Introduction

Aortic dissection (AD) is a life-threatening cardiovascular emergency associated with a high mortality rate. Up to 50% of patients may die within the initial 48 h without prompt treatment. Early misdiagnoses occur in approximately 25%–50% of cases due to the variable and non-specific clinical presentation of AD1,2. Despite substantial advancements in early detection and treatment of AD over the last decade, the overall survival rate remains limited, and an AD diagnosis is correlated with a poor prognosis3,4. While biomarkers, including tumor necrosis factor (TNF), C-reactive protein (CRP), and plasma D-dimer, demonstrate a certain degree of sensitivity in diagnosing AD, their absence of specificity renders them ineffective for early detection. As the current situation lacks effective early diagnostic tools, exploring more specific molecular biomarkers for AD diagnosis is urgent and crucial to developing targeted AD therapies.

Concurrently, plasma secretory proteins, released into the bloodstream from locally affected sites in the aorta or impacted endocrine organs, perform diverse functions, including signaling5,6, immune response7,8, inflammation9, and tissue repair6,10. These proteins encompass hormones, enzymes, blood proteins such as albumin and globulins, clotting factors, and cytokines. Notably, specific plasma secretory proteins, including TNF and CRP, play a crucial role in maintaining physiological homeostasis and are extensively employed as biomarkers for diagnosing and monitoring disease activity7 and assessing treatment efficacy in cardiovascular diseases, inflammatory conditions, and cancers11. Exploring blood diagnostic biomarkers has opened up new avenues for comprehending this condition beyond conventional approaches. Recent research has increasingly focused on plasma proteins to differentiate patients with AD from those without AD, especially in cases exhibiting atypical symptoms, and investigate their functions in cellular communication. This study aimed to identify biomarkers that can serve as screening tools for high-risk populations or indicators of prognosis. This shift towards investigating biomarkers in the blood aims to supplement imaging diagnostics and provide deeper insights into AD pathophysiology.

Increasing evidence from genome-wide association studies (GWAS) suggests a substantial correlation between genetic factors and the occurrence of non-familial AD. Identifying single nucleotide polymorphisms (SNP) in genes such as TSHA4, LRP112, and FBN113 are related to the disease. These genetic variants occur in noncoding regions, likely influencing AD via gene expression modifications14. Mendelian Randomization (MR) analysis, using genetic variants as instrumental variables (IVs), investigates potential causal relationships between exposures and outcomes in reducing confounding and reverse causation. It is superior to traditional methods in disease etiology studies15,16,17. The Summary databased MR (SMR) approach combines cis-expression quantitative trait loci (cis-eQTL) and GWAS data to identify gene expressions with pleiotropic effects or causal relationships with diseases such as cardiovascular disorders and systemic lupus erythematosus18,19,20,21, and highlighting its utility in the study of complex trait pleiotropy.

This study takes a comprehensive approach, integrating data from aortic eQTL and GWAS to pinpoint genes that may be pathogenic and related to AD. Following that, the human secretory proteome expression spectrum is used to identify plasma secretory proteins associated with AD, implying a possible causal link between AD and these proteins. An independent single-cell expression spectrum dataset was employed, examining various cell types in both AD and normal tissue to validate the differential expression of these proteins. Furthermore, clinical samples were used to determine the differential expression of plasma secretory proteins in tissues and plasma from healthy control (HC) and AD groups. The findings of this study highlight the potential of novel plasma secretory proteins as biomarkers for diagnosing AD, providing new avenues for further exploration and understanding of the diagnostic mechanisms associated with AD.

Results

Comprehensive GWAS and eQTL analysis for identifying AD-related plasma secretory proteins

Our entire research workflow is outlined in Supplementary Fig. S1.

After a rigorous screening process, SNPs demonstrating significant associations (P < 5 × 10 − 8, F-value > 10) with the exposure and maintaining independence (r2 < 0.001, within a 10,000 kb physical window) were selected (Fig. 1A). 8 secretory proteins related to AD were discovered after identified 194 genes associated with AD by MR analysis and linked them with 1039 human plasma proteins, (Fig. 1B). Genes such as DBNL, NPC2, SUMF2, and TFPI were identified as positive causal effects genes (Fig. 2B,C,E,G) using the inverse variance weighted (IVW)22 method (odds ratio [OR] 1.324, 1.393, 1.356, and 2.161; 95% confidence interval [CI] 1.045–1.687, 1.096–1.769, 1.094–1.681, and 1.188–3.930; p = 0.020, 0.007, 0.005, and 0.012), and the weighted Median method23 (OR 1.291, 1.405, 1.351, and 2.280; 95% CI 1.011–1.649, 1.093–1.805, 1.086–1.680, and 1.154–4.541; p = 0.041, 0.008, 0.007, and 0.019). The CCN3, PRCKSH, TEX264, and TGFBR3 were identified as negative causal effects genes (Fig. 2A,D,F,H) using the IVW method (OR 0.773, 0.597, 0.552, and 0.611; 95% confidence interval [CI] 0.603–0.991, 0.347–0.953, 0.347–0.813, and 0.379–0.985; p = 0.043, 0.031, 0.003, and 0.043), and the weighted median method (OR 0.644, 0.648, 0.551, and 0.611; 95% CI 0.459–0.094, 0.422–0.994, 0.367–0.828, and 0.349–1.069; p = 0.011, 0.047, 0.004, and 0.084) (Fig. 1C). There was no evidence of heterogeneity among the AD-related genetic IVs (Table1). Additionally, Supplementary Figs. S2-S3 presents the overall impact and sensitivity analysis of each SNP within the 8 genes on AD.

Fig. 1
figure 1

Mendelian randomization and plasma secretory proteins analysis. (A) Manhattan plot showing SNPs of GWAS result. (B) The Venn diagram shows the 8 intersecting genes related to AD from the plasma secretome proteins and MR analysis results. (C) Forest plot illustrating the odds ratios (ORs) and 95% confidence intervals (CIs) for the association between genetic variants and AD.

Full size image
Fig. 2
figure 2

Scatter plots for the causal association between the 8 proteins and AD.

Full size image
Table1 MR results for the relationship between 8 genes on AD.
Full size table

Single-cell RNA sequencing (scRNA-Seq) profiling and screening of AD-related plasma secretory proteins

Using the ‘singeR’ package in R, we identified eight different cell types in aortic tissues. These include endothelial cells (ECs), smooth muscle cells (SMC), T cells, B cells, macrophages, monocytes, dendritic cells (DC), and NK cells (Fig. 3A). The UMAP results of the AD and HC groups were consistent with those obtained from the overall analysis (Fig. 1C). The top 50 differentially expressed genes (DEGs) in each cell type are represented in a heatmap (Fig. 3B). By calculating the cell proportions in three normal samples and six AD samples, we discovered that SMC, and monocytes are the primary cells (Fig. 3C).

Fig. 3
figure 3

Single-cell transcriptomic profiling of AD and healthy control. (A) Uniform Manifold Approximation and Projection (UMAP) analysis shows 8 distinct clusters of AD (left) and healthy control (right) of ascending aorta samples. (B) Heatmap of the top 50 DEGs for the 8 main cell types. (C) Bar plot showing the proportions of the 8 main cell types in the AD and normal groups. (D) Dot plot showing the expression profiles of the 8 genes across eight major cell types.

Full size image

To investigate whether plasma secretory proteins are localized in vascular tissue cells, we used the 8 plasma secretory proteins (DBNL, NPC2, SUMF2, TFPI, CCN3, PRCKSH, TEX264, TGFBR3) to demonstrate their cellular localization landscape (Figs. 3D and 4). We found that CCN3 was mainly expressed in SMC. DBNL, and NPC2 primarily expressed in monocytes and macrophages, TFPI in ECs, and TGFBR3 was expressed primarily in SMC and ECs. PRCKSH, SUMF2, and TEX264 were expressed in different subtypes (Figs. 3D and 4).

Fig. 4
figure 4

Cellular localization of the eight secretory proteins. (A) CCN3 is primarily expressed in SMC. (B) DBNL is expressed in monocytes, macrophages, and SMC, with predominant expression in monocytes. (C) NPC2 is expressed in monocytes, macrophages, ECs, and SMC, with primary expression in monocytes, macrophages, and ECs. (D) PRKCSH is expressed in monocytes, macrophages, ECs, and SMC. (E) SUMF2 is expressed in monocytes, macrophages, ECs, SMC, and T cells. (F) TFPI is primarily expressed in ECs and SMC. (G) TEX264 is expressed in monocytes, macrophages, ECs, SMC, and T cells. (H) NPC2 is primarily expressed in ECs and SMC.

Full size image

We also compared the expression of these eight genes in various cell types between the AD and normal groups. We discovered that CCN3 was expressed differentially in ECs, T cells, macrophages, monocytes, DC, and NK cells but not in SMC and B cells. DBNL was expressed differentially in ECs and SMC but not in other cell types. NPC2 was expressed differentially in six cell types but not in DC and NK cells. PRCKSH was expressed differentially in ECs, macrophages, monocytes, and T cells but not in B cells, DC, NK cells, and SMC. SUMF2 was expressed differentially in macrophages, monocytes, and SMC but not in endothelial, T, B, DC, and NK cells. TEX264 was expressed differentially in ECs, monocytes, and T cells but not in SMC, B cells, macrophages, DC, and NK cells. TFPI was expressed differentially in ECs, SMC, macrophages, monocytes, and DC but not in B, NK, and T cells. TGFBR3 was expressed differentially in ECs, T cells, macrophages, monocytes, and NK cells but not in B cells, DC, and SMC (Fig. 5).

Fig. 5
figure 5

Expression analysis of the eight genes in single-cell datasets from 6 AD samples and 3 HC samples. Data are shown as the mean ± SEM. AD vs HC group (Student’s t-test). (* indicates p < 0.05, ** indicates p < 0.01, *** indicates p < 0.001, **** indicates p < 0.0001).

Full size image

Validation in gene expression omnibus (GEO) datasets

The differential expression of eight plasma secretory proteins was assessed using GEO datasets GSE52093, GSE147026, and GSE153434. CCN3 was found to have significant expression differences in all three datasets (Fig. 6A). DBNL demonstrated significant differential expression in the GSE147026 dataset (Fig. 6B). NPC2 was differentially expressed in the GSE147026 and GSE153434 datasets (Fig. 6C). PRKCSH and TEX264 both exhibited significant differential expression in the GSE52093 and GSE147026 datasets (Fig. 6D,E). SUMF2 was differentially expressed in the GSE52093 dataset (Fig. 6F), whereas TFPI was differentially expressed in the GSE153434 dataset (Fig. 6G). TGFBR3 did not display differential expression in any datasets (Fig. 6H).

Fig. 6
figure 6

The mRNA Expression Levels of the 8 Genes in 3 GEO datasets. (A) CCN3 shows significant expression differences in the GSE52093, GSE147026 and GSE153434 datasets. (B) DBNL is significantly differentially expressed in the GSE147026 dataset. (C) NPC2 is significantly differentially expressed in the GSE147026 and GSE153434 datasets. (D) PRKCSH is significantly differentially expressed in the GSE52093 and GSE147026 datasets. (E) TEX264 is significantly differentially expressed in the GSE52093 and GSE147026 datasets. (F) SUMF2 is significantly differentially expressed in the GSE52093 dataset. (G) TFPI is significantly differentially expressed in the GSE153434 datasets. (H) TGFBR3 does not show differential expression in any of the 3 datasets. GSE52093 datasets are shown as the mean ± SEM (n = 12), GSE147026 datasets are shown as the mean ± SEM (n = 8), GSE153434 datasets are shown as the mean ± SEM (n = 20). AD vs HC group (Student’s t-test). (* indicates p < 0.05, ** indicates p < 0.01, *** indicates p < 0.001, **** indicates p < 0.0001).

Full size image

qPCR Analysis of Differential Expression of the Eight Genes in AD and HC

We conducted qPCR on six AD and six normal aortic tissue samples to determine whether the eight genes identified are differentially expressed between AD and HC. Our findings revealed differential expression of CCN3, NPC2, SUMF2, TFPI, and TEX264 between the AD and HC groups (Fig. 7A,C,E–G). Specifically, CCN3 and TEX264 were downregulated in the disease group (Fig. 4A,F), whereas NPC2, TFPI, and SUMF2 were over-expressed (Fig. 7C,E,G). No differential expression was observed for DBNL, PRKCSH, and TGFBR3 (Fig. 7B,D,H). These findings are consistent with Mendelian analysis, which revealed that CCN3 and TEX264 are negative causal effects genes, while NPC2, TFPI, and SUMF2 are positive.

Fig. 7
figure 7

The mRNA Expression Levels of the 8 Genes in Clinical Samples from AD and HC (A–H). Data are shown as the mean ± SEM (n = 12). AD vs HC group (Student’s t-test). (* indicates p < 0.05, ** indicates p < 0.01, *** indicates p < 0.001, **** indicates p < 0.0001).

Full size image

Enzyme-linked immunosorbent assay (ELISA) analysis of differential expression of the 8 genes in plasma for AD and HC

Plasma derived from six AD patients and six healthy individuals was tested using ELISA to explore the protein expression of these eight genes in plasma. Our study observed a marked difference in the plasma expression of CCN3, PRKCSH, SUMF2, and TFPI (Fig. 8A,D,F). The AD group exhibited elevated protein levels of CCN3 and SUMF2, which aligned with our qPCR analysis from tissue samples. In contrast, TFPI protein levels were lower in the AD group (Fig. 8F), which contradicts our tissue-based qPCR findings. DBNL, NPC2, TEX264, and TGFBR3 protein levels exhibited insignificant variation between the disease and normal groups (Fig. 8B–C,G–H). Although the AD group displayed a pronounced increase in PRKCSH protein levels in plasma (Fig. 8D), the mRNA expression of PRKCSH in tissue samples was indifferent.

Fig. 8
figure 8

ELISA Analysis of Protein Expression Levels of the 8 Genes in Plasma Samples from AD and HC (A–H). Data are shown as the mean ± SEM (n = 12). AD vs HC group (Student’s t-test). (* indicates p < 0.05, ** indicates p < 0.01, *** indicates p < 0.001, **** indicates p < 0.0001).

Full size image

Enrichment function analysis

The Gene Ontology (GO) analysis revealed that the DEGs in CCN3-type cells were primarily enriched in biological processes (BP), including cell-substrate adhesion, positive regulation of cell adhesion, and extracellular structure organization. The enrichment of cellular components (CC) was observed in the collagen-containing extracellular matrix, cell-substrate junction, and endoplasmic reticulum lumen. Molecular function (MF) enrichment included terms like extracellular matrix structural constituent, collagen binding, and fibronectin binding (Supplementary Fig. S4). Pathway enrichment analysis revealed that commonly expressed genes were primarily involved in processes, such as focal adhesion, regulation of actin cytoskeleton, and the PI3K-Akt signaling pathway (Supplementary Fig. S4).

Discussion

In this study, we used GWAS and eQTL analysis and identified 194 genes associated with AD. After overlapping these genes with the human plasma secretome, eight plasma secretory proteins—CCN3, PRCKSH, TEX264, TGFBR3, DBNL, NPC2, SUMF2, and TFPI—were identified as having a potentially causal relationship with the occurrence of AD. After analyzing single-cell expression data, the study discovered that CCN3 was primarily expressed in SMC. DBNL was predominantly expressed in monocytes and macrophages, whereas NPC2 was primarily expressed in monocytes, macrophages, and ECs. PRKCSH, SUMF2, and TEX264 were broadly expressed across various cells, whereas TFPI was expressed mainly in endothelial and SMC. After validating the differential expression of these eight genes using three GEO datasets, qPCR, and ELISA, CCN3 was identified as a potential biomarker for AD. Furthermore, enrichment analysis of DEGs from cells revealed differential expression of CCN3, which was extracted from single-cell data.

Identifying and confirming GWAS is challenging because most associations (> 90%) are found in noncoding regions of the genome. Here, we used AD-related GWAS data combined with Genotype Tissue Expression (GTEx) eQTL analysis to identify genes causally related to AD. Multiple studies have used this method to identify genes associated with cardiovascular diseases, including ESR124, RPS9, RPS1825, IL-1ra, and MCSF26. These findings may serve as a foundation for further research into the molecular mechanisms of AD.

Single-cell sequencing analysis has successfully elucidated cellular heterogeneity. Utilizing this technique, we identified different cell types and investigated the cellular distribution of eight secretory proteins. The CCN3, DBNL, NPC2, TFPI, and TGFBR3 expressions exhibited tissue-specific patterns, unlike PRKCSH, SUMF2, and TEX264, which lacked such specificity. Research has shown that the genes CCN327, TFPI28, and TGFBR329 are linked with cardiovascular diseases. However, no studies have directly linked DBNL, NPC2, PRKCSH, SUMF2, and TEX264 genes to cardiovascular diseases. This differential expression suggests a link between the functional roles of these genes in disease occurrence. These findings serve as a basis for future research into the mechanisms by which these genes influence the target cells during AD. Plasma secretory proteins have been confirmed to be highly relevant and diagnostic in the study of cardiovascular diseases. In the context of AD, these proteins reflect the state of inflammation and tissue damage in the body through changes in their concentration in the blood, which is critical for the early diagnosis of AD, especially in patients without obvious symptoms30,31,32.

Our research has confirmed CCN3 as a biomarker for AD by validating GEO datasets, tissue qPCR, and plasma ELISA. CCN3, also known as the nephroblastoma overexpressed (NOV) gene, encodes a small, secreted, cysteine-rich protein, a member of the CCN family of regulatory proteins, which plays a role in ECs, SMC, monocytes, and lymphocytes, and influences processes, including proliferation, adhesion, migration, differentiation, and survival.33,34,35,36,37 Previous research has indicated that laminar shear stress induced the expression of CCN3 in ECs (ECs) in vitro, while inflammatory cytokines such as TNF-α suppressed it38. In vivo, expression of CCN3 was detected in the ECs of large vessels, where it was found to inhibit cytokine-induced vascular cell adhesion molecule-1 and the adhesion of leukocytes to ECs38. CCN3 inhibits the activation of NF-κB, resulting in the recruitment of monocytes38. It also inhibits the proliferation and migration of vascular SMC (VSMC), suggesting its protective properties against atherosclerosis39. While the role and mechanisms of CCN3 in AD have not been reported, atherosclerosis is one of the causes of AD, implying that CCN3 might play an important role in AD.

Our study discovered that CCN3 is significantly downregulated in tissue and plasma levels in AD patients, implying that CCN3 may play a protective role against AD, similar to its effects against atherosclerosis. In single-cell expression localization analysis, CCN3 was primarily expressed in SMC. Although there was an insignificant difference in expression between AD and HC groups in SMC, previous research has demonstrated that CCN3 can regulate the proliferation and migration of SMC, thereby affecting vascular homeostasis39. Single-cells enrichment analysis revealed that cells with differential CCN3 expression were primarily enriched in BP, including cell-substrate adhesion, positive regulation of cell adhesion, and extracellular structure organization. The CC participates in collagen-containing extracellular matrix, cell-substrate junction, and endoplasmic reticulum lumen, while MF enriched in extracellular matrix structural constituent, collagen binding, and fibronectin-binding. These functions are closely related to the occurrence of AD40,41,42,43. In the Kyoto encyclopedia of genes and genomes (KEGG) enrichment analysis, cells with differential expression of CCN3 were mainly related to focal adhesion, regulation of actin cytoskeleton, and PI3K-Akt signaling pathway, and the loss of SMC led to an increase in focal adhesion kinase signaling, resulting in MAPK activation and autonomous SMC proliferation44. TGF-β induces the PI3K-Akt signaling pathway, which regulates the phenotypic transformation of SMC45. All these findings suggest that CCN3 plays a vital role in AD development.

Indeed, our study has some limitations. First, the AD-related GWAS and eQTL data in the MR study are from Europe and are derived from blood samples. We also lacked data from other centers that specialize in tissue research. Unlike GWAS and eQTL data, the single-cell data is derived from Asian tissue samples. In our single-cell data analysis, we only identified eight cell types; a more detailed classification of cell types could potentially allow for more accurate localization of the eight secretory proteins, revealing the potential functions of the genes. We discovered CCN3 as a potential biomarker for AD, with significant differential expression confirmed in both plasma and tissue. However, additional experimental validation is required to verify CCN3’s role as a protective mechanism in AD. In our subsequent investigations, we plan to explore the association of CCN3 with the incidence and prognosis of AD and the molecular mechanisms of CCN3 involved in the onset and progression of AD.

In conclusion, we identified CCN3 as a biomarker for AD and confirmed its downregulation in AD using different methods. Additionally, we have preliminarily investigated the mechanisms by which CCN3 may contribute to AD, which provides new insights into the molecular mechanisms underlying AD pathogenesis.

Materials and methods

GWAS and QTL summarized data

GWAS summary statistics for AD were extracted from the FinnGen consortium, including 967 cases and 381,977 controls. The GWAS summarized data can be downloaded at https://storage.googleapis.com/finngen-public-data-r10/summary_stats/finngen_R10_I9_AORTDIS.gz. Cis-eQTL genetic variants were used as the IVs for gene expression46. We performed SMR analysis for blood. We used Version 8 of the eQTL summarized data from the GTEx project47. The eQTL data can be downloaded from https:// cnsgenomics.com/data/SMR/#eQTLsummarydata.

Human secretome data

The human secretome contains 1039 proteins predicted to translocate to peripheral blood. The data can be found in the Human Protein Atlas. (https://www.proteinatlas.org/).

scRNA-seq data

The scRNA-seq dataset was obtained from the GSE213740 datasets, including 6 AD patients and 3 HC cohorts48. The data can be found in GEO (https://www.ncbi.nlm.nih.gov/geo/).

Data acquisition and screening of DEGs

The data from AD studies in the GEO database were obtained by the R package ‘GEOquery’49. The six included expression profiling series were divided into training and validation cohorts (Table 2), and the batch effects were adjusted by the ‘ComBat’ function of R package ‘sva’ when consolidating all the training cohorts. The R package ‘limma’ was used to analyze differential expression between AD and normal samples. The thresholds were adjusted at p < 0.05 and |log2FC|> 1.

Table 2 Information about the research of AD in the GEO database.
Full size table

MR analysis

We initially quantified the goodness-of-fit of our model to the eQTL data by calculating the R2 value. Following that, an F-test was conducted. The F value was computed using the previously calculated R2 value and the parameter sample size-exposure (F = (N-2)*R2/(1-R2)). The eQTL data was then filtered using the computed F value50,51. Data points with an F value of ≤ 10 were identified as IVs and thus excluded from our dataset. This rigorous process ensured the robustness of our findings and the reliability of our subsequent analyses. We then conducted a two-sample MR analysis on AD-related GWAS and eQTL. After receiving the AD-related MR results, we filtered out results where the MR Egger, Weighted median, Inverse variance weighted, Simple mode, and Weighted mode OR values were consistent (all greater than 1 or all less than 1). Briefly, SMR applies MR principles to integrate GWAS and eQTL summary statistics, aiming to test for a pleiotropic association between gene expression and AD via a shared and potentially causal variant at a locus. The heterogeneity in dependent instruments (HEIDI) test was employed to investigate the presence of linkage in the observed association. If the null hypothesis is rejected (PHEIDI < 0.05), it suggests that the observed association might be attributed to two distinct genetic variants in high linkage disequilibrium with each other. The default settings in SMR were used in this study (P-eQTL < 5 × 10–8, minor allele frequency [MAF] > 0.01, excluding SNPs in strong linkage disequilibrium [LD, r2 > 0.9] with the top associated eQTL, and removing SNPs in low LD or not in LD [r2 < 0.05] with the top associated eQTL). The false discovery rate was used to adjust for multiple tests. This approach provides a comprehensive understanding of the potential genetic foundations of AD.

All detailed R codes used for the analyses can be found in the supplementary file 1.

Materials

In this study, custom primers were designed and acquired from TSINGKE Biotechnology Co., Ltd. The sequences for these primers are listed in Table 3. The study used the following reagents: Trizol solution (TaKaRa, Japan); reverse transcription kit Prime Script™RT (TaKaRa, Japan); qPCR kit SYBR ® Premix Ex Taq™ (TaKaRa, Japan). ELISA kit (Fankew, Shanghai Kexing Trading Co., Ltd, Shanghai, China; CCN3: F0291-HB, PRKCSH: F0295-HB, TEX264: F0299-HB, SUMF2: F0303-HB, NPC2: F0307-HB, DBNL: F0310-HB, TGFBR3: F0314-HB, TFPI: F0200-B.)

Table 3 The sequences of primers for RT-qPCR.
Full size table

Sample collection

Before starting the research, ethical approval was obtained from the Ethics Committee of the First Affiliated Hospital Chongqing Medical University. The research adhered strictly to the guidelines outlined in the Helsinki Declaration. Written informed consent was obtained from all participating patients or their relatives. Six normal and six AD ascending aorta tissues along with their peripheral blood were ultimately collected, the detailed participant information can be found in Supplementary Table S1.

The participants’ data in this study were collected through medical history interviews and relevant examinations upon admission. The sociodemographic variables collected included: gender (male, female), age, marital status (married, unmarried), smoking history (smoker, non-smoker), and alcohol consumption history (drinker, non-drinker). Anthropometric measurements included: height, weight, body mass index (BMI), heart rate (HR), and blood pressure (SBP, DBP) with an average of three measurements. Medical history included: hypertension, diabetes (including diabetes or elevated blood glucose), dyslipidemia, chronic obstructive pulmonary disease, and coronary heart disease.

Biochemical markers were measured from blood samples collected after overnight fasting. These markers included Blood Glucose, Total Cholesterol, Triglycerides, LDL, HDL, D-dimer, CRP, and PCT.

Inclusion and exclusion criteria for the recruitment of AD participants:

Inclusion criteria: patients diagnosed with aortic dissection based on clinical and imaging findings (CT scan, and angiography), aged over 40 years, who have provided written informed consent or consent through their legal representative, and are scheduled for surgical treatment of aortic dissection. Exclusion criteria: patients with severe other systemic diseases, such as severe respiratory conditions (e.g., chronic obstructive pulmonary disease, severe asthma), chronic kidney disease (stage 3 or higher), cancer (including any type of malignant tumor), and severe infections (e.g., sepsis, severe pneumonia). Additionally, patients with genetic and immune-related diseases, such as Marfan syndrome, bicuspid aortic valve, and autoimmune diseases (e.g., Ehlers-Danlos syndrome, systemic lupus erythematosus, rheumatoid arthritis), will be excluded. Patients undergoing immunosuppressive treatment (e.g., corticosteroids or other immunosuppressive agents) will also be excluded from the study.

Recruitment Process for Control Participants:

The control group participants in this study are sourced from body donations. The inclusion criteria are as follows: participants are aged over 40 years and have never been diagnosed with AD. All participants must have voluntarily signed informed consent before death or consent through their legal representative, allowing their bodies to be used for medical and scientific research. The exclusion criteria include participants with severe respiratory diseases (such as chronic obstructive pulmonary disease or severe asthma), chronic kidney disease (stage 3 or above), malignant tumors, severe infections (such as sepsis or severe pneumonia), Marfan syndrome, immune system diseases (such as Ehlers-Danlos syndrome, systemic lupus erythematosus, rheumatoid arthritis, etc.), or those receiving immunosuppressive therapy (such as steroids or other immunosuppressive medications). All sample collection procedures strictly adhere to ethical guidelines.

Total RNA isolation and qPCR

Total RNA was extracted from 50 mg of each aortic tissue sample using Trizol reagent (TaKaRa, Japan). For cDNA reverse transcription, 1 μg of RNA was used in a 20 μL reaction volume with the Prime Script™RT reagent Kit. Subsequently, qPCR was performed on a BioRad Real-Time PCR System (Bio-Rad, America) using SYBR® Premix Ex Taq™. Gene expression levels were quantified using the 2–ΔΔCt method, with 18S serving as the reference gene for normalization.

ELISA assay

In the ELISA experiment, a 96-well plate was initially set up with blank control wells, standard wells, and test sample wells. The standard substances were accurately added to the corresponding wells according to a concentration gradient. For the test sample wells, 40 µL of sample diluent was first added, followed by 10 µL of the test sample, resulting in a final dilution factor of 5. The plate was then sealed and incubated at 37 ℃ for 30 min. Subsequently, the concentrated washing solution was diluted with distilled water and used to wash the wells to remove unbound substances. The enzyme-labeled reagent was added to each well, and the plate was incubated again at 37 ℃. The wells were then rewashed, and the color reagent A was added to each well, followed by the color reagent B. The mixture was gently shaken to mix and incubated at 37 ℃ in the dark for 10 min for color development. The addition of a stop solution terminated the reaction. Finally, the blank control well was used to zero the instrument, and each well’s absorbance (OD value) was measured at a wavelength of 450 nm. All measurements were performed within 15 min after the addition of the stop solution.

scRNA-seq data pre-processing

We first downloaded the normalized 10X scRNA-seq dataset (GSE213740) with its expressional matrix, cell barcodes, and features. Then, we created the Seurat object by the Read10X of Seurat package52. Quality control measures for scRNA-Seq (gene count per cell, hemoglobin-related genes, percentage of mitochondrial genes) were calculated using the Seurat R package (version 4.2.1)53,54. Low-quality cells were excluded if they had fewer than 500 expressed genes or contained over 10% mitochondrial and hemoglobin-related genes. The DoubletFinder package55 was employed to eliminate doublet cells, using an expected doublet rate of 0.075. To correct for batch effects among individuals, the Harmony package56 and its RunHarmony function were utilized.

Clustering and differential gene expression analysis of scRNA-Seq data using the “Seurat”Package

We selected 3000 highly variable genes for PCA, then used the RunUMAP function to perform Uniform Manifold Approximation and Projection (UMAP) with 30 principal components for dimensionality reduction. Primary cell clusters were identified using the Seurat package with a resolution of 0.4. For subclustering target cells, a resolution of 0.3 was selected to reveal changes in subpopulations across all cell types in each sample. For two-dimensional (2D) visualization, the RunUMAP function in Seurat with dimension parameters (1:20) was employed. Cell cluster annotation was conducted using the computational tool ‘singleR’, which matches annotated reference data with selected marker genes to optimally differentiate cell phenotypes57. The FindAllMarkers function in Seurat was used to determine the DEGs in each cell cluster. Differential gene expression between clusters was analyzed using the Wilcox test in Seurat, which evaluated the significance of gene expression changes among various cell subsets.

GO and KEGG enrichment analysis

Firstly, we excluded cells with zero expression of CCN3 from the single-cell dataset (GSE213740). Subsequently, the remaining cells were categorized into two groups: the AD group and the HC group. The cells exhibiting significant differential expression of CCN3 (|Log FC|> 0.3, p < 0.05) were identified and named CCN3-type cells. Finally, we performed GO and KEGG enrichment analyses on the DEGs for CCN3-type cells. Functional enrichment analysis of GO terms and KEGG pathways related to the DEGs was conducted using the clusterProfiler (v3.14.3) R package and hypergeometric distribution58. The significantly enriched GO terms and KEGG pathways were identified based on the adjusted p < 0.05 criterion and visualized using bar plots59,60,61.