NFS1 Regulates IDH2 to Attenuate Abdominal Aortic Aneurysms via Interacting With SP2

TL;DR

This study shows that NFS1 helps prevent abdominal aortic aneurysms by regulating energy metabolism in smooth muscle cells.

Contribution

The study reveals a novel role for NFS1 in AAA progression via its interaction with SP2 and regulation of IDH2.

Findings

NFS1 is downregulated in AAA tissues and its loss promotes glycolysis in vascular smooth muscle cells.

NFS1 acts as a cofactor for SP2 to regulate the expression of IDH2, which is critical for AAA development.

Reduced NFS1 binding to SP2 leads to decreased IDH2 expression and worsens AAA progression.

Abstract

Abdominal aortic aneurysm (AAA) is a life‐threatening condition with limited pharmacological therapies. The pathological progression of AAA is closely attributed to the phenotypic switching of vascular smooth muscle cells (VSMCs). NFS1 is the rate‐limiting enzyme for the synthesis of iron‐sulfur proteins, and the roles of NFS1 in AAA initiation and development have not been explored. Angiotensin II (Ang II) infusion‐induced AAA animal model with Apoe
−/− mice combined with human thoracic aorta samples are used to analyze the role of NFS1 in AAA development. Gain or loss‐of‐function studies are conducted to investigate the regulatory roles of NFS1 on SMC phenotypic switching at both cellular and animal levels. CUT&Tag is further performed for identifying the targets of NFS1 involved in AAA progression. NFS1 is downregulated in the abdominal aortic tissues from both patients and mice. Defects in NFS1 in VSMCs led to enhanced glycolysis and impaired mitochondrial function, contributing to the phenotypic transformation of VSMCs. Mechanistically, NFS1 functions as a transcriptional cofactor of SP2 for inducing the transcription of Idh2. Inhibition of IDH2 partially attenuated the protective effect of NFS1 on AAA. This study uncovers a crucial role for NFS1 in the development and progression of AAA, suggesting that NFS1 may serve as a novel therapeutic and prognostic marker for this condition.

The loss of NFS1 leads to its reduced binding with the transcription factor SP2, resulting in decreased expression of the downstream gene Idh2. This causes a shift in energy supply toward glycolysis in VSMCs, ultimately contributing to the occurrence and progression of AAA.

Figures

NFS1 is downregulated in human and murine aortic aneurysm tissues. A) α‐SMA and CSE proteins expressions in human aortic aneurysm lesion areas and non‐lesion area were assessed by Western blotting (n = 6). B) OPN and NFS1 proteins expression in human aortic aneurysm lesion areas and non‐lesion area were assessed by Western blotting (n = 6). C) PCA analysis of human AAA bulk RNA‐seq data. D) Sample_distance_heatmap of human AAA bulk RNA‐seq data. E) Volcano plot of differentially expressed genes in human AAA bulk RNA‐seq data. F) Mouse AAA scRNA‐seq clustering results plot. Apoe −/− mice were infused with Ang II (1000 ng kg−1 min−1) subcutaneously for 4 weeks to induce AAA. A dose of 14 µmol kg−1 sodium hydrosulfide (NaHS) were injected via the abdomen three times a week for 2 weeks before Ang II infusion. G) Volcano plot of differentially expressed genes in VSMC of mouse AAA scRNA‐seq data. H) α‐SMA and CSE proteins in mouse AAA were assessed by Western blotting (n = 6). I) OPN and NFS1 proteins expressions in mouse AAA were assessed by Western blotting (n = 6). Data were analyzed by Welch ANOVA followed by the Dunnett`s T3 post hoc test (α‐SMA, OPN, NFS1) and Kruskal‐Wallis ANOVA followed by the Dunn`s post hoc test (CSE).

NaHS effectively attenuates Ang II‐induced AAA formation. Apoe −/− mice were infused with Ang II (1000 ng kg−1 min−1) subcutaneously for 4 weeks to induce AAA. A dose of 14 µmol kg−1 NaHS were injected via the abdomen three times a week for 2 weeks before Ang II infusion. A) Representative images of the macroscopic features of AAA formation in the indicated groups. Scale bar, 2 mm. B) Representative images of abdominal aortas visualized by MUI using B mode and M mode in the indicated groups. C) Infrarenal aortas maximal diameter measured by MUI B mode in the different groups (n = 6). Data were analyzed by Welch ANOVA followed by the Dunnett`s T3 post hoc test. D) Representative HE, EVG, Masson staining of abdominal aortas in the indicated groups. Scale bar, 100 µm. E) Semi‐quantification of the grade of elastin degradation and F) middle membrane collagen deposition in the abdominal aortas (n = 7). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test (elastin degradation), Kruskal‐Wallis ANOVA followed by the Dunn`s post hoc test (collagen deposition).

NaHS rescues SMC phenotype switching by targeting at NFS1. MOVAS ere stimulated with 3 µM Ang II for 24 h. NaHS (30 µM) was added 30 min prior to Ang II treatment where applicable. A–D) α‐SMA and OPN proteins expression and quantification in MOVAS (n = 6). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test (α‐SMA), Kruskal‐Wallis ANOVA followed by the Dunn`s post hoc test (OPN). E–G) Cyclin D1 and PCNA proteins expression and quantification in MOVAS (n = 6). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test. H) Relative EdU levels of MOVAS in different treated groups (n = 6). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test. I,J) NFS1 protein expression and quantification in MOVAS (n = 6). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test. K,L) Differential expression of α‐SMA and OPN proteins after NFS1 was knockdown. M,N) Differential expression of α – SMA and OPN proteins after NFS1 was overexpressed.

IDH2 acts as a target of NFS1. A,B) CUT & Tag was performed with NFS1 antibody in aortic tissues. NFS1 was enriched in the promoter region of numerous genes. These genes with similar distribution patterns were clustered together by a clustering algorithm to show the binding trend of NFS1 as a transcription co‐factor. (n = 3). C) Significant terms of biological process in gene ontology (GO) analyses were carried out in NFS1‐related genes. D) NFS1 peaks in the promoter regions of Idh2. MOVAS were stimulated with 3 µM Ang II for 24 h. NaHS (30 µM) was added 30 min prior to Ang II treatment where applicable. E,G) Western blotting assays and quantitative analysis of IDH2 in MOVAS (n = 6). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test. Apoe −/− mice were infused with Ang II (1000 ng kg−1 min−1) subcutaneously for 4 weeks to induce AAA. A dose of 14 µmol kg−1 NaHS were injected via the abdomen three times a week for 2 weeks before Ang II infusion. F,H) Western blotting and quantitative analysis of IDH2 in mouse AAA tissues (n = 6). Data were analyzed by Welch ANOVA followed by the Dunnett`s T3 post hoc test. I,J) Western blotting assays and quantitative analysis of IDH2 in human aortic aneurysm tissues (n = 6).

NFS1 regulates glycolytic pathway in AAA progression. MOVAS were stimulated with 3 µM Ang II for 24 h. NaHS (30 µM) was added 30 min prior to Ang II treatment where applicable. A) PFKP, HK2, ENO1, PGAM1, LDHA, ALDOA and PKM2 protein expressions assessed by Western blotting in MOVAS. B–H) Quantification of protein expressions measured by Western blotting in MOVAS (n = 6). All data was analyzed by one‐way ANOVA followed by the Tukey post hoc test. I) Differences in extracellular lactate contents between various treatments (n = 6). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test (extracellular lactate). J) Mitochondrial superoxide content in different groups measured by MitoSOXs staining, the results were showed as fluorescence images. Scale bar, 50 µm. K) Representative fluorescence images of mitochondrial permeability transition pore evaluated by mPTP staining. Scale bar, 50 µm. L) JC‐1 staining tested by microplate reader with different treatments (n = 6). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test.

Inhibition of IDH2 attenuates the protective effect of NFS1. MOVAS were stimulated with 3 µM Ang II for 24 h. NaHS (30 µM) was added 30 min prior to Ang II treatment where applicable. For transfection, 1.5 µg of si‐IDH2 was added to each well of a 6‐well plate. A) Differential expression of α‐SMA and OPN proteins by NFS1 overexpression and enasidenib treatment. B) The ECAR of glycolytic rate assay in different groups (n = 4). C) The OCR of cell mito stress test in different groups (n = 4). D) Representative images of the macroscopic features of AAA formation in indicated groups. Scale bar, 2 mm. AAV9 vectors was delivered via tail vein injection at a dose of 5 × 1012 genomic copies (GC) 2 weeks prior to modeling. For enasidenib treatment, mice were received intraperitoneal injections (15 mg kg−1 day−1) for the first 3 days of each week over 2 weeks prior to modeling. E) Representative HE, EVG, Masson staining of abdominal aortas in indicated groups. Scale bar, 200 µm. F) Representative images of abdominal aortas visualized by MUI using B mode and color mode in the indicated groups. G,H) Differential expression of α – SMA and OPN proteins of murine measured by Western blot after injected with AAV‐NFS1 together with enasidenib treatment (n = 6).

NFS1 works as a co‐factor of SP2 for regulating Idh2 transcription. MOVAS were stimulated with 3 µM Ang II for 24 h. NaHS (30 µM) was added 30 min prior to Ang II treatment where applicable. A) Motif analysis of peaks sequences after pull‐down with NFS1 antibody in CUT & Tag. B) Co‐IP assay for analyzing the transcription factors binding to NFS1 in MOVAS. C) Immunofluorescence co‐staining of SP2 and NFS1 to observe the co‐localization in MOVAS. Left figures scale bar, 50 µm. Right figures scale bar, 5 µm. D,E) SP2 protein expression assessed by Western blotting in MOVAS and mouse aortic tissue, respectively (n = 6). F,G) Quantification of SP2 protein expression measured by Western blotting in MOVAS and mouse aortic tissue, respectively (n = 6). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test. H) SP2 protein expression assessed by Western blot in human aortic aneurysm tissues (n = 6). I) Quantification of SP2 protein expression measured by Western blot in aortic tissue (n = 6). Data were analyzed by Student t test. J,K) Construction of full‐length and truncated reporter plasmids for analyzing the binding region of SP2 in IDH2 promoter (n = 6).

NFS1 directly binds to SP2. Apoe −/− mice were infused with Ang II (1000 ng kg−1 min−1) subcutaneously for 4 weeks to induce AAA. A dose of 14 µmol kg−1 NaHS were injected via the abdomen three times a week for 2 weeks before Ang II infusion. AAV9 was delivered via tail vein injection at a dose of 5 × 1012 genomic copies (GC) 2 weeks prior to modeling. A,B) Modeling the protein spatial structure of NFS1 and SP2 for predicting their binding sites. C) Co‐transfection with different NFS1 plasmids and full‐length SP2 plasmid into 293T for pull‐down experiments with anti–HA antibody. D) Different AAV9 vectors containing the full‐length or deleted NFS1 were constructed and injected into mice via tail vein, and the expressions of α‐SMA and OPN in AAA tissues were verified by Western blotting (n = 6). E) Representative HE, EVG, Masson staining of abdominal aortas in the indicated groups. Scale bar, 200 µm. F) Representative images of abdominal aortas visualized by MUI using B mode and color mode in the indicated groups.

Full text

Introduction

1

Abdominal aortic aneurysm (AAA) is the permanent dilation of abdominal aorta, which can lead to catastrophic rupture and sudden death. Notably, most AAA patients are asymptomatic prior to a fatal rupture, thus the prevention and treatment strategies for AAA are very complicating^[^ 1 ^]^ Although recent endovascular stent graft and open surgical repair therapies have significantly improved the prognosis of AAA patients, there are still no effective pharmacotherapy for limiting the progression and rupture of AAA. A profound understanding of the underlying molecular mechanisms involved in AAA formation is necessary for developing effective drugs to prevent and attenuate AAA progression and rupture.

The well‐known risk factors for AAA include smoking, dyslipidemia, hypertension, and age, etc.^[^ 2 ^]^ In addition, recent studies have highlighted that metabolic disorders may play crucial roles in the onset and progression of AAA.^[^ 3 ^]^ Mitochondria, as the center for amino acid, glucose and lipid metabolism, are vital for maintaining cellular energy homeostasis.^[^ 4 ^]^ Mitochondria are also the main source of intracellular reactive oxygen species (ROS).^[^ 5 ^]^ Impaired mitochondrial function is closely associated with the development of various diseases.^[^ 6 ^]^ However, it remains inadequately explored on the regulatory mechanisms by which impaired mitochondrial function affects cellular metabolism in AAA.

The iron‐sulfur (Fe‐S) cluster is a cofactor for all Fe‐S proteins, which play important roles in regulating energy metabolism, DNA repair, lipid biosynthesis, and iron homeostasis, etc.^[^ 7 ^]^ Deficiencies in the Fe‐S cluster can lead to various metabolic diseases.^[^ 8 ^]^ Cysteine desulfurase (NFS1) is essential for the initial assembly of Fe‐S cluster. Abnormal level of NFS1 often disrupts the assembly of Fe‐S cluster, leading to mitochondrial respiratory defects (due to a failures in complexes I, II, and III) and severe metabolic dysfunction.^[^ 9, 10 ^]^ The biosynthesis of Fe‐S cluster is closely related to persulfides, including hydrogen sulfide (H_2_S). H_2_S was traditionally known as a toxic gas with an unpleasant rotten egg odor.^[^ 11 ^]^ However, recent research has unveiled that H_2_S is an important cellular signaling molecule with important roles in both physiological and pathological processes.^[^ 12 ^]^ As reported, H_2_S donor AP39 attenuated ischaemia‐reperfusion (IR) injury by reducing harmful oxidative molecules and blocking the opening of mitochondrial transition pore through AMPK/UCP2 pathway.^[^ 13 ^]^ We previously found that H_2_S maintains the stability of Fe‐S cluster in mitochondria through S‐sulfhydration of NFS1, thereby protecting myocardium from ischemic injury.^[^ 14 ^]^ In addition, the improvement of energy metabolism by H_2_S should also not be overlooked. A recent study showed that H_2_S blocks Angiotensin II (Ang II)‐induced Warburg effect and endoplasmic reticulum stress, thus inhibiting the progression of atrial fibrosis.^[^ 15 ^]^ Another study revealed that H_2_S improves energy metabolism in cardiac hypertrophy by regulating SIRT3 expression.^[^ 16 ^]^ Given its importance in cellular metabolism, the roles of NFS1 and H_2_S in vascular smooth muscle cell (VSMC) biology and AAA progression have yet to be fully investigated.

Isocitrate Dehydrogenase (NADP^+^) 2 (IDH2), a type of isocitrate dehydrogenase, is subcellularly localized in mitochondria and catalyzes the conversion of isocitrate to α‐ketoglutarate (α‐KG) together with IDH3.^[^ 17 ^]^ In recent years, research on IDH2 has focused on tumor therapy, and a variety of inhibitors targeting IDH2 have been developed and shown to effectively treat cancers in several clinical studies.^[^ 18, 19, 20 ^]^ However, the molecular mechanism of IDH2 involvement in cardiovascular disease is not yet fully understood. SP2 is a member of the family of specificity proteins, a class of transcription factors with a zinc‐finger structure that regulate transcription of target genes by binding to GC‐rich elements in promoter regions.^[^ 21 ^]^ It has been demonstrated that SP2 drives interstitial cell osteogenic differentiation in bicuspid aortic valves by binding to miR‐195‐5p.^[^ 22 ^]^ The interplay between glycolytic enzyme and the transcriptional activity of SP2 in AAA remains to be elucidated.

By using both human thoracic AA (TAA) samples and mouse model of AAA, this current study investigated the involvement of NFS1 in the pathogenesis of AAA and the underlying mechanisms. It was found that NFS1 regulates the expression of IDH2 through direct binding to SP2, thereby maintaining normal mitochondrial functions in VSMCs and preventing the formation of AAA.

Results

2

NFS1 is Downregulated in Human and Murine Aortic Aneurysm Tissues

2.1

Research findings revealed a striking pathological resemblance between TAA and AAA, further uncovering that the differentially expressed genes identified in TAA displayed concordant expression changes in the AAA mouse model.^[^ 23, 24 ^]^ In this study, we selected human TAA and corresponding adjacent non‐diseased tissues as research subjects to investigate alterations in the expression levels of functionally distinct proteins.^[^ 25, 26 ^]^ By comparing NFS1 expression in human proximal ascending aortic aneurysm lesions with that in the adjacent non‐lesion aorta, we found a striking reduction in NFS1 abundance in the TAA lesions (Figures 1B and S1D, Supporting Information; patients’ information as shown in Table S1, Supporting Information). Given that SMC phenotypic transformation is a key characteristic of aneurysms, we next investigated the changes in protein levels of α‐SMA and OPN. As shown in Figures 1A and S1A,C, Supporting Information, α‐SMA was downregulated in lesions area while the OPN was upregulated. CSE, the key enzyme involved in endogenous H_2_S production in vascular tissues,^[^ 27 ^]^ was also downregulated in the diseased aortic tissue (Figures 1A and S1B, Supporting Information). To validate the findings obtained from human TAA samples, we established a BAPN‐induced mouse TAA model (Figure S1E, Supporting Information). In this model, aggravated disruption of aortic elastic fibers was observed (Figure S1F,G, Supporting Information). Furthermore, immunofluorescence analysis of mouse aortic tissues revealed decreased expression of NFS1 and α‐SMA in TAA, whereas OPN expression was elevated (Figure S1H,I, Supporting Information).

NFS1 is downregulated in human and murine aortic aneurysm tissues. A) α‐SMA and CSE proteins expressions in human aortic aneurysm lesion areas and non‐lesion area were assessed by Western blotting (n = 6). B) OPN and NFS1 proteins expression in human aortic aneurysm lesion areas and non‐lesion area were assessed by Western blotting (n = 6). C) PCA analysis of human AAA bulk RNA‐seq data. D) Sample_distance_heatmap of human AAA bulk RNA‐seq data. E) Volcano plot of differentially expressed genes in human AAA bulk RNA‐seq data. F) Mouse AAA scRNA‐seq clustering results plot. Apoe −/− mice were infused with Ang II (1000 ng kg−1 min−1) subcutaneously for 4 weeks to induce AAA. A dose of 14 µmol kg−1 sodium hydrosulfide (NaHS) were injected via the abdomen three times a week for 2 weeks before Ang II infusion. G) Volcano plot of differentially expressed genes in VSMC of mouse AAA scRNA‐seq data. H) α‐SMA and CSE proteins in mouse AAA were assessed by Western blotting (n = 6). I) OPN and NFS1 proteins expressions in mouse AAA were assessed by Western blotting (n = 6). Data were analyzed by Welch ANOVA followed by the Dunnetts T3 post hoc test (α‐SMA, OPN, NFS1) and Kruskal‐Wallis ANOVA followed by the Dunns post hoc test (CSE).

To validate whether NFS1 is similarly downregulated in AAA, we first analyzed the human AAA bulk RNA‐seq dataset obtained from the GEO database. Quality control analysis indicated good dispersion among samples derived from 7 AAA and 7 control groups (Figure 1C,D, and Figure S2A, Supporting Information). Differential gene expression analysis revealed a total of 1150 downregulated genes in AAA, including NFS1 and ACTA2, while 2743 genes were upregulated, among which SSP1 (OPN) was identified (Figure 1E). To further verify these findings, we analyzed the mouse AAA single‐cell RNA‐seq dataset from the GEO database. The results revealed a total of 17 distinct cell clusters (Figure 1F), among which three clusters were identified as SMCs. These included cluster 0, representing contractile SMCs (marked by Acta2 and Tagln), cluster 6, representing proliferating SMCs (marked by Mgp and Tpm4), and cluster 12, representing stressed SMCs (marked by Fos and Atf3) (Figure 1F, Figure S3A–C, Supporting Information).^[^ 6, 28 ^]^ Subsequently, all SMCs were extracted for differential gene expression analysis. Among the 10 497 genes analyzed, 1446 genes were found to be downregulated, with Nfs1 exhibiting significant downregulation (Figure 1G). Subsequently, we established Ang II‐induced mouse AAA models, in which the protein expression patterns were highly consistent with those observed in preceding analyses (Figure 1H,I, Figure S4A–D, Supporting Information). Overall, our data indicated that NFS1 expression is reduced in aortic aneurysmal lesions.

NaHS Effectively Attenuates Ang II‐Induced AAA Formation

2.2

For the induction of an AAA animal model, Apoe ^−/−^ mice were administered with Ang II for 28 days. Compared with the Ang II group, the Ang II + NaHS group exhibited significantly ameliorated AAA formation, as observed by both micro‐ultrasound and macroscopic examination (Figure 2A,B). Consistently, the maximal diameter of the suprarenal abdominal aorta was significantly reduced in the Ang II + NaHS group relative to the Ang II group (Figure 2C). Furthermore, we monitored blood pressure changes across different groups. The results demonstrated that NaHS supplementation significantly attenuated Ang II‐induced hypertension (Figure S5A, Supporting Information). Histological analyses, including H&E, EVG, and Masson staining, revealed that NaHS injection mitigated elastic fiber fracture and collagen deposition in the aortic tunica media induced by Ang II (Figure 2D–F). In addition, Ang II administration reduced H_2_S levels in both aortic tissues and serum, a trend that was attenuated upon NaHS supplementation (Figure S5B,C, Supporting Information). Analyses of adhesion molecules in aortic tissues (Figure S5D–F, Supporting Information) and inflammatory cytokines in serum (Figure S5G–I, Supporting Information) further indicated that Ang II markedly induced their expression and secretion, whereas NaHS supplementation significantly suppressed these responses.

NaHS effectively attenuates Ang II‐induced AAA formation. Apoe −/− mice were infused with Ang II (1000 ng kg−1 min−1) subcutaneously for 4 weeks to induce AAA. A dose of 14 µmol kg−1 NaHS were injected via the abdomen three times a week for 2 weeks before Ang II infusion. A) Representative images of the macroscopic features of AAA formation in the indicated groups. Scale bar, 2 mm. B) Representative images of abdominal aortas visualized by MUI using B mode and M mode in the indicated groups. C) Infrarenal aortas maximal diameter measured by MUI B mode in the different groups (n = 6). Data were analyzed by Welch ANOVA followed by the Dunnetts T3 post hoc test. D) Representative HE, EVG, Masson staining of abdominal aortas in the indicated groups. Scale bar, 100 µm. E) Semi‐quantification of the grade of elastin degradation and F) middle membrane collagen deposition in the abdominal aortas (n = 7). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test (elastin degradation), Kruskal‐Wallis ANOVA followed by the Dunns post hoc test (collagen deposition).

NaHS Rescues SMC Phenotype Switching by Targeting at NFS1

2.3

To further explore the mechanisms by which NFS1 affects SMC phenotypic transformation in AAA, we analyzed the effects of NaHS on α‐SMA and OPN protein expressions in Ang II‐treated MOVAS. It was found that Ang II significantly suppressed smooth muscle contractile protein α‐SMA but induced the proliferation marker protein OPN, both changes were attenuated upon NaHS supplementation (Figure 3A–D). Given that VSMC phenotypic transformation in AAA has been linked to cell cycle regulation,^[^ 29 ^]^ we next examined Cyclin D1 and PCNA, two key cell cycle–promoting proteins. Both were significantly elevated following Ang II treatment but suppressed with NaHS co‐incubation (Figure 3E–G). These results were further validated by EdU fluorescence staining in MOVAS (Figure 3H, Figure S6A,B, Supporting Information).

NaHS rescues SMC phenotype switching by targeting at NFS1. MOVAS ere stimulated with 3 µM Ang II for 24 h. NaHS (30 µM) was added 30 min prior to Ang II treatment where applicable. A–D) α‐SMA and OPN proteins expression and quantification in MOVAS (n = 6). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test (α‐SMA), Kruskal‐Wallis ANOVA followed by the Dunn`s post hoc test (OPN). E–G) Cyclin D1 and PCNA proteins expression and quantification in MOVAS (n = 6). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test. H) Relative EdU levels of MOVAS in different treated groups (n = 6). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test. I,J) NFS1 protein expression and quantification in MOVAS (n = 6). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test. K,L) Differential expression of α‐SMA and OPN proteins after NFS1 was knockdown. M,N) Differential expression of α – SMA and OPN proteins after NFS1 was overexpressed.

Next, we assessed NFS1 protein expression across different treatment groups. Consistent with the findings in mouse and human aortic aneurysm tissues, NFS1 was significantly downregulated in Ang II‐treated MOVAS but upregulated by NaHS supplementation (Figure 3I,J). Immunofluorescence staining performed on different treatment groups also indicated a decrease in intracellular expression of NFS1 and α‐SMA following Ang II stimulation (Figure S6C, Supporting Information). Protein S‐sulfhydration is one of the key pathways involved in the biological effects of H_2_S.^[^ 30 ^]^ We subsequently assessed the S‐sulfhydration levels of intracellular NFS1, and found that Ang II stimulation resulted in a reduction in NFS1 S‐sulfhydration, which was restored to normal levels upon NaHS supplementation (Figure S6D,E, Supporting Information). These findings suggest that NaHS prevents AAA by concurrently stimulating NSF1 at both translational and post‐translational levels.

To further investigate the involvement of NFS1 in the phenotypic switching of SMCs, NFS1 was either knocked down or overexpressed in MOVAS, respectively (Figures S6F,G,J,K, Supporting Information). After transfection with si‐NFS1 plasmids, MOVAS shifted from a contractile phenotype to a secretory phenotype (Figure 3K,L, Figure S6H, I, Supporting Information). In contrast, NFS1 overexpression in MOVAS significantly increased α‐SMA and decreased OPN following Ang II treatment, indicating a critical role of NFS1 in maintaining the contractile phenotype of VSMCs (Figure 3M,N, Figure S6L, M, Supporting Information). In subsequent studies, we investigated the role of NFS1 in mice. After constructing AAV9 vectors encoding si‐NFS1 for injection into the mice, we observed that the NFS1 knockdown group showed more severe lesions compared to the si‐NC group (Figure S7A–C, Supporting Information). EVG and masson staining results also revealed that NFS1 knockdown exacerbated elastic fiber disruption (Figure S7D,E, Supporting Information).

IDH2 Acts as a Target of NFS1

2.4

To identify the downstream signals of NFS1 in AAA, we performed CUT&Tag analysis on mouse AAA tissues using an anti‐NFS1 antibody. The data revealed that NFS1 was preferentially enriched at promoter and upstream regions of multiple genes (Figures 4A,B, and S8A, Supporting Information). To further elucidate the biological functions of NFS1 binding, GO enrichment analysis was conducted on genes located near the identified Peaks. We selected the top 20 most significant enriched biological process terms as shown in Figure 4C, among which 16 were associated with metabolism. Notably, IDH2—a pivotal TCA cycle enzyme that catalyzes the oxidative decarboxylation of isocitrate to α‐ketoglutarate, was identified as a result of peak calling.^[^ 31 ^]^ These findings suggest that NFS1 may act as a transcription factor for regulating cellular metabolism by targeting at Idh2 (Figure 4D).

IDH2 acts as a target of NFS1. A,B) CUT & Tag was performed with NFS1 antibody in aortic tissues. NFS1 was enriched in the promoter region of numerous genes. These genes with similar distribution patterns were clustered together by a clustering algorithm to show the binding trend of NFS1 as a transcription co‐factor. (n = 3). C) Significant terms of biological process in gene ontology (GO) analyses were carried out in NFS1‐related genes. D) NFS1 peaks in the promoter regions of Idh2. MOVAS were stimulated with 3 µM Ang II for 24 h. NaHS (30 µM) was added 30 min prior to Ang II treatment where applicable. E,G) Western blotting assays and quantitative analysis of IDH2 in MOVAS (n = 6). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test. Apoe −/− mice were infused with Ang II (1000 ng kg−1 min−1) subcutaneously for 4 weeks to induce AAA. A dose of 14 µmol kg−1 NaHS were injected via the abdomen three times a week for 2 weeks before Ang II infusion. F,H) Western blotting and quantitative analysis of IDH2 in mouse AAA tissues (n = 6). Data were analyzed by Welch ANOVA followed by the Dunnett`s T3 post hoc test. I,J) Western blotting assays and quantitative analysis of IDH2 in human aortic aneurysm tissues (n = 6).

To further explore the possibility that IDH2 may directly contribute to AAA, we examined the changes in IDH2 expression in mouse and human aortic aneurysm tissues. A significant decrease in IDH2 expression was observed in AAA tissues from Ang II‐treated mice, which was returned to normal level following NaHS treatment (Figure 4E–H). IDH2 expression was also lower in the lesion area from human aortic aneurysm tissues but remained high in the surrounding normal area (Figure 4I,J). Interestingly, NFS1 knockdown also reduced IDH2 expression in abdominal aortic tissues, implying that IDH2 may be regulated by NFS1 and involved in AAA progression (Figure S8B,C, Supporting Information).

NFS1 Regulates Glycolytic Pathway in AAA Progression

2.5

Alterations in glycolytic activity within immune cells are known to play a pivotal role in AAA progression.^[^ 32 ^]^ Given that IDH2 is integral to the TCA cycle, we hypothesized that the changes in carbohydrate metabolism in VSMCs might contribute to AAA progression. To explore this hypothesis, we first examined the expressions of key glycolytic enzymes.^[^ 33 ^]^ We found that most glycolytic enzymes, including PFKP, HK2, ENO1, PGAM1, LDHA, and ALDOA, were upregulated following Ang II treatment and returned to levels similar to the control group after NaHS treatment (Figure 5A–G). While only PKM2 was downregulated in the Ang II group but increased following NaHS administration (Figure 5A,H). No significant changes were observed in LDHB and PGK1 with or without NaHS treatment (Figure S9A‐‐C, Supporting Information).

NFS1 regulates glycolytic pathway in AAA progression. MOVAS were stimulated with 3 µM Ang II for 24 h. NaHS (30 µM) was added 30 min prior to Ang II treatment where applicable. A) PFKP, HK2, ENO1, PGAM1, LDHA, ALDOA and PKM2 protein expressions assessed by Western blotting in MOVAS. B–H) Quantification of protein expressions measured by Western blotting in MOVAS (n = 6). All data was analyzed by one‐way ANOVA followed by the Tukey post hoc test. I) Differences in extracellular lactate contents between various treatments (n = 6). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test (extracellular lactate). J) Mitochondrial superoxide content in different groups measured by MitoSOXs staining, the results were showed as fluorescence images. Scale bar, 50 µm. K) Representative fluorescence images of mitochondrial permeability transition pore evaluated by mPTP staining. Scale bar, 50 µm. L) JC‐1 staining tested by microplate reader with different treatments (n = 6). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test.

Since the expression of most of the enzymes in the glycolytic pathway was upregulated in the AngII group, we hypothesized that lactate accumulation might occur in VSMCs under these conditions. To validate our conjecture, we next accessed the extracellular and intracellular lactate content. Lactate levels were elevated after Ang II treatment but were attenuated following NaHS introduction (Figure 5I and Figure S9D, Supporting Information). Given that alterations in glucose metabolism are known to impair mitochondrial function and thereby drive the pathogenesis of multiple diseases.^[^ 34, 35 ^]^ Thus, we next explored mitochondrial function by conducting a mitochondrial membrane potential (JC‐1) assay. The results showed a decrease in normally polarized mitochondria (J‐aggregates, red) and an increase in abnormally depolarized mitochondria (J‐monomers, green) in the Ang II group, while NaHS co‐incubation attenuated this membrane potential decrease (Figure 5L and Figure S10A, Supporting Information).

A decrease in membrane potential is typically associated with ROS accumulation and excessive mitochondrial permeability transition pore (mPTP) opening.^[^ 36, 37 ^]^ Consistently, Ang II stimulation increased both mitochondrial ROS levels (as detected by MitoSOX) (Figure 5J, Figure S10B,C, Supporting Information) and mPTP opening (Figure 5K, Figure S10D, E, Supporting Information), both of which were markedly attenuated by NaHS treatment.

To further investigate whether NFS1 influences mitochondrial function, NFS1 was either knocked down or overexpressed in MOVAS. In NFS1 knockdown cells, we observed higher MitoSOXs levels and mPTP opening compared to the si‐NC group (Figure S11A–F, Supporting Information). JC‐1 staining also revealed a significant decrease in membrane potential following NFS1 knockdown (Figure S11G,H, Supporting Information). In contrast, NFS1 overexpression reduced MitoSOXs accumulation (Figure S12A–C, Supporting Information), decreased mPTP opening (Figure S12D–F, Supporting Information) and normalized JC‐1 staining (Figure S12G,H, Supporting Information), in comparison to the Ang II‐treated group alone.

Inhibition of IDH2 Attenuates the Protective Effect of NFS1

2.6

Enasidenib is a reversible inhibitor of IDH2.^[^ 38, 39, 40 ^]^ To validate that IDH2 acts as a downstream effector of NFS1, we administered enasidenib both in vitro and in vivo to observe its effects on AAA. In MOVAS, enasidenib treatment induced a phenotype switch from the contractile to the secretory state, in contrast to the phenotype observed with NFS1 overexpression (Figures 6A and S13A,B, Supporting Information). Subsequently, both pharmacological and biological inhibition of IDH2 led to an increase in the extracellular acidification rate (ECAR) in MOVAS, indicating a shift in the cellular energy source from oxidative phosphorylation (OXPHOS) to glycolysis (Figures 6B and S13C–E, Supporting Information). Similarly, in the cellular mitochondrial stress test, the results with oxygen consumption rate (OCR) showed that biological and pharmacological inhibition against IDH2 resulted in a significant reduction in mitochondrial respiration compared with other groups. (Figures 6C and S13F–H, Supporting Information).

Inhibition of IDH2 attenuates the protective effect of NFS1. MOVAS were stimulated with 3 µM Ang II for 24 h. NaHS (30 µM) was added 30 min prior to Ang II treatment where applicable. For transfection, 1.5 µg of si‐IDH2 was added to each well of a 6‐well plate. A) Differential expression of α‐SMA and OPN proteins by NFS1 overexpression and enasidenib treatment. B) The ECAR of glycolytic rate assay in different groups (n = 4). C) The OCR of cell mito stress test in different groups (n = 4). D) Representative images of the macroscopic features of AAA formation in indicated groups. Scale bar, 2 mm. AAV9 vectors was delivered via tail vein injection at a dose of 5 × 1012 genomic copies (GC) 2 weeks prior to modeling. For enasidenib treatment, mice were received intraperitoneal injections (15 mg kg−1 day−1) for the first 3 days of each week over 2 weeks prior to modeling. E) Representative HE, EVG, Masson staining of abdominal aortas in indicated groups. Scale bar, 200 µm. F) Representative images of abdominal aortas visualized by MUI using B mode and color mode in the indicated groups. G,H) Differential expression of α – SMA and OPN proteins of murine measured by Western blot after injected with AAV‐NFS1 together with enasidenib treatment (n = 6).

In vivo studies further revealed that NFS1 overexpression attenuated AngII‐induced AAA in mice, whereas enasidenib treatment caused aneurysms in these mice to be deteriorated. It was observed that application of enasidenib aggravated the pathological changes of AAA in mice. (Figure 6D–F, Figure S14A–C, Supporting Information). Western blotting analysis of aortic tissues from AAA model mice also demonstrated that enasidenib abolished the protective role of NFS1 overexpression against the phenotypic transformation of VSMCs in the AAA tissues (Figure 6G,H, FigureS14D–G, Supporting Information). To confirm the role of IDH2 in AAA progression, we first constructed si‐IDH2 for IDH2 knockdown (Figure S15A,B, Supporting Information). Subsequent results confirmed that IDH2 knockdown induced phenotypic switching in SMCs (Figure S15C–E, Supporting Information) and impaired mitochondrial function (Figures S15F–I, Supporting Information). We then generated an AAV9 virus for IDH2 knockdown (Figures S16B–D, Supporting Information), injected it into mice, and induced the AAA model. The results indicated that IDH2 deficiency significantly exacerbated AAA lesions (Figure S16A,E,F, Supporting Information).

NFS1 Works as a Co‐Factor of SP2 for Regulating Idh2 Transcription

2.7

To investigate the mechanisms how NFS1 regulates Idh2, we performed motif analysis on the Peaks obtained from the CUT & Tag assay. Four transcription factors were identified for potentially regulating IDH2 transcription, including SP1, SP2, SP5, and KLF14 (Figure 7A). To validate these findings, we transfected MOVAS with plasmid overexpressing NFS1‐Flag followed by co‐immunoprecipitation (Co‐IP) using an anti‐Flag antibody. The results showed that only SP2 was able to bind to NFS1 (Figure 7B). Immunofluorescence co‐staining under confocal microscopy further confirmed the direct interaction between SP2 and NFS1 outside the nucleus (Figure 7C). In subsequent PLA experiments, we directly observed that the interaction between SP2 and NFS1 occurred not only in the perinucleus but also within the nucleus (Figure S17A, Supporting Information).

NFS1 works as a co‐factor of SP2 for regulating Idh2 transcription. MOVAS were stimulated with 3 µM Ang II for 24 h. NaHS (30 µM) was added 30 min prior to Ang II treatment where applicable. A) Motif analysis of peaks sequences after pull‐down with NFS1 antibody in CUT & Tag. B) Co‐IP assay for analyzing the transcription factors binding to NFS1 in MOVAS. C) Immunofluorescence co‐staining of SP2 and NFS1 to observe the co‐localization in MOVAS. Left figures scale bar, 50 µm. Right figures scale bar, 5 µm. D,E) SP2 protein expression assessed by Western blotting in MOVAS and mouse aortic tissue, respectively (n = 6). F,G) Quantification of SP2 protein expression measured by Western blotting in MOVAS and mouse aortic tissue, respectively (n = 6). Data were analyzed by one‐way ANOVA followed by the Tukey post hoc test. H) SP2 protein expression assessed by Western blot in human aortic aneurysm tissues (n = 6). I) Quantification of SP2 protein expression measured by Western blot in aortic tissue (n = 6). Data were analyzed by Student t test. J,K) Construction of full‐length and truncated reporter plasmids for analyzing the binding region of SP2 in IDH2 promoter (n = 6).

Western blotting analysis revealed that SP2 expression was significantly decreased in both in vitro and in vivo Ang II‐induced models, but returned to control levels with NaHS treatment (Figure 7D–G). In human aortic aneurism tissues, SP2 expression was reduced in the lesion areas, while remaining high in the surrounding normal tissue (Figure 7H,I). To explore whether SP2 would bind to Idh2 promoter and determine the specific binding region, we constructed full‐length (−2000 to 0 bp) and truncated (−600 to 0 bp) luciferase reporter plasmids containing Idh2 promoter (Figure 7J).^[^ 41 ^]^ SP2 overexpression significantly induced the luciferase activity of the full‐length plasmid but had no effect in the truncated plasmid, suggesting that SP2 may bind to the −2000 to −600 bp region upstream of the Idh2 promoter for promoting Idh2 transcription (Figure 7K). Subsequently, we performed ChIP‐PCR using an anti‐NFS1 antibody. The results revealed that Idh2 transcription was reduced in SMCs following SP2 knockdown, further validating our hypothesis that NFS1 and SP2 collaboratively regulate Idh2 (Figure S17B, Supporting Information).

NFS1 Directly Binds to SP2

2.8

To investigate the interaction between NFS1 and SP2, we first performed molecular docking to predict the binding sites between NFS1 and SP2. The results, as shown in Table S2, Figures 8A,B, and S18A, Supporting Information, indicated that SP2 binds within the 1–300 amino acid region of NFS1. To further examine these predictions, we constructed full‐length and truncated NFS1 constructs together with a full‐length SP2 plasmid (Figure S18B, Supporting Information). The Co‐IP results showed that deletion of the NFS1 amino acid fragment 301–459 did not affect its interaction with SP2, whereas deletion of either residues 18–200 or 201–300 abolished the interaction with SP2 (Figure 8C).

NFS1 directly binds to SP2. Apoe −/− mice were infused with Ang II (1000 ng kg−1 min−1) subcutaneously for 4 weeks to induce AAA. A dose of 14 µmol kg−1 NaHS were injected via the abdomen three times a week for 2 weeks before Ang II infusion. AAV9 was delivered via tail vein injection at a dose of 5 × 1012 genomic copies (GC) 2 weeks prior to modeling. A,B) Modeling the protein spatial structure of NFS1 and SP2 for predicting their binding sites. C) Co‐transfection with different NFS1 plasmids and full‐length SP2 plasmid into 293T for pull‐down experiments with anti–HA antibody. D) Different AAV9 vectors containing the full‐length or deleted NFS1 were constructed and injected into mice via tail vein, and the expressions of α‐SMA and OPN in AAA tissues were verified by Western blotting (n = 6). E) Representative HE, EVG, Masson staining of abdominal aortas in the indicated groups. Scale bar, 200 µm. F) Representative images of abdominal aortas visualized by MUI using B mode and color mode in the indicated groups.

We next constructed AAV9 vectors encoding NFS1 deletion mutants for in vivo studies. Western blotting analysis of the aortic tissues revealed that mice with NFS1 del (18–200) or NFS1 del (201–300) exhibited more pronounced phenotypic changes of VSMCs, including reduced expression of α‐SMA and increased expression of OPN (Figure 8D, Figure S18C–E, Supporting Information). Immunohistochemical analysis showed similar degrees of elastin fiber fragmentation in the aneurysms of NFS1 del (18–200) or NFS1 del (201–300) mice compared to Ang II‐treated mice (Figure 8E, Figure S18F, Supporting Information). Ultrasound results confirmed that NFS1 del (18–200) or NFS1 del (201–300) overexpression was unable to attenuate aortic aneurysm formation induced by Ang II infusion, in contrast to ad‐NFS1 wt transfected mice (Figure 8F, Figure S18G, Supporting Information). Collectively, these findings suggest that the 18–300 amino acid region of NFS1 is critical for its interaction with SP2, and that deletion of this domain abolishes the protective effects of NFS1 on AAA progression and VSMC phenotypic modulation, likely by impairing its ability to regulate Idh2 transcription through SP2 binding.

Discussions

3

The development of AAA is a chronic and multistep pathological process driven by multiple factors, characterized by persistence inflammation, progressive loss of VSMCs, and disruption of ECM.^[^ 42 ^]^ The loss of structural integrity in the aorta leads to vessel weakening, which results in aortic dilation and ultimately aortic rupture. Unfortunately, despite the high mortality rate of ruptured AAA, current clinical interventions are limited to endovascular or surgical repair, with no effective drug therapies available.^[^ 43, 44 ^]^ This underscores the urgent need to elucidate the underlying mechanisms of AAA pathogenesis and to identify novel therapeutic targets.

In this study, we identified NFS1 as a crucial regulator of glycolysis and mitochondrial functional homeostasis with potential therapeutic relevance for AAA. First, NFS1 was found to be adaptively downregulated in aortic aneurysm lesions in both human and murine models. Functionally, overexpression of NFS1 in VSMCs mitigated vascular remodeling, whereas its deficiency aggravated aneurysm progression. Mechanistically, NFS1 functions as a transcriptional cofactor in metabolic regulation, directly targeting the rate‐limiting TCA cycle enzyme IDH2 to sustain mitochondrial function. Moreover, we demonstrated that NFS1 regulates Idh2 transcription and protects against AAA by interacting with SP2.

As a rate‐limiting enzyme in iron‐sulfur cluster synthesis, abnormal NFS1 expressions often cause various diseases.^[^ 9 ^]^ It has been shown that the NFS1‐OPA1 axis attenuates doxorubicin‐induced myocardial toxicity by inhibiting ferroptosis, thereby reducing cardiovascular damage associated with cancer treatment.^[^ 45 ^]^ Another study found that knockdown of NFS1 increased the sensitivity of colorectal cancer cells to Oxaliplatin, promoting apoptosis and ferroptosis through ROS accumulation.^[^ 46 ^]^ In this present study, we observed that NFS1 expression was significantly reduced in aortic aneurysm tissues, and restoration of NFS1 protected from AAA progression. Interestingly, changes in NFS1 expression were correlated with the alterations in mitochondrial functions in SMCs. Through NFS1 knockdown and overexpression experiments, we demonstrated the essential roles of NFS1 in precisely regulating SMC phenotypic transformation and blood vessel adaptability via modulation of mitochondrial functions.^[^ 47 ^]^

Mitochondrial functions are inextricably linked to energy metabolic state.^[^ 48 ^]^ Glucose metabolism, including glycolysis and OXPHOS, plays a critical role in mitochondrial activity.^[^ 49 ^]^ Previous studies have shown that PHB2 deficiency causes a shift from PKM1 to PKM2, enhancing glycolysis in VSMCs and leading to their phenotypic transformation.^[^ 50 ^]^ Another study suggested that inhibiting macrophage glycolysis increases NAD^+^ levels, ensuring the continuation of OXPHOS and ultimately reducing intravascular thrombosis.^[^ 51 ^]^ Our study observed that the glycolytic pathway in VSMCs was activated in AAA, while OXPHOS was suppressed. The TCA cycle is a key component of OXPHOS, and we demonstrated that IDH2, a rate‐limiting enzyme in the TCA cycle, was inhibited in AAA. In addition, Idh2 was identified as a downstream gene regulated by NFS1, co‐contributing to the phenotypic transformation in AAA. Pharmacological inhibition of IDH2 partially attenuated the protective effects of NFS1 overexpression on AAA and exacerbated disease progression. Consistent with our findings, sirtuin3‐mediated deacetylation of IDH2 has been shown to attenuate myocardial ischemia‐reperfusion (I/R) injury by protecting mitochondrial functions.^[^ 52 ^]^ Additionally, IDH2 deficiency has been reported to impair the NADPH‐glutathione antioxidant system and activate the Renin‐Angiotensin System, ultimately leading to hypertension, which is a major risk factor for AAA.^[^ 53 ^]^

To further elucidate the molecular mechanism by which NFS1 regulates downstream IDH2, we analyzed the CUT & Tag data on NFS1, and identified SP1, SP2, SP5, and KLF14 as transcription factors involved in Idh2 regulation. Using co‐IP assay, we demonstrated that only SP2 forms a complex with NFS1.^[^ 21 ^]^ Previous studies have shown that SP2 knockdown reduced its binding to the promoter region of miR‐195‐5P, leading to increased expression of SMAD7 and contributing to aortic valve calcification.^[^ 22 ^]^ Additionally, SP2 enhances the genomic load of the PBX1 complex, which binds to promoter regions of highly expressed genes.^[^ 54 ^]^ Interestingly, SP2 has been found to interact with NFYa to activate genes involved in metabolism and proliferation, highlighting its role in mitochondrial metabolism during cardiac development and regeneration.^[^ 55 ^]^ Moreover, this study discovered that NFS1 acts as a transcriptional cofactor, synergizing with SP2 to regulate Idh2 transcription for protecting AAA. Through multiple approaches, we demonstrated that NFS1 binds to SP2 at the 18–200 and 201–300 amino acid sequences. These findings provide further evidence for the involvement of the NFS1‐SP2 complex in the regulation of Idh2.

However, this study has several limitations. First, while we based our research on tissue collection from patients with dilated thoracic aorta, validation in AAA patient tissues is still necessary, despite the reliability of the animal experiments. Second, although our initial CUT&Tag analysis lacked sufficient biological replicates and was exploratory in nature, the predicted regulatory relationship was subsequently validated by multiple complementary experiments, ensuring the robustness of our conclusions. Third, the discovery of decreased NFS1 expression in AAA mice suggests that NFS1 may serve as an effective therapeutic target for AAA, which warrants further confirmation in human AAA patient cohorts. Last, the lack of an effective agonist for the IDH2 target in this study highlights an area for improvement in future experiments by directly using Idh2 knockout mice.

Conclusion

4

In conclusion, our study has identified a previously unrecognized role for the NFS1‐SP2‐IDH2 axis, which is crucial for maintaining the contractile phenotype of VSMCs and, consequently, mitigating AAA progression. Therefore, NFS1 may serve as a novel therapeutic and prognostic marker for AAA.

Experimental Section

5

Human Thoracic Aortic Aneurysm Samples

Human TAA samples were obtained from the Department of Cardiovascular Surgery, Renji Hospital, Shanghai Jiao Tong University School of Medicine with the approval of the Institutional Ethical Review Board (Document No. KY2024‐18‐13). The patients and families of the donors provided informed consent with approval by the Ethics Committee; and the study was conducted according to the criteria set by the Declaration of Helsinki (2013). The medical information of patients is detailed in Table S1, Supporting Information.

Single‐Cell RNA Sequence

Using the Aspera, this work retrieved the raw single‐cell RNA sequencing datasets of mouse AAA diseased and normal tissues from the GEO database (GSE191226, GSE193265, GSE239620). After obtaining the raw sequencing data, this work converted the FASTQ files into.h5 format using the Cell Ranger package in Python. Quality control for individual sample.h5 files was performed using the Scanpy and rpy2 packages. Following quality control, the sample data were integrated and clustered using the Scvi algorithm within the Scanpy package. For the annotation of cell populations, highly variable genes from different clusters were manually annotated using the CellMarker 2.0 website. Specifically, all SMC clusters identified in the clustering results were selected for downstream differential gene expression analysis. The resulting differential gene expression data were imported into R, and volcano plots were generated for visualization using the ggplot2 package.

Bulk RNA‐Seq Analysis

Publicly available bulk RNA‐seq data were obtained from GEO under accession GSE183464, including 7 AAA samples and 7 control samples. Differential expression analysis was performed in R version 4.2.0. Low‐count genes were filtered, normalization and dispersion estimation were carried out with DESeq2, and log2 fold changes were shrunk with apeglm. Significance was defined as FDR less than 0.05 and absolute log2 fold change greater than one. Plots were generated for visualization using the ggplot2 package.

Animal Experiment and Treatment

All animal care and experimental protocols were approved by the Ethics Committee on Animal Experiments of Shanghai Youshu Life and Technology Company and complied with the Animal Management Rules of the Chinese Ministry of Health (Document No. YS – m202404001). All animal experiments were in accordance with the Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85‐23, revised 1996).

All male Apoe ^−/−^ mice (6–8 weeks old, 18–22 g) were purchased from Cyagen Biosciences (Guangzhou, China). All mice were housed in standard cages in a specific pathogen‐free environment and kept on a 12‐h light/12‐h dark cycle with free access to food and water. In the in vivo study, the Apoe ^−/−^ mice were randomly divided as follows: (1) saline; (2) Ang II; (3) Ang II + NaHS; (4) NaHS; (5) Ang II + control adeno‐associated virus 9 (AAV9) vector (AAV‐NC); (6) Ang II + AAV9 vector with NFS1 cDNA (AAV‐NFS1); (7) Ang II + AAV‐NFS1 + enasidenib (IDH2i); (8) Ang II + AAV‐NFS1 del (18–200); (9) Ang II + AAV‐NFS1 del (201–300); (10) Ang II + AAV‐NFS1 del (301–459); (11) Ang II + NaHS + AAV9 vector with negative control si‐RNA (AAV‐siNC); (12) Ang II + NaHS + AAV9 vector with NFS1 si‐RNA (AAV‐siNFS1); (13) Ang II + AAV‐NFS1 + AAV‐siNC; (14) Ang II + AAV‐NFS1 + AAV‐siIDH2.

All mice were treated for 4 weeks with continuous subcutaneous infusion of an osmotic pump (Alzet model 2004, Alza Corp., USA) with simultaneous high‐fat feeding at the same time. The saline group received an infusion of saline, whereas the Ang II group received an infusion of 1000 ng kg^−1^ min^−1^ Ang II (A9525, Sigma‐Aldrich, St. Louis, MO) for 4 weeks as reported elsewhere.^[^ 56 ^]^ A dose of 14 µmol kg^−1^ NaHS (Sigma‐Aldrich) were injected via the abdomen three times a week for 2 weeks before Ang II infusion and continued for the entire experiment. Systolic blood pressure (SBP) was measured and recorded by the tail‐cuff method on days 0, 7, 14, 21, and 28 following subcutaneous implantation of Ang II–infusing osmotic minipumps. All in vivo interventions in animal models were performed via tail vein injection of AAV9, which were purchased from Bioegene Co. (Shanghai, China). To overexpress or silence the target genes in mouse aortic SMCs, 5 × 10^12^ genomic copies (GC) per mouse. At the end of second week after the AAV9 infection, mice were implanted with osmotic pump containing Ang II as described previously. For the treatment with enasidenib, after embedding osmotic pump, 15 mg kg^−1^ enasidenib was injected through abdominal, once daily for 3 days, followed by a 4 days interval, the administrations were sustained 2 weeks.^[^ 38 ^]^ After 4 weeks, the osmotic pump was removed, and Ang II infusion was stopped.

In TAA model, 3 weeks‐old male C57BL/6 J mice were fed with a standard diet or the standard diet supplemented with 0.4% BAPN (β‐aminopropionitrile monofumarate; 0.4 g/100 g) (Cat#A3134, Sigma‐Aldrich, St. Louis, MO, USA) for 28 days.

Mice were anaesthetized by isoflurane inhalation (1.5%) and then euthanized by cervical dislocation, and the aortas were collected. Aortic tissues were embedded in 4% paraformaldehyde for histological analysis or snap‐frozen in liquid nitrogen for protein expression studies.

Cell Culture and Treatment

Mouse aorta SMCs (MOVAS) were purchased from Sunncell (Wuhan, China) and cultured in Dulbeccos modified Eagles medium (DMEM)‐F12 (MA0595, Meilunbio, Dalian, China) supplement with 1% penicillin‐streptomycin (10 378 016, Thermo Fisher Scientific, Waltham, MA, USA) and 10% fetal bovine serum (F0193, Sigma‐Aldrich). The in vitro interventions on MOVAS were performed using small interfering RNA (Obio Co. Ltd, Shanghai, China) or adenoviruses carrying target gene cDNA fragments (BioeGene Co. Ltd, Shanghai, China). In the in vitro study, the cells were randomly divided as follows: (1) PBS(Control); (2) Ang II; (3) Ang II + NaHS; (4) NaHS; (5) small interfering negative control (si‐NC); (6) small interfering NFS1 (si‐NFS1); (7) adenoviruses carrying negative control cDNA (ad‐NC); (8) adenoviruses carrying NFS1 cDNA (ad‐NFS1); (9) Ang II + ad‐NFS1; (10) Ang II + ad‐NFS1 + enasidenib (IDH2i); (11) Ang II + NaHS + small interfering IDH2 (si‐IDH2); (12) ad‐NFS1 + si‐NC; (13) ad‐NFS1 + si‐IDH2; (14) ad‐NC + si‐NC; (15) ad‐NFS1 + si‐NC; (16) ad‐NFS1 + small interfering SP2 (si‐SP2).

293T cells were purchased from Sunncell (Wuhan, China) and cultured in DMEM (MA0212, Meilunbio, Dalian, China) supplement with 1% penicillin‐streptomycin and 10% fetal bovine serum. Both cells were used for formal experiments when they reached 70–80% confluence. The in vitro interventions on 293T cells were performed using the pcDNA3.1 (+) plasmid carrying target gene cDNA fragments (BioeGene Co. Ltd, Shanghai, China). To be more precise, full‐length and delete‐length mouse NFS1 cDNA and full‐length SP2 cDNA were synthesized and cloned into the plasmid to generate (1) pcDNA3.1‐m_NFS1 wt‐3 × Flag (NFS1 wt‐Flag); (2) pcDNA3.1‐m_NFS1 del (18–200)‐3 × Flag (NFS1 del(18–200)‐Flag); (3) pcDNA3.1‐m_NFS1 del (201–300)‐3 × Flag (NFS1 del(201–300)‐Flag); (4) pcDNA3.1‐m_NFS1 del (301–459)‐3 × Flag (NFS1 del(301–459)‐Flag); (5) pcDNA3.1‐m_SP2 wt‐3 × HA (SP2 wt‐HA).

In Vivo Multimodal Imaging Analysis

To exhibit the morphological features of abdominal aortas, a Vevo2000 High‐Resolution Micro‐Ultrasound Imaging (MUI) System with a 550‐MHz transducer (VisualSonics Inc., Ontario, Canada) was used. Anesthesia was induced using a 2% isoflurane mixture in the presence of 100% oxygen in an induction chamber. Data were collected from mice lightly anesthetized with 1–2% isoflurane, which maintained a heart rate of >450 beats min^−1^. MUI images were acquired using the B mode in accordance with the manufacturer's instructions. Three measurements of the abdominal aortic diameter on both short and long axis were performed by a blinded investigator to determine the maximal aortic dilation.

Hematoxylin and Eosin, Elastin Van Gieson, and Masson's Trichrome Staining

H&E, EVG, and Masson's trichrome staining were performed according to the standard procedures. First, the slides (5 µm) were deparaffinized in xylene I and II (for 20 min each) at 25 °C. After deparaffinization, xylene was removed using 100% ethanol for 10 min, followed by fixation in 95%, 90%, 80%, and 70% ethanol for 5 min at 25 °C. The slides were then stained with H&E solution (G1005, Servicebio, China), EVG (G1042, Servicebio) and Masson's trichrome (G1006, Servicebio).

S‐Sulfhydration Assay

The biotin switch assay was performed as previously described with minor modifications.^[^ 57 ^]^ Briefly, MOVAS were homogenized in 250 mM HEN buffer containing 250 mM HEPES, 1 mM EDTA, 0.1 mM neocuproine, and 100 µM deferoxamine, followed by centrifugation at 13 000 g for 30 min at 4 °C. A total of 500 µg of protein lysate was then incubated in blocking buffer (HEN buffer supplemented with 2.5% SDS and 20 mM MMTS) at 50 °C for 25 min with frequent vortexing. To remove excess MMTS and precipitate proteins, cold acetone was added and samples were incubated at −20 °C for 1 h. Proteins were pelleted and resuspended in HEN buffer containing 1% SDS, followed by incubation with 4 mM biotin‐HPDP at 25 °C for 3 h. Biotinylated proteins were subsequently captured using streptavidin–agarose beads, washed with HEN buffer, and analyzed by Western blotting.

Measurements of Mitochondrial Membrane Potential

To estimate the changes of mitochondrial membrane potential, MOVAS were stained with JC‐1 Mitochondrial Membrane Potential Assay Kit (C2006, Beyotime, Shanghai, China) in accordance with the manufacturer's instructions. Then, the nuclei were counterstained with DAPI. The images were captured using a TCS SP8 Confocal Microscope (Leica, Wetzlar, Germany). The mean fluorescence intensity was measured using a Synergy H1 Multimode Reader (BioTek, Winooski, Vermont, USA).

Assessment of mPTP Opening

mPTP opening was assessed using the Mitochondrial Permeability Transition Pore Assay Kit (C2009s, Beyotime). After treatment, the cells were incubated with calcein‐AM staining solution and CoCl_2_ (1 ×) at 37 °C in the dark for 35 min. The cells were washed with PBS and cultured in DMEM‐F12 for 30 min at 37 °C. The data were processed using a FACSCelesta Flow Cytometer (BD, San Jose, CA, USA), and TCS SP8 Confocal Microscope (Leica).

Detection of Mitochondrial Reactive Oxygen Species

For the detection of mitochondrial ROS in MOVAS, MitoSOXs Red (M36008, Invitrogen, USA) was used according to the manufacturer's protocol. MOVAS were washed with cold PBS and cultured in a serum‐free medium supplemented with 5 µM MitoSOXs Red for 30 min at 37 °C, protected from light. Following staining, the mean fluorescence intensity and images were recorded using a microplate reader, flow cytometry, and confocal microscopy, respectively.

Cell Proliferation Analysis with 5‐Ethynyl‐2′‐Deoxyuridine Labelling

An EdU detection reagent kit (C00755, Beyotime) was used to determine the cell proliferation status. All the processes were based on the manufacture`s protocol. Briefly, after different exposure for 24 h, MOVAS were treated with 50 mM EdU medium for 2 h. Then, the cells were washed with PBS three times and then stained with Hoechst for nucleus. The mean fluorescence intensity was measured using a Synergy H1 Multimode Reader (BioTek, Winooski, Vermont, USA) and confocal microscopy (Leica TCS SP8, Leica, Germany).

Lactate Measurement

For the detection of lactate, a LD kit (A01921, Nanjingjiancheng, Nanjing, China) was used according to the manufacturer's protocol. The cell culture medium or cell lysates were added to the working solution for incubation at 37 °C for 10 min. The absorbance was read at 530 nm with a Synergy H1 Multimode Reader.

H2S Concentrations

H_2_S levels in plasma and tissue homogenates were measured using a commercial colorimetric assay kit (E‐BC‐K355‐M, Elabscience, Wuhan, China) following the manufacturer's instructions. Absorbance was recorded at 665 nm, and concentrations were calculated from standard curves and normalized to protein content.

Serum ELISA Assay

Mouse serum was collected and inflammatory factor levels (IL‐1β, IL‐6, TNF‐α) after different treatments were analyzed using a Mouse ELISA Kit (E‐EL‐M0037, E‐EL‐M0044, E‐EL‐H0109, Elabscience, Wuhan, China) according to the manufacturer's guidance.

Immunofluorescence Staining

For immunofluorescence staining, MOVAS were cultured in 6 cm plates for different treatments. Subsequently, the plates were co‐incubated with anti‐NFS1 (1:50, 67021‐1‐lg, Proteintech, Wuhan, China) or anti‐SP2 (1:50, 25000‐1‐ap, Proteintech) primary antibodies overnight at 4 °C. After washing with PBS for three times, Alexa Fluor 488/555 conjugated secondary antibodies (1:200, Invitrogen, USA) were applied for 1 h at 37 °C in the dark. Sections were mounted with ProLong Gold anti‐fade reagent with 4′,6‐diamidino 2‐phenylindol (DAPI, P36931, Invitrogen, USA) for observation of fluorescence by a confocal microscopy (Leica TCS SP8, Leica, Germany).

As for tissue, the sections were incubated with primary antibodies against the proteins of interest at 4 °C overnight, followed by incubation with appropriate fluorophore‐conjugated secondary antibodies for 1 h at room temperature in the dark. Nuclei were counterstained with DAPI for 5 min, and the sections were mounted with antifade mounting medium. Images were acquired using a fluorescence or confocal microscope (Leica TCS SP8, Leica, Germany).

Dual Luciferase Reporter Assay

The binding locus site of SP2 on the IDH2 promoter was predicted (5‐GGGCGGGAC‐3) as previously reported.^[^ 41 ^]^ The wild type mouse IDH2 promoter (−2000–0 bp) and a mutated IDH2 promoter (−600–0 bp) were subcloned into pRP‐basic luciferase reporter vectors, termed as pRP‐hRluc‐m_wtIDH2 promoter‐Luci and pRP‐hRluc‐m_mutIDH2 promoter‐Luci, respectively. These plasmids were constructed and packed by VectorBuilder Co. Ltd (Guangzhou, China).

The 293T were co‐transfected with 1 µg pcDNA3.1‐m_SP2 wt‐3 × HA plasmid (pSP2^oe^) or pcDNA3.1‐NC plasmid (pNC^oe^)with 1 µg pRP‐hRluc‐m_wtIDH2 promoter‐Luci (pIDH2 wt) or pRP‐hRluc‐m_mutIDH2 promoter‐Luci (pIDH2 del) for 48 h. Luciferase activities were measured with the Dual Luciferase Reporter Assay System (RG027, Beyotime, Shanghai, China) using the SpectraMaxi3 reader.

Co‐Immunoprecipitation

The 293T cells were transfected with different plasmids for 48 h and then lysed with NP40 cell lysis buffer. After centrifuge, the cell supernatants were incubated with an anti‐Flag primary antibody (14 793, Cell Signaling Technology, Danvers, MA, USA) at 4 °C overnight. Anti‐IgG (2729, Cell Signaling Technology, Danvers, MA, USA) was served as a negative control. The immune complexes were then purified by 20 µL of protein G magnetic beads (88 847, ThermoFisher, Waltham, MA, USA) at 25 °C for 2 h, centrifuged and washed with NP40 cell lysis buffer for four to five times. The immunoprecipitated protein was further analyzed by Western blotting with target antibodies.

Proximity Ligation Assay

Proximity Ligation Assay (PLA) was performed using NaveniFlex Cell MR Atto647N (NC.MR.100 Atto647N, Univ, Shanghai, China) to detect endogenous protein–protein interactions based on spatial proximity. After fixation and permeabilization following an optimized immunofluorescence protocol for the primary antibodies used, samples were blocked at 37 °C for 60 min. Primary antibodies against the proteins of interest, previously validated by immunofluorescence and western blotting, were incubated with cells either for 60 min at 37 °C or overnight at 4 °C, followed by washes in TBS‐T buffer. Subsequently, Navenibody probes were added and incubated at 37 °C for 60 min, after which enzymatic ligation and amplification steps were carried out according to the manufacturer's protocol. Nuclear staining was performed with DAPI, and slides were mounted with antifade medium. Imaging was conducted on a fluorescence or confocal microscope equipped with DAPI, Atto647 filter sets.

Chromatin Immunoprecipitation and qPCR

ChIP was performed using a standard protocol with minor modifications. Briefly, cells were crosslinked with 1% formaldehyde at room temperature for 10 min, and the reaction was quenched with 125 mM glycine. Nuclei were isolated and chromatin was sheared to an average size of 200–500 bp by sonication. The sheared chromatin was immunoprecipitated overnight at 4 °C with antibodies against NFS1, followed by incubation with protein A/G magnetic beads. After sequential washes, crosslinks were reversed, and DNA was purified using phenol–chloroform extraction or a commercial purification kit. Quantitative PCR (qPCR) was performed using primers specific for the IDH2 promoter region, with results normalized to input DNA.

Western Blotting Analysis

MOVAS and aortic tissues were lysed in RIPA buffer (P0013C, Beyotime, Shanghai, China) containing a protease inhibitor cocktail and phosphatase (P1045, Beyotime, Shanghai, China). The cell or aortic tissue lysates were centrifuged at 12 000 rpm at 4 °C for 10 min, and the protein concentration of the supernatant was quantified using a Bicinchoninic Acid Protein Assay Kit (A55861, ThermoFisher, Waltham, MA, USA). The protein samples were boiled for 10 min, separated using 10% SDS‐PAGE (M00656, GenScript, NJ, USA) and transferred to polyvinylidene fluoride membranes. For immunoblotting, the membranes were blocked in blocking buffer (P30500, NCM Biotech, Suzhou, China) for 40 min at 25 °C, and the proteins were detected after incubating the membranes with primary antibodies overnight at 4 °C as follows: anti‐αSMA (1:1000, ab184705, Abcam, Cambridge, MA, USA), anti‐OPN (1:1000, 22952‐1‐AP, Proteintech, Wuhan, China), anti‐NFS1 (1:1000, 67021‐1‐lg, Proteintech, Wuhan, China), anti‐CSE (1:1000, 602341‐1‐lg, Proteintech, Wuhan, China), anti‐Cyclin d1 (1:1000, A11022, Abclonal, Wuhan, China), anti‐PCNA (1:1000, A12027, Abclonal, Wuhan, China), anti‐IDH2 (1:1000, A7190, Abclonal, Wuhan, China), anti‐PFKP (1:1000, 8164t, CST, Danvers, MA, USA), anti‐HK2 (1:1000, 2867t, CST, Danvers, MA, USA), anti‐ENO1 (1:1000, 3810t, CST, Danvers, MA, USA), anti‐PGAM1 (1:1000, 12098s, CST, Danvers, MA, USA), anti‐LDHA (1:1000, 3582t, CST, Danvers, MA, USA), anti‐LDHB (1:1000, 56298s, CST, Danvers, MA, USA), anti‐ALDOA (1:1000, 8060s, CST, Danvers, MA, USA), anti‐PKM2 (1:1000, 4053s, CST, Danvers, MA, USA), anti‐HA‐Tag (1:1000, 3724t, CST, Danvers, MA, USA), anti‐SP1 (1:1000, 21962‐1‐ap, Proteintech, Wuhan, China), anti‐SP2 (1:1000, 25000‐1‐ap, Proteintech, Wuhan, China), anti‐SP5 (1:1000, 10774‐1‐ap, Proteintech, Wuhan, China), anti‐ZNF532 (1:1000, A305‐442A, Bethyl Laboratories, Montgomery, TX, USA). The internal controls for normalizing protein abundance were anti‐α‐tubulin (1:5000, 11224‐1‐AP, Proteintech, Wuhan, China), anti‐β‐actin (1:10 000, ab8226, Abcam, Cambridge, MA, USA), anti‐COX IV (1:1000, 11242‐1‐ap, Proteintech, Wuhan, China), anti‐VCAM1 (1:1000, ab174279, Abcam, Cambridge, MA, USA), and anti‐ICAM1 (1:1000, ab222736, Abcam, Cambridge, MA, USA). After washed three times with Tris‐buffered saline (TBS), the membranes were incubated for 1 h at room temperature with horseradish peroxidase (HRP)‐conjugated anti‐mouse IgG secondary antibodies (1:10 000, ab205719, Abcam, Cambridge, MA, USA) or anti‐rabbit IgG secondary antibodies (1:10 000, ab205718, Abcam, Cambridge, MA, USA) followed by incubation with Pierce ECL Western blotting Substrate (SB‐WB004, ShareBio, China). The signals were quantified and analyzed using ImageJ software (vision:2.9.0, NIH, Bethesda, MD, USA) after normalization to α‐tubulin or β‐actin.

Molecular Docking

Protein amino acid sequences were obtained from UNIPROT database (www.uniprot.org) and protein structures were predicted using the Swiss Model (swissmodel.expasy.org). PyMOL (version 3.0.3, DeLano Scientific LLC, AZ, USA) and HDOCK (hdock.phys.hust.edu.cn) were further used for analyzing and docking the protein interactions.

CUT & Tag

After different treatments, equal number of cells (1 × 10^5^) were incubated with 10 µL of concanavalin A coated magnetic beads and incubated at room temperature for 10 min. The cells were then sequentially incubated with primary antibody anti‐NFS1 (15370‐1‐AP, Proteintech, Wuhan, China), secondary antibody (Goat anti‐rabbit IgG H&L, AB206‐01‐AA) and hyperactive PG‐TN5/PA‐TN5 Transposo. Then, the samples are washed three times in Dig‐300 buffer containing the protease inhibitor mixture and digitalis saponins, resuspended in fragmentation buffer, and incubated at 37 °C for 1 h to allow the pA/G‐Tnp fragmentation reaction to complete. The fragmented DNA was extracted and amplified by PCR. CUT & Tag libraries were next constructed using HyperactiveTM In Situ ChIP Library Prep Kit (TD901, Vazyme, China) and sequenced on an Illumina NovaSeq platform, and 150 bp paired‐end reads were generated.

As for data analysis, fastp v 0.20.0 was used to remove adapter and low‐quality reads. The paired‐end reads were aligned using Bowtie2 version 2.3.4.3 with options: end‐to‐end sensitive. Duplicated reads were removed using Picard v2.18.17 with the parameter: REMOVE_DUPLICATES  =  true. SEACR v1.3 was used for analyzing peak calling with threshold: 0.01. Scatterplots, correlation plots, and heatmaps were displayed using deepTools v2.27.1. Annotation of peaks was performed using an R package ChIPseeker v1.12.1. MEME‐ChIP v 5.0.5 was used to search for the binding sites. Peaks with M > 0.2 and P < 0.05 were defined as specific peaks.

Statistical Analysis

All statistical analyses were conducted using GraphPad Prism 8.0.1. The Shapiro‐Wilk test was applied to assess the normality of the data distribution initially. Comparisons between 2 groups were performed by the Student t test (equal variances) or Welch t test (unequal variances) for normally distributed variables. Non‐normally distributed variables were analyzed using the Mann‐Whitney U test. For multiple comparisons involving more than 2 groups, the Brown‐Forsythe test was applied to assess variance homogeneity. If the data passed the equal variance test, a one‐way ANOVA analysis was used followed by the Tukey post hoc test; otherwise, a Welch ANOVA test was performed followed by the Dunnett`s T3 method. Unless indicated otherwise, parametric data are expressed as mean ± SEM and nonparametric data are represented as median (interquartile range). A value of p < 0.05 was considered statistically significant.

Conflict of Interest

The authors declare no conflict of interest.

Author Contributions

L.Z., D.W., and Y.Z. contributed equally to this perspective paper. L.Z., Y.W., and S.X. designed experiments. L.Z. completed the in vitro and in vivo sections. Y.Z., D.W., and X.Q. performed the data analysis. H.W. and Y.S. contributed to the discussion. L.Z., Y.W., and G.Y. wrote and revised the manuscript. S.X. and Y.W. provided research funding. This work was supported by the National Natural Science Foundation of China (Grant No. 81873462, 82200454, and 82371584). Shanghai Municipal Science and Technology Commission (Grant No. 17DZ1930203). Shanghai Jiao Tong University School of Medicine Affiliated Renji Hospital Cultivation Fund (Grant No. RJPY‐DZX‐005). Pudong New District Health Industry Special Fund (Grant NO. PW2021E‐04).

Supporting information

Supporting Information