Comprehensive analysis for the role of macrophage-driven genes in abdominal aortic aneurysm

Introduction

Abdominal aortic aneurysm (AAA) is a chronic vascular disease that often progresses asymptomatically and is characterized by the persistent dilation of the abdominal aorta. If left untreated, AAA carries a high risk of rupture, with a post-rupture mortality rate of as high as 90% (1). It is estimated to cause approximately 175,000 deaths globally each year and affects 3–9% of the population aged 65 years and older (2). Major risk factors for AAA include age, sex, smoking, hypertension, and atherosclerosis (3,4). Although surgical interventions, such as endovascular aneurysm repair (EVAR), have become increasingly refined, patients remain at risk for postoperative rupture and thus require continuous surveillance (5). Current clinical strategies predominantly rely on aneurysm diameter for treatment decisions, yet lack accurate tools for assessing disease progression and rupture risk (6,7). Therefore, there is an urgent need to identify molecular biomarkers that can facilitate early diagnosis and risk stratification.

Chronic inflammation is widely regarded as a key driver in the pathogenesis of AAA, with macrophages playing a central role in the inflammatory processes within the aortic wall (8). Studies have shown that macrophages infiltrate the tunica media and adventitia during the early stages of aneurysm formation (9), in which they secrete matrix metalloproteinases (MMPs), pro-inflammatory cytokines, and reactive oxygen species (ROS), contributing to extracellular matrix degradation and vascular wall remodeling (10-12). Moreover, macrophage polarization into either the M1 (pro-inflammatory) or M2 (anti-inflammatory/repair) phenotype significantly influences the direction of AAA progression (13). A predominance of M1 macrophages has been associated with elastic fiber fragmentation, vascular smooth muscle cell apoptosis, and compromised vascular wall integrity, thereby accelerating aneurysm expansion (14,15). Recent histological and transcriptomic studies have further revealed that macrophage-rich regions within AAA tissues exhibit distinct gene expression profiles, suggesting that macrophages are not only pivotal pathological mediators but also potential diagnostic biomarkers and therapeutic targets (16). However, most existing studies focus on isolated aspects of macrophage function and lack a comprehensive understanding of their interactions with the immune microenvironment.

To address this gap, the present study integrates single-cell RNA sequencing (scRNA-seq) and bulk RNA-seq data to identify macrophage-driven diagnostic biomarkers in AAA using least absolute shrinkage and selection operator (LASSO) regression and immune infiltration analyses. Furthermore, these key genes were identified in human AAA tissue. This approach aims to uncover potential biomarkers for early detection and provide mechanistic insights into disease progression, thereby offering a theoretical basis for precision diagnosis and targeted therapy in AAA. We present this article in accordance with the STARD reporting checklist (available at https://cdt.amegroups.com/article/view/10.21037/cdt-2025-365/rc).

Methods

Public data acquisition and preprocessing

Three AAA-related datasets were downloaded from the GEO database (https://www.ncbi.nlm.nih.gov/geo/): GSE166676, GSE47472, and GSE57691. GSE166676 is a scRNA-seq dataset containing six samples; GSE47472 is a bulk RNA-seq dataset comprising 14 AAA samples and 8 normal controls; GSE57691 includes 49 AAA and 10 normal samples. Probe-to-gene symbol conversion for GSE47472 and GSE57691 was performed using the idmap3 R package (https://github.com/jmzeng1314/idmap3). These datasets were selected based on predefined criteria, including: (I) publicly available human AAA transcriptomic data; (II) inclusion of both AAA and non-aneurysmal control samples; and (III) adequate sample size for model construction and validation. All available samples meeting these criteria within each dataset were included, and no samples were excluded after quality assessment.

scRNA-seq analysis

Single-cell analysis was conducted using the Seurat package in R version 4.3.2. Quality control (QC) was performed by filtering cells with 200< nFeature_RNA <5,000 and mitochondrial gene percentage (percent.mt) <10 to remove low-quality cells. Data normalization and batch effect correction across multiple samples were carried out using NormalizeData and IntegrateData, respectively (GSE166676). Principal component analysis (PCA) was applied, followed by t-distributed stochastic neighbor embedding (t-SNE) based on the top 30 principal components for dimensionality reduction. Cell clustering was performed using FindClusters with a resolution of 0.5. Cell type annotation was based on known markers from the CellMarker2.0 database.

Identification of macrophage-related differentially expressed genes (DEGs)

DEG analysis on the GSE47472 dataset was performed using the limma package. DEGs associated with macrophages were identified by intersecting these DEGs with macrophage marker genes obtained from the single-cell analysis. A Venn diagram was plotted using the VennDiagram package to visualize the overlapping genes.

Enrichment analysis

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed on the intersected genes using the clusterProfiler package. Terms with a P value <0.05 were considered significantly enriched. The results were visualized using the ggplot2 package.

Construction and validation of the diagnostic model

To assess the diagnostic potential of the intersected genes, the GSE47472 dataset was used as the training set. LASSO regression was performed using the glmnet package, and the optimal model was selected under the lambda.1se parameter. The LASSO regression was applied to this set of DEGs to identify the most predictive features for AAA diagnosis. At the optimal penalty parameter (logλ =−2.2), five hub genes were retained in the final diagnostic model. The model was validated using GSE57691. The diagnostic performance was quantified by drawing receiver operating characteristic (ROC) curves with the pROC package. The classification threshold was derived from the ROC curve based on the Youden index to identify the optimal cut-off value for discriminating AAA from controls. No participants had missing covariate or gene expression data in the included GEO datasets; therefore, no imputation procedures were required.

Correlation analysis between marker genes and immune cell infiltration

Immune infiltration analysis was conducted using the CIBERSORT algorithm implemented in the IOBR package. The proportion of immune cells between the AAA and normal groups was compared, and the correlation between marker gene expression and immune cell infiltration levels was evaluated (using a merged dataset of GSE47472 and GSE57691).

Correlation analysis between marker genes and pathway scores

The hallmark gene sets were obtained from the MSigDB database (https://www.gsea-msigdb.org/gsea/index.jsp). Gene set variation analysis (GSVA) was performed using the GSVA package to calculate pathway enrichment scores. Differences in pathway scores between AAA and control groups were analyzed, and correlations between pathway scores and marker gene expression levels were evaluated (using a merged dataset of GSE47472 and GSE57691).

Mendelian randomization (MR) analysis between marker genes and AAA

Genome-wide association study (GWAS) summary data for both AAA and the marker genes were obtained from the IEU OpenGWAS database (https://gwas.mrcieu.ac.uk/). Two-sample MR analysis was conducted using the TwoSampleMR package to investigate the causal relationship between marker gene expression and AAA risk (using a merged dataset of GSE47472 and GSE57691).

MR analytical framework and assumptions

(I) Relevance: genetic variants used as instrumental variables (IVs) are strongly associated with the exposure (marker gene expression). (II) Independence: IVs are independent of confounders affecting both the exposure and the outcome. (III) Exclusion restriction: IVs influence the outcome (AAA) only through the exposure, not through alternative biological pathways. Single-nucleotide polymorphisms (SNPs) significantly associated with each marker gene (P<5×10⁻⁸) were selected as candidate IVs. Linkage disequilibrium (LD) clumping was conducted with an r²<0.001 and a 10,000 kb window to ensure independence between SNPs. The strength of IVs was assessed using F-statistics, with F >10 indicating adequate instrument strength and minimizing weak-instrument bias. Horizontal pleiotropy was evaluated using the MR-Egger intercept test and the MR-PRESSO global test. Cochran’s Q test was applied to assess heterogeneity across SNPs. When pleiotropy or heterogeneity was detected, MR-Egger or weighted median estimators were used as sensitivity analyses in addition to the inverse-variance weighted (IVW) primary analysis. To account for potential shared biological pathways or correlated gene expression, multivariable MR analyses were conducted when more than one marker gene demonstrated significant association signals. MVMR allowed simultaneous adjustment for multiple exposures to estimate the direct effect of each gene on AAA risk. This study employed a classical two-sample MR design. The exposure-GWAS and outcome-GWAS datasets originated from independent cohorts within the IEU OpenGWAS resource. No participant overlap was reported. The two samples were derived from comparable populations of predominantly European ancestry, ensuring consistency in baseline genetic architecture. All selected genetic variants that predicted the marker gene expression in the exposure dataset remained valid predictors in the outcome dataset after harmonization. A schematic figure illustrating the three MR assumptions is shown in Figure S1 (17).

Tissue samples and RNA quantification

These specimens were collected at Ningxia Medical University General Hospital between January 2020 and December 2025. Patients with infrarenal AAA who underwent elective open repair were included, whereas those with prior aortic surgery, active infection, autoimmune or systemic inflammatory diseases, malignant tumors, or incomplete clinical information were excluded. Normal abdominal aortic tissues were obtained from patients undergoing aorto-iliac surgery for non-aneurysmal conditions or from organ donors, and all satisfied the same exclusion criteria. The collection and use of human AAA tissues and normal abdominal aortic specimens were approved by the Ethics Committee of General Hospital of Ningxia Medical University (No. KYLL-2023-0293). Written informed consent was obtained from all participants or their legal representatives. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Immediately after excision, tissues were rinsed in ice-cold PBS to remove residual blood, snap-frozen in liquid nitrogen, and stored at −80 ℃ until processing.

Total RNA was extracted from approximately 50–100 mg of tissue using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) following the manufacturer’s instructions. Tissue homogenization was performed with a sterile glass grinder on ice, and phase separation was achieved by adding chloroform, followed by centrifugation at 12,000 ×g for 15 min at 4 ℃. The aqueous layer was carefully collected, precipitated with isopropanol, and washed twice with 75% ethanol. RNA pellets were air-dried and dissolved in RNase-free water. RNA concentration and purity were determined using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), and samples with an A260/A280 ratio between 1.8 and 2.0 were considered suitable for downstream assays.

Complementary DNA (cDNA) was synthesized from 1 µg of total RNA using a reverse transcription kit (Vazyme, Nanjing, China). Real-time quantitative polymerase chain reaction (RT-qPCR) was conducted using SYBR Green Master Mix (Takara, Kusatsu, Shiga, Japan) on a QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal control. The expression levels of the selected hub genes ARG2, S100A6, VASH1, PI3, and SMU1 were quantified, and relative expression was calculated by the 2^−ΔΔCt method.

Statistical analysis

All statistical analyses were performed using R version 4.3.2. For comparison between two groups, Student’s t-test or Wilcoxon rank-sum test was applied depending on data distribution. Correlation analysis was conducted using Pearson’s methods. For ROC curve analysis, area under the curve (AUC) values were calculated to assess diagnostic performance. A two-tailed P<0.05 was considered statistically significant.

Results

Single-cell transcriptomic heterogeneity analysis and cell type annotation in AAA

First, a detailed analysis of the scRNA-seq dataset from AAA samples was performed. As shown in Figure S2A, post-QC processing led to a more concentrated distribution of nFeature_RNA and nCount_RNA, with reduced variability among samples, indicating effective removal of low-quality cells and technical noise. Violin plots further demonstrated improved data uniformity after QC, providing a solid foundation for downstream analyses (Figure S2B).

A total of 19 cell clusters were identified (Figure 1A), and their cell types were annotated based on the expression profiles of characteristic marker genes. Seven major cell types were annotated: B cells, endothelial cells, epithelial cells, macrophages, NK cells, CD34^– pre-B cells, and T cells, among which macrophages were the most abundant (Figure 1B). A heatmap illustrated the top five significantly upregulated marker genes for each cell type (Figure 1C). The distinct Z-score distribution across clusters confirmed the reliability of transcriptional signatures specific to each cell type. A new supplementary table with DEGs for each cluster has been added (Table S1).

Figure 1 Single-cell transcriptomic heterogeneity analysis and cell type annotation of AAA. (A) t-SNE plot based on unsupervised clustering, showing the spatial distribution of 18 cell clusters. Different colors represent distinct clusters. (B) Cell type annotation results based on marker gene expression, with different colors indicating different cell clusters. (C) Heatmap of Z-score expression levels for characteristic marker genes across clusters. Red indicates high expression, and blue indicates low expression. AAA, abdominal aortic aneurysm; t-SNE, t-distributed stochastic neighbor embedding.

Identification and functional enrichment of macrophage-related DEGs

Next, we performed bulk transcriptome analysis to identify DEGs in AAA. Heatmap and volcano plots revealed 747 upregulated and 2,078 downregulated genes in AAA samples compared with normal controls (Figure 2A,2B). From the single-cell dataset, we extracted 573 macrophage-related genes (Table S2). By intersecting these genes with the DEGs, 59 overlapping genes were identified, defined as macrophage-related DEGs (Figure 2C and Table S3). Heatmap visualization demonstrated distinct expression patterns of these genes, such as TREM2, S100A8, and SMU1 were markedly upregulated in AAA, whereas CALB1, GLRX, CCL18, PLTP, and BMP2 were downregulated (Figure 2D).

Figure 2 Differential gene expression analysis of macrophages. (A) Heatmap of DEGs. Rows represent genes, and columns represent samples (AAA vs. normal). Red indicates high expression; green indicates low expression. (B) Volcano plot showing the distribution of DEGs. Pink dots indicate upregulated genes, green dots indicate downregulated genes, and gray dots indicate non-significant genes. (C) Venn diagram showing the overlap between DEGs and macrophage-related genes. (D) Heatmap of macrophage-related DEGs. Red indicates high expression; green indicates low expression. AAA, abdominal aortic aneurysm; DEG, differentially expressed gene; FC, fold change.

Functional enrichment analysis of these 59 genes was then conducted. GO analysis showed significant enrichment in pathways related to lipid metabolism and immune response (Figure 3A). Specifically, in the biological process (BP) category, terms such as “positive regulation of lipid localization”, “defense response to bacterium”, and “cellular response to molecule of bacterial origin” were highly enriched. Cellular component (CC) terms included “membrane microdomain”, “secretory granule lumen”, and “vesicle lumen”, while molecular function (MF) terms included “high-density lipoprotein particle binding”, “chemokine activity”, and “calcium-dependent protein binding”. KEGG pathway analysis further revealed significant enrichment in the “NOD-like receptor signaling pathway”, “IL-17 signaling pathway”, and “PI3K-Akt signaling pathway” (Figure 3B).

Figure 3 Pathway enrichment analysis of macrophage-related DEGs. (A) Bar plot of GO enrichment results. (B) Bubble plot of KEGG pathway enrichment analysis. BP, biological process; CC, cellular component; DEG, differentially expressed gene; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function.

Construction and validation of a diagnostic model for AAA

To further investigate the diagnostic potential of macrophage-related DEGs in AAA, LASSO regression analysis was applied, and five key genes associated with AAA progression were identified to construct a diagnostic model. As shown in Figure 4A, increasing the log value of the penalty parameter λ led to progressive shrinkage of the regression coefficients. The partial likelihood deviance curve (Figure 4B) indicated optimal model performance at logλ =−2.2, achieving a balance between model simplicity and predictive accuracy. ROC curve analysis was then used to assess the model’s performance. In the training set GSE47472, the model achieved an AUC of 0.981 (95% CI: 0.951–0.993) (Figure 4C). In the validation set GSE57691, the AUC remained at 0.935 (95% CI: 0.903–0.998) (Figure 4D), demonstrating strong predictive performance across cohorts. 59 macrophage-related genes along with the five core genes selected through LASSO regression analysis as Table S4.

Figure 4 Construction and validation of the AAA diagnostic model. (A) LASSO regression coefficient profiles as a function of logλ. (B) Partial likelihood deviance plotted against logλ. The dotted line indicates the optimal λ with minimal model deviance. (C) ROC curve for the training dataset GSE47472. (D) ROC curve for the validation dataset GSE57691. AAA, abdominal aortic aneurysm; AUC, area under the curve; CI, confidence interval; LASSO, least absolute shrinkage and selection operator; ROC, receiver operating characteristic.

Validation of key model gene expression and cell type-specific analysis

Next, the expression of the five model genes in AAA and normal samples was analyzed. In the training set, compared with the normal group, ARG2 and SMU1 were significantly downregulated in AAA (P<0.01), whereas VASH1, PI3, and S100A6 were significantly upregulated (P<0.01, Figure 5A). The validation set further confirmed these expression patterns of key genes in AAA (Figure 5B). Cell type-specific analysis revealed a preferential enrichment of these five model genes in macrophages, particularly SMU1 (Figure 5C).

Figure 5 Expression validation and cell type-specific analysis of key model genes. (A) Boxplots showing the expression levels of five key genes (ARG2, S100A6, VASH1, PI3, SMU1) in AAA and normal samples in the training set (GSE47472). (B) Validation of gene expression in the combined validation dataset (GSE47472 + GSE57691). (C) Violin plots showing the expression levels of key genes across different cell types. *, P<0.05; **, P<0.01; ***, P<0.001. AAA, abdominal aortic aneurysm.

Enrichment of key signaling pathways and gene-pathway interaction analysis

To further explore key pathways associated with AAA, GSVA was performed. The results showed significant enrichment of inflammatory response, PI3K-AKT-mTOR signaling, and apoptosis pathways (Figure 6A). These three pathway scores were then compared between AAA and normal samples, revealing significantly higher activity in the AAA group (P<0.001, Figure 6B). Correlation analysis between genes and pathways (Figure 6C) showed that all genes except ARG2 and S100A6 were significantly positively correlated with the inflammatory response pathway; notably, SMU1 was significantly positively correlated with all three pathways. These results suggest that these genes may drive AAA progression by modulating inflammation-related signaling.

Figure 6 Enrichment of key signaling pathways and gene-pathway interaction analysis. (A) Bar plot of hallmark gene set enrichment results. (B) Boxplots of GSVA scores for the top three pathways. (C) Scatter plot of gene-pathway correlations. Dot color represents correlation coefficient (red = positive, blue = negative), and dot size represents −log₁₀(P value). ***, P<0.001. AAA, abdominal aortic aneurysm; GSVA, gene set variation analysis.

Immune cell infiltration differences and their correlation with five key genes

To investigate the potential immune regulatory roles of the five key genes in AAA, immune cell infiltration analysis was conducted. Significant differences were observed between the immune microenvironments of AAA patients and healthy individuals. As shown in Figure 7A, CD4 naïve T cells, resting NK cells, and M1 macrophages were significantly increased in the AAA group (P<0.05), while M2 macrophages were significantly decreased (P<0.05). Correlation analysis between genes and immune cell types (Figure 7B) revealed that ARG2 was positively correlated with M2 macrophages but negatively correlated with CD4 naïve T cells and M1 macrophages. PI3 showed a significant positive correlation with M1 macrophages, while SMU1 was negatively correlated with M2 macrophages. These findings suggest that the key genes may participate in AAA pathogenesis by regulating macrophage polarization.

Figure 7 Immune cell infiltration differences and correlation with key genes. (A) Boxplots of immune cell infiltration levels. (B) Scatter plot of gene-immune cell correlations. Dot size indicates −log₁₀(P value) (larger means more significant), and color represents Pearson correlation coefficient (red = positive, blue = negative). *, P<0.05; **, P<0.01; ns, not significant. AAA, abdominal aortic aneurysm.

MR analysis of the association between five key genes and AAA

Two-sample MR was performed to evaluate whether genetically predicted expression levels of the five candidate genes exert a causal effect on AAA risk. Among these genes, only SMU1 demonstrated nominal evidence of a positive causal association with AAA. The IVW approach yielded an OR of 1.338 (95% CI: 0.914–1.763, P=0.04), while MR-Egger regression produced an OR of 1.172 (95% CI: 0.868–1.584, P=0.003) (Figure 8A). To assess directional pleiotropy, we examined the MR-Egger intercept, which was 0.004 (P=0.46), indicating no significant horizontal pleiotropy. Cochran’s Q test similarly showed no marked heterogeneity across SNPs (IVW: Q=5.87, P=0.44; MR-Egger: Q=5.51, P=0.48). MR-PRESSO did not detect any outlier variants (global test P=0.51), suggesting that the causal estimates were not driven by pleiotropic SNPs. Sensitivity analyses demonstrated partial inconsistency across methods. The weighted median and weighted mode estimators produced ORs below 1, which diverged from the IVW and MR-Egger directions (Figure 8A,8B). The discrepancy between the reported P values and confidence intervals results from the small number of available SNP instruments and the associated instability of MR-Egger and IVW estimates under low instrumental strength. Under such conditions, the P value may reach nominal significance due to model fitting, whereas the confidence interval remains wide and crosses 1, reflecting imprecision. Nevertheless, leave-one-out analyses (Figure 8C) showed that the exclusion of any single SNP did not materially alter the direction of the IVW estimate, supporting basic robustness. A discrepancy emerged between MR findings and tissue-level expression: genetic predisposition to higher SMU1 expression was associated with increased AAA risk, whereas RT-qPCR demonstrated reduced SMU1 expression in human AAA tissues. This pattern may reflect disease-stage-dependent regulation, compensatory transcriptional suppression in advanced lesions, or cell-type-specific expression patterns that are not captured in bulk tissue assays. Collectively, these results indicate that although SMU1 may participate in AAA pathogenesis, additional experimental and genetic evidence is required to confirm its causal role.

Figure 8 MR analysis of the association between five key genes and AAA. (A) Forest plot of MR analysis results for the five genes and AAA. Significant results (P<0.05) are highlighted in red. (B) Scatter plot showing the effect of SMU1 on AAA. (C) Leave-one-out sensitivity analysis. AAA, abdominal aortic aneurysm; CI, confidence interval; MR, Mendelian randomization; OR, odds ratio; SNP, single-nucleotide polymorphism.

Verification of the expression of the top 5 hub genes in clinical samples

To verify the bioinformatics findings, we assessed the expression of five hub genes (ARG2, SMU1, VASH1, PI3, and S100A6) in three AAA tissues and three normal abdominal aortic specimens by RT-qPCR. Among them, SMU1 was significantly downregulated in AAA tissues compared with controls (P<0.01). In contrast, VASH1, PI3, and S100A6 did not show statistically significant differences between AAA and control groups (Figure 9). These results highlight SMU1 as the most robustly associated gene with AAA pathogenesis in clinical samples. The qPCR primer sequences are shown in Table S5.

Figure 9 Validation of hub gene expression in human AAA tissues. RT-qPCR analysis of ARG2, SMU1, VASH1, PI3, and S100A6 in three AAA tissues and three normal abdominal aortic controls. SMU1 was significantly downregulated in AAA, whereas VASH1, PI3, and S100A6 showed no statistically significant differences. Data are shown as mean ± SEM. *, P<0.05; **, P<0.01; ns, not significant. AAA, abdominal aortic aneurysm; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; RT-qPCR, reverse transcription quantitative polymerase chain reaction; SEM, standard error of the mean.

Discussion

This study integrated scRNA-seq, bulk RNA-seq, and MR analyses to comprehensively investigate the potential molecular mechanisms and diagnostic biomarkers of AAA, with a specific focus on macrophage-associated key genes. Among them, five genes—ARG2, S100A6, VASH1, PI3, and SMU1—were identified as potential diagnostic indicators for AAA. Notably, SMU1 consistently exhibited distinctive characteristics across multiple analyses, suggesting it may serve as a novel focus for future research.

Macrophages play a central role in the pathogenesis of AAA by mediating inflammatory responses that affect extracellular matrix degradation and oxidative stress imbalance, thereby promoting disease progression (13). In addition to immune-driven proteolysis, oxidative stress has been increasingly recognized as a critical and complementary factor in the formation of aneurysms. Excessive production of ROS—originating from activated nicotinamide adenine dinucleotide phosphate (NADPH) oxidases, dysfunctional mitochondria, and other enzymatic sources—can directly damage extracellular matrix components and facilitate the conversion of pro-MMPs into their active forms, particularly MMP2 and MMP9. ROS also act as second messengers to activate redox-sensitive transcriptional programs (for example, NF-κB and AP-1), thereby amplifying pro-inflammatory cytokine expression and protease production (18,19). Recent mechanistic work further links mitochondrial oxidative injury and disturbed redox homeostasis with vascular remodeling and aneurysm progression (20), while genetic and epigenetic studies highlight oxidative pathways as modulators of vascular cell phenotype and matrix stability (21). Importantly, oxidative stress and inflammation can form a feed-forward loop: macrophage activation increases ROS production, which in turn promotes further macrophage recruitment and pro-inflammatory polarization, perpetuating matrix degradation and aneurysm expansion, thereby promoting matrix degradation and aneurysm expansion. Although our scRNA-seq analysis identified 15 macrophage subsets within AAA tissues, for this study, we adopted the classical M1/M2 framework as a simplified model to illustrate the imbalance between pro- and anti-inflammatory phenotypes, a concept widely used in the field of vascular inflammation. Nevertheless, characterizing the 15 macrophage subsets beyond M1/M2 polarization represents an important direction for future research. Integrating additional scRNA-seq datasets, combined with experimental validation, will be essential to unravel their distinct functional roles in AAA. Our scRNA-seq analysis revealed that macrophages are the most abundant immune cell type within AAA tissues, consistent with previous findings (22). M0 macrophages can polarize into either pro-inflammatory M1 or anti-inflammatory M2 phenotypes, a process known to regulate vascular inflammation in AAA (23). Induction of M1 polarization and M1-type inflammation has been shown to promote AAA formation (24). Our immune infiltration analysis further demonstrated a higher level of M1 macrophages and a lower level of M2 macrophages in AAA tissues compared to normal controls, reinforcing the notion that macrophage polarization imbalance may be a key driver of persistent inflammation and pathological progression in AAA. Notably, our data also revealed that SMU1 expression was restricted to a subset of macrophages, suggesting potential heterogeneity in SMU1-associated macrophage functions. Characterizing these subsets in more detail will be a key focus of our future studies, which may help clarify the precise role of SMU1-positive macrophages in AAA pathogenesis.

Previous studies have demonstrated that several critical genes are involved in regulating macrophage polarization and modulating the expression of their downstream target genes, thereby contributing to local aortic inflammation and AAA development (8,25-27). Moreover, Wu et al. recently performed a transcriptomic analysis to characterize macrophage-related gene expression patterns in AAA and developed a diagnostic model with moderate predictive performance (AUC =0.906) (28). Building on this, our study identified 59 macrophage-associated DEGs by integrating scRNA-seq and bulk RNA-seq data, aiming to achieve a higher-resolution understanding of macrophage involvement in AAA. We subsequently constructed a novel diagnostic model based on five selected genes using LASSO regression. This model demonstrated excellent predictive performance in two independent datasets (GSE47472 and GSE57691), with AUC values of 0.981 and 0.968, respectively, significantly outperforming the model established by Wu et al. (28). These findings not only underscore the diagnostic potential of macrophage-related genes in AAA but also highlight the advantage of multi-omics integration in biomarker discovery for complex vascular diseases.

Among the five genes included in the diagnostic model, SMU1 demonstrates substantial clinical translational potential due to its consistently low expression in AAA tissues, specific enrichment in macrophages, and close association with inflammatory signaling pathways. SMU1 is a splicing-associated factor primarily involved in mRNA splicing and cell cycle regulation (29,30). Previous studies have suggested the biomarker potential of SMU1 in various malignancies. For instance, Cai et al. identified SMU1 as a potential biomarker for ovarian cancer (31), and Qian et al. reported that SMU1 overexpression is correlated with aggressive phenotypes and poor prognosis in gastric cancer (32). In this study, we confirmed that SMU1 expression is significantly downregulated in human AAA tissues compared with non-aneurysmal abdominal aorta (n=3 vs. 3). Furthermore, MR analysis provided preliminary evidence supporting a potential causal relationship between SMU1 and AAA. Although some MR results slightly missed conventional significance thresholds, the consistent direction of effect and stability in leave-one-out sensitivity analysis enhanced the credibility of SMU1’s involvement in AAA pathogenesis.

More importantly, our findings suggest that SMU1 may play a role in vascular inflammation and immune regulation. GSVA analysis revealed that SMU1 expression is strongly positively correlated with multiple inflammation-related pathways, including the PI3K-AKT-mTOR signaling pathway and apoptosis pathway—both of which have been widely implicated in AAA pathogenesis (33-35). Regulation of the PI3K-Akt pathway has been shown to suppress inflammatory responses associated with AAA (36). In addition, immune infiltration analysis showed that SMU1 expression is negatively correlated with the abundance of M2 macrophages, suggesting that SMU1 may promote macrophage polarization toward the pro-inflammatory M1 phenotype, thereby exacerbating inflammation and vascular wall destruction. Prior studies have also demonstrated that activation of the PI3K pathway facilitates macrophage recruitment and survival (37,38). We acknowledge prior reports linking SMU1 to AAA pathogenesis via stabilization of MMP2 and MMP9 mRNA (39). Our data extend these findings by linking SMU1 to macrophage-related immunoinflammatory signatures in AAA, thus providing complementary insight into its epigenetic and transcriptional roles in aneurysm remodeling. Moreover, clinical validation in human AAA tissues further demonstrated that SMU1 was significantly downregulated, whereas VASH1, PI3, and S100A6 showed no statistically significant changes, highlighting SMU1 as the most robust candidate among the identified hub genes.

Future studies are warranted to expand our findings in larger cohorts of human AAA patients to validate the diagnostic and prognostic value of SMU1. In addition, more detailed mechanistic investigations, including in vivo models and molecular assays, are needed to clarify how SMU1 regulates inflammatory signaling, extracellular matrix degradation, and macrophage heterogeneity in AAA. Moreover, the potential of SMU1 as a therapeutic target should be further explored by employing genetic manipulation strategies (e.g., siRNA, AAV-mediated knockdown, or overexpression) or small-molecule inhibitors in animal models. These efforts may ultimately provide new insights into the clinical translation of SMU1 as both a biomarker and a therapeutic target in AAA.

Despite the strong data support, several limitations should be acknowledged. First, although the diagnostic model showed high AUC values, the relatively small sample sizes and class imbalance in the public datasets raise concerns regarding potential overfitting, and the representativeness of these cohorts limits the generalizability of the findings to broader populations. Second, although the MR analysis provides preliminary evidence for a potential causal relationship between SMU1 and AAA risk, causal inference based solely on genetic instruments is inherently limited. A more cautious interpretation is warranted, and future studies should adopt a triangulation framework integrating genetic, observational, and experimental evidence. Third, our conclusions rely primarily on transcriptomic data; although RT-qPCR offered initial validation, further protein-level confirmation and functional assays are needed to clarify the mechanistic role of SMU1. Additional in vivo and in vitro studies will also be required to elucidate how SMU1 contributes to macrophage-related inflammatory pathways in AAA development.

Conclusions

In summary, SMU1 is a novel AAA-associated gene with promising diagnostic value, potentially involved in disease progression by modulating macrophage polarization and inflammatory signaling. These findings not only broaden our understanding of the immunopathological mechanisms of AAA but also lay the theoretical groundwork for future early diagnostic and therapeutic strategies targeting SMU1.

Acknowledgments

We would like to thank AHMED (A British national pursuing a graduate degree at Beijing Language and Culture University) for his help in polishing our paper.

Footnote

Reporting Checklist: The authors have completed the STARD reporting checklist. Available at https://cdt.amegroups.com/article/view/10.21037/cdt-2025-365/rc

Data Sharing Statement: Available at https://cdt.amegroups.com/article/view/10.21037/cdt-2025-365/dss

Peer Review File: Available at https://cdt.amegroups.com/article/view/10.21037/cdt-2025-365/prf

Funding: This study was supported by the Natural Science Foundation of Ningxia Province, China (Nos. 2024AAC03555, 2022AAC03512, and 2022AAC03527).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://cdt.amegroups.com/article/view/10.21037/cdt-2025-365/coif). L.Y., G.Z. and Z.H. report that this study was supported by the Natural Science Foundation of Ningxia Province, China (No. 2022AAC03512). The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was approved by the Ethics Committee of Ningxia Medical University General Hospital (No. KYLL-2023-0293). Written informed consent was obtained from all participants or their legal representatives. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Schanzer A, Oderich GS. Management of Abdominal Aortic Aneurysms. N Engl J Med 2021;385:1690-8. [Crossref] [PubMed]
2. Liu S, Long C, Hong Y, et al. Prevalence of risk factors associated with rupture of abdominal aortic aneurysm (AAA): a single center retrospective study. PeerJ 2023;11:e15752. [Crossref] [PubMed]
3. Toghill BJ, Saratzis A, Bown MJ. Abdominal aortic aneurysm-an independent disease to atherosclerosis? Cardiovasc Pathol 2017;27:71-5. [Crossref] [PubMed]
4. Tang W, Yao L, Roetker NS, et al. Lifetime Risk and Risk Factors for Abdominal Aortic Aneurysm in a 24-Year Prospective Study: The ARIC Study (Atherosclerosis Risk in Communities). Arterioscler Thromb Vasc Biol 2016;36:2468-77. [Crossref] [PubMed]
5. Antoniou GA, Georgiadis GS, Antoniou SA, et al. Late Rupture of Abdominal Aortic Aneurysm After Previous Endovascular Repair: A Systematic Review and Meta-analysis. J Endovasc Ther 2015;22:734-44. [Crossref] [PubMed]
6. Lu H, Sun J, Liang W, et al. Cyclodextrin Prevents Abdominal Aortic Aneurysm via Activation of Vascular Smooth Muscle Cell Transcription Factor EB. Circulation 2020;142:483-98. [Crossref] [PubMed]
7. Wanhainen A, Verzini F, Van Herzeele I, et al. Editor's Choice - European Society for Vascular Surgery (ESVS) 2019 Clinical Practice Guidelines on the Management of Abdominal Aorto-iliac Artery Aneurysms. Eur J Vasc Endovasc Surg 2019;57:8-93. [Crossref] [PubMed]
8. Cai D, Sun C, Murashita T, et al. ADAR1 Non-Editing Function in Macrophage Activation and Abdominal Aortic Aneurysm. Circ Res 2023;132:e78-93. [Crossref] [PubMed]
9. Zhu J, Meganathan I, MacAruthur R, et al. Inflammation in Abdominal Aortic Aneurysm: Cause or Comorbidity? Can J Cardiol 2024;40:2378-91. [Crossref] [PubMed]
10. Tsai SH, Hsu LA, Tsai HY, et al. Aldehyde dehydrogenase 2 protects against abdominal aortic aneurysm formation by reducing reactive oxygen species, vascular inflammation, and apoptosis of vascular smooth muscle cells. FASEB J 2020;34:9498-511. [Crossref] [PubMed]
11. Pincemail J, Defraigne JO, Courtois A, et al. Abdominal Aortic Aneurysm (AAA): Is There a Role for the Prevention and Therapy Using Antioxidants? Curr Drug Targets 2018;19:1256-64. [Crossref] [PubMed]
12. Kuivaniemi H, Ryer EJ, Elmore JR, et al. Understanding the pathogenesis of abdominal aortic aneurysms. Expert Rev Cardiovasc Ther 2015;13:975-87. [Crossref] [PubMed]
13. Hou N, Zhou H, Li J, et al. Macrophage polarization and metabolic reprogramming in abdominal aortic aneurysm. Immun Inflamm Dis 2024;12:e1268. [Crossref] [PubMed]
14. Wang J, Da X, Chen Y, et al. Glutamine Protects against Mouse Abdominal Aortic Aneurysm through Modulating VSMC Apoptosis and M1 Macrophage Activation. Int J Med Sci 2024;21:1414-27. [Crossref] [PubMed]
15. Silvestro M, Rivera CF, Alebrahim D, et al. The Nonproteolytic Intracellular Domain of Membrane-Type 1 Matrix Metalloproteinase Coordinately Modulates Abdominal Aortic Aneurysm and Atherosclerosis in Mice-Brief Report. Arterioscler Thromb Vasc Biol 2022;42:1244-53. [Crossref] [PubMed]
16. Le S, Wu J, Liu H, et al. Single-cell RNA sequencing identifies interferon-inducible monocytes/macrophages as a cellular target for mitigating the progression of abdominal aortic aneurysm and rupture risk. Cardiovasc Res 2024;120:1351-64. [Crossref] [PubMed]
17. Haycock PC, Burgess S, Wade KH, et al. Best (but oft-forgotten) practices: the design, analysis, and interpretation of Mendelian randomization studies. Am J Clin Nutr 2016;103:965-78. [Crossref] [PubMed]
18. Kasprzak MP, Gryszczyńska B, Olasińska-Wiśniewska A, et al. Blb-NRF2-PON1 Cross-Talk in Abdominal Aortic Aneurysm Progression. Antioxidants (Basel) 2023;12:1568. [Crossref] [PubMed]
19. Gutowski Ł, Gutowska K, Pioruńska-Stolzmann M, et al. Systems Approach to Study Associations between OxLDL and Abdominal Aortic Aneurysms. Int J Mol Sci 2019;20:3909. [Crossref] [PubMed]
20. Das AA, Waldeck-Weiermair M, Yadav S, et al. Differential aortic aneurysm formation provoked by chemogenetic oxidative stress. J Clin Invest 2025;135:e188743. [Crossref] [PubMed]
21. Liu T, Wang P, Guo C, et al. Screening and investigating the regulatory mechanisms of oxidative stress-related biomarkers in thoracic aortic aneurysms. J Appl Genet 2025; Epub ahead of print. [Crossref]
22. Zhao G, Lu H, Chang Z, et al. Single-cell RNA sequencing reveals the cellular heterogeneity of aneurysmal infrarenal abdominal aorta. Cardiovasc Res 2021;117:1402-16. [Crossref] [PubMed]
23. Cheng Z, Zhou YZ, Wu Y, et al. Diverse roles of macrophage polarization in aortic aneurysm: destruction and repair. J Transl Med 2018;16:354. [Crossref] [PubMed]
24. Song H, Yang Y, Sun Y, et al. Circular RNA Cdyl promotes abdominal aortic aneurysm formation by inducing M1 macrophage polarization and M1-type inflammation. Mol Ther 2022;30:915-31. [Crossref] [PubMed]
25. Wu Y, Jiang D, Zhang H, et al. N1-Methyladenosine (m1A) Regulation Associated With the Pathogenesis of Abdominal Aortic Aneurysm Through YTHDF3 Modulating Macrophage Polarization. Front Cardiovasc Med 2022;9:883155. [Crossref] [PubMed]
26. Li D, Li J, Liu H, et al. Pathogenic Tconvs promote inflammatory macrophage polarization through GM-CSF and exacerbate abdominal aortic aneurysm formation. FASEB J 2022;36:e22172. [Crossref] [PubMed]
27. Cai D, Li C, Zhang Y, et al. CircHipk3 serves a dual role in macrophage pyroptosis by promoting NLRP3 transcription and inhibition of autophagy to induce abdominal aortic aneurysm formation. Clin Transl Med 2024;14:e70102. [Crossref] [PubMed]
28. Wu Z, Yu W, Luo J, et al. Comprehensive transcriptomic analysis unveils macrophage-associated genes for establishing an abdominal aortic aneurysm diagnostic model and molecular therapeutic framework. Eur J Med Res 2024;29:323. [Crossref] [PubMed]
29. Keiper S, Papasaikas P, Will CL, et al. Smu1 and RED are required for activation of spliceosomal B complexes assembled on short introns. Nat Commun 2019;10:3639. [Crossref] [PubMed]
30. Shah VJ, Maddika S. CRL7(SMU1) E3 ligase complex-driven H2B ubiquitylation functions in sister chromatid cohesion by regulating SMC1 expression. J Cell Sci 2018;131:jcs213868. [Crossref] [PubMed]
31. Cai J, Qiu J, Wang H, et al. Identification of potential biomarkers in ovarian carcinoma and an evaluation of their prognostic value. Ann Transl Med 2021;9:1472. [Crossref] [PubMed]
32. Qian M, Liang X, Zeng Q, et al. SMU1 Knockdown Suppresses Gastric Carcinoma Growth, Migration, and Invasion and Modulates the Cell Cycle. Cancer Control 2024;31:10732748241281716. [Crossref] [PubMed]
33. Wang Z, Guo J, Han X, et al. Metformin represses the pathophysiology of AAA by suppressing the activation of PI3K/AKT/mTOR/autophagy pathway in ApoE(-/-) mice. Cell Biosci 2019;9:68. [Crossref] [PubMed]
34. Jia L, Jing Y, Wang D, et al. Through network pharmacology and molecular docking to explore the underlying mechanism of Artemisia annua L. treating in abdominal aortic aneurysm. Front Physiol 2022;13:1034014. [Crossref] [PubMed]
35. Ni XQ, Zhang YR, Jia LX, et al. Inhibition of Notch1-mediated inflammation by intermedin protects against abdominal aortic aneurysm via PI3K/Akt signaling pathway. Aging (Albany NY) 2021;13:5164-84. [Crossref] [PubMed]
36. Xiong JM, Liu H, Chen J, et al. Curcumin nicotinate suppresses abdominal aortic aneurysm pyroptosis via lncRNA PVT1/miR-26a/KLF4 axis through regulating the PI3K/AKT signaling pathway. Toxicol Res (Camb) 2021;10:651-61. [Crossref] [PubMed]
37. Zhao SJ, Kong FQ, Jie J, et al. Macrophage MSR1 promotes BMSC osteogenic differentiation and M2-like polarization by activating PI3K/AKT/GSK3β/β-catenin pathway. Theranostics 2020;10:17-35. [Crossref] [PubMed]
38. Linton MF, Moslehi JJ, Babaev VR. Akt Signaling in Macrophage Polarization, Survival, and Atherosclerosis. Int J Mol Sci 2019;20:2703. [Crossref] [PubMed]
39. Zhou H, Wang L, Liu S, et al. The role of phosphoinositide 3-kinases in immune-inflammatory responses: potential therapeutic targets for abdominal aortic aneurysm. Cell Cycle 2022;21:2339-64. [Crossref] [PubMed]