Machine learning combined with omics-based approaches reveals T-lymphocyte cellular fate imbalance in abdominal aortic aneurysm

Abstract

Background

Abdominal aortic aneurysm (AAA) is typically an asymptomatic disease closely associated with immune mechanisms. A deep understanding of cellular responses within AAA tissues, particularly the molecular changes in T-cell populations, is critical for disease diagnosis and treatment. However, the specific mechanisms inducing T-lymphocyte fate imbalance in AAA remain to be elucidated.

Results

The analysis revealed the core mechanisms driving T-lymphocyte fate imbalance in AAA. We successfully established a comprehensive regulatory map encompassing T-cell infiltration regulatory features, critical transcription factors, and dysregulated immune signaling pathways. Machine learning algorithms identified transcription factors FOSB and JUNB as key biomarkers. Validation across multiple independent datasets and clinical samples confirmed the feasibility and accuracy of FOSB and JUNB as clinical diagnostic biomarkers for AAA.

Conclusions

Through the analysis of single-cell and bulk data, hallmarks of human AAA cellular landscape and T-cell comprehensive developmental relationships were recapitulated. This study identified important roles of T-cell and the molecular mechanisms for the dynamic T-cell infiltrating process, which could characterize disease status and landscape of human AAA microenvironment. Using the deep learning algorithms, FOSB and JUNB were demonstrated as pivotal biomarkers of AAA, together with screening the potential pharmacologic agents targeting T-cell polarization. Taken together, this expands the current understanding of AAA pathogenesis and may provide a feasible immune-targeted therapeutic strategy.

Background

Abdominal aortic aneurysm (AAA) is an age-associated chronic inflammatory aortic disease that is typically induced by smoking, atherosclerosis, and genetic variants. AAA grows slowly without noticeable signs and symptoms until rupture occurs. The mortality associated with ruptured AAA can be as high as 80–90% [1]. In-hospital mortality remains high, despite the progression of surgical interventions such as endovascular and open repairs [2]. The lack of accurate diagnostic biomarkers, combined with limited pharmacological therapies to slow or prevent the progression AAA, contributes to the challenges in AAA clinical management [3, 4]. Therefore, understanding the complicated pathophysiological mechanisms of AAA is essential to the development of diagnostic biomarkers and targeted therapeutics. Previous studies indicate that AAA is induced by complex pathological processes, including active ongoing immune inflammation, oxidative stress, smooth muscle loss, extracellular matrix (ECM) degeneration, and matrix proteinase activation [5, 6]. Recent evidence suggests that AAA is an antigen-driven T-cell autoimmune disease [7, 8]. Both CD4⁺ and CD8⁺ T cells have been reported to play essential roles in promoting AAA formation and growth. In response to various microenvironment stimuli, T cells can be developed into different subsets, such as T helper cells, regulatory T cells (Tregs), and cytotoxic T cells [9]. The imbalance of polarized T-cell phenotypes consequentially exacerbates AAA progression. Therefore, exploring the subpopulations and imbalances in T-cell infiltrates may lead to the development of novel immune-targeted therapeutic strategies [4].

A few previous attempts have identified the predominant types of aortic cells using aneurysmal mouse AAA and single-cell RNA sequencing (sc-RNA seq) technology [10,11,12,13], but the heterogeneous T-cell subpopulations and regulatory mechanisms that dominate T-cell fate during the progression of human AAA remain elusive. Due to the species difference, none of the current animal AAA models reflects precisely the complex pathophysiological environment of human AAA patients. As a result, the comprehensive analysis of human samples and identifying the T-cell interactions in AAA are timely and desirable for the development of new diagnostic and therapeutic applications. As modern approaches to study the cellular state transitions under different pathophysiological conditions, single-cell transcriptomic analysis allowed for an in-depth characterization of complex molecular and functional alterations on a cell-by-cell resolution and artificial intelligence based on machine learning could capture the unpredictability and complexity of human physiology, improving the accuracy of cell analysis. With the help of machine learning algorithms, the most critical characteristics of AAA can be optimally extracted from high-dimensional signatures assessed by single-cell analysis.

Herein, we performed the first comprehensive analysis using human AAA sc-RNA seq integrated with bulk transcriptomic data to provide new insights into the heterogeneous identities, diverse functional states, and immune microenvironment landscape in human AAA aortic wall. We revealed a roadmap of the cellular fate program in T cells as well as dynamic infiltration state-related signature and regulation patterns that may provide potential targets for AAA early intervention. The transcription factors FosB proto-oncogene (FOSB) and JunB proto-oncogene (JUNB) that had been overlooked previously in aneurysm were identified as key biomarkers through multiple machine learning algorithms and validated at both single-cell and population levels. This study might provide novel targets and mechanisms for clinical intervention and therapy in AAA.

Results

Single-cell transcriptome profiles of human AAA

As shown in the analysis process (Fig. 1A), data from 4 aneurysmal abdominal aorta samples in the database were included in this research. Following data integration and quality filtering, 7343 cells were projected onto a UMAP plot based on integrative unsupervised clustering analysis (Fig. 1B), of which 16 populations were obtained and 9 main cell types were identified and assigned to each population based on CellMarker database and SingleR algorithm (Fig. 1C). The levels of cell type-specific canonical markers were presented in Fig. 1D.

Fig. 1

Overview of single cells derived from AAA aortic tissue. A Schematic of the analysis processes. B UMAP visualization of all the single cells of 4 samples from aortic aneurysmal tissue. The different colors indicated different samples (n = 7343 cells). C UMAP visualization of unsupervised clustering identified 16 cell populations and 9 main cell types. D Cleveland dot plot showing single-cell transcript level of cell type-specific canonical marker genes and stacked bar plot displaying the distribution of different cell types. E The histogram showing the cell number and percentage of each cell type in different samples. F The visualization for the cellular cross-talk strength identified T-cell ranked the first for communication weight. G The violin plots of cell cycle scores (G2M score and S score) for each cell type

The highest cellular population was T cells followed by monocytes/macrophages, whereas smooth muscle cell (SMC) and endothelial cell (EC) populations were present in small numbers (Fig. 1E), in agreement with previous findings of cellular responses of AAA cells at the tissue and single-cell level in mice [11, 12]. We applied the cell-chat analysis to infer intercellular signal transduction during AAA progression. T lymphocytes, with the heaviest weights, played central roles in the intersected cellular communication network (Fig. 1F, Additional file 1: Fig. S1D). Cell cycle phase scores were assigned based on the gene expression signatures for the G2, S, and G2/M phases. Immune cells presented high proliferation scores (Fig. 1G), and the highest proliferating cells were observed in the T subpopulation (Additional file 1: Fig. S1F). In contrast, SMCs and ECs showed reduced proliferation activities, evidenced by the lowest proliferation scores (Additional file 1: Fig. S1F).

The top markers (sorted by average Log (fold change)) for a cell type were identified relative to all other types. The identified neural cells featured the high expression of WFDC2, FXYD2, and KLK1 (Additional file 1: Fig. S1A). We demonstrated that the immune cell lineages are comprised: (i) B cell (clusters 2, 5); (ii) CD4 + T (clusters 0, 1, 3); (iii) CD8 + T (clusters 7, 10); (iv) mast cell (cluster 15); (v) NK (cluster 11); (vi) monocyte/macrophage (clusters 4, 6, 8). Moreover, non-immune cell lineages were observed, namely endothelial cell (EC) (cluster 12), smooth muscle cell (SMC) (cluster 13), and neural cell (Schwann cell) (cluster 9) (Additional file 1: Fig. S1B) expressing NGFR associated protein [14]. Furthermore, the significant correlation of the 2000 most variable genes was identified to delineate the relationship among cell clusters in AAA. Histiocytes were clustered together, and the close relationships between B and macrophages as well as T and NK were demonstrated (Additional file 1: Fig. S1C). The functional analysis suggested proliferation was suppressed in histiocytes, and the immune state was activated in immune cell lineages for AAA patients (Additional file 1: Fig. S1E). The sustained immune response inhibited SMC proliferation and upregulated cellular impairment, promoting AAA progression [15]. Moreover, SMC showed high expression of collagen and proteoglycan genes (such as FN1, COL1A2, DCN) (Additional file 1: Fig. S1A). SMC and EC both exhibited cellular damage and collagen-containing extracellular matrix degeneration hallmarks, suggesting they underwent phenotypic transformation in AAA (Additional file 1: Fig. S1E). After hypothetical ligand-receptor pairing among each subpopulation was sorted, a large number of TIMP1-CD63/TIGB1-collagen genes fibrosis axis [16] between T-cell and SMC were revealed (Additional file 1: Fig. S1D), indicating the aggravated microenvironment within the vessel wall. An additional supplementary table (Additional file 2: Table S1) catalogs the identified cell populations and specifies their details.

T lymphocyte transcriptomes revealed phenotypic and functional heterogeneity in AAA

All T lymphocyte populations were isolated and combined (Fig. 2A). Initially, singleR identified 5 clusters of T cells, and unsupervised re-clustering further revealed 8 phenotypes (Fig. 2B). Each phenotype exhibited varying gene expression patterns, and top phenotype-specific markers were presented in Fig. 2C. The activation status propensities were further assessed, revealing 5 CD4⁺ clusters (T_1, T_2, T_4, T_5, T_7) and 3 CD8⁺ clusters (T_3, T_6, T_8) (Additional file 1: Fig. S2A). T_1, T_2, and T_7, with strong correlation (Fig. 2D), were characterized by the upregulated expression of chemokine (C–C motif) receptor 7 (CCR7), lymphoid-enhancer-binding factor 1 (LEF1), recombinant human CD62L/SELL protein (SELL), and transcription factor 7 (TCF7), indicating the stem-cell and central-memory-like phenotypes. T_1 and T_2, with the largest proportion (Additional file 1: Fig. S2B), were related to functions associated with V(D)J recombination and lymphocyte proliferation (Fig. 2E), displaying a naïve state. T_7 mediated important cellular functions such as cell recruitment and differentiation (Fig. 2E). This cellular expressed tumor necrosis factor (TNF) and Th1-related cytokine (Additional file 1: Fig. S2A), representing the Th1-like central-memory subcluster. Based on the expression of specific classical markers forkhead box P3 (FOXP3), interleukin 2 receptor subunit alpha (IL2RA), T cell immunoreceptor with Ig and ITIM domains (TIGIT), and cytotoxic T-lymphocyte associated protein 4 (CTLA4), T_4 was labeled as Treg, whose protective role has been demonstrated in animal models of AAA [17]. Effector-memory Th1-like subcluster was observed in T_5 that expressed C-X-C motif chemokine receptor 3 (CXCR3), C-X-C motif chemokine receptor 4 (CXCR4), C–C motif chemokine receptor 6 (CCR6), and TNF, interferon gamma (IFNG) (Additional file 1: Fig. S2A). T_5 had the most extensive and strong interactions among all T subclusters (Fig. 2D). Because of its high levels of granzyme K (GZMK), granzyme A (GZMA), killer cell lectin like receptor K1 (KLRK1), and CXCR4, T_3 was identified as effector-memory CD8⁺ T cells, featuring cell killing and response to INFG and TNF. Terminally differentiated cytotoxic CD8⁺ T profile was present in T_6 that expressed granzyme B (GZMB), natural killer cell granule protein 7 (NKG7), Fc gamma receptor IIIa (FCGR3A), and fibroblast growth factor binding protein 2 (FGFBP2). T_8, a small pool of cells preferentially releasing noktochor (NKT), has acquired cytotoxicity and killing ability after activation (Fig. 2E, Additional file 1: Fig. S2A).

Fig. 2

Heterogeneity of T lymphocytes in human AAA. A UMAP visualization of T lymphocytes of 4 samples from aortic aneurysmal tissue. Color denoted different samples (n = 3841 cells). B UMAP visualization of singleR roughly identified 5 T subclusters (left), and unsupervised clustering revealed 8 distinct T phenotypes (right). C Heatmap of relative expression of top markers per phenotype. D Cell–cell interaction scores between T clusters. Asterisks corresponded to a significant difference (*, P < 0.05; **, P < 0.01; ***, P < 0.001). E Bubble plot depicting the feature of each phenotype. The abscissa represented the − log₁₀(P) value. F Violin plots showing cell cycle scores for each phenotype. G The iTALK network showing the intersected cellular communication in T populations

Furthermore, T_3 and T_5 exhibited a high proliferating state, indicating that they played a critical role in AAA pathogenesis (Fig. 2F). The intracellular communication network was established to investigate the complex T-cell phenotype interaction, which revealed the strong cross-talk between T_5 and the remaining populations (Fig. 2G). Pathway-integrated analyses demonstrated that T_1 and T_2 were mainly involved in Wnt-signaling and proliferation-associated pathways. NF-ĸB, transforming growth factor beta (TGFB), and mitogen activated kinase-like protein (MAPK) signalings and oxidative-phosphorylation-hallmark were upregulated in T_5. T_3 was implicated in TNF signaling and T_6 showed selective enrichment in fluid shear stress-associated pathway (Additional file 1: Fig. S2C, D). Upregulated TNFRSF4 co-expression in T_4 was detected (Additional file 1: Fig. S2A), which has been reported to be a key negative regulator of Treg [18, 19]. Furthermore, T_4 was significantly involved in P53 signaling and apoptosis pathways (Additional file 1: Fig. S2C). Collectively, these results suggested a potential correlation between the human AAA state and the impairment of immune suppressor ability of Treg.

Trajectory reconstruction recapitulated T-cell fate decision and expression kinetics

To fully understand the dynamics of T-cell infiltration over the disease progression, we performed a pseudotime analysis based on the transcriptional changes during the cell continuous development process using Monocle2 algorithm, with the tendency of transition among different phenotypes revealed (Fig. 3A, Additional file 1: Fig. S3A). Differentiation trajectory revealed a main bifurcation event that led to 2 cell infiltration fates (Fig. 3B). Fate_2 comprised of T_3 and T_6, while fate_1 comprised of T_4 in the pathogenesis of AAA (Fig. 3C). The details of infiltration fates were shown in the supplementary document (result and discussion, Additional file 1: Fig. S3B–D). On the basis of cell development and differentiation, all T-cells can also be categorized into 8 subsets by monocle dimensionality reduction algorithm, supporting our findings that the presence of 8 major different cell phenotypes in heterogeneous T-cell populations during human AAA progression (Additional file 1: Fig. S3B). In AAA, T cells bifurcated into vastly 3 infiltration states at various time points (Additional file 1: Fig. S3C). T_1 and T_2, endowed with high plasticity, primarily dominate the root of chronological trajectory (starting state 3), which was assigned as the least mature cellular state in the pseudotime and precursors of other phenotypes. The remaining phenotypes were predominantly in terminal infiltration states (Additional file 1: Fig. S3D).

Fig. 3

High-dimensional single-cell lineage mapping tracks T phenotype infiltration states. A AAA T cells developmental pseudotime mapping within UMAP plot (n = 3841 cells). B Pseudotime ordering of T cells along a bifurcated developmental trajectory produced by the Monocle algorithm. C Identified T phenotypes contributing to distinct infiltration states were superimposed on the trajectory. D Heatmap displaying expression pattern of branch-specific fate genes, ordered based on their common kinetics through pseudotime. The pseudotime clusters for fate genes were generated using unsupervised clustering. E, F Dynamical expression of representative genes plotted as a function of pseudotime, colored by infiltration state (up) and phenotypes (bottom)

BEAM demonstrated dynamic expression of genes governing diversification of T-cell fates and construction of the Lineage trajectories. Unsupervised clustering of these cell fate genes revealed 3 clusters with different kinetics along the pseudotime (Fig. 3D). Genes in cluster_2 possessed decisive roles in fate_1 specification while cluster_1 determined Fate_2 as infiltration tendency. Assessment of the canonical markers based on BEAM revealed a shift in T cells toward immune-potentiating or immunosuppressive fate with progression (Fig. 3E, F). The proinflammatory interferon response and cytotoxic makers such as C–C motif chemokine Ligand 3 (CCL3), interferon gamma (FNG), interferon regulatory factor 1 (IRF1), perforin 1 (PRF1), GZMA, NKG7, and AHNAK nucleoprotein (AHNAK) were expressed mainly in Fate_2. Inhibitory marker TIGHT and cytotoxic T-lymphocyte associated protein 4 (CTLA4) were expressed only in Fate_1 with a reduced acceleration compared to the Fate_2’s markers. The modest expression of CCR7 and selectin L (SELL) straddled both pre-branch and Fate_1 branches, suggesting a relatively immature state and insignificant activation in T cells at this regulatory cell fate.

AAA immune microenvironment characterization

We used a large number of independent RNA-seq datasets, including all genome enrichment analysis (GSEA) results, to verify scRNA-seq-derived derived hypotheses in subsequent analyses (Figs. 4, 5, and 6, Additional file 1: Figs. S4 and S5). GSEA-calculated statistics revealed significant activation of specific pathways in AAA patients, including “antigen processing and presentation,” “autoimmune thyroid disease,” “T-cell receptor (TCR) signaling,” and “Th1/2 cell differentiation” (Fig. 4A). This result suggested that the specific Ag–driven T-cell autoimmunity is responsible for the pathogenesis. Cardiac muscle contraction, fatty acid elongation, and oxidative phosphorylation were suppressed (Fig. 4A). Disease ontology analysis also unveiled a strong link between AAA and few disease types, including vasculitis, collagen disease, and autoimmune disease (Additional file 1: Fig. S4A). Hallmark signaling computed via GSVA indicated that the immunity/inflammation-related pathways were upregulated while metabolism-related pathways were inhibited. Interferon (IFN), KRAS proto-oncogene (KRAS), TNF, NF-κB, IL6-JAK-STAT3, and IL2–STAT5 were the highly expressed cytokines and signaling molecules (Fig. 4B). Based on the PCA analysis, the immune infiltration landscape revealed a significant difference in the immune status (Additional file 1: Fig. S4B). The outcomes indicated that AAA was associated with elevated infiltration of immunocytes, especially activated T cells. T-helper cells, especially Th1, had a higher and more significant infiltration profile than other immune cells. Infiltration of activated B cell, mast cell, and eosinophil was also found in AAA samples (Fig. 4C, Additional file 1: Fig. S4C). Abdominal aortic wall tissues contained high levels of immune functional factors and enhanced immune response (Fig. 4D, Additional file 1: Fig. S4C). Activated T cells were most likely to be regulated by IFN-gamma-response, TNF-signaling-via-NF-κB, and IL6-JAK-STAT3-signaling (P < 0.001) (Fig. 4E). Genes of major categories, including MHC molecules, immunomodulators/checkpoints (CP), effector cells, and immunosuppressive cells, were assessed to identify immunophenoscore (IPS)—an effective predictor of immune response and therapy (Fig. 4F, Additional file 1: Fig. S4D).

Fig. 4

Immune microenvironment and molecular and functional characteristics for AAA. A Activated or inhibited KEGG pathways of GSEA results in patients relative to controls (AAA n = 80 patients, control n = 10 healthy individuals). B Up- or downregulated hallmark pathways of GSVA result in patients relative to controls. C, D Boxplot illustrating the comparison of infiltrating level of immunocyte subpopulations and immune function scores between AAA and control groups. E Correlation matrix of different immune cells and signaling hallmarks. F Heatmap illustrating the weight distribution from each component of immunophenoscore. G Boxplot illustrating IPS score (up) and IPS z-score (down). H Boxplot illustrating the antigen presentation, effector cell, immunosuppressive cell, and checkpoint scores

Fig. 5

Comprehensive analyses of T lymphocyte infiltration state-related gene set. A Regulatory interaction network governing diversification of cell fates. The circles are represented by TIRS, the left half of which corresponds to cell fates, while the right shows the up-/downregulation in AAA patients relative to controls (AAA n = 80 patients, control n = 10 healthy individuals). B Bar plot showing top terms enriched in up-/downregulated TIRS based on ORA analysis (adj. P value < 0.05). C Circle plot showing the chromosomal locations and expression level of TIRS. D, E PCA score plot based on TIRS signature. F The box plot illustrating significant differences in the PTS of individual status by Mann–Whitney U test (P < 0.001) (high PTS n = 45 individuals, low PTS n = 45 individuals). G Pearson’s chi-squared test demonstrated the percent weight of different status varied significantly between low- and high-PTS group (P < 0.001). H The relationships between TIRS expression and subgrouping based on immune cell infiltration. I Scatter plots revealing negative trend between PTS and IPS, and the relationship between them and immune infiltration. Samples on the abscissa were arranged according to PTS, and the box plot showing strong relationships between PTS grouping and IPS as well as each component

Fig. 6

Identification of TIRS regulatory mechanisms and key biomarkers. A LASSO-based feature selection, with the optimal lambda determined when the partial likelihood deviance reached the minimum value (left). SVM-RFE-based feature selection, with root mean square error (RMSE) reached the minimum value and R-squared reached the max value (mid). Venn diagram presented the intersection of key biomarkers obtained through both algorithms (right). B Aberrant expression profiles for key biomarkers in Abdominal Aortic Wall Dataset 1 (AAA n = 80 patients, control n = 10 healthy individuals; Student’s t-test). C ROC curve demonstrating the diagnostic efficacy of FOSB, JUNB, CST7, and TBC1D4 in Abdominal Aortic Wall Dataset 1. D Clinical impact plot illustrating the clinical utility of key biomarkers. The “Number high risk” curve closely aligns with the “Number high risk with the event” curve at each threshold probability, indicating exceptional predictive power. E Aberrant expression profiles for key biomarkers in Abdominal Aortic Wall Dataset 2 (AAA n = 9 patients, control n = 10 healthy individuals; Student’s t-test). F ROC curve validating the diagnostic efficacy of FOSB, JUNB, CST7, and TBC1D4 in Abdominal Aortic Wall Dataset 2. G Clinical impact plot demonstrating the clinical utility of key biomarkers. Again, the “Number high risk” curve is closely aligned with the “Number high risk with the event” curve at each threshold probability, highlighting the biomarkers’ strong predictive power. H Aberrant expression profiles for key biomarkers in Perivascular Adipose Tissue Dataset 3 (dilated n = 30, non-dilated n = 30; Student’s t-test). I ROC curve verifying the diagnostic efficacy of FOSB, JUNB, CST7, and TBC1D4 in Perivascular Adipose Tissue Dataset 3. J Impact plots reiterated superior predictive performance probability, indicating outstanding predictive capability

Aneurysmal tissue showed a higher IPS than that of control samples (P < 0.01), indicating increased potency of immunogenicity in AAA (Fig. 4G). AAA showed a higher score for antigen presentation and effector cells and lower score for immunosuppressive cells than those of controls, suggesting a disorder of immunity regulation in AAA, especially immunosuppressive cell dysfunction in disease progression (Fig. 4H). This hypothesis agreed with the results from previous analysis that enhanced immunity with deregulated T regulatory function was identified in aneurysm wall microenvironment [20].

Biological implications underlying TIRS

To bring further insight into the role of genes as cell fate determinants, following the significant cellular state-related differential gene set (968 genes, P < 0.0001), BEAM cell fate leader gene set (437 genes), and differentiation-trajectory ordering gene set (375 genes) extracted out, a total of 374 T-cell infiltration state-related genes were obtained which were differentially expressed across pseudotime through the intersection of the above three gene sets. Further, eBayes-moderated ANOVA t-test was conducted on the basis of abdominal aorta wall bulk data. Eventually, the most dysregulated gene set (FDR < 0.05) in aneurysm tissues shared 50 genes overlapping with 374 T-cell infiltration state-related genes (Additional file 1: Fig. S5A), which were integrated as final T-cell infiltration regulatory gene signature (TIRS) (Additional file 1: Fig. S5B).

A coregulatory network was established to map TIRS to bulk-cell level. A high density of interactions was determined among TIRS (P < 0.001), showing the regulatory mechanisms underlying the pathogenesis and T-cell fate based on this unique gene set. The vast majority of TIRS components were co-expressed and upregulated in the aneurysmal region during the progression of infiltrated T-cell cellular differentiation. TIRS that governs the different cell fates was further clustered into 3 modules. Most components (red) were identified as the determinant module responsible for the transition of T-cell into Fate_2 (immune-potentiating). Fate_2 module genes (red) showed more statistically significant expression dysregulation than Fate_1 module genes (orange) (Fig. 5A).

To investigate the functional significance of TIRS, we performed the over-representation analysis (ORA) by applying all MSigDB sets. TIRS was involved in modules related to T-cell datasets, cellular activation and differentiation, autoimmunity, and inflammatory responses (Fig. 5B). Figure 5C illustrates the chromosomal locations and average expression of TIRS in aortic bulk tissue. As an unsupervised machine learning technique, PCA was performed to summarize TIRS into a smaller number of high-order components that outlined significant differences between AAA samples and controls (Fig. 5D). To better understand the correlation between TIRS and patient status, we characterized further the TIRS by extracting significant components to generate a continuous variable PCA-T-score (PTS) for each sample. Integrative analysis proved the significant difference in PTS between healthy controls and disease status despite the different aneurysmal size (P < 0.001), where lower PTS could be an indicator of AAA status (Fig. 5E). Bar plot based on the non-parametric tests further demonstrated the percent weight of different arterial dilation degree between low-PTS group (P < 0.00001) and high-PTS group (P < 0.05). Samples with greater disease severity were primarily distributed in low-PTS group compared to high-PTS group (large size: 44% vs 36%, small size: 42% vs 36%, health sample: 0% vs 22%, P < 0.01), suggesting a potential association between TIRS and individual status (Fig. 5F). Similar findings were also drawn in another aneurysm sample dataset (Additional file 1: Fig. S5C).

A significant connection between PTS and infiltrating degree of multiple immunocyte subpopulations, especially activated CD4⁺T and EM_CD8⁺T in abdominal aorta tissue, was identified (Additional file 1: Fig. S5D). Global TIRS expression levels between high- and low-immune infiltration tissues also showed significant differences (Additional file 1: Fig. S5E, Fig. 5G). The scatter plot further demonstrated that PTS decreased with elevated level of immune infiltration. A noticeable trend for an inverse correlation between PTS and IPS were also identified (Fig. 5H). IPS showed statistically significant differences between the high- and low-score groups (Fig. 5I), highlighting a close association between TIRS and the status of aneurysm immune microenvironment such as antigen recognition/presentation and cellular activation. T-cell activation (the process of proliferation, differentiation, and transformation of T cells stimulated by antigens) increases cell surface PD-1 expression [21]. PD-1 neutralizing antibodies and inhibitors suppressed aortic tissue inflammation and decreased vascular SMC apoptosis and vessel wall calcification in murine AAA, thereby alleviating AAA progression [22]. Here we also showed that PD-1 was expressed significantly higher in the low-PTS group than that in the high-PTS group (P < 0.0001) (Additional file 1: Fig. S5F). Taken together, TIRS may correlate with the therapeutic response to immunotherapy and serve as a promising pathological biomarker in AAA.

Detection and evaluation of key biomarkers

TIRS was subjected to the machine learning-based process for feature selection. Taking the intersection of LASSO and SVM-RFE, 4 key biomarkers were identified, including FOSB, JUNB, cystatin F (CST7), and TBC1 domain family member 4 (TBC1D4) (Fig. 6A). The expression dysregulations of these key biomarkers were verified in the abdominal aortic wall, perivascular adipose tissue from AAA patients. In Abdominal Aortic Wall Dataset 1, the expressions of FOSB, JUNB, TBC1D4, and CST7 were significantly higher in the abdominal aortic wall compared to the control aortic wall (Fig. 6B). The biomarker discrimination efficacy was evaluated using ROC curves, with AUCs for FOSB, JUNB, CST7, and TBC1D4 recorded at 0.911, 0.917, 0.926, and 0.955, respectively (Fig. 6C). The clinical impact curve (CIC) suggested that AAA patients could benefit from FOSB, JUNB, and a two-gene combination at high-risk thresholds from 0 to 1 (Fig. 6D). In Abdominal Aortic Wall Dataset 2, similar results were observed, with higher expressions of FOSB, JUNB, TBC1D4, and CST7 compared to the control aortic wall. The AUCs were 0.982 for FOSB, 0.911 for JUNB, 0.893 for CST7, and 0.857 for TBC1D4. The CIC also indicated potential benefits for AAA patients from FOSB, JUNB, and the two-gene combination at high-risk thresholds (Fig. 6E–G). In Perivascular Adipose Tissue Dataset 3, FOSB and JUNB expressions were elevated in dilated perivascular adipose tissue compared to non-dilated tissue, while CST7 and TBC1D4 showed no significant differences (Fig. 6H). AUCs for FOSB, JUNB, CST7, and TBC1D4 were 0.956, 0.887, 0.506, and 0.514, respectively (Fig. 6I). The CIC indicated that AAA patients could also benefit from FOSB, JUNB, and the two-gene combination at high-risk thresholds in this dataset (Fig. 6J).

We also confirmed the dysregulation of key biomarkers in clinical samples from AAA patients, analyzing both abdominal aortic wall and peripheral blood (Fig. 7A). In the Abdominal Aortic Wall Inhouse Dataset 1, FOSB and JUNB levels were significantly higher in the AAA aortic wall compared to the control, while CST7 and TBC1D4 showed no significant difference (Fig. 7B). The AUCs for FOSB, JUNB, CST7, and TBC1D4 were 1.000, 1.000, 0.600, and 0.800, respectively (Fig. 7C). The clinical impact curve (CIC) indicated that AAA patients could benefit from FOSB, JUNB, and the two-gene combination at high-risk thresholds from 0 to 1 (Fig. 7D). In the Peripheral Blood Inhouse Dataset 2, FOSB, JUNB, CST7, and TBC1D4 expressions were elevated in AAA patients compared to controls (Fig. 7E). The AUCs for these biomarkers were 0.989 for FOSB, 0.906 for JUNB, 0.933 for CST7, and 0.736 for TBC1D4 (Fig. 7F). The CIC also suggested potential benefits for AAA patients from FOSB, JUNB, and the two-gene combination at high-risk thresholds in this dataset (Fig. 7G). Overall, individuals with elevated FOSB and JUNB levels may possess a high risk of AAA onset or progression and require special attention and timely interventional therapy. We tested this hypothesis in a mouse model of AAA. We produced mouse AAA model using the BAPN + angII method. The gross anatomical images showed obvious production of auxiliary aortic aneurysm. We confirmed the existence of AAA using the hematoxylin and eosin (H&E) or Verhoeff’s dye liquor and Van Gieson’s dye liquor (EVG) staining. H&E staining results showed that the AAA model group mice had abdominal aorta whose three-layer structure of the tube wall is disordered, severely damaged, with inflammatory cell infiltration and endometrial hyperplasia. H&E and EVG stainings displayed elastic fibers in the abdominal aortic aneurysm wall of AAA model group mice obvious degradation and appear fracture (Fig. 7H, I). Using immunohistochemistry to measure FOSB and JUNB signals, we found that the expression of FOSB and JUNB in AAA mice increased compared with those in control mice (Fig. 7J–L). ELISA results also showed that serum FOSB and JUNB were higher in AAA than the control group (Additional file 1: Fig. S6A, B).

Fig. 7

Verification of key biomarkers. A Abdominal aortic wall and peripheral blood samples obtained from AAA patients. B Aberrant expression profiles for key biomarkers in the abdominal aortic wall (Inhouse Dataset 1; AAA n = 5 patients, control n = 4 healthy individuals). C ROC curve validating the diagnostic efficacy of FOSB, JUNB, CST7, and TBC1D4 in the abdominal aortic wall (Inhouse Dataset 1). D Clinical impact plot demonstrating the clinical utility of key biomarkers. The “Number high risk” curve closely aligns with the “Number high risk with the event” curve at each threshold probability, indicating exceptional predictive power. E Aberrant expression profiles for key biomarkers in peripheral blood (Inhouse Dataset 2; AAA n = 24 patients, control n = 15 healthy individuals). F ROC curve validating the diagnostic efficacy of FOSB, JUNB, CST7, and TBC1D4 in peripheral blood (Inhouse Dataset 2). G Clinical impact plot illustrating the clinical utility of key biomarkers. Again, the “Number high risk” curve remains closely aligned with the “Number high risk with the event” curve at each threshold probability, highlighting the biomarkers’ strong predictive capability. H Mice were infused with saline or Ang II (1000 ng/kg/min) + BAPN. Gross abdominal aorta images were shown. Scale bar is 1 cm. I Representative images of immunohistochemical stains for elastin fiber (Van Gieson) and representative photomicrographs of hematoxylin and eosin (H&E) staining. Scale bar is 200 μm. J–L Representative immunohistochemical staining of FOSB and JUNB in aortic cross sections. Scale bar is 50 μm. Data are expressed as mean ± SEM (control n = 3 mice, AAA n = 5 or 6 mice). Student’s t-test was utilized to compare continuous variables between the two groups

Discussion

AAA is an autoimmune disease where T cells infiltrate to the lesions following microenvironmental stimulation and contribute to AAA formation and progression [8]. It is well documented that T cells with unbalanced phenotypes appear in the aneurysmatic vascular wall, but the respective proportion, function, and activation status remain elusive [23]. Given the fact that there is currently no effective pharmacological therapies to prevent AAA progression [4], understanding the role of bioactive molecules governing T-cell fate trajectory may help identify novel therapeutic targets and develop new intervention strategies. To date, limited single-cell sequencing studies have been conducted on aneurysmal samples. The existing data from AAA patients and mouse AAA model [11,12,13, 24, 25] lacks in-depth analysis of T-cell population heterogeneity and the genetic mechanism underlying divergent cellular fate decisions.

Our results indicated that T cells serve as the center of cellular communication network in human AAA lesion aortic wall, where the highest degree of infiltration of T cells and macrophages and the lowest level of SMCs and ECs were identified. Prior study showed that AAA formation required the presence of CD4⁺ T cells [26]. CD4⁺ T-cell density correlated negatively with vascular wall collagen and elastin contents, and CD4⁺ T-cell number increased with aneurysm size, suggesting a role for T cells in AAA lesion instability [27].

A strong communication was concluded between T cells and SMCs via multiple signaling axes, such as TIMP metallopeptidase inhibitor 1 (TIMP1)-CD63, major histocompatibility complex, class I, B (HLA_B)-CD3 delta subunit of TCR complex (CD3D), CXCR4), and collagen molecules-integrin subunit beta 1 (ITGB1), indicating the profound interactions induced by autoimmunity (Fig. S1D). Indeed, T-cell cytokines contribute to SMC death [28] and TIMP1 modulates myocardial fibrosis by mediating CD63 activity in de novo collagen synthesis [16], suggesting a relationship between T-cell and myofibroblast activation in AAA [16]. Cell cycle and functional analyses showed that cell proliferation was suppressed in histiocytes but increased in T cells, consistent with previous finding of significant T-cell infiltration and clonal expansion in AAA lesions [7].

Eight distinct subpopulations were identified from the T lineages, characterizing the transcriptional profiles and their original and functional heterogeneities (Table 1). The majority of cells were in the naïve or resting state. However, 2 distinct subpopulations (T_3 and T_5) exhibited the highest score of cell–cell cross-talk among all clusters. T_3 and T_5, characterized by a specific high level of IFNG and TNF, could promote the proinflammatory milieu and inhibit collagen synthesis, further enhancing the continuous dilation and aneurysm rupture [29]. The cytotoxic mediators were colocalized to T_6, which contributed to the apoptotic death of histiocytes and the impaired repair and maintenance of the arterial ECM [30]. Limited anti-inflammatory Th2 signaling was detected, suggesting the predominance of proinflammatory Th1-like cells and minimal Th2 involvement.

Table 1 Eight distinct subpopulations of T cells in AAA

The trajectory analysis further revealed the signatures of T-cell heterogeneity, which were mapped to 3 infiltration states and 2 fates (immune-potentiating: Fate_2 and immunosuppressive: Fate_1) along with disease progression. The extent of proinflammatory and cytotoxic genes was significantly upregulated in FATE_2 as compared to the regulatory genes in FATE_1. Treg-like T_4, dominating in FATE_1 end-stage, exhibited a high degree of the negative regulator and apoptosis signaling, preventing the programming of such immune-modulatory cells into the resolution of inflammation under this infiltration state [18, 19]. Modest expression of initial cell state-related genes straddled FATE_1, suggesting a relatively potential immature state of T cells. The polarization in T-cell subset balance occurs during chronic inflammatory AAA progression, leading to the loss of Treg function and the upregulation of proinflammatory-like phenotypes. The results from our bulk-cell immunological analysis supported the observations from the single-cell analysis, where enhanced immunogenicity, higher immune activation, and stronger response to immunotherapy were identified in AAA aortic tissue compared with control aortic tissue, indicating the immunomodulatory dysfunction in human AAA lesion.

Developing new strategies that target cellular phenotype and function may provide novel therapeutic targets for AAA. Therefore, we assessed TIRS underlying the dynamic program of T-cell infiltration state. Subsequent comprehensive analyses at the bulk level demonstrated a strong connection between TIRS and disease status as well as immunotherapy response. We identified benfotiamine as a potential target-specific agent, a vitamin B1 analog that significantly reversed the aberrant expression pattern of molecules involved in a multicomponent coregulatory network centered on TIRS, and blocked AAA T-cell polarization to an inflammatory-potentiating state. Previous studies reported that benfotiamine modulated several immunocyte activities, including microglial cells, dendritic cells, and macrophages, and shifted these cells from proinflammatory toward quiescent cell state under immunological complication states [31,32,33,34].

Early diagnosis of AAA before rupture will reduce the risk of patient sudden death, but remains a clinical challenge. Current detection and risk assessment rely on imaging and morphological features and lack effective pathological biomarkers. Here we utilized multiple machine learning algorithms to screen key biomarkers for AAA based on the T-cell infiltration signature. Incorporation of machine learning facilitated the recognition of the most critical biological features from high-dimensional signatures that may improve the performance of precision diagnosis and therapy. Our identified biomarkers FOSB and JUNB are highly expressed in the arterial wall, PVAT, and peripheral blood with convenient accessibility. These biomarkers are associated with various AAA immune signalings such as TCR-calcium pathway and antigen presentation. We also revealed the kinetic pattern of FOSB and JUNB over the AAA developmental timing, which induced a T-cell inflammatory state in AAA. The transcriptional activity and abundance of FOSB and JUNB were dependent on the cell type and differentiation state [35, 36]. The dimeric transcription factor activator protein-1 (AP-1) participates in lymphoid cellular inflammation, differentiation and proliferation, and can be activated by inflammatory cytokines and antigenic pathogen, leading to adaptive immunity [37]. Stimulant-induced T-cell activation is accompanied by activation of AP-1 immediate genes. AP-1 dimeric complexes composed of FOSB and c-JUN that are detectable in primary activated T cells upon TCR/CD3 stimulation [38]. Inappropriate activation of AP-1 genes contributed to T-cell transformation and dysregulated phenotype in adult T-cell leukemia [37]. Here we demonstrated that FOSB and JUNB exhibited a close correlation with proinflammatory signalings of IFN response, TNF, and NF-κB pathways that have been studied as AAA progression biomarkers [4]. T-cell-derived IFN and TNF cytokines are known to promote aortic dilatation in experimental AAA [26]. However, it is important to note that while FOSB and JUNB demonstrate specific associations with T-cell functional subsets in AAA in this study, their upregulation as AP-1 family members in other immune-related diseases (e.g., cancer and cardiovascular diseases) necessitates cautious interpretation of their diagnostic value, which is primarily considered within the context of serving as screening adjuncts rather than standalone diagnostic biomarkers and should be integrated into multimodal diagnostic frameworks to avoid overemphasis on single indicators.

Benfotiamine targets IFN and TNF by its activity to downregulate the release of master cytokines TNF and interleukins by blocking the NF-κB pathways in several inflammatory cells [31, 34]. Pretreatment of benfotiamine prevented antigen-induced activation of FOS and JUN [33]. These findings and results from this study demonstrated that benfotiamine exhibited therapeutic effect in targeting identified proinflammatory phenotypic switching-related signature. Vitamin B1 supplements are safe for human use, supporting the possibility of using benfotiamine to target T cells in AAA therapy. Li et al. reported that the increase in dietary intake of vitamin B1 is significantly associated with a decrease in the risk of severe abdominal aortic calcification [39]. This finding further supports the potential role of vitamin B1 metabolism in aortic pathologies, while experimental evidence for the direct involvement of vitamin B1 deficiency in the pathogenesis of AAA remains limited. Benfotiamine’s efficacy as a therapeutic agent for AAA remains to be validated through rigorous preclinical studies and clinical trials.

Several limitations exist in this study. Information and procedure bias from a computational biology approach are inevitable, potentially limiting the ability to fully recapitulate the developmental trajectories and diversities of T-cell infiltration states. Inclusion of functional experiments may help understand the pathogenic role of identified infiltration state driver signature and the regulatory mechanism. The challenge in sample collection led to the limited number of standard-compliant data. Clinical variations with larger samples for key biomarkers merit future exploration. Further in vitro and in vivo experimental validation to test the viability of TIRS-specific agents will also serve as future research directions.

Conclusions

Through the analysis of single-cell and bulk data, hallmarks of human AAA cellular landscape and T-cell comprehensive developmental relationships were recapitulated. This study identified important roles of T-cell and the molecular mechanisms for the dynamic T-cell infiltrating process, which could characterize disease status and landscape of human AAA microenvironment. Using the deep learning algorithms, FOSB and JUNB were demonstrated as pivotal biomarkers of AAA, together with screening the potential pharmacologic agents targeting T-cell polarization. Taken together, this expands the current understanding of AAA pathogenesis and may provide a feasible immune-targeted therapeutic strategy.

Methods

Data collection

For bulk-cell level, our study retrospectively recruited 4 independent AAA datasets from the Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/) database, including GSE98278 (aneurysmal abdominal aortic wall tissues from 31 patients) [40], GSE57691 (aneurysmal abdominal aortic wall tissues from 49 AAA patients and nonaneurysmal abdominal aortic wall tissues from 10 control individuals) [41], GSE7084 (aneurysmal abdominal aortic wall tissues from 9 AAA patients and nonaneurysmal abdominal aortic wall tissues from 10 control individuals) [42], and GSE119717 (30 perivascular adipose tissues (PVATs) around AAA and 30 PVATs around not-dilated abdominal aorta) [6]. Among these datasets, a total of 80 samples, generated by integrating the GSE57691 and GSE98278, were denoted as dataset1; GSE7084 as dataset2; GSE119717 as dataset3. The bulk raw data were all processed and corrected using limma and sva packages [43, 44]. The Ensembl database was used to obtain gene annotations for each probe set. If multiple probe sets correspond to the same gene, the probe set with the highest mean intensity across all samples was retained. Single-cell RNA-sequence (scRNA-seq) data of AAA patients’ abdominal aortic tissues from aneurysmal (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE166676) was downloaded from the GEO [45]. The list of TFs was derived from the Cistrome database (http://cistrome.org/) [46]. C1–C8 and hallmark datasets were retrieved from the Molecular Signatures Database (MSigDB) (https://www.gsea-msigdb.org/gsea/msigdb/index.jsp).

Single-cell RNA-seq data processing

All single-cell samples were harvested at the time of surgical intervention. Samples were minced, digested in 0.2% collagenase II (Life Technologies) and 0.2% collagenase V (Sigma) in plain medium for 1 h at 37 °C, and strained through a 70-μM mesh. The preliminary data processing of single-cell RNA-seq data started from the Cell Ranger Single Cell Software Suite 3.3.1 (http://10xgenomics.com/). The pair-ended reads fastq files were trimmed to remove template switch oligo (TSO) sequence and poly-A tail sequence. Then, command of “cellranger count” was used to quantify the clean reads, aligned to the GRCH37 human genome. The quantified gene expression matrices were analyzed using Seurat pipeline (version: 4.1.0; https://satijalab.org/seurat/) for further processing [47]. Briefly, cells with more than 25% of reads from mitochondria genes, and less than 200 or more than 2500 genes were removed while genes expressing in more than 3 single cells were included into further analysis. The “IntegrateData” function was used to integrate all samples into one Seurat object, which was then scaled and standardized using the “ScaleData” function. Moreover, top 2000 variable genes were initially identified via “vst” selection, considered the input features for dimensionality reduction using principal component analysis (PCA). To reduce dimensionality, the first 20 significant PCs determined by jackstraw analysis were incorporated into graph-based Louvain clustering and Uniform Manifold Approximation and Projection (UMAP) analysis.

Differential expression analysis

For single-cell data, the findAllMarkers function with “wilcox” method was performed to identify differentially expressed genes (DEGs) from top 2000 variable genes. Only genes with log2Fold Change (FC) > 1.0 or < − 1.0 and false discovery rate (FDR) < 0.05 were defined as cluster markers. The “limma” package was used for DEGs and DETFs between AAA samples and controls based on microarray data, with |Log2 FC|> 1.0 and FDR < 0.05 as the screening criteria [43, 48].

From clustered cell mapping to corresponding cell types

To identify potential AAA cell populations within unsupervised clusters, DEGs of all subclusters were used as the potential reference that was combined with canonical cell type-specific surface markers derived from CellMarker for comprehensive annotation of cell type [49]. Moreover, the computational tool SingleR was also pursued to confirm the inferred cell types in an unbiased fashion [50]. The known cell surface biomarkers of T cell (CD3D, CD3E, CD8A), B cell (CD79A, CD79B, MS4A1), monocyte/macrophage (LYZ, CD68, CD14), neural cell/Schwann cell (NGFR), smooth muscle cell (TAGLN, ACTA2, CALD1), mast cell (KIT, HDC, TPSAB1), endothelial cell (VWF, CD34, FABP4), and NK cell (FGFBP2, KLRF1, NKG7) were selected for annotation of the cells aforementioned.

Integrated analyses for single-cell data

The discrimination analysis and quantification were performed through Cell Cycle Scoring function based on previously defined cell cycle-related genes, with cells projected onto UMAP space for visualization and colored according to cycle clustering for visualization. Next, to elucidate potential cellular communication patterns and ligand-receptor pairs among various cell clusters in AAA tissues, we utilized CellChat analysis [51] combined with the iTALK approach [52] to conduct cellular communication analysis. Different algorithms ensure the accuracy and reliability of the results. To analyze the heterogeneity of infiltration state in AAA T lymphocyte lineages, Monocle2 [53, 54] was carried out to identify the translational relationships among T clusters. In summary, genes for trajectory ordering were filtered from the top genes differentially expressed among T cell subclusters using the differentialGeneTest function. Following the selection of top 2000 ordering genes through DDRTress algorithm (q < 0.001), single cells were projected onto the lower dimensional space reduced from expression profiles and ordered along pseudotime with “reduceDimension” function. To further detect leader genes playing essential roles in T-cell fate decisions, branched expression analysis modeling (BEAM) from Monocle2 was carried out to identify genes with branch-dependent expression kinetic and represented with the “plot_genes_branched_heatmap” function. Prior to the identification of final TIRS, to minimize the noise-induced error and improve the practicability and operability of biomarkers, features were further screened by integrating multiple algorithms. We retained only molecules as IRSGs that satisfied the conditions as follows: (1) genes used for ordering cells along the trajectory from DDRTress algorithm (q < 0.001); (2) differential genes among different infiltration states (q < 0.001); (3) cell fate driver genes identified through BEAM analysis; (4) genes that included in top 2000 variable genes among different T-cell phenotypes; (5) genes that significantly differentially dysregulated in abdominal aorta wall between AAA patients and healthy controls from eBayes test. T-cells were scored for the activity of key signaling pathways highly associated with key biomarkers using the UCell algorithm [55, 56].

Biological function analysis

Biological functions and signaling pathways enrichment analyses were performed using the “cluster Profiler” R package with thresholds of P < 0.01 and FDR < 0.05 [57]. Gene sets comprising C1–C8 and hallmark were downloaded from the MSigDB database (http://software.broadinstitute.org/gsea/msigdb) [58, 59], where the functional characteristics of TIRS were identified by performing gene sets over-representation analysis (GSORA) [60]. Moreover, GSVA and GSEA analyses were conducted to reveal the hallmark signaling and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways [61,62,63].

Identification of immune microenvironment characterization

The immune infiltration level and immune function activity in each sample were evaluated through the ssGSEA algorithm based on the expression levels of immune-specific marker genes [62]. The marker genes were obtained from a previously published work [64]. Differences in immune landscape between AAA patients and healthy controls were calculated with a significance criterion of P < 0.05. Correlations between infiltration levels by different immune cell types were identified using Pearson’s correlation. The immunophenoscore (IPS) was identified to measure the immune state and immunotherapy response for each sample [65]. IPS-related genes, classification, and associated weights were obtained with the available R-script deposited on GitHub (https://github.com/icbi-lab/Immunophenogram). With a panel of marker genes regarding immune response or immune toleration, we quantified 4 different immunophenotypes (antigen presentation molecules, effector cells, suppressive cells, and selected immunomodulators) and also constructed immunophenogram for each sample. The z-score was generated to summarize 4 classes, and the higher z-score of IPS represented the more immunogenic the sample.

Principal component analysis (PCA)

PCA was used to achieve dimension reduction of all samples for further analysis. Based on 28 types of immune cells, we performed PCA to identify the difference in immune infiltration level between AAA samples and healthy controls. To explore the molecular pathological characteristics of TIRS, we calculated the top 2 principal components of TIRS as the signature score of each sample: PCA-T-score (PTS = ∑PC1 + ∑PC2). The correlations between TIRS signature and disease status as well as immune microenvironment were further revealed.

Clinical samples

The present study was endorsed by the Ethics Committee of the First Affiliated Hospital of Zhengzhou University (2021-KY-1260–002). This research was conducted in accordance with the 2008 Declaration of Helsinki. All subjects signed informed consent documents. A total of 39 peripheral venous blood samples were collected from 24 AAA inpatients and 15 control healthy individuals (peripheral blood sample, Inhouse dataset2). Frozen human AAA lesion specimens were obtained from 5 patients. Control specimens of healthy abdominal aortas were obtained from 4 donors via autopsy with causes of death unrelated to aortic disease, as previously described in our referenced methodology [66]. Discarded and decoded human aortas were reused according to the protocol 2010P001930 pre-proved by the Human Investigation Review Committee at the Brigham and Women’s Hospital, Boston, MA.

Quantitative real-time PCR

qRT-PCR was performed to determine the expression levels of FOSB, JUNB, CST7, and TBC1D4 in tissue and blood samples. Total RNA was isolated from human abdominal aortic wall tissues and peripheral blood samples using RNAiso Plus (Takara, Dalian, China) according to the manufacturer’s protocol. The RNA quality was assessed using a NanoDrop One C (Thermo Fisher Scientific, Waltham, USA) ultramicro UV spectrophotometer, and the RNA integrity was evaluated based on agarose gel electrophoresis. cDNA was synthesized using HiScript III RT SuperMix for qPCR (Vazyme Biotech, Nanjing, China). The product was immediately stored at − 80 °C until analyses. The qRT-PCR was performed with ChamQ Universal SYBR qPCR Master Mix (Vazyme Biotech, Nanjing, China) on a QuantStudio 3 Real-Time PCR System (Applied Biosystems, Foster City, USA). Internal reference gene GAPDH was employed. qRT-PCR assays were performed in triplicate with the following conditions: (1) 95 °C for 30 s, (2) 40 cycles of 95 °C for 10 s and 60 °C for 30 s, (3) 95 °C for 15 s, 60 °C for 60 s, 95 °C for 15 s. Human abdominal aortic wall tissues were homogenized using TissueLyser II (QIAGEN) according to the manufacture’s instruction. Total RNA was isolated using TRIzol® reagent (Cat#15,596,018, Thermo Fisher Scientific) from homogenized tissues. High-Capacity cDNA Reverse Transcription Kit (Cat#4,368,813, Thermo Fisher Scientific) was used to generate cDNA. The relative mRNA levels of target genes were quantified using the iTaq UniverSYBR Green SMX 5000 (Cat#1725125, Bio-Rad) with an ABI PRISM 7900 Sequence Detector system (Applied Biosystems Co, Foster City, CA) following the manufacturer’s instructions. The ΔCT (Ct mRNA − Ct GAPDH) method was used to calculate mRNA relative expression. The relative quantification values for mRNA were calculated by the 2 − ΔΔCt method (Table 2).

Table 2 The forward and reverse primers for qRT-PCR

Screening and verification of key biomarkers

Least absolute shrinkage and selection operator (LASSO) logistic regression and support vector machine-recursive feature elimination (SVM-RFE) were conducted to perform feature selection to screen AAA key biomarkers from TIRS based on the Cohort1. LASSO regression has a contractile penalty function on variables that induces sparsity of predictors in the expression profile. The “glmnet” package was utilized to fulfill this purpose [67,68,69]. Moreover, SVM-RFE, a machine learning method based on a support vector machine, was utilized to identify the best variables via deleting SVM-generated eigenvectors. SVM module was constructed to further screen the diagnostic value of these biomarkers in AAA through “e1071” package [70, 71]. Eventually, the genes from both machine learning algorithms were combined for further analysis. The diagnostic and predictive value of biomarkers was estimated by receiver operating characteristic (ROC) curve and area under the ROC (AUC) in the multiple cohorts (dataset1 (GSE57691 + GSE98278), dataset2 (GSE7084), dataset3 (GSE119717), Inhouse dataset1, Inhouse dataset2). The decision curve analysis (DCA) and clinical impact curve (CIC) were performed to assess the clinical net benefit of biomarkers.

Establishment of abdominal aortic aneurysm model

Referring to the method in the previous paper [72], in simple terms, 10–12 week old C57BL/6 mice (purchased from Weitong Lihua Experimental Animal Technology Co. Ltd (Beijing, China)) were anesthetized by isoflurane inhalation (induction dose: 3–5%, maintenance dose: 1–2%) and then osmotic mini-pumps (Alzet micro-osmotic mini pump 1004) containing AngII were implanted subcutaneously into the dorsum of the mouse. The mice were administered daily at a rate of l μg/kg/min through the drug pump buried in their backs for 4 weeks. At the same time, during the first 2 weeks of pumping angiotensin, BAPN was administered to mice in the surgery group by feeding them water containing 0.625 mg/mL BAPN. The sham surgery group drinking normal water was used as the control group. After study, all mice were sacrificed with an inhalation overdose (5%) of isoflurane and cervical dislocation and then abdominal aortas were collected. Serum samples from mice were collected and cryopreserved at − 80 °C for downstream analyses including ELISA. All mice operations complied with the Institutional Animal Care and Use Committees of Zhengzhou University and followed the National Institutes of Health and US Department of Agriculture Guidelines for Care and Use of Animals in Research (ZZU-LAC20240906 [15]).

ELISA

Serum FOSB and JUNB concentrations were quantified according to the manufacturer’s protocols for the ELISA kits (FOSB: SAB #EK11596; JUNB: BIOESN BES4857K). The assay procedure involved sequential addition of standards, test samples, primary antibodies, and HRP-conjugated streptavidin to microplate wells, followed by incubation at 37 °C. Chromogenic and stop solutions were then introduced stepwise. Optical density values were determined using a microplate reader, with protein concentrations derived from standard curve calculations.

Histology and immunohistochemistry

The mouse abdominal aorta was fixed with 4% PFA, dehydrated, and embedded in paraffin. Slice (5 μm), remove wax, stain with hematoxylin and eosin (H&E), or Verhoeff’s dye liquor and Van Gieson’s dye liquor (EVG) to observe the structure. The expression of FOSB and JUNB in tissue sections was evaluated by immunohistochemical staining as previous studies [73, 74]. Simply speaking, the slices were rehydrated and blocked with 5% goat serum for 30 min. The primary antibodies against FOSB (1:500, catalog No. ab184938, Abcam) and JUNB (1:50, catalog No. 10486–1-AP, Proteintech) were incubated overnight at 4 °C, followed by incubation with horseradish peroxidase conjugated secondary antibodies.

Images were captured using a microscope (Nikon).

Statistical analysis

In the present work, two-tailed P value < 0.05 and FDR < 0.05 was suggested to be statistically significant. Mann–Whitney U test or Student’s t-test was utilized to compare continuous variables between the two groups when appropriate. Pearson or Spearman correlation analysis was utilized when applicable to identify bivariate relationships. For descriptive statistics, the mean ± standard deviation was used for continuous variables with a normal distribution [75]. Moreover, the median (range) was used only for continuous variables with an abnormal distribution. All data processing, statistical analysis, and plotting were conducted with R 4.1.3 software (Institute for Statistics and Mathematics, Vienna, Austria; www.r-project.org).