ATP6V0D2 Promotes Atherosclerosis‐Associated Vascular Smooth Muscle Cell Dysfunction by Modulating Ferroptosis‐Related Signaling

ABSTRACT

Background

Atherosclerosis (AS) underlies a substantial proportion of cardiovascular morbidity and mortality, yet the molecular events that drive vascular cell injury remain only partly resolved.

Methods

We first compared aortic transcriptomes from apolipoprotein E‐deficient (ApoE−/−) mice maintained on control or high‐fat diets. Candidate‐gene effects were then examined in human vascular smooth muscle cells (HVSMCs) challenged with oxidized low‐density lipoprotein (ox‐LDL).

Results

The aortic dataset contained 2832 differentially expressed genes (1800 increased and 1032 decreased). Of the five genes taken forward, ATP6V0D2 displayed a graded increase as the ox‐LDL concentration rose. Network and enrichment results connected ATP6V0D2 to proton movement, acidification of intracellular organelles, and iron transport. Silencing ATP6V0D2 limited the ox‐LDL‐evoked accumulation of neutral lipid and restrained cell‐cycle entry, motility, and invasion. At the molecular level, knockdown lowered acyl‐CoA synthetase long‐chain family member 4 (ACSL4) and recovered solute carrier family 7 member 11 (SLC7A11) and glutathione peroxidase 4 (GPX4). It also raised glutathione (GSH) abundance and superoxide dismutase (SOD) activity while decreasing malondialdehyde (MDA), reactive oxygen species (ROS), mitochondrial contraction, and cristae damage. Ferrostatin‐1 shifted the molecular, redox, and ultrastructural readouts in the same direction.

Conclusions

These data identify ATP6V0D2 as a contributor to the abnormal HVSMC phenotypes produced by ox‐LDL. Its depletion appears to blunt this response, in part through attenuation of changes compatible with ferroptosis and warrants further evaluation as a mechanistic target for AS.

This study identifies upregulated ATP6V0D2 as a key regulator of ox‐LDL‐induced pathological changes in human vascular smooth muscle cells. ATP6V0D2 knockdown alleviates atherosclerosis‐related dysfunction by suppressing ferroptosis, providing a potential target for atherosclerosis.

1. Introduction

Atherosclerosis (AS) is the primary pathological foundation of numerous cardiovascular disorders, making cardiovascular disease the foremost cause of mortality globally [1, 2]. Atherosclerotic plaques arise through a prolonged interaction among retained lipoproteins, inflammatory signaling, oxidative injury, and maladaptive behavior of vascular‐wall cells. Because these lesions develop in the aorta as well as coronary, cerebral, carotid, renal, mesenteric, and peripheral arteries, their clinical consequences extend across multiple cardiovascular territories [3]. Various interrelated mechanisms, including lipid accumulation, chronic inflammation, endothelial dysfunction, and oxidative stress, contribute to the initiation and progression of AS [4, 5, 6]. Despite considerable progress, the molecular mechanisms underlying AS remain incompletely understood. Therefore, identifying previously unrecognized molecular regulators may provide new mechanistic insights and potential therapeutic targets for AS. The emerging clinical potential of biological agents targeting inflammatory mediators, including interleukin‐1β and interleukin‐6 signaling, further supports the importance of inflammation as a mechanism‐based therapeutic target in AS, although concerns regarding patient selection, safety, and long‐term efficacy remain [7]. Identifying previously unrecognized molecular regulators linking lipid metabolism, oxidative injury, and vascular inflammation may therefore provide additional mechanistic insights and potential therapeutic opportunities.

Vacuolar H+‐ATPase (V‐ATPase) couples ATP hydrolysis to proton translocation and thereby establishes the acidic milieu required by several intracellular compartments [8]. Its catalytic V1 sector supplies the energy, whereas the membrane‐integrated V0 sector provides the route for proton passage [9]. ATP6V0D2 encodes the d2 isoform within V0 and supports both assembly and cell‐type‐specific functions of the pump [10]. Reported roles for this subunit span osteoclast‐mediated extracellular acidification, host defense, allergic inflammation, viral responses, and cancer biology [11, 12, 13, 14, 15]. Whether the same acidification machinery participates in smooth‐muscle‐cell dysfunction during AS has not been determined.

Ferroptotic injury develops when iron‐dependent membrane lipid oxidation outpaces cellular repair. One central protective route imports cystine through SLC7A11, sustains the GSH pool, and enables GPX4 to detoxify phospholipid hydroperoxides. By contrast, ACSL4 enriches membranes with polyunsaturated fatty‐acyl substrates and increases their vulnerability to oxidation [16, 17]. Evidence from vascular endothelium, macrophages, and smooth muscle increasingly places this redox imbalance within atherosclerotic biology [18, 19]. A mechanistic link between ATP6V0D2‐dependent organelle acidification and this ferroptosis‐related response in VSMCs, however, remains unexplored.

VSMCs also exhibit marked phenotypic plasticity within plaques: depending on local cues, their descendants may acquire matrix‐producing, inflammatory, or macrophage‐like features and can influence either plaque stability or progression [20]. Cell survival pathways are therefore especially relevant to lesion composition. Autophagy, senescence, and regulated VSMC death are intertwined with fibrous‐cap integrity and necrotic‐core development [21]. These observations provide a vascular‐cell context in which an organelle‐acidification factor such as ATP6V0D2 could intersect with oxidative lipid injury.

We therefore used aortic RNA sequencing in atherosclerotic ApoE−/− mice to identify disease‐associated genes and selected Atp6v0d2 for functional study. In ox‐LDL‐exposed HVSMCs, small interfering RNA (siRNA) was used to reduce ATP6V0D2, after which lipid storage, cell‐cycle distribution, migration, invasion, oxidative status, ferroptosis‐related proteins, and mitochondrial ultrastructure were evaluated.

2. Material and Methods

2.1. Experimental Animals and AS Model

Specific‐pathogen‐free ApoE−/− mice were enrolled at 8 weeks of age (18–19 g; Cavens Lab Animal, China). Animals were allocated to control and AS cohorts; six mice per cohort contributed material for RNA sequencing. The control cohort remained on standard chow, whereas the AS cohort consumed a high‐fat diet for 16 weeks. Housing was maintained at 22°C ± 1°C and 50% ± 5% relative humidity under specific‐pathogen‐free conditions, with unrestricted access to water and food. Aortas were harvested after euthanasia at the end of the dietary intervention. The protocol complied with the ARRIVE guidelines and was a by the Ethics Committee of the Second Affiliated Hospital of Anhui Medical University (YX2024‐207).

2.2. Transcriptome Analysis

Aortic total RNA (n = 6) was prepared for transcriptomic profiling. Yield and integrity were checked with the RNA Nano 6000 Assay Kit on an Agilent 2100 Bioanalyzer (Agilent Technologies, CA, USA), and only qualified material entered library preparation. Poly‐T oligo‐coupled magnetic beads were used to isolate mRNA. The recovered RNA was fragmented and converted successively into first‐ and second‐strand cDNA, followed by end repair, 3′ adenylation, adaptor attachment, and PCR enrichment. Library concentration and quality were assessed by Qubit 2.0 fluorometry, Agilent 2100 analysis, and qRT‐PCR. Paired‐end 150 bp reads were generated on an Illumina NovaSeq 6000. Each sample yielded approximately 41–48 million raw reads and, after adaptor and low‐quality sequence removal, approximately 41–47 million clean reads. Quality metrics ranged from 96.63% to 97.76% for Q20, 91.01% to 93.65% for Q30, and 44.48% to 50.10% for GC content. HISAT2 (Version 2.0.5) mapped clean reads to the reference genome, and featureCounts (Version 1.5.0‐p3) produced gene‐level abundance estimates. DESeq2 (Version 1.20.0) defined differential expression at adjusted p value < 0.05 and |log2(fold change)| ≥ 1. Functional enrichment for GO terms and KEGG pathways was evaluated with clusterProfiler.

2.3. Differential Gene Analysis

DESeq2 (Version 1.20.0) was used for group‐wise expression testing, with Benjamini–Hochberg correction of the resulting p values. A gene was classified as differentially expressed when adjusted p < 0.05 and |log2 fold change| > 1. GO and KEGG over‐representation analyses were run in clusterProfiler (Version 3.8.1), applying adjusted p < 0.05 as the significance boundary. STRING supplied the protein–protein interaction (PPI) network. Candidate ranking jointly considered effect size, statistical support, representation in enriched terms, network connectivity, biological relevance, and practical feasibility of validation. Five genes were tested in ox‐LDL‐challenged HVSMCs; ATP6V0D2 was advanced because its induction was reproducible and tracked the ox‐LDL dose.

2.4. Cell Culture and AS Model Establishment In Vitro

HVSMCs from the American Type Culture Collection (ATCC, USA) were maintained at 37°C and 5% CO₂ in Dulbecco's modified Eagle medium (Gibco, USA) containing 20% fetal bovine serum (Gibco) and 1% penicillin–streptomycin. To reproduce an AS‐related lipid stress in vitro, cells received 100 μg/mL ox‐LDL (Solarbio, China) for 24 h [22]. For ferrostatin‐1 (Fer‐1, HY‐100579, MedChemExpress, USA) treatment, hVSMCs were pretreated with Fer‐1 (1μM) for 4 h and then co‐incubated with ox‐LDL for additional 24 h.

2.5. Cell Transfection

Three siRNAs directed against ATP6V0D2 and one nonspecific control were delivered to HVSMCs with NanoTrans reagent (Biomedical, Anhui, China). For each transfection, 100 pmol siRNA and 3‐μL reagent were diluted separately in serum‐free medium, mixed, and left for 15–20 min before application. Culture medium was renewed 8 h later, and downstream measurements were obtained 48 h after transfection. Reduction of ATP6V0D2 was confirmed by qRT‐PCR and immunoblotting. The sequences for three siRNAs of ATP6V0D2 were as follows:

• siATP6V0D2‐1 sense, 5′‐GCAGUUAUAUGAUAGACAATT‐3′, antisense, 5′‐UUGUCUAUCAUAUAACUGCTT‐3′. siATP6V0D2‐2 sense, 5′‐GGUUCACAAGUGUUG

• GAAATT‐3′, antisense, 5′‐UUUCCAACACUUGUGAACCTT‐3′. siATP6V0D2‐3 sense, 5′‐CGAGAGACCCUCUAUCCAATT‐3′, antisense, 5′‐UUGGAUAGAGGGU

• CUCUCGTT‐3′. siCtrl sense, 5′‐UUUGUACUACACAAAAGUACUG‐3′, antisense, 5′‐CAGUACUUUUGUGUAGUACAAA‐3′.

2.6. Oil Red O Staining

For neutral‐lipid staining, cells were rinsed in phosphate‐buffered saline (PBS) and exposed for 10 min to 0.15% Oil Red O prepared in 60% isopropanol (Beyotime, China). Hematoxylin counterstaining and a further wash preceded image acquisition with a Zeiss microscope (Germany).

2.7. Cell Cycle

Cell‐cycle distribution was assessed using a propidium iodide/RNase staining kit (Biomedical, Anhui, China) according to the manufacturer's instructions. At 48 h after transfection, cells were harvested, washed, stained for 1 h at room temperature, and analyzed using a NovoCyte flow cytometer (Agilent, USA). The proportions of cells in G0/G1, S, and G2/M phases were calculated.

2.8. Wound Scratch Healing Assay

Two horizontal lines were drawn on the underside of each well of a 12‐well plate. Infected HVSMCs were seeded at a density of 1.5 × 10⁵ cells per well and culture until they reached > 90% confluence. The culture medium was then replaced with a low concentration of serum. A linear scratch perpendicular to the reference lines was then created using a sterile 10‐μL pipette tip. The cells were gently washed two to three times with PBS to remove detached cells and subsequently cultured in fresh low‐serum medium at 37°C in a humidified incubator containing 5% CO₂. Images of the same marked fields were captured at the 0 and 24 h using a microscope (Mshot, Guangzhou, China), and cell migration was assessed by measuring scratch closure.

2.9. Transwell Migration and Invasion

Invasion inserts were prepared by adding 50‐μL Matrigel (Corning, USA), diluted 1:8 with serum‐free medium, to the upper compartment and allowing it to gel overnight at 37°C. The coating was then rehydrated for 2 h at 37°C with 500‐μL serum‐free medium placed in both compartments. Next, 100 μL of a serum‐free cell suspension was loaded above the membrane and 600‐μL medium containing 20% fetal bovine serum was placed below it. Following 48 h at 37°C, cells that had not traversed the membrane were cleared from its upper face with a cotton swab. Inserts were washed three times in PBS before fixation and staining of the invaded cells. Counts were obtained from three random fields per insert at ×100 magnification. Migration was measured with the identical workflow but without Matrigel coating.

2.10. Biochemical Assessment of Glutathione (GSH), Malondialdehyde (MDA), and Superoxide Dismutase (SOD)

GSH content, MDA abundance, and SOD activity were assayed with kits bc1175 (Solarbio, Beijing, China), S0131 (Beyotime, Shanghai, China), and A001‐3 (Nanjing Jiancheng Bioengineering Institute, Nanjing, China), respectively. Cells from each condition were recovered, rinsed with ice‐cold PBS, disrupted in the extraction solution specified for each kit, and centrifuged at 4°C. Assays were performed on the supernatants according to the corresponding kit protocols: a colorimetric reaction for GSH, a thiobarbituric acid‐based lipid‐peroxidation readout for MDA, and an inhibition colorimetric method for SOD. Absorbance was acquired at the wavelengths prescribed by each manufacturer. A bicinchoninic acid assay provided total protein values used to normalize GSH, MDA, and SOD measurements.

2.11. Flow‐Cytometric Measurement of Intracellular Reactive Oxygen Species (ROS)

The fluorescent precursor 2′,7′‐dichlorodihydrofluorescein diacetate (DCFH‐DA) was used to quantify cellular ROS. A 1:1000 dilution in serum‐free medium produced a working concentration of 10 μM. After treatment, cells were collected at 1 × 10⁶ cells/mL in this solution and kept at 37°C in darkness for 20 min, with gentle inversion at 5‐min intervals. Three washes with serum‐free medium removed probe remaining outside the cells. When used, the positive‐control population was exposed to the ROS‐inducing reagent for 20–30 min. Oxidized DCF fluorescence was acquired immediately by flow cytometry and summarized as the fraction of ROS‐high cells.

2.12. Transmission Electron Microscopy

Cells were fixed in 2.5% glutaraldehyde, post‐fixed with osmium tetroxide, dehydrated through a graded ethanol series, embedded in resin, sectioned, and contrasted with uranyl acetate and lead citrate. Mitochondrial ultrastructure was examined by transmission electron microscopy. Condensed mitochondria, increased membrane density, and reduced or disrupted cristae were considered morphology consistent with ferroptotic injury.

2.13. Quantitative Real‐Time Polymerase Chain Reaction (QRT‐PCR)

Total RNA was recovered from HVSMCs with AG RNAex Pro RNA reagent (Accurate Biotechnology, Hunan, China), quantified, and checked for purity. The Evo M‐MLV reverse‐transcription kit (Accurate Biotechnology) generated cDNA. Amplification used SYBR Green Pro Taq HS master mix (Accurate Biotechnology) and a StepOnePlus Real‐Time PCR System (Applied Biosystems, USA). Transcript abundance was referenced to GAPDH and calculated by the 2^−ΔΔCt approach. Primer pairs were ATP6V0d2, forword, 5′‐TCTCACCTATATGACGTGCAGT‐3′, reverse, 5′‐GGTGGCACTTCCCCAGAATTT‐3′. GDF3, forword, 5′‐ACCCCAGAAGTTCCAACCTG‐3′, reverse, 5′‐AGAAAGCGAAGTACATTCCCG‐3′. ACOD1, forword, 5′‐TGCTGCTGGCTTTCAATGTT‐3′, reverse, 5′‐CATGTCATTGGCCTCCTTGG‐3′. GPNMB, forword, 5′‐AAGATTGCCACTTGATGCCG‐3′, reverse, 5′‐TCCCTCATGTAAGCAGAAGGTC‐3′. HAVCR2, forword, 5′‐CTGCTGCTACTACTTACAAGGTC‐3′, reverse, 5′‐GCAGGGCAGATAGGCATTCT‐3′.GAPDH, forword, 5′‐AGAAGGCTGGGGCTCATTTG‐3′, reverse, 5′‐AGGGGCCATCCACAGTCTTC‐3′. Reaction conditions: 95°C, 30 s; 40 cycles (95°C, 5 s, 60°C, 30 s) 60°C to 95°C.

2.14. Western Blot

Total proteins were isolated and solubilized in RIPA buffer containing phenylmethylsulfonyl fluoride (PMSF; Solarbio, Beijing, China), and their concentration was determined with a BCA kit (Beyotime, Shanghai, China). Equal amounts of protein (40 μg per lane) were separated by 8%–12% SDS–polyacrylamide gel electrophoresis (Epizyme Biotech, Shanghai, China) and transferred onto polyvinylidene difluoride membranes (PVDF, Millipore, USA). After blocking with 5% nonfat milk, the membranes were incubated with primary antibodies against ATP6V0D2 (1:1000; ab236375, Abcam, UK), ACSL4 (1:1000; ab155282, Abcam), GPX4 (1:1000; 67763‐1‐Ig, Proteintech, Wuhan, China), and GAPDH (1:10,000; Proteintech) overnight at 4°C. The membranes were subsequently incubated with the appropriate horseradish peroxidase‐conjugated goat anti‐mouse or goat anti‐rabbit IgG secondary antibodies. Protein bands were visualized using an enhanced chemiluminescence detection kit (Proteintech), and band intensities were quantified using Fusion software. Target‐protein expression was normalized to that of GAPDH.

2.15. Statistical Analysis

Data are reported as mean ± standard deviation (SD) from three biologically independent experiments. GraphPad Prism 9.0 (GraphPad Software, USA) was used for analysis. Distributional normality and equality of variance were examined with Shapiro–Wilk and Brown–Forsythe tests. When their assumptions were met, comparisons involving at least three groups used one‐way analysis of variance with Tukey correction; otherwise, Kruskal–Wallis analysis with Dunn post hoc testing was selected. Every test was two‐sided, and p < 0.05 denoted statistical significance. Exact p values are given when available, with threshold symbols explained in the figure legends.

3. Results

3.1. Aortic Transcriptomes Distinguish the AS and Control Groups

Relative to controls, aortas from the AS cohort contained 2832 DEGs: 1800 transcripts were higher and 1032 were lower. In Figure 1A, the volcano plot uses red for significantly increased genes, green for significantly decreased genes, and blue for genes failing to meet differential‐expression criteria. Those criteria were adjusted p < 0.05 together with |log2 fold change| > 1. Within the clustering map, replicates from the same condition grouped together, whereas AS and control profiles separated from one another (Figure 1B). GO results were organized into biological process, cellular component, and molecular function, with the 10 leading terms from each domain displayed in Figure 1C. Figure 1D presents the 20 most strongly enriched KEGG pathways, summarizing signaling and metabolic themes represented among the DEGs.

FIGURE 1.

DEGs in aortic tissues was analyzed by transcriptome analysis (n = 6). (A) Volcano plot of differentially expressed genes (DEGs). Red dots represent significantly upregulated genes, green dots represent significantly downregulated genes, and blue dots represent genes without significant changes (|log₂ fold change| > 1 and adjusted p < 0.05). (B) Heatmap showing the expression patterns of DEGs. Red indicates relatively high expression, whereas green indicates relatively low expression. (C) Gene Ontology (GO) functional annotation of the DEGs. (D) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of the DEGs.

3.2. Five Candidate DEGs Were Detected by QRT‐PCR

Based on the transcriptomic analysis of aortic tissues, five candidate genes—ATP6V0D2, GDF3, ACOD1, GPNMB, and HAVCR2—were selected for subsequent validation in an ox‐LDL‐induced HVSMC model. qRT‐PCR analysis showed that ATP6V0D2 expression was significantly increased in ox‐LDL‐treated HVSMCs compared with control cells, whereas the expression levels of GDF3, ACOD1, GPNMB, and HAVCR2 were significantly decreased (p < 0.05; Figure 2A–E). Based on these findings, ATP6V0D2 was selected as the candidate hub gene for further investigation. To determine whether ATP6V0D2 expression was associated with the ox‐LDL concentration, HVSMCs were treated with 0, 10, 20, 40, 80, or 100 μg/mL ox‐LDL for 24 h under otherwise identical culture conditions. Both the mRNA and protein levels of ATP6V0D2 increased progressively following ox‐LDL treatment, with the highest expression observed at 100 μg/mL (Figure 2F,G).

FIGURE 2.

DEGs mRNA expression levels in HVSMCs were detected by QRT‐PCR. The expression level of (A) Atp6v0d2, (B) GDF3, (C) ACOD1, (D) GPNMB, and (E) HAVCR2 mRNA in different groups; compared to Control group, *p < 0.05, **p < 0.01. (F) The expression level of ATP6V0D2 mRNA after different concentration of ox‐LDL treatment. (G) The expression level of ATP6V0D2 protein after different concentration of ox‐LDL treatment. Data are presented as the mean ± SD from three independent experiments. The results were analyzed using an unpaired two‐tailed Student's t test. Compared to 0‐μg/mL ox‐LDL group, *p < 0.05, **p < 0.01.

3.3. Interaction‐Network and Enrichment Profile of ATP6V0D2

The PPI network contained 51 connected nodes, among them ATP6V0D2 and several components related to V‐ATPase (Figure 3A). For every GO domain, Figure 3B reports the four strongest enrichment signals. The biological‐process set included lowering of pH, transport of ferric iron and transferrin, and acidification of phagosomes. Vacuolar membranes and proton‐translocating V‐type ATPase complexes dominated the cellular‐component output, while transporter and proton‐coupled ATPase activities characterized molecular function. Five KEGG pathways met the enrichment criterion: rheumatoid arthritis, epithelial‐cell signaling during Helicobacter pylori infection, Vibrio cholerae infection, synaptic vesicle cycling, and collecting‐duct acid secretion (Figure 3C). Thus, the computational results converge on proton handling, acidic organelles, and iron‐trafficking biology around ATP6V0D2.

FIGURE 3.

PPI network and enrichment analysis. (A) PPI network was constructed for ATP6V0D2, then (B) GO and (C) KEGG pathway annotation were conducted for the downstream pathway of Atp6v0d2.

3.4. ATP6V0D2 Knockdown Reduced Lipid Accumulation and Proliferation‐Associated Cell‐Cycle Progression

Three small interfering RNAs targeting ATP6V0D2 (siATP6V0D2‐1, siATP6V0D2‐2, and siATP6V0D2‐3) were transfected into HVSMCs. Based on the qRT‐PCR and Western blot results, siATP6V0D2‐3 exhibited the greatest knockdown efficiency and was therefore selected for subsequent experiments (p < 0.05; Figure 4A–C). The four experimental conditions were control, ox‐LDL, siCtrl + ox‐LDL, and siATP6V0D2 + ox‐LDL. Ox‐LDL caused a prominent rise in Oil Red O‐positive lipid, which was diminished when ATP6V0D2 was silenced (Figure 4D). In cell‐cycle profiles, ox‐LDL and siCtrl + ox‐LDL shifted cells away from G0/G1 relative to control. Knockdown countered that shift, increasing G0/G1 versus both ox‐LDL‐containing comparison groups (p < 0.05). Ox‐LDL also expanded the S‐phase fraction, an effect reduced by ATP6V0D2 siRNA (p < 0.05; Figure 4E,F). G2/M proportions did not differ significantly. These results place ATP6V0D2 upstream of the lipid‐storage and proliferative responses elicited by ox‐LDL in HVSMCs.

FIGURE 4.

ATP6V0D2 knockdown suppressed lipid accumulation and HVSMCs cell‐cycle progression. (A) The mRNA expression level of ATP6V0D2 after transfection was measured by quantitative real‐time PCR (qRT‐PCR). (B) ATP6V0D2 protein expression after transfection was determined by Western blotting. (C) Densitometric quantification of the Western blot results. (D) Oil Red O staining was performed to evaluate lipid accumulation in HVSMCs from the indicated groups, scale bar = 50 μm. (E) Cell‐cycle distribution in the different groups was analyzed by flow cytometry. (F) Quantification of the proportions of HVSMCs in the G0/G1, S, and G2/M phases. *p < 0.05 and **p < 0.01 versus the si‐NC group. Data are presented as the mean ± SD (n = 3). One‐way ANOVA with Tukey's test. Compared to the Control group,*p < 0.05, **p < 0.01; compared to the ox‐LDL group, ^# p < 0.05, ^## p < 0.01; compared to the siCtrl + ox‐LDL group, ^{^} p < 0.05, ^{^^} p < 0.01.

3.5. Atp6v0d2 Knockdown Restrains HVSMC Motility and Invasion

Both scratch closure and Transwell measurements indicated greater migration and invasion after ox‐LDL exposure, irrespective of control‐siRNA transfection, than in untreated controls (p < 0.05; Figure 5A–F). Introducing siATP6V0D2 significantly reduced each response relative to the ox‐LDL and siCtrl + ox‐LDL conditions (p < 0.05). ATP6V0D2 induction therefore accompanies—and its depletion curtails—the motile and invasive phenotype triggered by ox‐LDL.

FIGURE 5.

ATP6V0D2 knockdown suppressed HVSMCs migration and invasion. (A) Wound healing assay was used to investigate the HVSMCs migration in different groups, scale bar = 50 μm. (B) Bar graph shows the quantification of migration cells. Transwell assay was used to detect the (C) cells migration and (E) invasion ability, scale bar = 20 μm. Bar graphs presents the quantification of (D) migration and (F) invasion cells. Data are presented as the mean ± SD (n = 3). One‐way ANOVA with Tukey's test. Compared to the Control group,*p < 0.05, **p < 0.01; compared to the ox‐LDL group, ^# p < 0.05, ^## p < 0.01; Compared to the siCtrl + ox‐LDL group, ^{^} p < 0.05, ^{^^} p < 0.01.

3.6. ATP6V0D2 Knockdown Altered Ferroptosis‐Related Protein Expression

Based on the GO, KEGG, and PPI analyses, together with evidence from previous studies, we investigated the ferroptosis‐related regulators ACSL4, SLC7A11, and GPX4. In the ox‐LDL and siCtrl + ox‐LDL groups, ATP6V0D2 and ACSL4 were elevated, whereas SLC7A11 and GPX4 were lower than in controls (p < 0.05). Following ATP6V0D2 knockdown, ATP6V0D2 and ACSL4 were reduced compared to the control group. SLC7A11 and GPX4 moved in the opposite direction and exceeded their levels in both ox‐LDL comparison groups (p < 0.05; Figure 6A–I). This coordinated pattern is consistent with reduced ACSL4‐associated susceptibility and recovery of the SLC7A11–GPX4 protective axis after ATP6V0D2 silencing.

FIGURE 6.

ATP6V0D2 knockdown suppressed the expression of ACSL4 while promoted GPX4 and SLC7A11 expression. QRT‐PCR was utilized to determine the expression levels of (A) ATP6V0D2, (B) ACSL4, (C) GPX4, and (D) SLC7A11. (E) Western blot was utilized to detect the expression levels of ATP6V0D2, ACSL4, GPX4, and SLC7A11. Bar graph shows the gray value analysis of (F) ATP6V0D2, (G) SLC7A11, (H) GPX4, and (I) ACSL4. Data are presented as the mean ± SD (n = 3). One‐way ANOVA with Tukey's test. Compared to the Control group,*p < 0.05, **p < 0.01; compared to the ox‐LDL group, ^# p < 0.05; compared to the siCtrl + ox‐LDL group, ^{^} p < 0.05.

3.7. Oxidative and Ultrastructural Changes Are Attenuated After ATP6V0D2 Knockdown

Relative to ox‐LDL and siControl + ox‐LDL, ATP6V0D2 silencing raised cellular GSH and SOD activity while lowering MDA and the proportion of ROS‐positive cells (p < 0.05; Figure 7A–E). Electron microscopy further showed less mitochondrial shrinkage and cristae loss than after ox‐LDL alone (Figure 7F). Fer‐1 shifted these oxidative indices and mitochondrial features in the same direction. Immunoblotting revealed lower ATP6V0D2 and ACSL4 with higher GPX4 and SLC7A11 after knockdown; Fer‐1 likewise favored an anti‐ferroptotic marker pattern (Figure 7G–K).

FIGURE 7.

ATP6V0D2 knockdown attenuated oxidative‐stress and ferroptosis associated changes in ox‐LDL‐treated HVSMCs. Cells were assigned to ox‐LDL, siControl + ox‐LDL, siATP6V0D2 + ox‐LDL, and ox‐LDL + ferrostatin‐1 (Fer‐1) groups. (A–C) Cellular GSH, MDA, and SOD activity. (D,E) Representative DCFH‐DA flow‐cytometry histograms and percentages of ROS‐positive cells. (F) Representative transmission electron micrographs; arrows indicate mitochondria with altered (red) or relatively preserved (blue) ultrastructure (n = 5), scale bars = 500 nm. (G–K) Representative immunoblots and densitometric quantification of ATP6V0D2, ACSL4, GPX4, and SLC7A11 normalized to GAPDH. Blots are representative of three independent experiments. Data are presented as the mean ± SD (n = 3). One‐way ANOVA with Tukey's test. Compared to the ox‐LDL group, *p < 0.05 and **p < 0.01; compared to the siCtrl + ox‐LDL group, ^# p < 0.05 and ^## p < 0.01; compared to the siATP6V0D2 + ox‐LDL group, ^{^} p < 0.05 and ^{^^} p < 0.01 versus siATP6V0D2 + ox‐LDL.

4. Discussion

RNA sequencing identified Atp6v0d2 as an AS‐associated DEG, and subsequent validation showed a concentration‐dependent increase in ATP6V0D2 in ox‐LDL‐treated HVSMCs. ATP6V0D2 was prioritized using integrated differential‐expression, enrichment, PPI‐connectivity, biological relevance, and validation criteria rather than expression change alone. Although several other candidates were validated, ATP6V0D2 was selected because of its reproducible upregulation and its established links to lysosomal acidification, autophagy, immune regulation, and cholesterol handling [14, 15, 23].

A previous study reported that ATP6V0D2 was associated with an increased susceptibility to acute coronary syndrome and chronic coronary heart disease [24]. In the present study, transcriptomic analysis of aortic tissues identified 2832 differentially expressed genes, comprising 1800 upregulated and 1032 downregulated genes. Based on their expression patterns and the results of GO and KEGG enrichment analyses, Atp6v0d2, Gdf3, Acod1, Gpnmb, and Havcr2 were selected as candidate genes. These genes participate in several biologically relevant processes. GDF3 is a member of the transforming growth factor‐β superfamily [25], whereas ACOD1, also known as IRG1, produces itaconate and regulates oxidative stress and immune responses [26]. GPNMB is a transmembrane glycoprotein implicated in neurodegenerative and inflammatory disorders [27, 28, 29]. HAVCR2, also known as TIM‐3, is an immune‐checkpoint molecule involved in immune tolerance and T‐cell exhaustion [30]. Subsequent qRT‐PCR validation in ox‐LDL‐treated HVSMCs demonstrated that GDF3, ACOD1, GPNMB, and HAVCR2 were downregulated, whereas ATP6V0D2 was significantly upregulated. Considering its consistent expression pattern, potential cardiovascular relevance, and central position in the bioinformatics analyses, ATP6V0D2 was selected as the candidate hub gene for further functional investigation.

To investigate the role of ATP6V0D2 in AS‐associated cellular dysfunction, we first examined its expression in HVSMCs exposed to increasing concentrations of ox‐LDL. ATP6V0D2 expression increased in a concentration‐dependent manner following ox‐LDL treatment. Subsequent siRNA‐mediated ATP6V0D2 knockdown reduced intracellular lipid accumulation, increased the proportion of cells in the G0/G1 phase, and decreased the proportion in the S phase, suggesting suppression of ox‐LDL‐induced cell‐cycle progression. ATP6V0D2 knockdown also attenuated the enhanced migration and invasion of ox‐LDL‐treated HVSMCs. Collectively, these findings indicate that ATP6V0D2 knockdown alleviates several AS‐related phenotypic changes in the HVSMC. Previous studies have shown that ATP6V0D2 regulates macrophage polarization, inflammasome activation, and tumor progression in a context‐dependent manner [11, 12, 15]. To explore its potential mechanism in AS, we performed PPI, GO, and KEGG analyses, which highlighted associations with ferric iron transport, transferrin transport, and organelle acidification. Based on these findings and previous evidence linking iron metabolism to ferroptosis, the ferroptosis‐related regulators ACSL4, GPX4, and SLC7A11 were selected for further investigation.

GPX4 is a selenium‐dependent phospholipid hydroperoxidase that uses GSH to reduce membrane lipid hydroperoxides and thereby constitutes a major defense against ferroptosis [31, 32]. Ferroptosis is an iron‐dependent form of regulated cell death characterized by excessive phospholipid peroxidation and has increasingly been implicated in AS progression [33, 34]. SLC7A11, a functional component of the cystine/glutamate antiporter system Xc^−, imports cystine to support intracellular GSH synthesis and maintain GPX4 activity [35]. In contrast, ACSL4 promotes ferroptosis susceptibility by facilitating the incorporation of polyunsaturated fatty acids into membrane phospholipids, thereby providing substrates for lipid peroxidation [16].

In the present study, ox‐LDL treatment increased ACSL4 expression while decreasing SLC7A11 and GPX4 expression in HVSMCs. ATP6V0D2 knockdown reversed these molecular changes, increased cellular GSH levels and SOD activity, and reduced MDA and ROS levels. Moreover, ATP6V0D2 silencing attenuated ox‐LDL‐associated mitochondrial shrinkage and cristae disruption, which are morphological features commonly associated with ferroptosis. Treatment with the ferroptosis inhibitor ferrostatin‐1 produced directionally similar effects on oxidative‐stress indices, mitochondrial morphology, and ferroptosis‐related protein expression, further supporting the involvement of ferroptosis in the effects of ATP6V0D2. These observations are consistent with the findings of Bai et al., who reported that ferrostatin‐1 reduced iron accumulation and lipid peroxidation and restored SLC7A11 and GPX4 expression in high‐fat diet‐fed ApoE−/− mice [36]. Collectively, our results suggest that ATP6V0D2 knockdown protects HVSMCs against ox‐LDL‐induced injury, at least partly by limiting ACSL4‐associated lipid peroxidation and restoring the SLC7A11/GSH/GPX4 antioxidant defense system. This study has several limitations. First, mechanistic experiments were conducted in a single ox‐LDL‐treated HVSMC model and therefore do not capture endothelial, macrophage, and immune‐cell interactions. Second, although Fer‐1 and canonical oxidative‐stress and ultrastructural readouts strengthened the ferroptosis‐related interpretation, genetic epistasis or rescue experiments involving SLC7A11 or GPX4 were not performed. Third, ApoE−/− mice were used for transcriptomic discovery only; in vivo ATP6V0D2 gain‐ or loss‐of‐function studies are required. Finally, independent clinical cohorts are needed to determine translational relevance.

In conclusion, ATP6V0D2 is high expression in the AS cell model. ATP6V0D2 knockdown may suppress lipid accumulation, as well as cell proliferation, migration and invasion. The underlying mechanism is related to the SLC7A11/GPX4 pathway. The current study reveals that ATP6V0D2 has crucial roles in the initiation and progression of AS supporting further investigation of this molecule as a candidate regulator.

Author Contributions

Conceived and designed the study: Chaohui Zhang, Zheng Huang, and Feng Gao; performed the literature search and data extraction: Chaohui Zhang, Zheng Huang, Feng Gao, Mengli Li, Jie Wu, and Zhangle Hu; analyzed the data: Chaohui Zhang, Mengli Li, Jie Wu, and Xiaochen Wang; drafted the manuscript: Chaohui Zhang, Zheng Huang, and Feng Gao. All authors confirmed the final version.

Funding

This study was supported by Natural Science Research Project for Universities of Anhui Province (2022AH050770 and 2022AH050771).

Ethics Statement

The Ethics Committee of the Second Affiliated Hospital of Anhui Medical University approved this study (Approval number: YX2024‐207).

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The dataset used and/or analyzed in this study is available from the corresponding author on reasonable request. The raw RNA‐sequencing data have been deposited in the NCBI Sequence Read Archive (SRA) database under accession number PRJNA1514040.