Neuronatin Influences Endothelial Homeostasis and Atherosclerosis via ER–Mitochondria Interactions

Abstract

Background

Atherosclerosis (AS) is driven by endothelial dysfunction, lipid dysregulation, and inflammation, yet its molecular mechanisms remain incompletely understood. Neuronatin (NNAT), an endoplasmic reticulum (ER)–associated protein, has an unclear role in AS. ER–mitochondria contact sites have been implicated in endothelial cell function. This study aimed to investigate the role of NNAT and its association with ER–mitochondria interactions in AS.

Methods

NNAT expression was analyzed in human AS datasets and in ApoE^−/− mice fed a Western-type diet (WD). NNAT-deficient and double-knockout mice were evaluated for plaque burden, lipid accumulation, and vascular inflammation. Human umbilical vein endothelial cells (HUVECs) treated with oxidized LDL (oxLDL) were used to assess NNAT’s effects on endothelial function, lipid metabolism, and ER–mitochondria-related features.

Results

NNAT expression was markedly reduced in human AS lesions and in WD-fed ApoE^−/− mice. Loss of NNAT accelerated plaque formation, lipid deposition, and vascular inflammation. In HUVECs, oxLDL suppressed NNAT, inducing endothelial dysfunction, cholesterol accumulation, reactive oxygen species generation, and altered ER–mitochondria interactions. NNAT overexpression restored nitric oxide production, attenuated adhesion molecule expression and NF-κB activation, and improved lipid homeostasis. These effects were associated with improved ER–mitochondria contact and mitochondrial function. Disruption of ER–mitochondria interactions markedly attenuated the protective effects of NNAT, suggesting that ER–mitochondria interactions contributes to NNAT-associated endothelial effects.

Conclusion

NNAT is associated with endothelial homeostasis and attenuation of AS progression, potentially involving ER-mitochondria interactions. NNAT deficiency aggravates atherosclerotic phenotypes, highlighting its potential relevance in cardiovascular disease.

Introduction

Atherosclerosis (AS) is the leading pathological basis of cardiovascular and cerebrovascular diseases, characterized by lipid accumulation, chronic inflammation, and endothelial dysfunction, which ultimately leading to plaque formation and vascular complications such as coronary heart disease and stroke.¹ Nevertheless, the underlying molecular mechanisms remain incompletely defined.²

Neuronatin (NNAT), an endoplasmic reticulum (ER)-associated protein originally identified as a regulator of neural development and pancreatic β-cell function,³ has been increasingly studied in the context of cardiovascular disease. In high-fat diet and diabetic mouse models, NNAT expression is markedly elevated, suggesting a role in AS-related inflammation.⁴ Mechanistic studies indicate that the PPARγ/NNAT/NF-κB axis promotes inflammatory cytokine secretion in adipose tissue, thereby accelerating AS progression.⁵ Clinical evidence further supports this link, as NNAT expression in epicardial adipose tissue of patients with coronary artery disease may influence atherosclerosis through metabolic regulation.⁶ These findings suggest a potential role for NNAT in AS; however, its specific function in vascular pathology remains unclear.

ER–mitochondria contact sites (also referred to as mitochondria-associated membranes, MAM) represent important subcellular structures involved in cellular homeostasis. They participate in calcium (Ca²⁺) transfer, mitochondrial dynamics, mitophagy, and ER stress responses.⁷ Through these processes, ER–mitochondria interactions contribute to metabolic and cellular homeostasis. Disruption of ER–mitochondria interactions has been implicated in metabolic disorders and proposed as a potential therapeutic target.^8,9 Importantly, accumulating evidence suggests that impaired ER–mitochondria interactions are associated with AS-related events, including Ca²⁺ dysregulation, lipid imbalance, apoptosis, and mitochondrial dysfunction.^10–12

Emerging evidence suggests that NNAT regulates Ca²⁺ signaling through interaction with the ER calcium pump SERCA (sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPase). By inhibiting SERCA activity, NNAT reduces ER Ca²⁺ uptake, elevates cytosolic Ca²⁺, and modulates downstream signaling pathways.¹³ In cancer and adipocyte models, NNAT has been shown to respond to oxidative stress, regulate Ca²⁺ homeostasis, and influence cellular proliferation or differentiation.^14,15 These observations suggest that NNAT integrates Ca²⁺ signaling with metabolic and stress responses, raising the possibility that it may influence AS development through mechanisms related to ER–mitochondria interactions. However, direct experimental evidence remains limited.

In this study, we investigated the role of NNAT in atherosclerosis progression and its association with endothelial ER–mitochondria interactions. Using complementary in vivo and in vitro approaches, we investigated the role of NNAT in AS pathogenesis and its relationship with ER–mitochondria -associated structural and functional alterations in endothelial cells.

Materials and Methods

Public Data Ethical Statement

Human data analyzed in this study were obtained from publicly available, de‑identified gene expression datasets deposited in the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). These datasets (GSE100927, GSE153555, GSE51472, and GSE40156) consist of fully de‑identified human tissues and clinical annotations, and contain no personally identifiable information. According to institutional policy at Central People’s Hospital of Zhanjiang, secondary analysis of fully anonymized and publicly available human data does not constitute research involving identifiable human subjects and is therefore exempt from formal Institutional Review Board (IRB) approval; this determination was confirmed by the Ethics Committee of Central People’s Hospital of Zhanjiang. The study was conducted in accordance with the ethical principles of the World Medical Association’s Declaration of Helsinki. All original GEO datasets were collected under appropriate informed consent and ethical approvals by the contributing investigators.

Animals and Diets

Wild-type (WT) C57BL/6J and ApoE^−/− mice were obtained from Cyagen (Guangzhou, China) and maintained under standard pathogen-free conditions with a 12-h light/dark cycle at 22–24 °C. NNAT-deficient (Nnat^−/−) mice were generated using CRISPR/Cas9-mediated gene targeting (Cyagen, Guangzhou, China), with deletion of the genomic region encompassing exons 1–3 of the Nnat gene. Genotyping was performed by PCR using genomic DNA extracted from mouse tail biopsies, followed by sequencing analysis to confirm the targeted deletion. Double knockout (Nnat^−/−ApoE^−/−) mice were produced by crossbreeding Nnat^−/− with ApoE^−/− mice.

Male mice were used in all experiments. All experimental procedures were carried out according to the Care and Use of Laboratory Animals (National Research Council, 8th edition, 2011), and approved by the Institutional Animal Care and Use Committee of Central People’s Hospital of Zhanjiang (ZJDY2022-16). Mice were fed either a standard chow diet or a Western-type diet containing 21% fat and 0.2% cholesterol (Cat. D12079B, Research Diet, New Brunswick, NJ, USA) for 14 weeks starting at 8 weeks of age.

Each experimental group consisted of n = 6 mice unless otherwise specified. Due to the use of genetically defined mouse models, group allocation was determined by genotype rather than random assignment. To minimize bias, mice were included without preselection, and all analyses were performed in a blinded manner.

Assessment of Atherosclerotic Lesions

After 14 weeks of dietary intervention, mice were anesthetized with isoflurane inhalation (3% for induction and 1.5–2% for maintenance, delivered in oxygen at a flow rate of 1 L/min) and euthanized by transcardial perfusion with phosphate-buffered saline (PBS) prior to tissue collection. Aortas and aortic roots were harvested, fixed in 4% paraformaldehyde, and embedded in OCT or paraffin for histological analyses. Plaque formation was evaluated using Oil Red O (ORO) staining for lipid content and hematoxylin and eosin (HE) for lesion morphology. En face preparations of the aorta were also stained with Oil Red O to quantify lipid accumulation. Elastin fibers and collagen content were assessed using Elastica van Gieson (EVG) and Masson’s trichrome staining, respectively. Fibrous cap thickness, intimal thickness, plaque area, and necrotic core size were measured using Image J v2.1.4.7 software (National Institutes of Health, Bethesda, MD, USA).

Immunohistochemistry and Immunofluorescence

Paraffin-embedded sections were deparaffinized, rehydrated, and subjected to antigen retrieval. Sections were incubated overnight at 4°C with primary antibodies: α-smooth muscle actin (α-SMA) (Cat. GB111364, Servicebio, Wuhan, China), F4/80 (Cat. GB11027, Servicebio), and CD31 (Cat. GB11063-2, Servicebio). For immunofluorescence, colocalization of NNAT and CD31 was performed to assess endothelial localization. The NNAT antibody was purchased from Affinity (Cat. DF13349, Changzhou, China), and the CD31 antibody was from Servicebio (Cat. GB11063-2, Wuhan, China). Secondary antibody incubation was performed as per the manufacturer’s instructions. Images were acquired using identical imaging settings across all experimental groups, including consistent laser intensity, detector gain, and exposure time. Sequential scanning mode was employed to minimize spectral overlap between channels, ensuring that the imaging settings were identical for all groups. Nnat^−/− tissues were included as negative controls to confirm the specificity of NNAT staining. Images were analyzed with Image Pro Plus 6.0 software (Minneapolis, MN, USA).

Serum Biochemical Analysis

Blood samples were collected via retro-orbital bleeding and centrifuged at 3,000 g for 15 min at 4 °C to obtain serum. Total cholesterol (TC, Cat. A110), low-density lipoprotein cholesterol (LDL-C, Cat. A113), monocyte chemoattractant protein-1 (MCP-1, Cat. H115), tumor necrosis factor-α (TNF-α, Cat. H052), and interleukin-6 (IL-6, Cat. H007) levels were measured using commercially available kits (Jiancheng, Nanjing, China) according to the manufacturer’s instructions.

Cell Culture and Treatments

Human umbilical vein endothelial cells (HUVECs, Cat. FH1122, Fuheng Biology, Shanghai, China) were cultured in endothelial cell growth complete medium (ECGM, Cat. FH-HUVEC, Fuheng Biology). Cells were maintained at 37 °C in a humidified atmosphere with 5% CO₂. For experiments, HUVECs were exposed to oxidized LDL (oxLDL, 0–100 µg/mL, Cat. L34357, Thermo Fisher Scientific, Waltham, MA, USA) for 48 hours. The effects on cell viability were assessed using the Cell Count Kit 8 (CCK-8, Cat. G4103, Servicebio, Wuhan, China), following the manufacturer’s instructions.

Adenoviral Transduction for Gene Overexpression and Knockdown

Recombinant adenoviral vectors expressing human NNAT (Ad-NNAT), short hairpin RNA targeting NNAT (Ad-shNNAT), and short hairpin RNA targeting PACS-2 (Ad-shPACS-2) were purchased from Cyagen (Guangzhou, China). These vectors were used to transduce HUVECs according to the manufacturer’s instructions. HUVECs were seeded and allowed to reach approximately 60–70% confluence prior to infection. For transduction, cells were exposed to adenoviral vectors at a multiplicity of infection (MOI) of 50 in serum-free medium for 6–8 hours, after which the medium was replaced with complete medium. Following infection, cells were incubated for an additional 24–48 hours before downstream assays. The target sequences for the shRNAs used in this study are listed in Supplementary Table 1. The efficiency of gene overexpression and knockdown was confirmed by Western blot analysis (Supplementary Figure 1).

Lipid Accumulation and Cholesterol Measurement

Intracellular lipid droplets in HUVECs were visualized by Oil Red O staining. Cells were fixed with 4% paraformaldehyde and stained with Oil Red O solution (Cat. 60723ES, Yeasen, Shanghai, China) for 1 hour at room temperature, then washed with PBS. Stained lipids were extracted with isopropanol, and absorbance was measured at 510 nm to quantify lipid accumulation.

Total cholesterol (TC, Cat. 60723ES, Yeasen, Shanghai, China) and free cholesterol (FC, Cat. 60724ES, Yeasen, Shanghai, China) levels were quantified using the corresponding assay kits according to the manufacturer’s instructions. Absorbance was measured at ~500 nm with a microplate reader, and sample concentrations were calculated against a standard curve and expressed in μM.

Endothelial Function Assays

Expression of adhesion molecules Intercellular Adhesion Molecule 1 (ICAM-1) and Vascular Cell Adhesion Molecule 1 (VCAM-1), as well as phosphorylation of Endothelial Nitric Oxide Synthase (p-eNOS) (Ser1177), in HUVECs was assessed by Western blotting. Nitric oxide (NO) production was measured using a detection kit (Cat. 50107ES50, Yeasen, Shanghai, China)) and normalized to protein content.

Reactive Oxygen Species (ROS) and Mitochondrial Function

Intracellular ROS levels were detected using dihydroethidium (DHE, Cat. G1746, Servicebio, Wuhan, China). Mitochondrial membrane potential was assessed using a JC-1 assay kit (Cat. G009, Jiancheng, Nanjing, China). Adenosine triphosphate (ATP) levels were measured using a luminescent ATP assay kit (Cat. A095, Jiancheng, Nanjing, China). Antioxidant enzyme activities, including superoxide dismutase (SOD, Cat. A001), glutathione peroxidase (GSH-Px, Cat. A005), and malondialdehyde (MDA, Cat. A003), were determined using commercial kits (Jiancheng, Nanjing, China) according to the manufacturer’s instructions.

Proximity Ligation Assay (PLA) and Transmission Electron Microscopy (TEM)

PLA was performed to assess ER–mitochondria contact sites using the Duolink^® in situ kit (Cat. DUO92101, Sigma-Aldrich, St. Louis, MO, USA) according to the manufacturer’s instructions. TEM was used to visualize ER–mitochondria contact structures in HUVECs. Cells were fixed in 2.5% glutaraldehyde, postfixed in 1% osmium tetroxide, dehydrated, embedded in resin, and sectioned for TEM observation (Hitachi, Tokyo, Japan).

Quantitative PCR (qPCR)

Total RNA was extracted using TRIzol reagent (Cat. 15596026, ThermoFisher, Waltham, MA, USA). RNA concentration and purity were determined using a NanoDrop spectrophotometer. Equal amounts of RNA were reverse-transcribed into cDNA using a cDNA synthesis kit (Cat. G3333, Servicebio, Wuhan, China) according to the manufacturer’s instructions. QPCR was performed using SYBR Green Master Mix (Cat. G3326, Servicebio) on a real-time PCR system (ABI 7500, Applied Biosystems, USA). The PCR conditions were as follows: initial denaturation at 95 °C for 10 min, followed by 40 cycles of denaturation at 95 °C for 10s and annealing/extension at 60 °C for 30s. β-actin (ACTB) was used as an internal control. Relative gene expression levels were calculated using the 2⁻ΔΔCt method. All reactions were performed in triplicate. Primer sequences are listed in Supplementary Table 2.

Western Blotting

Cells were lysed in RIPA buffer supplemented with protease and phosphatase inhibitors. BCA assay was used to detect protein concentrations, and equal amounts of protein (30 μg) were separated by SDS-PAGE and transferred onto PVDF membranes.

Membranes were blocked with 5% non-fat milk in TBST for 1 h at room temperature and incubated overnight at 4 °C with primary antibodies against NNAT (Cat. DF13349), Phosphofurin Acidic Cluster Sorting Protein 2 (PACS-2) (Cat. DF12441), Mitofusin 2 (MFN2) (Cat. DF8106), Lectin-like Oxidized Low-Density Lipoprotein Receptor 1 (LOX-1) (Cat. DF6522), ATP-binding Cassette Subfamily A Member 1 (ABCA1) (Cat. DF8233), ICAM-1 (Cat. AF6088), VCAM-1 (Cat. DF6082), p-eNOS (Cat. AF3247), eNOS (Cat. AF0096), p-eNOS (Ser1177) (Cat. XXX), p-p65 (Ser536) (Cat. AF2006) and p65 (AF5006) (all from Affinity, Changzhou, China). After washing, membranes were incubated with HRP-conjugated secondary antibodies for 1 h at room temperature. Protein bands were visualized using an enhanced chemiluminescence (ECL) detection system (Cat. G2020, Servicebio, Wuhan, China) and quantified using ImageJ software (NIH, USA). β-actin (Cat. GB15001, Servicebio) was used as a loading control.

Statistical Analysis

Data in this work are presented as mean ± standard deviation (SD). Normality of data distribution was assessed using the Shapiro–Wilk test prior to statistical analysis. For normally distributed data, statistical significance was determined using unpaired Student’s t-test for two-group comparisons or one-way ANOVA followed by Tukey or Tamhane’s T2 post hoc test for multiple comparisons. A P-value < 0.05 was considered statistically significant. GraphPad Prism 9.0 was used for all analyses.

Results

NNAT Deficiency Exacerbates Atherosclerosis in WD-Fed ApoE^−/− Mice

Analysis of GEO datasets revealed that NNAT expression was significantly reduced in aortas from patients with AS (Figure 1A), stenotic aortic tissues (Figure 1B), and calcified aortic valve regions (Figure 1C) compared with healthy controls. Consistent findings were observed in WD-fed ApoE^−/− mice, where NNAT expression was markedly downregulated in the aorta (Figure 1D) and adventitia (Figure 1E).

Figure 1.

Effect of NNAT deficiency on atherosclerotic progression in WD–fed ApoE^−/− mice. (A) Analysis of NNAT mRNA expression in carotid atherosclerotic plaques (n = 30) and non-atherosclerotic arteries (n = 12) from the GSE100927 dataset. (B) NNAT expression in aortic valve tissues from patients with aortic stenosis (n = 10) and healthy controls (n = 10) based on the GSE153555 dataset. (C) NNAT expression in calcified aortic valve tissues (n = 5) and normal valves (n = 5) derived from the GSE51472 dataset. (D and E) Temporal changes in aortic NNAT expression during atherosclerosis progression in ApoE^−/− mice using data from the GSE40156 dataset. (F) Representative macroscopic images showing aortic arch plaque formation after 14 weeks of WD feeding in mice with different genotypes. (G) Representative Oil Red O staining of aortic arches. (H) Quantification of lesion area percentage based on (G). (I and J) Representative en face Oil Red O staining images of whole aortas and corresponding quantitative analysis of lesion area. (K–O) Serum levels of LDL-C, TC, MCP-1, TNF-α, and IL-6 in mice of each genotype. Data are presented as mean ± SD (n = 6 per group). *P < 0.05 vs groups of Control, Normal or WT, ^#P < 0.05 vs ApoE^−/− group.

After 14 weeks of WD feeding, wild-type (WT) mice showed minimal plaque deposition in the aortic arch, while ApoE^−/− mice developed substantial plaques. NNAT knockout alone did not result in significant atherosclerotic lesion formation; however, combined NNAT and ApoE deficiency led to a striking increase in plaque burden (Figure 1F). Oil Red O staining of aortic arch sections demonstrated that NNAT deficiency significantly enhanced lipid deposition and lesion formation in ApoE^−/− mice (Figure 1G and H). Consistently, en face Oil Red O staining of the aorta revealed a significantly greater lipid burden in double-knockout mice compared with single-knockout controls (Figure 1I and J). Serum biochemical analysis showed that double-knockout mice exhibited significantly higher levels of total cholesterol, LDL-C, and pro-inflammatory markers (MCP-1, TNF-α, and IL-6) compared with ApoE^−/− mice (Figure 1K–O).

NNAT Deficiency Aggravates Aortic Root Lesions in WD-Fed ApoE^−/− Mice

Given its critical role in atherosclerosis development, the aortic root was further analyzed. Oil Red O staining showed that NNAT deficiency significantly increased lipid deposition in the aortic root of ApoE^−/− mice (Figure 2A and B; P < 0.05). Intimal thickness, plaque area, and necrotic core size were also significantly greater in Nnat^−/−ApoE^−/− mice compared with ApoE^−/− controls (HE staining; Figure 2C–F). Histological analyses revealed more severe elastin fiber disruption (EVG staining; Figure 2G) and reduced collagen content with thinner fibrous caps (Masson’s staining; Figure 2H–J). Immunohistochemistry demonstrated a significant reduction in α-SMA–positive smooth muscle cell content (Figure 2K and L) and increased macrophage infiltration (F4/80 staining; Figure 2M and N) in NNAT-deficient plaques. Together, these results demonstrate that NNAT deficiency is associated with increased lesion burden and features consistent with a more vulnerable plaque phenotype in WD-fed ApoE^−/− mice.

Figure 2.

Effect of NNAT deficiency on aortic valve lesions in WD–fed ApoE^−/− mice. (A) Representative Oil Red O staining images of aortic valve tissues from mice of different genotypes. (B) Quantitative analysis of lipid deposition in aortic valve plaques based on Oil Red O staining. (C) Representative HE staining images of aortic valve tissues from mice of different genotypes. (D) Intima-to-media thickness ratio (I/M ratio) of aortic roots calculated from HE staining. (E) Quantification of aortic root plaque area based on HE staining. (F) Percentage of necrotic core area within plaques based on HE staining. (G) Representative EVG staining images of aortic valve lesions. (H) Representative Masson’s trichrome staining images. (I) Quantification of collagen content in plaques based on Masson staining. (J) Quantitative analysis of fibrous cap thickness based on Masson staining. (K) Representative immunohistochemical staining of α-SMA in aortic root plaques. (L) Quantification of α-SMA expression in plaques. (M) Representative immunohistochemical staining of F4/80 in aortic root plaques. (N) Quantification of F4/80 expression in plaques. Data are presented as mean ± SD (n = 6 per group). *P < 0.05 vs ApoE^−/− group.

oxLDL Suppresses NNAT Expression in HUVECs

Colocalization analysis showed that NNAT is predominantly expressed in endothelial cells of the aortic root, as indicated by its overlap with CD31 (Figure 3A). To model lipid-induced toxicity, oxLDL was used to treat HUVECs in vitro. Cell viability was assessed after 48 hours of treatment with varying oxLDL concentrations. Results indicated that concentrations of 50 and 100 μg/mL significantly reduced HUVEC viability (P < 0.05), suggesting substantial cell damage (Figure 3B). Interestingly, while these higher concentrations of oxLDL led to reduced cell viability, they also caused a dose-dependent increase in lipid accumulation, as observed in Oil Red O staining (Figure 3C and D), along with higher levels of total and free cholesterol (Figure 3E and F). These findings highlight the dual effects of oxLDL treatment: at higher concentrations, oxLDL not only induces significant cell damage but also exacerbates lipid accumulation, reinforcing its role in endothelial dysfunction and atherogenesis. Both qPCR and Western blot analyses demonstrated a significant reduction in NNAT expression at both the mRNA and protein levels following oxLDL treatment (Figure 3G–I).

Figure 3.

Effect of oxLDL stimulation on NNAT expression in HUVECs. (A) Representative immunofluorescence images showing CD31 and NNAT colocalization in aortic root plaques of WD-fed mice. (B) Effects of different doses of oxLDL on HUVECs cell viability after 48 hours of treatment. (C) Representative Oil Red O staining of HUVECs exposed to varying concentrations of oxLDL for 48 h. (D) Quantitative analysis of lipid droplet accumulation based on extracted Oil Red O absorbance (OD value). (E and F) TC and Fch levels in each group of HUVECs. (G) Quantitative PCR analysis of NNAT mRNA expression. (H, I) Western blot analysis of NNAT protein levels in HUVECs under different doses of oxLDL treatment. Data are presented as mean ± SD and n=6. *P < 0.05 vs 0 μg/mL oxLDL.

NNAT Overexpression is Associated with Improved Endothelial Function and Reduced Inflammatory Signaling in HUVECs

To evaluate the role of NNAT in endothelial cells, HUVECs were transduced with adenovirus to overexpress or knock down NNAT and then exposed to oxLDL. OxLDL treatment markedly increased ICAM-1 and VCAM-1 expression while reducing p-eNOS levels, effects that were reversed by NNAT overexpression and further exacerbated by NNAT knockdown (Figure 4A–D). Consistently, NNAT overexpression restored NO production, whereas knockdown further suppressed it (Figure 4E).

Figure 4.

Effect of NNAT overexpression on lipid metabolism in oxLDL-treated HUVECs. (A–D) HUVECs were infected with NNAT-overexpressing or NNAT-silencing adenoviruses for 48 h, followed by 48 h of oxLDL exposure. Protein levels of ICAM-1, VCAM-1 (normalized to β-actin), and phosphorylated eNOS (normalized to total eNOS) were detected by Western blotting. (E) NO concentration in culture supernatants. (F and G) Intracellular TC and Fch levels in each group. (H) Representative Oil Red O and Bodipy staining images. (I) Quantitative analysis of lipid droplet accumulation based on Oil Red O absorbance. (J) Quantification of Bodipy fluorescence intensity. (K–O) Western blot analysis of LOX-1, ABCA1 and IκBα protein expression normalized to β-actin and p-p65 protein expression normalized to total p65. Control: HUVECs cultured under normal conditions without treatment; oxLDL: oxLDL treatment at 100 μg/mL; NNAT-OE: NNAT overexpression followed by oxLDL treatment; NNAT-KD: NNAT knockdown followed by oxLDL treatment. Data are presented as mean ± SD and n=6. *P < 0.05 vs control group, ^#P < 0.05 vs oxLDL group.

Next, lipid metabolism was assessed. NNAT overexpression significantly reduced intracellular total and free cholesterol levels as well as lipid droplet accumulation, while knockdown increased these parameters (Figure 4F–J; P < 0.05). Mechanistically, NNAT overexpression decreased LOX-1 expression and upregulated ABCA1, whereas knockdown produced the opposite effects (Figure 4K and M).

In addition, we assessed NF-κB activation and found that oxLDL significantly down-regulated expression of IκBα and increased p65 phosphorylation, which were attenuated by NNAT overexpression and further enhanced by NNAT knockdown (Figure 4K–O). These findings suggest that NNAT is associated with changes in endothelial inflammatory responses through NF-κB–related signaling.

NNAT Overexpression is Associated with ER–Mitochondria Interactions and Mitochondrial Function in HUVECs

Overexpression of NNAT restored mitochondrial membrane potential and ATP production, whereas knockdown aggravated mitochondrial dysfunction (Figure 5A–C). ROS levels, assessed by DHE staining, were markedly increased by oxLDL treatment but were reduced by NNAT overexpression and further elevated upon knockdown (Figure 5D and E). Overexpression of NNAT also significantly increased the activity of antioxidant enzymes SOD and GSH-Px while reducing MDA levels; these effects were reversed by NNAT knockdown (Figure 5F–H).

Figure 5.

Effect of NNAT overexpression on ER–mitochondria interactions in oxLDL-treated HUVECs. (A) Representative JC-1 staining images of HUVECs. (B) Quantitative comparison of mitochondrial membrane potential based on JC-1 fluorescence. (C) Cellular ATP content in each group. (D) Representative DHE staining images. (E) Quantitative analysis of DHE fluorescence intensity. (F–H) Measurement of SOD, GSH-PX, and MDA levels in each group. (I) Representative PLA images showing ER–mitochondria contact sites between IP3R and VDAC1. (J) Quantification of PLA puncta per cell based on (I). (K–M) Western blot analysis of PACS-2 and MFN2 protein expression normalized to β-actin. Control: HUVECs cultured under normal conditions without treatment; oxLDL: oxLDL treatment at 100 μg/mL; NNAT-OE: NNAT overexpression followed by oxLDL treatment; NNAT-KD: NNAT knockdown followed by oxLDL treatment. Data are presented as mean ± SD and n=6. *P < 0.05 vs control group, ^#P < 0.05 vs oxLDL group.

PLA analysis revealed that oxLDL disrupted ER–mitochondria contact sites, which was partially rescued by NNAT overexpression and exacerbated by knockdown (Figure 5I and J). Consistently, Western blotting showed that NNAT overexpression increased the expression of PACS-2 and MFN2, whereas knockdown reduced their levels (Figure 5K–M). PACS-2 and MFN2 were used as representative markers of ER–mitochondria contact sites rather than exhaustive indicators of all related proteins.

Collectively, these results indicate that NNAT is associated with ER–mitochondria interactions and mitochondrial function under atherogenic stress conditions.

Disruption of ER–Mitochondria Interactions Attenuates NNAT-Associated Endothelial Effects

To further assess the role of ER–mitochondria interactions, PACS-2 knockdown was used to modulate ER–mitochondria contact in HUVECs. TEM and PLA analyses confirmed reduced ER–mitochondria contact sites in PACS-2–deficient cells (Figure 6A–C). Under these conditions, the endothelial protective effects of NNAT overexpression—including suppression of ICAM-1 and VCAM-1, restoration of eNOS phosphorylation, and NO production—were markedly reduced (Figure 6D and E). Similarly, NNAT-mediated improvements in lipid metabolism, such as reduced total and free cholesterol, decreased lipid droplet accumulation, and modulation of LOX-1 and ABCA1 expression, were significantly impaired when ER–mitochondria interactions was disrupted (Figure 6F–M). Together, these findings suggest that ER–mitochondria interactions are associated with NNAT-associated endothelial protective effects.

Figure 6.

Effect of PACS-2 knockdown on NNAT-mediated regulation of ER–mitochondria interactions in HUVECs. (A) Transmission electron microscopy images showing ER–mitochondria contact morphology after PACS-2 knockdown. (B and C) PLA analysis evaluating the ER–mitochondria contact sites following PACS-2 silencing. (D and E) Western blot detection of ICAM-1, VCAM-1, and p-eNOS protein expression in HUVECs overexpressing NNAT and/or with PACS-2 knockdown under oxLDL stimulation. (F) NO concentration in culture supernatants. (G and H) Intracellular TC and Fch levels in each group. (I) Representative Oil Red O and Bodipy staining images. (J) Quantitative analysis of lipid droplet accumulation based on Oil Red O absorbance. (K) Quantification of Bodipy fluorescence intensity. (L and M) Western blot analysis of LOX-1 and ABCA1 protein expression normalized to β-actin. Control: HUVECs cultured under normal conditions without treatment; oxLDL: oxLDL treatment at 100 μg/mL; NNAT-OE: NNAT overexpression followed by oxLDL treatment; PACS-2-KD: PACS-2 knockdown followed by oxLDL treatment. NNAT-OE + PACS-2-KD: NNAT overexpression combined with PACS-2 knockdown followed by oxLDL treatment. Data are presented as mean ± SD and n=6. *P < 0.05 vs control group, ^#P < 0.05 vs oxLDL group, ^&P < 0.05 vs NNAT-OE group.

Discussion

NNAT is an imprinted gene that plays an important role in metabolic regulation, influencing food intake, insulin secretion, and adipocyte differentiation. Alternative splicing of NNAT mRNA produces three isoforms (α, β, and γ), each containing hydrophobic and hydrophilic domains. Whether these isoforms function redundantly or possess distinct biological activities remains uncertain.¹⁶ Previous studies have shown that NNAT loss reduces whole-body energy expenditure in mice and increases susceptibility to diet-induced obesity and impaired glucose tolerance.¹⁷ Millership et al further demonstrated that whole-body or β-cell–specific deletion of NNAT impairs glucose-stimulated insulin secretion under nutrient excess, resulting in glucose intolerance without changes in body weight or food intake.¹⁸ These findings establish NNAT as a key regulator of energy metabolism, though its contribution to AS has not been fully defined.

In this study, public dataset analyses revealed that NNAT expression is significantly decreased in carotid lesions and stenotic or calcified aortic tissues from AS patients, as well as in the aortas of ApoE-deficient mice. Experimental validation confirmed that NNAT deficiency markedly aggravated lipid deposition and lesion severity in WD-fed ApoE^−/− mice, supporting an association between NNAT and atheroprotective effects in the context of AS and suggesting its potential relevance for further mechanistic investigation in the context of AS. It is important to note that this study utilized only male mice, which may limit the generalizability of the findings, as sex differences can influence the progression of atherosclerosis. Future studies should incorporate both male and female mice to explore potential sex-specific effects of NNAT in AS. Additionally, we acknowledge that a formal power calculation was not performed, which constitutes a limitation of the study. Future research with larger sample sizes and predefined power calculations will be crucial to enhance the reliability and validity of our findings.

Endothelial cells, including valve endothelial cells, serve as the “first barrier” in vascular and valvular pathology. By sensing disturbed blood flow and metabolic stress, they initiate inflammation, promote lipid accumulation and calcification, and interact with valve interstitial cells to drive lesion progression.^19,20 Earlier work by Mzhavia et al showed that NNAT is primarily localized in endothelial cells of the murine aorta and upregulated under conditions of obesity and diabetes.⁴ In contrast, our finding of NNAT downregulation suggests that its expression is context-dependent and influenced by specific pathological environments. Using immunofluorescence, we observed that NNAT is abundantly expressed in endothelial cells within the aortic root, supporting a potential contribution of endothelial NNAT to AS-related vascular alterations, although a definitive causal role cannot be established and contributions from other NNAT-expressing tissues cannot be excluded.

Our in vitro findings further support this role. In HUVECs, oxLDL treatment progressively reduced NNAT expression while increasing cholesterol accumulation and lipid droplet formation. NNAT overexpression alleviated oxLDL-induced endothelial dysfunction and lipid deposition, whereas knockdown worsened these effects. Our in vitro findings indicate that NNAT is associated with endothelial protective effects under oxLDL exposure, which may be related to the aggravated atherosclerotic phenotype observed in vivo. In addition, our results indicate that NNAT is associated with modulation of NF-κB signaling in endothelial cells. OxLDL-induced NF-κB activation was attenuated by NNAT overexpression and enhanced by NNAT knockdown, consistent with our previous findings.⁵

Because of its sequence homology to phospholamban, NNAT was initially proposed to regulate sarco/endoplasmic reticulum Ca²⁺-ATPase (SERCA) pump activity.¹⁷ Later studies confirmed that NNAT localizes to the ER and antagonizes SERCA, limiting Ca²⁺ reuptake into the ER.¹³ NNAT expression is also sensitive to oxidative stress, and by modulating Ca²⁺ homeostasis and downstream signaling, it has been implicated in diverse diseases including breast cancer, Lafora disease, obesity, diabetes, and psychiatric disorders.^14,21–24

The role of NNAT in ER–mitochondria interactions remains less understood. Given its involvement in ER calcium handling and lipid metabolism—both relevant to ER–mitochondria interactions—it is plausible that NNAT is associated with ER–mitochondria interactions-related processes. ER–mitochondria contact sites coordinate calcium (Ca²⁺) homeostasis, lipid transfer, mitochondrial dynamics, and apoptotic signaling, which are important for cellular metabolism and stress responses in cardiovascular cells.^25,26 These processes have been implicated in cardiovascular pathophysiology, including endothelial dysfunction and inflammation, supporting the relevance of ER–mitochondria interactions in AS progression.^26,27 Other ER-membrane–associated proteins, such as SERCA regulators and the Sigma-1 receptor, are known to influence calcium signaling and mitochondrial function at ER–mitochondria contact sites.^28,29 NNAT may participate in Ca²⁺ signaling and mitochondrial bioenergetics within this cellular context.

Our results showed that oxLDL induced oxidative stress, mitochondrial dysfunction, accompanied by altered ER–mitochondria interactions in HUVECs. In this study, ER–mitochondria interactions were evaluated using representative markers (PACS-2 and MFN2) together with functional assays, rather than a comprehensive profiling of all associated components. These observations differ from previous work by Yu et al, in which oxLDL was reported to increase MAM formation in endothelial cells through enhanced PACS2‑mediated ER–mitochondria contact, facilitating Ca²⁺ transfer and apoptotic signaling.³⁰

ER–mitochondria contact sites are dynamic interorganelle contact sites that coordinate calcium homeostasis, lipid exchange, ROS handling, and apoptotic signaling, and their functional states varies depending on cellular stress conditions.^25,31 Disruption of ER–mitochondria interactions has been implicated in cardiovascular diseases and oxidative stress-related injury, suggesting that both increases and decreases in ER–mitochondria contacts may occur under different pathological conditions.²⁵ In our model of sustained oxLDL challenge, prolonged metabolic and oxidative stress may impair ER–mitochondria interactions, leading to mitochondrial dysfunction and increased ROS production. Importantly, NNAT overexpression attenuated ROS production, improved mitochondrial function, and was associated with preservation of ER–mitochondria contacts, while disruption of ER–mitochondria interactions via PACS-2 knockdown markedly attenuated these protective effects. These observations support the notion that the endothelial protective effects of NNAT are associated with preserved ER–mitochondria interactions, although a direct structural role of NNAT in ER–mitochondria contact sites was not demonstrated in this study.

Several limitations of this study should be acknowledged. First, an important limitation is the use of a global NNAT-deficient mouse model. Given that NNAT is expressed in multiple metabolically active tissues, systemic deletion may lead to broader metabolic and inflammatory alterations that could indirectly influence atherosclerosis development. Therefore, the observed phenotypes cannot be exclusively attributed to endothelial-specific NNAT deficiency, and contributions from other NNAT-expressing tissues cannot be excluded. Second, the relatively small sample size in animal experiments may limit the statistical power of the in vivo findings. Future studies with larger cohorts and predefined power calculations are needed to further validate these results. Third, although our data support a role for NNAT in modulating ER–mitochondria interactions, it remains unclear whether NNAT directly regulates MAM components or acts indirectly through upstream signaling pathways. Direct evidence of NNAT subcellular localization at ER–mitochondria interfaces or its interaction with established tethering proteins was not addressed in the present study. Collectively, these limitations highlight the need for future studies using tissue-specific genetic models, larger sample sizes, and high-resolution mechanistic approaches to further define the role of NNAT in atherosclerosis.

In summary, this study provides evidence that NNAT is associated with vascular homeostasis under atherogenic stress conditions. NNAT deficiency is associated with exacerbated atherosclerotic phenotypes in ApoE^−/− mice, whereas NNAT overexpression attenuates oxLDL-induced endothelial dysfunction and lipid dysregulation in vitro. Importantly, beyond our previous work on NNAT-related inflammatory signaling, the present study suggests an association between NNAT and ER–mitochondria interactions, as assessed by ER–mitochondria contact integrity (PLA and TEM), representative proteins (PACS-2 and MFN2), and mitochondrial functional parameters. These findings extend the understanding of NNAT from inflammatory regulation to a broader role in endothelial homeostasis and AS. Future studies are needed to further clarify the relationship between NNAT and ER–mitochondria interactions in cardiovascular disease.

Funding Statement

This work was supported by the Natural Scientific Foundations of China (82270478) and Disciplinary Construction Project of Central People’s Hospital of Zhanjiang (2020A14).

Data Sharing Statement

Data will be made available on request from the corresponding author.

Author Contributions

Wenkai Yang: Conceptualization, Project Administration, Writing – Review & Editing, Funding Acquisition. Yinghui Huang: Methodology, Visualization, Formal Analysis, Writing – Original Draft. Kai Tang: Methodology, Validation, Writing – Review & Editing. Kun Tian: Investigation, Methodology, Writing – Review & Editing. Bowen Zhang: Methodology, Software, Writing – Review & Editing. Tongju Han: Methodology, Supervision, Resources, Writing – Review & Editing. All authors gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare that there are no conflicts of interest in this study.