxide, ferrous iron, and bilirubin (Fig. S3A) [24]. HMOX1 is widely accepted to be an effective cardioprotective protein [25]. Previous studies have demonstrated that genetic defects in HMOX1 render vascular cells vulnerable to oxidative injury [26]. In addition, the pharmacological induction or gene transfer of Hmox1 can alleviate atherosclerotic lesions [27]. Thus, HMOX1 induction via endogenous factors or pharmaceutical approaches is considered a defence mechanism against oxidative stress [27]. However, HMOX1 induction has also been reported to be correlated with cell dysfunction since iron retention induced by HMOX1 can aggravate iron overload and drive ferroptotic oxidative stress [26,[28], [29], [30]]. For example, Hmox1 induction can drive cardiac ferroptosis and induce cardiomyopathy in mice with sickle cell disease [31]. HMOX1 may play both protective and detrimental roles in the cardiovascular system [28,32]. This difference may depend on the expression level of Hmox1 [33]. Recent studies have demonstrated that low Hmox1 expression prevents retinal degeneration but that high Hmox1 expression has the opposite effect [33]. To explore the paradoxical effects of Hmox1, MDFCs were treated with different concentrations of Haemin. MDFCs treated with haemin at relatively low concentrations (≤50 μM) presented a slight increase in Hmox1 expression (Fig. S3B). No significant differences were observed in terms of cell viability, the GSH/GSSG ratio or ROS levels (Figs. S3C–E). Although lower levels of haemin induced increased iron content, it did not have an obvious effect on intracellular lipid peroxidation (Fig. S3F). However, high concentrations of haemin (≥75 μM) significantly increased Hmox1 expression, with considerable decreases in cell viability and the GSH/GSSG ratio. In addition, high doses of haemin led to excessive ROS accumulation, increased iron overload and lipid peroxidation. Our findings suggest that high expression levels of Hmox1 may induce MDFC ferroptosis.

3.3. High levels of Hmox1 promote MDFC ferroptosis and exacerbate plaque instability

Further loss- and gain-of-function experiments were performed to explore the effects of Hmox1. Hmox1 silencing in MDFCs (Fig. S4A) was implemented with siRNA duplex oligonucleotides (siHmox1). The administration of siHmox1 inhibited Hmox1 protein expression in MDCFs under hypoxia and antagonized hypoxia-mediated expression of Gpx4. (Fig. S4B). In addition, Hmox1 knockdown increased cell viability and the GSH/GSSG ratio, reduced intracellular ROS accumulation and Fe²⁺ overload, and alleviated lipid peroxidation (Figs. S4C–F). MDFCs transduced with an Hmox1-overexpressing vector (oeHmox1) presented significantly increased Hmox1 expression and decreased Gpx4 expression (Fig. 3A). High levels of Hmox1 induced ferroptotic cell death, which was characterized by ROS overproduction, increased iron retention, and lipid peroxidation (Fig. 3B and C).

Fig. 3.

Hmox1 overexpression drives ferroptotic cell death and exacerbates plaque instability.

A. Western blot analysis of Hmox1 and Gpx4 expression in MDFCs transduced with a control or Hmox1 overexpression vector (oeHmox1). B. Measurement of cell viability and intracellular ROS levels in control and Hmox1-overexpressing MDFCs. C. Measurement of intracellular iron levels and lipid peroxidation levels in control and Hmox1-overexpressing MDFCs. Scale bar, 50 μm. D. The top 20 upregulated and downregulated DEGs in the GSE28829 (stable and unstable plaque) and GSE163154 (non-IPH and IPH) datasets. Pearson correlation analysis showing the correlations of Hmox1 with CCLs and MMPs. E. The mRNA levels of Ccl3, Ccl4, Ccl8, Ccl19, Mmp7, Mmp9 and Mmp12 were assessed in Hmox1-overexpressing MDFCs. F. Representative images of transverse sections of brachiocephalic artery (BCA) lesions stained with haematoxylin and eosin, Masson and Oil Red O. Scale bar, 250 μm. MDFCs, macrophage-derived foam cells; DEGs, differentially expressed genes; IPH, intraplaque haemorrhage; CCL, C–C motif chemokine ligand. ns P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Since that HMOX1 is positively correlated with immune infiltration, we aimed to identify related genes and explore the potential mechanisms involved. The top 20 upregulated and downregulated DEGs in GSE163154 and GSE28829 are shown in Fig. 3D. Both heatmaps underscore the importance of C–C motif chemokine ligands (CCLs) and matrix metalloproteinases (MMPs). Correlation analysis suggested that HMOX1 was positively associated with CCLs and MMPs. Hmox1 overexpression promoted the production of Ccl3, Ccl4, Ccl19, and Mmp9 (Fig. 3E). The interactions among proteins were predicted with the STRING database, revealing interactions of Hmox1 with Mmp9 and Ccl3 (data not shown). ELISA revealed increased expression levels of Mmp9 and Ccl3 in the culture supernatant of the oeHmox1 group (data not shown). To investigate whether excessive Hmox1 levels are deleterious to plaque stability, ApoE^−/− mice (fed a high-fat diet) were treated with high doses of AAV2/9-F480-m-Hmox1-3xflag-ZsGreen (AAV-Hmox1, Fig. 3F). The results of the animal experiments revealed that high expression of Hmox1 aggravated the plaque burden in brachiocephalic arteries compared with that in the control group (AAV-Con). Although there was no significant difference in collagen volume between the 2 groups, the plaque lesions in the AAV-Hmox1 group presented rupture-prone characteristics with increased lipid deposition.

3.4. Hif1a stabilization under Egln3 degradation promotes Hmox1 transcription

In further mechanistic experiments, we performed RNA-seq in MDFCs exposed to hypoxia and identified Egln3 as a candidate gene (Fig. 4A). Consistently, GSEA revealed that the HIF-1 signalling pathway and ferroptosis pathway were activated in hypoxia-treated MDFCs. The western blot results revealed decreased expression of Egln3 under hypoxia, with increased Hif1a and Hmox1 activity (Fig. 4B). Hif1a is a transcription factor that is mainly hydroxylated by Egln3 under normoxia. This process is inhibited under hypoxia due to Egln3 degradation, which stabilizes Hif1a and facilitates Hif1a accumulation [34]. Similarly, Egln3 silencing promoted intracellular Hif1a accumulation and upregulated Hmox1 expression (Fig. 4C) via the induction of ferroptosis (Fig. 4D). Our data suggest a pro-transcriptional role of Hif1a in hypoxia-treated MDFCs. The interaction between the Hif1a protein and the Hmox1 promoter was predicted via the JASPAR database (Table S4). To understand the mechanisms by which HIF1A affects HMOX1 expression, 293T cells were cotransfected with HIF1A pcDNA3.1 and HMOX1 promoter-reporter plasmids. Our results revealed that HIF1A enhanced the activity of the 2000-, 1500-, 1300-, 1000- and 100-bp promoters (Fig. 4E), indicating that the binding site for HIF1A may be located within 100 bp of the HMOX1 promoter region. Mutation of this binding site significantly inhibited the luciferase activity activated by HIF1A. Direct binding of Hif1a to the Hmox1 promoter was also validated via ChIP analysis (Fig. 4F).

Fig. 4.

Hypoxia-inducible factor (Hif)1a stabilization under Egln3 degradation promotes Hmox1 transcription

A. The top 20 DEGs between the normoxia and hypoxia groups and GSEA of the top 10 upregulated and downregulated signalling pathways. B. The protein levels of Egln3, Hif1a, and Hmox1 in MDFCs under normoxia and hypoxia. C. Western blot analysis of Egln3, Hif1a, and Hmox1 expression in MDFCs transduced with siEgln3. D. Representative fluorescence images showing the iron content and lipid peroxidation levels in MDFCs. Scale bar, 50 μm. E. 293T cells were transfected with 2000-, 1500-, 1300-, 1000- or 100-bp Hmox1 promoter-luciferase plasmids and further transfected with pcDNA3.1 or pcDNA3.1_HIF1A before the relative changes in luciferase activity were assessed. The binding site was mutated from ACGTGA to ATTTTA, and 293T cells were transfected with the mutant promoter and subsequently transfected with pcDNA3.1_HIF1A, followed by assessment of luciferase activity. F. ChIP assay to assess the binding of Hif1a to the Hmox1 promoter in MDFCs under normoxia and hypoxia. N.D., not detected. MDFCs, macrophage-derived foam cells; DEGs, differentially expressed genes; ns P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001.

Oxidative stress is a common cause of AS and plays a pivotal role in plaque progression. As a redox-regulated enzyme, HMOX1 is induced in AS and acts against oxidative stress but can also promote cell death, a phenomenon that may be due to HMOX1 expression levels [33]. Our study revealed that high levels of Hmox1 promote MDFC ferroptosis and exacerbate plaque instability. Hypoxia is a key element in AS and is involved in the regulation of iron-related proteins. A previous study showed that Hif1a regulated Hmox1 transcription by binding to the Hmox1 promoter region [35]. Consistently, our findings suggest that Hif1a stabilization due to Egln3 degradation under hypoxia activates the transcription of Hmox1 in MDFCs, which subsequently drives ferroptotic cell death. Our study has several limitations. First, the proferroptotic role of Hmox1 was not validated in Hmox1^−/− mice. Second, the concept of stable vs unstable plaques is often very difficult and in addition, the concept of superficial erosion was introduced [36] Furthermore, unstable plaques do not necessarily lead to symptomatic carotid artery stenosis.

4. Conclusion

The effects of Hmox1 are level dependent, and excessive pharmacological induction of Hmox1 drives ferroptotic oxidative stress in MDFCs. Our data revealed that hypoxic conditions induced Hmox1 overexpression in MDFCs, which aggravated ferroptotic cell death and affected plaque stability. Further analysis suggested that Egln3 degradation-mediated Hif1a stabilization increased Hmox1 activity. The present work demonstrated that high Hmox1 expression levels are deleterious to MDFC viability and plaque stability, providing a reference for the management of AS.

Funding

This study was supported by the Natural Science Foundation of Shanghai (Grant 22ZR1412100), the Shanghai Science and Technology Innovation Action Plan (No. 22Y11909500), and the Outstanding Resident Clinical Postdoctoral Program of Zhongshan Hospital Affiliated with Fudan University.

CRediT authorship contribution statement

Zhenyu Guo: Writing – original draft, Visualization, Formal analysis, Data curation, Conceptualization. Wan Zhang: Writing – original draft, Funding acquisition, Formal analysis, Conceptualization. Hongxia Gao: Writing – original draft, Validation, Conceptualization. Yang Li: Writing – original draft, Formal analysis, Data curation. Xu Li: Writing – review & editing, Resources, Methodology, Data curation. Xiaohu Yang: Writing – review & editing, Resources, Project administration, Investigation, Conceptualization. Longhua Fan: Writing – review & editing, Supervision, Resources, Project administration, Methodology, Funding acquisition, Conceptualization.

Declaration of competing interest

None.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Fig. S1

Identification and cluster analysis of DE-FRGs associated with atherosclerosis.

A. Volcano plot of DEGs and enrichment analysis (GO of DEGs in terms of BP, CC, and MF, and GSEA of ferroptosis signalling pathways) in GSE100927. B. Identification of DE-FRGs by overlapping FRGs and DEGs in GSE100927. C. Expression landscapes of 12 DE-FRGs in GSE100927. D. Distribution of cophenetic values, residuals, RSSs, and silhouettes with ranks of 2–10 and a consensus map of NMF clustering with a rank of 2. E. Principal component analysis of the C1 and C2 clusters. F. Expression profiles of 12 DE-FRGs in the C1 and C2 clusters. DEGs, differentially expressed genes; GSEA, gene set enrichment analysis; BP, biological process; CC, cellular component; MF, molecular function; FRGs, ferroptosis-related genes; DE-FRGs, differentially expressed FRGs; NMF, nonnegative matrix factorization. *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. S2

Immune infiltration analysis and Hmox1 expression in macrophages

A. Different hallmarks between the C1 and C2 clusters according to GSEA. B. Infiltration of 28 types of immune cells between the two clusters. C. The infiltration of 28 types of immune cells in unstable and stable plaques (GSE28829). D. Venn diagram showing the intersection of key DE-FRGs obtained by three algorithms (RF, Lasso, and SVM-RFE) and hypoxia-related genes with the expression levels of Hmox1 in IPH (GSE163154) and unstable plaques (GSE28829). E. Pearson correlation analysis of the correlations of Hmox1 with 28 types of immune cells in the GSE163154 dataset. F. Assessment of Hmox1 expression in macrophages via scRNA-seq datasets. **P < 0.01, ***P < 0.001. FRGs, ferroptosis-related genes; DE-FRGs, differentially expressed FRGs; RF, random forest; Lasso, least absolute shrinkage and selection operator; SVM-RFE, support vector machine recursive feature elimination; IPH, intraplaque haemorrhage.

Fig. S3

Pharmacological induction of excessive Hmox1 expression drives MDFC ferroptosis.

A. Schematic representation of how Hmox1 catalyses the degradation of haem b to carbon monoxide, ferrous iron, and bilirubin. B. Western blot analysis of MDFCs exposed to gradient concentrations of haemin (0, 25, 50, 75, and 100 μM). C. Cell viability was assessed with Cell Counting Kit-8 (CCK-8) assays. D. The relative GSH/GSSG ratios were assayed. E. Intracellular ROS levels were assessed in MDFCs treated with different concentrations of haem. F. Iron levels and lipid peroxidation levels in MDFCs in response to gradient concentrations of haem. Scale bar, 50 μm. MDFCs, macrophage-derived foam cells. ns P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. S4

Loss-of-function analyses of Hmox1 in the regulation of ferroptotic oxidative stress

A. siRNA transfection efficacy was determined via RT‒qPCR. B. Western blot analysis of Hmox1 and Gpx4 expression in MDFCs transduced with siHmox1 under hypoxic conditions. C. Cell viability was assessed with Cell Counting Kit-8 (CCK-8) assays. D. The relative GSH/GSSG ratios were assayed. E. Intracellular ROS levels were assessed in MDFCs transfected with siHmox1 under hypoxic conditions. F. Measurement of intracellular iron and lipid peroxidation levels. Scale bar, 50 μm. MDFCs, macrophage-derived foam cells. ns P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001.

Data availability

Data will be made available on request.