Heterogeneity of common carotid artery and abdominal aorta: an angle of sphingolipid metabolism in vascular smooth muscle cells Authors: Bowen Li, Kaixiang Zhou, Kun Chen, Shuai Jiang, Xichen Wang, Zhenghui Gu, Chunhu Gu, Yuehu Han, Yonghong Liu, Yaoming Chang, Junxiang Bao Published: 2025 ABSTRACT AbstractBackgroundAtherosclerosis and aortic aneurysms are prevalent aortic disorders. Significantly, atherosclerosis frequently impacts the common carotid artery (CA), whereas aortic aneurysms typically involve the abdominal aorta (AA), indicating possible heterogeneity between CA and AA with an ambiguous underlying mechanism. Sphingolipids, a crucial branch of lipid metabolism, has increasingly garnered attention in vascular diseases by influencing the phenotypic regulation of vascular smooth muscle cells (VSMCs). Nevertheless, whether sphingolipids play a role in the heterogeneity between CA and AA and in disease susceptibility remains uncertain.Methods and resultsPublic transcriptomics were employed to clarify the distinctions in contractility and calcification of VSMCs in relation to the vulnerability of human AA and CA to aortic aneurysms and atherosclerosis, respectively. Bulk RNA-seq revealed transcriptomic variances in the biology of VSMCs between AA and CA from rats. Primary VSMCs from AA (AASMC) and CA (CASMC) were isolated for further validation. The specific variations in sphingolipid metabolism (comprising 8 classes of sphingolipids with 169 species) between AA and CA from rats were further characterized using UPLC–QTOF-MS-based lipidomics. Most sphingolipids, except sphingomyelin, were significantly elevated in CA compared to AA. Ceramides were the major contributor to these differences, which was further confirmed by in situ immunofluorescence of AA and CA from rats and healthy humans. The differential expression of genes involved in ceramide biosynthesis (Cers1-4), transport (Cert1), and metabolic processes (Smpd1, Sgpp1, Sphk2, and Sgms1/2), along with variations in subcellular organelles in VSMCs, contributed to the heterogeneity of sphingolipid metabolism between AA and CA. The effects of ceramides on VSMC biology were subsequently assessed using primary VSMCs. Specifically, AASMC exhibited greater contractility and lower susceptibility to calcification compared to CASMC. Exogenous ceramides heightened the susceptibility to calcification in both CASMC and AASMC, whereas the inhibition of ceramide synthases engendered an opposing consequence. Notably, in CA from patients with atherosclerosis but not AA from aortic aneurysms, the activation of sphingolipid metabolism (including ceramides) was positively correlated with calcification and negatively correlated with the regulatory processes of VSMC contraction. Furthermore, the ceramide metabolism was activated along with calcification in CA, which corresponded with the accumulation of ceramide in atherosclerotic plaques of human.ConclusionsIn this study, we identified ceramides, a major class of sphingolipid metabolites, as a promising determinant in unequal biology and susceptibility to calcification of CA and AA. Our results reveal the previously unappreciated role of sphingolipid metabolism, particularly ceramides, in regional vascular pathology. These findings provide novel insights that inform our understanding of disease mechanisms and highlight potential ways for future therapeutic exploration.Graphical AbstractSphingolipids' role in vascular heterogeneity between common carotid artery (CA) and abdominal aorta (AA) have been elucidated through a comprehensive approach involving public transcriptomics, RNA sequencing, and lipidomics. Ceramides, key in sphingolipid metabolism, modulate contractility and calcification differences in vascular smooth muscle cells (VSMCs) between CA and AA. Supplementary InformationThe online version contains supplementary material available at 10.1186/s12938-025-01473-7.Keywords: Common carotid artery, Abdominal aorta, Vascular heterogeneity, Multiomics, Sphingolipids, Ceramides Introduction Regional specificity of blood vessels has been generally acknowledged, including the distinct functional and structural features as well as susceptibility to cardiovascular diseases (CVD) [1]. Normally, the aorta appears more prone to dilatation and peripheral arteries more to occlusive disorders [2]. Even for elastic artery with similar structural features, calcification tends to emerge on the ascending aorta than descending aorta [3]. Common carotid artery (CA) and abdominal aorta (AA) are indispensable channels for blood supply to the brain and abdominal organs. Specifically, the dilatation becomes more pronounced in AA (23%) than in CA (11%) during aging [4]. Stiffness is higher in CA than AA before the age of 20, while the age-related degenerative remodeling favors AA compared to CA [4]. These functional and structural differences highlight the heterogeneity between these two arteries, which warrants further exploration. Thus far, both the distinct embryonic origins and hemodynamic microenvironment in various regions contribute to the divergent fates of these two arteries [5]. Meanwhile, a growing number of signaling molecules have been identified that contribute to vascular heterogeneity [6], among which lipids represent a complex class of molecules. On the one hand, blood lipids, including high density lipoprotein-cholesterol (HDL-C), low density lipoprotein-cholesterol (LDL-C), triglyceride (TG) and total cholesterol (TC), relate closely to the onset of CVD [7, 8]. For another, the in situ TG supplies energy through oxidative phosphorylation, phospholipids (PLs) and sphingolipids compose membrane structures, while cholesterol (Ch) serves as a precursor for the synthesis of cholic acid and steroid hormones [9]. Sphingolipid species also play a crucial role in the course of vascular remodeling [10]. Certain metabolites such as sphingomyelin (SM), ceramides, sphingosines (So), and sphingosine-1-phosphate (S1P) function divergently in cellular death, growth, differentiation, stress response, and inflammation [11, 12]. The general view asserts that the ceramides induce apoptosis [13], whereas S1P promotes cell survival and differentiation [14]. S1P works through binding a family of G protein-coupled receptors after secretion from cells [15]. In the meantime, sphingolipids and Co forge the basic composition of lipid rafts, modulating cell signaling transduction [16]. Even with the above understandings, the role of sphingolipids in vascular heterogeneity concerning CA and AA has not yet been elucidated. In this study, the abundance of sphingolipid species as well as key enzymes regulating their metabolism were compared between CA and AA. Subsequently, the role of sphingolipid metabolism in vascular heterogeneity under physiological or pathological stimuli was examined. The main aim was to further clarify the underlying mechanisms determining the distinct biology and pathological fates of CA and AA. Results 1. Human VSMCs of AA or CA tissues exhibit different changes of contractility and calcific phenotypic switching for aortic aneurysms and atherosclerosis stimuli based on public transcriptomics To explore the potential heterogeneity for AA and CA, we firstly analyzed the transcriptomic characteristics of AA tissues from abdominal aortic aneurysm patients (AA-AAA; GSE47472) and CA tissues from atherosclerotic patients (CA-AS; GSE100927) from GEO public database. There were apparent differences in transcriptome between patients and healthy donors both in AA-AAA and CA-AS (Figure S1A-B). Among 878 differentially expressed genes (DEGs) in AAA patients, 235 were upregulated, while 643 were downregulated (Fig. 1A). Among 914 DEGs in AS patients, 579 were upregulated, while 335 were downregulated (Fig. 1B). However, Venn analysis showed that only 4 genes were co-upregulated both in AA-AAA and CA-AS, while only 6 genes were co-downregulated (Fig. 1C-D). Gene Ontology (GO) based Gene Set Enrichment Analysis (GSEA) demonstrated the distinct differences in AA-AAA and CA-AS. The top ten gene sets activated in AA-AAA were mainly enriched in T cell toxicity, mitochondrial localization, and nitric oxide metabolism. The top ten gene sets of inhibition were mainly enriched in lamellipodium organization and anoikis. However, the top ten gene sets activated in CA-AS were mainly enriched in adaptive immune pathways, such as adaptive immune response, leukocyte mediated immunity. The top ten gene sets of inhibition were mainly enriched in muscle cell contraction (Figure S1C-D). Human abdominal aorta (AA) and carotid artery (CA) tissues extracted from abdominal aortic aneurysm (AAA) and atherosclerosis (AS) lesions, respectively, exhibit distinct VSMCs biology. A Volcano-map analysis of AA tissues from 14 AAA patients and 8 donors (bulk RNA-seq; GSE47472). B Volcano-map analysis of CA tissues from 29 AS patients and 12 donors (bulk RNA-seq; GSE100927). C Venn analysis of the upregulated genes from GSE47472 or GSE100927. D Venn analysis of the downregulated genes from GSE47472 or GSE100927.E Expression of key genes involved in VSMCs phenotypic switching from GSE47472 and GSE100927. F–G GSEA plots showing normalized enrichment scores (NES) and normalized P-value for vascular smooth muscle contraction and positive regulation of calcium-mediated signaling from GSE47472 (F) and GSE100927 (G). P-value was calculated with an empirical phenotype-based permutation test. H Correlation heatmap between the expressions of VSMC contraction genes, VSMC differentiation genes, and calcification genes from GSE47472. P value was based on Pearson’s correlation. (***P < 0.001). I Correlation analysis between the expressions of VSMC contractile genes and calcification genes from GSE47472. Correlation coefficient (R) and P value were based on Pearson’s correlation. J Correlation heatmap between the expressions of VSMC contraction genes, VSMC differentiation genes, and calcification genes from GSE100927. P value was based on Pearson’s correlation. (***P < 0.001). K Correlation analysis between the expressions of VSMC contractile genes and calcification genes from GSE100927. Correlation coefficient (R) and P value were based on Pearson’s correlation Given the critical role of VSMCs and phenotypic switching in vascular diseases, we further analyzed the expression of the key genes involved in VSMCs phenotypic switching. Compared with AA-AAA, CA-AS demonstrated a significant phenotypic switching, which was manifested in the significant downregulation of VSMCs contractile genes and the upregulation of calcification genes (Fig. 1E). These results were further confirmed by GSEA analysis (Fig. 1F). We then analyzed the expression correlation between the calcification genes and VSMC contraction (or VSMC differentiation) genes in both AA-AAA and CA-AS. As shown in Fig. 1H-I, calcification genes had no significant correlations with VSMC contraction or differentiation genes in AA-AAA. Intriguingly, in CA-AS, calcification genes were negatively correlated with VSMC contraction and differentiation genes (Fig. 1J-K). These results suggest that human AA or CA tissues exhibit distinct responses to aortic aneurysms and atherosclerosis stimuli according to transcriptomics, indicating underlying heterogeneity of these two arteries. In addition, the phenotypic switching from contractile to calcific of VSMCs plays a potential role in this different response of AA and CA. 2. VSMCs in AA and CA from rats have different degrees of contractility and calcified susceptibility To elucidate the potential heterogeneity of AA and CA, bulk RNA-seq was performed to analyze the transcriptomics of AA and CA isolated from SD rats (AA as control) (Fig. 2A). There were great differences in transcriptomics between AA and CA (Figure S2A). Among 2743 DEGs, 1975 were upregulated, while 768 were downregulated (Fig. 2B). Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis revealed that as compared to AA, the downregulated DEGs in CA were enriched in cytoskeletal and contraction-related pathways, such as VSMC contraction, tight junction, focal adhesion, and cytoskeleton in muscle cells (Fig. 2C). Additionally, heatmap analysis showed the lower expression of the VSMC’s contractile genes in CA relative to AA, including Acta2, Tagln, Myh11, and Cnn1 (Fig. 2D). These results indicate a distinct contractile feature of VSMCs between AA and CA. VSMCs in AA and CA from rats have different degrees of contractility and calcified susceptibility. A Schematic diagram of sample source for bulk RNA-seq. B Volcano-map analysis of RNA-seq for AA and CA tissues from rats. C KEGG pathway enrichment analysis of downregulated genes of RNA-seq for AA and CA tissues from rats. Statistical significance is calculated with a hypergeometric test. D Heatmap analysis involved in vascular smooth muscle cells contraction of RNA-seq. Representative confocal microscopy images of α-SMA (red) of AASMC and CASMC cultured with or without serum for 48h. Scale bar = 50 μm (20 μm in zoom-in panels). F–H Relative mRNA expression levels of Acta2 (F), Tagln (G), and Myh11 (H) of AASMC and CASMC cultured with or without serum for 48h. Statistical significance is calculated with a two-way ANOVA followed by the Tukey’s multiple comparisons test (n = 3 biological replicates;*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). I Immunoblot analyses of α-SMA and SM22α of AASMC and CASMC cultured with or without serum for 48h. Band quantification is normalized to the first lane. J Heatmap analysis involved in calcification of RNA-seq. K Representative images of Alizarin Red staining of AASMC and CASMC after Na3PO4 (3μM) treatment for 0, 7, and 14 days. Scale bar = 200μm. The quantifications of Alizarin Red area are shown in right panel. (n = 3; Student t tests; ***P < 0.001). L–M Relative expression level of Bmp2 (L) and Runx2 (M) in AASMC and CASMC after DPBS or Na3PO4 (3μM) treatment for 7 days. (n = 3; Student t tests; ***P < 0.001). N Representative images of cell contraction assay of AASMC and CASMC after DPBS or Na3PO4 (3μM) treatment for 7 days (left panel). Quantifications of cell contraction diameters of AASMC and CASMC after DPBS or Na3PO4 (3μM) treatment for 7 days are shown in the right panel. Scale bar = 2 mm. (n = 3; two-way ANOVA and Tukey’s multiple comparisons test; *P < 0.05, **P < 0.01, ***P < 0.001). Then, we isolated primary VSMCs from AA (AASMC) and CA (CASMC) by using collagen digestion to elucidate the contractile heterogeneity of AA and CA. The positive signal of α-SMA indicated a high purity of the isolated primary smooth muscle cells (Figure S2A). We found that CASMC had a lower abundance and looser subcellular distribution of α-SMA compared with AASMC (Figure S2A; 2E). Furthermore, serum deprivation (a trigger of the transformation to contractile phenotype) induced higher expression and more dense distribution of α-SMA in both cells (Fig. 2E). Nevertheless, the content and distribution of α-SMA in CASMC were still lower and looser than those in AASMC (Fig. 2E). The expression of VSMC’s contractile genes (Acta2, Tagln, and Myh11) and the cell contractility tests further confirmed that CASMC had lower contractility with or without serum deprivation compared with AASMC, which was consistent with the results of bulk RNA-seq (Fig. 2C-I). We next investigated the difference in calcification susceptibility between AA and CA. Firstly, heatmap analysis showed higher expression of the calcification genes in CA, including Klf4, Bmp2, and Runx2 (Fig. 2J). We then used Na3PO4 to induce calcification of cell models. Alizarin red staining results indicated that CASMC showed higher susceptibility to calcification than AASMC (Fig. 2K). These results were further confirmed by the upregulation of calcification-response genes (Bmp2 and Runx2) (Fig. 2L-M), the downregulation of contractile genes (Acta2, Myh11, and Tagln) (Figure S2E-H), and the lower cell contractility of CASMC compared with AASMC (Fig. 2N). Together, these findings indicate that VSMCs in AA and CA from rats have different degrees of contractility and calcification susceptibility. AA and CA represent different profiles of sphingolipid metabolism Given the critical role of lipids in VSMC phenotypic switching and vascular homeostasis, we focused on exploring the role of lipid metabolism in vascular heterogeneity. Notably, many DEGs (AA vs. CA) were enriched in lipid metabolism. Intriguingly, five essential processes involving sphingolipid metabolism were significantly enriched by GO analysis (Fig. 3A). Lipid classes profiles of sphingolipid metabolism for AA and CA from rats. A GO enrichment analysis of all differentially expressed genes (DEG) of RNA-seq for AA and CA tissues from rats. Statistical significance is calculated with a hypergeometric test. B Summary of all the lipid molecules distributed across 8 classes involving sphingolipid metabolism detected by lipidomics. Specific number is shown above the columns. C Lipid concentration (Log10, mg/g) analysis of all the samples both in AA and CA tissues. D PCA analysis of lipidomics. Lipidomics data of AA, CA, and qualitative control (QC) are indicated by gray, blue, and red dots, respectively. E Abundance (μg) of 8 lipid classes involving sphingolipid metabolism detected by lipidomics in AA and CA tissues. F Relative abundance (%) of 8 lipid classes involving sphingolipid metabolism detected by lipidomics in AA and CA tissues. G Bubble map analysis of differential expression lipid molecules of CA tissues compared with AA tissues. H Heatmap analysis of lipidomics of AA and CA tissues. The specific lipid classes are shown on the left Then, UPLC–QTOF-MS based lipidomics was performed to identify the differences of sphingolipid metabolism for these two tissues from 10 male rats around 8 weeks old (weighing about 300 g) (Supplemental Table 1). In total, 8 classes of sphingolipids involving 169 lipid species were identified. We detected a diverse array of lipids in AA and CA of rats (Fig. 3B, Supplemental Table 3). Among the 8 classes (169 species), SM, ceramides, and sphingosine (So) were the three most abundant classes in the arterial tissues. Ceramides and SM were the two most prevalent classes, comprising 76 and 66 species, respectively (Fig. 3B-C). Furthermore, differences in lipid classes between AA and CA were assessed based on fold change (FC) in Univariate Statistical Analysis, while differences in individual lipid species were determined based on both FC and Variable Importance for the Projection (VIP) in OPLS-DA model of Multivariate Statistical Analysis. The significant levels of FC and VIP were set at > 1.5 or < 0.67 (P < 0.05) and > 1 (P < 0.05), respectively. PCA analysis showed that the sphingolipid metabolism in AA and CA had significant differences (Fig. 3D). As shown in Fig. 3E-H, most sphingolipids classes were significantly higher in CA than AA, including ceramides (P < 0.01), 2-glucocerebroside ceramide (CerG2, P < 0.05), ganglioside/monosialo trihexosyl ceramide (GM3, P < 0.01), phytosphingosine (phSM, P < 0.01), and So (P < 0.01). Notably, SM, CerG1, and CerG3 showed no significant difference between these two tissues (P > 0.05). Taken together, these data demonstrate that there is a significant difference in the lipid classes of sphingolipid metabolism between AA and CA. Ceramides contribute significantly to the difference of sphingolipid metabolism between AA and CA Despite all falling under sphingolipid metabolism, these lipid classes have complex metabolic linkages as substrates or products of biochemical reactions based on fatty-acyl chains, which prompted us to outline the specific differences between AA and CA (Figure S3A). As shown in Supplemental Table 3 and Fig. 3F-G, among the 169 sphingolipids species detected, 76 were ceramides with 60 significantly higher in CA than AA, of which the top five were ceramides (d34:0) (P < 0.001), ceramides (d32:0) (P < 0.05), ceramides (d34:1) (P < 0.001), ceramides (d38:4) (P < 0.01), and ceramides (d42:2) (P < 0.01) (Fig. 4A). As for CerG, 8 CerG1, 3 CerG2 and 2 CerG3 were identified with just CerG1(d42:2) (P < 0.01) and CerG2(d18:1/24:0) (P < 0.05) significantly higher in CA than AA (Figure S3B). There were 4 GM3 perceived, all of which were higher in CA relative to AA (P < 0.001) (Fig. 4D). Similarly, 2 phSM, 66 SM and 7 So were revealed in the present study with 1, 9 and 6 significantly higher in CA than AA, respectively (Figure S3H, S3K, 4E). Ceramides contribute significantly to the sphingolipid metabolism difference between AA and CA. A Heatmap analysis of ceramides (Cer) in AA and CA tissues according to the order of saturation and length of carbon chain. The detected intensities are shown on the right panel. B Differences in the abundance of ceramides with different carbon chain lengths in AA and CA tissues. (n = 10; Student t tests; *P < 0.05, **P < 0.01, ***P < 0.001). C Differences in the abundance of ceramides with the number of double bonds (DBs) in AA and CA tissues. (n = 10; Student t tests; *P < 0.05, **P < 0.01). D Heatmap analysis of ganglioside/monosialo trihexosyl ceramide (GM3) in AA and CA tissues according to the order of saturation and length of carbon chain. The detected intensities are shown on the right panel. E Heatmap analysis of sphingosines (So) in AA and CA tissues according to the order of saturation and length of carbon chain. The detected intensities are shown on the right panel. F Differences in the abundance of GM3 with different carbon chain lengths in AA and CA tissues. (n = 10; Student t tests; ***P < 0.001, ****P < 0.0001). G Differences in the abundance of So with different carbon chain lengths in AA and CA tissues. (n = 10; Student t tests; ****P < 0.0001). H The abundance changes of the 8 lipid classes in sphingolipid metabolism of CA relative to AA tissues. Red circles represent the significant upregulated lipid classes in CA compared with AA The length and saturation of fatty-acyl chains determine the properties and functions of sphingolipids metabolites, which was explored in AA and CA of rats in the present study. Firstly, the vast majority of sphingolipids metabolites contain long fatty acyl chains with the number of carbon atoms greater than 12 (total carbon chain length greater than 30). The chain length distribution was similar between AA and CA. For ceramides, the range of total chain length was from 18 to 44 carbon atoms and the top 5 were 34, 42, 32, 38, 36. The amounts of ceramides with such chain lengths was lower in AA than CA (P < 0.01 or P < 0.001). Ceramides with shorter than 32 and longer than 42 carbon atoms were present in minimal amounts (Fig. 4B). For CerG, the total chain length was from 40 to 42 while only CerG2 with 42 carbon atoms showed an increasing trend in CA, although it was not significantly significant (P > 0.05) (Figure S3F). The chain length of GM3 was from 34 to 42, all of which showed less content in AA (P < 0.001) (Fig. 4F). For phSM, the main chain length was 37, significantly higher in CA than AA (P < 0.001) (Figure S3H). For SM, the range of total chain length was from 22 to 47, with the top 6 being 34, 36, 38, 40, 42, 44. SM with shorter than 34 and longer than 44 carbon atoms were in small amount. All the SM species showed no significant difference between AA and CA (P > 0.05) (Figure S3K). The chain of So species consisted of 16, 17 or 18 carbon atoms, with 17 in a very minimal quantity. The chain length of 16 and 18 in So was significantly higher in CA than AA (P < 0.0001) (Fig. 4G). The number of double bonds (DBs) are used to indicate the saturation of fatty-acyl chains. For ceramides, DBs were from 0 to 6 with the majority from 0 to 4. All these types of ceramides were in lower amount in AA than CA (P < 0.05) (Fig. 4C). The CerG, and GM3 contained 1 or 2 DBs, among which CerG1 with 2 DBs (P < 0.05) and GM3 with 1 or 2 DBs (P < 0.001) were more in CA than AA (Figure S3B, F). The So contained 0 or 1 DBs, all of which were in lower abundance in AA compared to CA (P < 0.0001) (Figure S3C). The phSM contained 4 or 5 DBs, of which just that with 4 DBs was higher in CA than AA (P < 0.001) (Figure S3I). For SM, the number of DBs were from 0 to 8 with the majority from 1 to 5 which showed no significant difference between CA and AA (P > 0.05) (Figure S3L). Together, lipidomics data suggest a vastly diverse profile of sphingolipid metabolites in these two arteries (Fig. 4H) and indicate that ceramides contribute significantly to the sphingolipids’ metabolic difference between AA and CA. Gene transcription of enzymes and subcellular organelles controlling ceramide metabolism in AA and CA As the different contents of ceramides contributed to the sphingolipids’ heterogeneity in AA and CA by lipidomics, we further investigated whether this difference is species-conserved and explored the potential causes. Immunofluorescence staining results suggested that ceramides were mainly distributed in VSMCs (Fig. 5A). Although ceramides were present in both AA and CA, the content of ceramides in CA was higher than that in AA, which was consistent with the lipidomics and transcriptomics results (Fig. 3A, 4A, 5A). Additionally, these results were also confirmed in human AA and CA tissues (Figure S4A). Gene transcription of enzymes and subcellular organelles controlling ceramide metabolism in AA and CA. A Representative immunofluorescence staining of α-SMA (green) and ceramides (red) in AA and CA tissues from rats. “L” indicates the side of vascular lumen. Scale bar = 100 μm (20 μm for zoom-in panels). B Relative expression levels of the synthetic and catabolic genes of ceramides in AA and CA tissues. (n = 5; Student t tests; *P < 0.05, **P < 0.01). B Relative expression levels of the synthetic and catabolic genes of SM in AA and CA tissues. (n = 5; Student t tests; *P < 0.05). C Relative expression levels of the catabolic genes of So in AA and CA tissues. (n = 5; Student t tests; *P < 0.05, **P < 0.01). D Selected sphingolipid metabolic reactions from KEGG, with indications of quantified lipid classes, acyl-chains (circles) and related genes (rectangles). Different colors represent the significant changes of CA relative to AA. E Representative transmission electron microscopy images of vascular smooth muscle cells in CA and AA tissues. Scale bar = 20 μm. The white arrows indicate lysosomes. The number of lysosomes per cell is quantified in right panel. (n = 21; Student t tests; ****P < 0.0001). F Transcription levels of genes associated with the contents of mitochondria (Tomm20), endoplasmic reticulum (Hspa5), and Golgi apparatus (Bet1) in AA and CA. (n = 5; Student t tests; *P < 0.05). G Representative transmission electron microscopy images of vascular smooth muscle cells in CA and AA tissues. “N” indicates nucleus; “M” indicates mitochondrion; “*” indicates Golgi apparatus; white arrows indicate endoplasmic reticulum connections. Scale bar = 5 μm To investigate the causes of sphingolipids’ heterogeneity in CA and AA, the expression of key genes governing the sphingolipid metabolism was examined. As shown in Figure S4B, the expression of Sptlc1 and Sptlc2 was similar between the two tissues (P > 0.05). The expression of different Cers (CERS encoding genes catalyzing the synthesis of ceramides) showed inconsistent changes (Fig. 5B). Compared with AA, Cers1, Cers2, Cers3, and Cers4 (mainly involved in synthesis of ceramides with 18–34 of fatty acyl chain length), but not Cers5 and Cers6 (mainly involved in synthesis of ceramides with 14–16 of fatty acyl chain length) were expressed at higher levels in CA than in AA (Fig. 5B). In addition, the expression of Asah1 (encoding gene of ASAH1 catalyzing the degradation of ceramides) exhibited no difference between the two tissues (Fig. 5B). Furthermore, CA showed a higher expression of Sgms1/2 (encoding gene of SMS catalyzing the synthesis of SM) and Smpd1 (encoding gene of ASM catalyzing the degradation of SM) compared with AA. CERT (ceramide transporter), which is encoded by Cert1, mediates the transport of ceramides from the endoplasmic reticulum (ER) to the Golgi apparatus. Cert1 expression was lower in CA than in AA (Fig. 5C). In addition, the expression of Sphk1, Sphk2 (encoding gene of SPHK1/2 catalyzing the phosphorylation of sphingosine to form S1P), Sgpp1, Sgpp2 (encoding gene of SGPP1/2 catalyzing the dephosphorylation of S1P), and Sgpl1 (encoding gene of SGPL1 catalyzing the decomposition of S1P to ethanolamine phosphate and hexadecanal) was examined. As compared to AA, CA had lower expression of Sgpp1 and Sphk2 (Fig. 5D). The expression of Sgpp2, Sphk1, and Sgpl1 exhibited no significant difference between CA and AA (Fig. 5D). A schematic summarizing the differential regulation of sphingolipid metabolism in AA versus CA is presented in Fig. 5E. Given the vital role of organelles in the synthesis, transformation, and decomposition of ceramides, TEM was applied to observe the subcellular structure of VSMCs in these two tissues (Fig. 5F, H). TEM results indicated that CA had more lysosomes than AA (Fig. 5F). The marker genes’ expression of Golgi apparatus (Bet1), but not ER (Hspa5) or mitochondria (Tomm20), was higher in AA as compared to CA (Fig. 5G). Similarly, the amount of ER and mitochondria counted on TEM images showed no difference between the two tissues (Fig. 5H). All these data reveal that the different ceramide metabolism between AA and CA is evolutionarily conserved in rats and human, which can be attributed to the differential expression of key genes and distinct subcellular organelle features in AA and CA. Ceramides promote the contractile inhibition and calcified susceptibility of VSMCs To further demonstrate whether ceramides are involved in the contractile heterogeneity and calcification sensitivity of VSMCs, we supplemented ceramides directly in the process of inducing contraction and calcification. Intriguingly, exogenous addition of ceramides inhibited the contractility of AASMC and CASMC, and CASMC showed more sensitivity than AASMC (Fig. 6A). Treatment with P053 (inhibitor of CERS) significantly enhanced the contractility due to serum deprivation (Fig. 6B). Additionally, exogenous ceramides accelerated the calcification of both AASMC and CASMC, with a stronger effect observed in CASMCs (Fig. 6C). We further investigated the potential mechanisms of the different calcification sensitivity between AASMC and CASMC after ceramides supplement. Calcification stimulation upregulated the expression of ceramides synthase genes in both VSMCs (Fig. 6D). Nevertheless, the expression levels of these Cers genes in CASMC were higher than that in AASMC (Fig. 6D, S5A-B). Moreover, in addition to Cers, calcification stimulation also accelerated the transition between ceramides and SM, as evidenced by upregulated expression of Sgms1/2 and Smpd1 (Figure S5C-E). Ceramides contribute to the contractile inhibition and calcified susceptibility of VSMCs as well as the response heterogeneity for aortic aneurysms and atherosclerosis stimuli of VSMCs in CA and AA. A Representative images of cell contraction assay of AASMC and CASMC after different treatments for 7 days (left panel). Quantifications of cell contraction diameters are shown in the right panel. Scale bar = 2 mm. (n = 3; two-way ANOVA and Tukey’s multiple comparisons test; *P < 0.05, **P < 0.01, ***P < 0.001). B Representative images of cell contraction assay of AASMC and CASMC after different treatments for 7 days (left panel). Quantifications of cell contraction diameters are shown in the right panel. Scale bar = 2 mm. (n = 3; Two-way ANOVA and Tukey’s multiple comparisons test; **P < 0.01, ***P < 0.001, ****P < 0.0001). C Representative images of Alizarin Red staining of AASMC and CASMC after different treatments for 7 days. The quantifications of Alizarin Red area are shown in right panel. Scale bar = 200 μm. (n = 3; two-way ANOVA and Tukey’s multiple comparisons test; *P < 0.05, **P < 0.01, ***P < 0.001). D Relative expression level of Cers1 and Cers2 in AASMC and CASMC after DPBS or Na3PO4 (3 μM) treatment for 7 days. (n = 3; Student t tests; **P < 0.01, ***P < 0.001). E GO-based GSEA analysis of RNA-seq from GSE47472 (AA-AAA) and GSE100927 (CA-AS). F Correlation heatmap between the expressions of ceramide metabolism genes, VSMC contraction genes, VSMC differentiation genes, and calcification genes from GSE47472. P value was based on Pearson’s correlation. (***P < 0.001). G Correlation heatmap between the expressions of ceramide metabolism genes, VSMC contraction genes, VSMC differentiation genes, and calcification genes from GSE100927. P value was based on Pearson’s correlation. (***P < 0.001). H Representative immunofluorescence staining of α-SMA (green) and ceramides (red) in carotid artery plaque tissue of an atherosclerosis patient. “L” indicates the side of vascular lumen. The plaque area is outlined by white dotted lines. Scale bar = 200 μm (100 μm for zoom-in panels) We then focused on exploring whether ceramide metabolism is involved in the pathological process of aortic aneurysms or atherosclerosis. Based on GSEA analysis, we found that multiple biological processes of sphingolipid metabolism were significantly activated in CA-AS but not in AA-AAA, including ceramide metabolic process, sphingolipid metabolic process, and lysosome organization (Fig. 6E, S5F-G). Similarly, processes involving calcification and VSMCs phenotypic switching were also significantly enriched in CA-AS but not in AA-AAA, such as the activation of positive regulation of calcium-mediated signaling, the inhibition of smooth muscle contraction and response to transforming growth factor beta, eta (Fig. 6E). Moreover, we analyzed the expression correlation of the ceramide metabolism genes, VSMC contraction (or VSMC differentiation) genes, and calcification genes in both AA-AAA and CA-AS. As shown in Fig. 6F and S5H, ceramide metabolism genes had no significant correlations with calcification genes, VSMC contraction or differentiation genes in AA-AAA. Intriguingly, in CA-AS, ceramide metabolism genes were negatively correlated with VSMC contraction and differentiation genes, and were positively correlated with calcification genes (Fig. 6G, S5I). Notably, the distribution of ceramides was significantly distinct in different regions of CA from the patient with atherosclerosis (Fig. 6H). This was reflected in the higher signals of ceramides in plaque areas (marked by lower α-SMA signals), but the lower levels of ceramides in normal vascular middle layer (marked by higher α-SMA signals) (Fig. 6H). These results indicate that ceramides facilitate the loss of contractility and increase the calcification susceptibility in VSMCs. Furthermore, they underscore the heterogeneous response of CA and AA VSMCs to the pathological stimuli of atherosclerosis and aortic aneurysms. Discussion The diversity of arteries is evident in many ways, including structure, function, and fate in a pathological milieu. By analyzing public transcriptomics and carrying out experiments in both human specimens and rat tissues, we revealed the unique characteristics of CA and AA in terms of contractility and calcification. Notably, in comparison to AA, CA exhibited reduced expression of contractile genes and demonstrated greater vulnerability to calcification. Lipidomics analysis disclosed a variation in sphingolipid metabolism between AA and CA, particularly concerning ceramides, which significantly contributed to the vascular differences in physiological regulation and pathological outcomes (summarized in Fig. 7). Heterogeneity of common carotid artery and abdominal aorta: an angle of sphingolipid metabolism in vascular smooth muscle cells Cardiovascular system is a carefully orchestrated mosaic of cell types derived from various developmental origins. VSMC progenitors arise from distinct embryonic sources including splanchnic mesoderm [17], somitic mesoderm [18, 19], neural crest (NC) [19, 20], mesothelia [21], and others. Prior research has demonstrated that vessels susceptible to atherosclerosis maintain their vulnerability even when situated in regions that typically do not develop atherosclerosis [22]. Additionally, differentiation of VSMCs from NC- or mesoderm-lineages was programmed in line with embryonic origins [21]. Thus, the disparities in VSMC characteristics and calcification susceptibility between CA and AA (Fig. 1 and Fig. 2) are likely primarily attributable to their distinct embryonic origins. UPLC–QTOF-MS has been demonstrated as a powerful tool for qualitative characterization of chemical components, facilitating the acquisition of MS spectra with high resolution, sensitivity and mass accuracy [23]. In the present lipidomic analysis, this technique enabled efficient and rapid metabolite profiling [23]. In recent years, lipidomics technologies have provided crucial tools for characterizing vascular pathological features [24, 25]. In our analysis, 169 sphingolipids species from 8 classes were identified in AA and CA, the majority of which were ceramides (76 species) and SM (66 species), indicating these are the principal active sphingolipids in vascular tissues. In prior investigations, six common sphingolipids were identified in carotid plaques, including dihydroceramide (DhCer), ceramides, S1P, SM, lactosylceramide (LacCer), and CerG, which modulate plaque inflammation in various paths [26]. DhCer, ceramides, SM and S1P mediate synergistic functions, while GluCer and LacCer primarily exhibit atheroprotective effects [26]. However, in the current study, DhCer and S1P were either undetectable or present at minimal levels in normal CA and AA tissues. This unevenness further suggests the predominant role of ceramides in comparison to S1P in the physiological and pathological modulation of the two arteries. Furthermore, the lower abundance of ceramides, sphingosine (So), CerG2, GM3, and phSM—but not SM—in AA compared to CA strongly implies an association between sphingolipid profiles and vascular heterogeneity. In mammalian cells, ceramides represent a significant category of hydrophobic lipids capable of forming ceramide-enriched gel domains or platforms [27, 28], containing a sphingoid base (routinely sphingosine) and a fatty acyl chain [28]. Various acyl groups exhibit distinct functions in cells due to their specific biophysical and physicochemical characteristics [29, 30]. In comparison to ceramide-free controls, the melting temperature of C4-C10 ceramide-rich domains decreased, whereas that of C2, C12, or longer chain ceramide-rich domains increased [30]. To replace cholesterol from SM-rich regions, the N-linked acyl chain of ceramides should have at least 8 methylene units [30]. Long acyl chain ceramides influence the dynamics of the membrane outer leaflet by regulating cholesterol and cytoskeleton assembly [31]. In the present work, the fatty acyl chain lengths of ceramides ranged from 6 to 26 (total chain length of 24–44), with the vast majority falling between 14 and 24 (total chain length of 32–42), indicating that fatty acids with long chains (> 12) are involved in ceramide synthesis, while medium (6–12) or short (1–6) chains are not. Furthermore, ceramides with fatty acyl chains of 22–24 carbons (total chain length of 40–42) were associated with hexose (Figure S3E), whereas those with chains of 16–24 carbons (total chain length of 34–42) were linked to sialic acid (Fig. 4F), suggesting a substrate selectivity in glycosylation. In contrast, phosphocholine connected ceramides to form sphingomyelin without a preference for chain length (Figure S3K). To our knowledge, this remarkable distinction has not been previously documented. The saturation of fatty acyl chains also influences the bioactivity of ceramides, particularly regarding membrane characteristics [32, 33]. In fluid lipids, saturated ceramides exhibit poor miscibility, necessitating minimal amounts of ceramides for the formation of gel domains [33]. Furthermore, saturated ceramides with varying acyl chain lengths displayed only slight differences in the melting properties of the gel domains, whereas the chain length of unsaturated ceramide species significantly impacted their capacity for gel–fluid phase separation [33, 34]. In particular, ceramides (C18:1) have difficulty being integrated into gel domains at physiological temperature [33]. Ceramides with short, medium, very long saturated acyl chains (C6:0, C12:0, C24:0), or with unsaturated acyl chains of any length [C6:1(3), C6:1(5), C18:1, and C24:1] tend to destabilize the lipid domains by melting at lower temperatures [35]. In the present study, the saturated ceramides or those with more than 2 double bonds failed to be glycosylated (Fig. 4C, S3D). Saturated ceramides also evaded the binding with phosphocholine to form SM (Figure S3L). Given that glycosylated ceramides are less toxic than ceramides, this finding offers compelling evidence for the assertion that saturated ceramides (C18:0 and C16:0) are more toxic than their unsaturated counterparts (C18:1 and C24:1) [36]. As for the biophysical features, the mix of free saturated and unsaturated ceramides, with the former being dominant, maintains both the strength and liquidity of cell membranes (Fig. 4C). Among the enzymes responsible for sphingolipid metabolism, CERSs are crucial for the synthesis of ceramides via both de novo and salvage pathways, situated on the endoplasmic reticulum (ER) [37, 38]. In mammalian cells, there are six CERS isoforms (CERS1–6), each of which is expressed in specific tissue and demonstrates a substrate preference for fatty acids [37–42]. CERS1 is predominantly expressed in mouse brain and skeletal tissues. CERS2 is broadly expressed across most mouse tissues, abundant in liver and kidney. CERS3 is enriched in mouse testis and skin tissues. CERS4 exhibits the highest expression in mouse skin, leukocytes, heart, and liver tissues. CERS5 is expressed in minimal quantities in most tissues, with the exception of skeletal muscle, testis, and kidneys. CERS6 also shows a low level in most tissues, with a higher expression in intestine and kidney tissues [40]. This tissue-specific expression pattern of CERS isoforms likely underlies the variation in ceramide species and their relative concentrations across different structures. Furthermore, dysregulation of CERS is implicated in the pathogenesis of various diseases. In our study, the expression of Cers1-4 was diminished, while Cers5-6 remained similar in AA compared to CA, providing further evidence for the tissue distribution of ceramides, which is consistent with the reduced abundance of ceramides in AA compared to CA. (Fig. 5B). Meanwhile, ceramides generation relied also on the SM hydrolysis, mostly catalyzed by ASM at lysosomes. The present study unveiled lower Smpd1 expression and fewer lysosomes in AA, indicating a relatively reduced capacity for this pathway (Fig. 5F-H). The rate-limiting step of the de novo ceramide synthesis pathway is initiated by SPT, the enzyme encoded by the by Sptlc genes [43]. The present study failed to detect the significant difference in Sptlc expression between the two arteries, suggesting that this pathway may not be a major contributor to the heterogeneity of CA and AA. Moreover, the degradation of ceramides produces So, which serves as a substrate for the enzymes of sphingosine kinase 1 and 2 (SPHK1/2). Sphk1 and Sphk2 have distinct subcellular localizations and functions [44]. Upon activation, the former predominantly binds to the plasma membrane, while the latter primarily localizes to the nucleus. In cells overexpressing Sphk2, accumulation of S1P within nuclear compartment inhibits the activity of histone deacetylase (HDAC), enhancing the acetylation of histone lysine residues [45]. Here, we observed that Sphk2 but not Sphk1 gene expression was elevated in AA compared to CA, partly elucidating the reduced levels of ceramides and So in AA tissues. This finding warrants further examination to determine whether the higher Sphk2 contributes to the pathological response of AA by modulating histone acetylation. SGPP1/2 and SGPL1 catalyze the dephosphorylation and decomposition of S1P, respectively. The elevated expression of Sgpp1 and Sgpl1 in AA (Fig. 5D) implies active catabolism of sphingolipids, which diminished the abundance of ceramides and So, or potentially the whole sphingolipids even further. SMS1 and SMS2 (encoded by Sgms1 and Sgms2) control the SM synthesis, with SMS1 mainly located in Golgi apparatus [46] while SMS2 on plasma membrane [47]. SMS2 is positively correlated with SM levels of lipid raft [48–50], and is essential for maintaining the fluidity of plasma membrane microdomains [51]. This is why SMS2 is directly associated with cell membrane lipid messengers and is crucial for cell survival and apoptosis [48, 52]. In the present study, contrary to the generally stable SM levels in CA and AA, both SMS1 and SMS2 exhibited lower levels in AA compared to CA (Fig. 5C). This apparent discrepancy, however, may be mitigated by the reduced SM degradation catalyzed by ASM in lysosomes. Due to the markedly hydrophobic feature of ceramides, they stay in the cellular membranes under physiological conditions, unable to diffuse freely inside cells [52]. Typically, following their synthesis via a de novo pathway in ER, ceramides migrate to the Golgi apparatus where they are converted into complex sphingolipids, such as GluCer and SM. This translocation is facilitated by either CERT, an ATP-dependent non-vesicle transporter, or ATP-independent vesicles [53, 54]. CERT has been reported to efficiently transfer ceramides with C14:0, C16:0, C18:0, and C20:0 acyl chains [55]. In muscle cells, diminished CERT levels trap ceramides in ER, hindering SM biosynthesis and reducing insulin sensitivity [56]. The subcellular localization of CERT determines its functional outcome by directing ceramides to specific organelles. Selective transport of ceramides to the mitochondria rather than the Golgi apparatus by CERT leads to apoptosis of cells [57]. In our study, higher Cert1 expression in AA compared to CA was accompanied by a more pronounced presence of the Golgi marker Bet1 (Fig. 5G). As previously mentioned, this enhanced CERT-mediated transport may help compensate for the lower SMS-catalyzed SM synthesis in AA. Advances in plasma lipidomics have highlighted circulating ceramides as risk indicators for cardiovascular diseases [27]. Ceramides, as a class of sphingolipids, exert multiple effects beyond mere structural components of biomembranes. They are also bioactive molecules engaged in diverse cellular processes, including insulin resistance, apoptosis, and mitochondrial damage [58, 59]. Although the correlations between plasma ceramides and atherosclerotic cardiovascular disease are robust, a direct causal relationship remains to be established [59]. In the present study, differences in ceramide levels were identified as the major contributor to the heterogeneity in sphingolipid metabolism between AA and CA. The ceramide levels in the plaques of atherosclerosis patients were significantly elevated compared to those in normal regions, indicating their potential role in atherosclerosis. In vitro, exogenous ceramide supplementation enhanced, whereas inhibition of ceramide synthesis diminished, the susceptibility of VSMCs to calcification. Our findings demonstrate a mechanism through which ceramides contribute to vascular disease by enhancing the calcification susceptibility of VSMCs. Recent studies have demonstrated that ceramides play a pivotal role in atherosclerosis through multiple interconnected mechanisms. In VSMCs, ceramides induce apoptosis through p38 MAPK/JNK activation and ER stress-mediated CHOP signaling, promoting plaque necrotic core formation [60, 61]. Concurrently, they drive VSMC osteogenic differentiation by activating the TLR4/NF-κB and p38 MAPK cascades, which upregulates osteogenic markers (e.g., MSX2, CBFA1, ALPL) and accelerates vascular calcification [62, 63]. Ceramides also sustain vascular inflammation by stimulating IL-6 release from VSMCs and promoting inflammatory cell migration [64]. Notably, the acid sphingomyelinase (aSMase)/ceramide pathway forms a self-amplifying cycle in which ceramides act as key signaling molecules and trigger a positive feedback loop to enhance their own production [65]. Whether ceramides modulate the differential calcification susceptibility in AASMCs and CASMCs through these signaling pathways requires further investigation. Moreover, given that MYOCD/SRF and RUNX2 serve as master transcriptional regulators of VSMC contractility and calcification, respectively [66, 67], a key future direction is to determine whether ceramides influence VSMC phenotypic switching through epigenetic modifications or by regulating the nuclear translocation of these fate-determining transcription factors. Several limitations of this study should be considered when interpreting the findings. First, although VSMCs are the predominant cell type in the CA and AA, further research is needed to investigate the potential influence of other cell types (e.g., vascular endothelial cells) on the observed heterogeneity in sphingolipid metabolism. Second, while we demonstrated that direct modulation of ceramide levels (via supplementation or synthesis inhibition) regulates VSMC calcification susceptibility, targeting upstream regulators of ceramide metabolism may represent another viable therapeutic strategy. Finally, although our study identifies ceramides as key mediators, the precise downstream signaling cascades and transcription factors that directly mediate the loss of contractility and gain of calcification sensitivity warrant further elucidation in future studies. Conclusions This study uncovered distinct biological features and differential calcification susceptibility between the AA and CA. Integrated multi-omics analyses revealed disparate sphingolipid metabolism, with ceramides identified as a key driver of VSMC calcification susceptibility. Our findings provide novel insights into the role of lipid metabolites in vascular diversity and thus highlight sphingolipid pathways as promising therapeutic targets for site-specific aortic diseases. Materials and methods Human subjects Participants experiments were carried out with Research Ethics Board approval from the XiJing Hospital of the Fourth Military Medical University. The vascular samples used in this study were obtained from one arteriosclerosis patient (male, 45 years old) and one healthy donor (dying from a non-cardiovascular disease). Informed written consent to participate in research was obtained from the patient and the donor. Animals Male Sprague–Dawley male rats (7–8 weeks old, body weight: ~300 g) were obtained from laboratory animal center of Fourth Military Medical University, and individually caged within a 12:12 h light/dark cycle at 23 ± 1 °C with standard lab chow and water available free to access. The rats were randomly assigned to two groups using a random number generator. Investigator blinding was implemented throughout the study. Inclusion criteria required animals to be healthy and of appropriate age and weight. All animal experiments were carried out following the rules of the American Association for the Accreditation of Laboratory Animal Care International and the Care and Use of Laboratory Animals. All procedures were approved by the Animal Care Committee of the Fourth Military Medical University accredited by AAALAC International. Euthanasia was performed using carbon dioxide, according to AVMA Guidelines for the Euthanasia of Animals (2020) and approved by local animal welfare committees. Public data processing For public transcriptomic analysis, GEO datasets of AA tissues from abdominal aortic aneurysm patients (AA-AAA; GSE47472) and CA tissues from atherosclerotic patients (CA-AS; GSE100927) were used in the present study. RNAs differential expression analysis was performed by DESeq2 software between two different groups. The genes with a P-value < 0.05 and fold-change ≥ 2 or ≤ 1/2 were considered differentially expressed genes. Pathway enrichment analysis and GO enrichment analysis was performed using the KEGG (https://www.kegg.jp/) and Gene Ontology database (http://www.geneontology.org/), Bioinformatic analysis was performed by R (V.4.4.1). RNA-seq For bulk RNA-seq of AA and CA tissues from rats, total RNA was extracted using Axypre™ Multisource Total RNA Miniprep kit (Axygen) according to the manufacturer’s protocol. RNA quality was evaluated by Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA) and verified using RNase-free agarose gel electrophoresis. After treatment with Oligo(dT) beads, the enriched mRNA was then fragmented and reverse-transcribed into cDNA using NEBNext Ultra RNA Library Prep Kit for Illumina (NEB #7530, New England Biolabs, Ipswich, MA, USA). The double-stranded cDNA products were then purified by QiaQuick PCR extraction kit (Qiagen, Venlo, The Netherlands), end repaired, added poly(A), and ligated with Illumina sequencing adapters. The adapter-ligated products were size selected by agarose gel electrophoresis. Finally, PCR amplification and sequencing were completed using Illumina Novaseq6000 by Gene Denovo Biotechnology Co. (Guangzhou, China). Raw reads were further filtered by fastp (V.18.0) to get high quality clean reads. Short reads alignment tool Bowtie2 (V.2.2.8) was used for mapping reads to ribosome RNA (rRNA) database. The rRNA mapped reads then will be removed. The remaining clean reads were further used in assembly and gene abundance calculation. An index of the reference genome was built, and paired-end clean reads were mapped to the reference genome (Ensembl_release100) using HISAT V.2.2.4 and other parameters set as a default. The mapped reads of each sample were assembled by using StringTie V.1.3.1 in a reference-based approach. For each transcription region, a FPKM (fragment per kilobase of transcript per million mapped reads) value was calculated to quantify its expression abundance and variations, using RSEM software. RNAs differential expression analysis was performed by DESeq2 software between two different groups. The genes with a P-value < 0.05 and fold-change ≥ 1.5 or ≤ 1/1.5 were considered differentially expressed genes. Pathway enrichment analysis and GO enrichment analysis was performed using the KEGG (https://www.kegg.jp/) and Gene Ontology database (http://www.geneontology.org/), Bioinformatic analysis was performed by Gene Denovo Biotechnology Co., Ltd (Guangzhou, China). Lipidomics 1) Sample pretreatment and lipids separation The rats were randomly assigned to two groups (n = 10). Following anesthesia, the common carotid arteries or abdominal aortas were harvested. After thawing the samples at 4°C, 200 µL of water was added for vortex. Then 240 µL of pre-cooled methanol was added, followed by 800 µL of methyl tert-butyl ether (MTBE) addition, vortex and 20 min of ultrasonic vibration. Then the samples were placed at room temperature for 30 min. After 15 min of centrifugation at 14,000 g and 10°C for 15 min, the upper layer organic phases were collected and blown dry with N2. The quality control (QC) samples are a composite prepared by pooling equal volumes of all the samples to be tested. Ultra-performance liquid chromatography (UPLC) Nexera LC-30A was used to separate lipids from serum samples. The temperature of C18 column was set at 45℃ and the flow rate was 300 μL/min. The mobile phase was composed of acetonitrile/water (A, 6:4, v/v) and acetonitrile isopropanol (B, 1:9, v/v). During the gradient elution, the samples were placed in automatic injector at 10°C. All the samples were analyzed continuously in random order to minimize the disturbance from signal fluctuation. 2) Lipidomic analysis After the separation with UPLC–QTOF-MS, the lipid species were measured by using quadrupole time-of-fight mass spectrometry (QTOF-MS) with negative and positive ion modes on Q Exactive (Thermo Scientific™). The following parameters were employed: heater temp 300 °C, sheath gas flow rate 45 arb, aux gas flow rate15 arb, sweep gas flow rate 1arb, spray voltage 3.0kV, capillary temp 350 °C, S-Lens RF Level 50%, MS1 scan ranges: 200–1800. The mass spectral (M/Z) of lipid species and lipid fragments were obtained with resolution of 70,000 and 17,500, respectively. 3) Data processing of lipidomic study Automated identification and relative quantitation of lipid species were fulfilled by LipidSearch™ software (Thermo Scientific™). The pretreated data were first under multivariate statistical analysis to reveal the general profile of lipid variation, which included principal component analysis (PCA), partial least squares-discriminant analysis (PLS-DA) and orthogonal partial least squares-discriminant analysis (OPLS-DA). By through PCA, the clustering of lipid species was evaluated while the outliers were identified. OPLS analysis maximized the distinction of case and control samples while minimizing the irrelevant variability. The proper number of orthogonal components was determined following sevenfold cross-validation. The R2Y denoted the variability justified by the model while the cross-validated Q2Y indicated performance of the model in cross-validation datasets. With the statistical modeling, top ranking lipid species with FDR (False Discovery Rate correction using the Benjamini–Hochberg procedure) adjusted P-value ≤ 0.05, went forward for further identification. Next, the univariate statistical analysis was performed for the comparison of the lipid species’ abundance between the two groups, which included Student’s t test, fold change analysis (FC) and correlation analysis. P < 0.05 was regarded statistically significant. Primary VSMC isolation Rat primary abdominal aorta smooth muscle cells (AASMC) or carotid artery smooth muscle cells (CASMC) were isolated from the abdominal aorta (AA) or carotid artery (CA) of adult male Sprague Dawley rats. Anaesthetize the rats by 2% isoflurane, dissect the neck and excise their AA or CA. After removing the adventitia, the tissues were cut open and scraped off the intima gently. Then, tissues were washed with cold DPBS (Sigma), minced into ~ 1mm3 pieces, and digested using 1mg/ml collagenase I (BioFROXX) at 37 °C in water bath for ~ 30 min. After digestion, the supernatant was collected and mixed with DMEM (Gibco). Cells were collected by centrifuging at 800 rpm for 5 min and further resuspended in DMEM supplemented with 10% FBS (Gibco), streptomycin (100 μg/ml) and penicillin (100 IU/ml) (Gibco). The purity of primary VSMCs was tested by immunofluorescence staining of α-SMA (Proteintech, Cat. #14,395–1-AP, 1:200). Isolated AASMC or CASMC were maintained in low-glucose DMEM (Gibco) containing 10% FBS (Gibco), and cells at passages 4 to 8 were used for further experiments. Before treatment, SMCs were seeded in 6-well plates (2 × 105 cells) or 20 mm confocal dishes (8 × 104 cells) and cultured for 12h. Then, serum-free DMEM were used for serum deprivation study. For inducing calcification, primary SMCs were cultured with Na3PO4 (3μM) and serum-free DMEM. Ceramides (MCE, Cat#HY-113679, 30μM) and P053 (MCE, Cat# HY-126015, 0.5μM) were used for intervening ceramides contents. All cells were maintained in a humidified 5% CO2 incubator at 37°C. Cell contraction assay The contractile ability of the cells was evaluated by Cell Contraction Assay Kit (Cell Biolabs, CBA-021) according to the manufacturer's instructions. Cells were harvested and resuspended in culture medium. A collagen lattice was prepared by mixing the cell suspension and ice-cold collagen gel solution in a volume ratio of 1:4. Then, 0.1 mL of the cell-collagen mixture (containing 104 cells) was added to a 96-well plate and incubated for 1 h at 37°C. After collagen polymerization, 0.1 mL of culture medium was added on top of each collagen gel lattice. 0, 6, 12 or 24 h later, the plates were scanned. For data analysis, the area of the gel in each well was analyzed by Fiji Image J (NIH, Bethesda, MD, United States). Immunofluorescence and confocal microscopy For cells, AASMC and CASMC were seeded on confocal dishes at an appropriate density. After treatments, the cells were washed with DPBS twice and then fixed with 4% formaldehyde buffer (PFA) at 37 °C for 20 min. After washing 2–3 times with PBS, the cells were permeabilized and blocked with 5% BSA and 15% goat serum in 1 × PBS for 1 h at 25 °C. Then, the primary antibodies were incubated at 4°C overnight. For tissues, samples were fixed in 4% formaldehyde buffer (PFA) at 4 °C for 24 h and embedded in paraffin. For rat samples, after euthanasia, rats were transcardially perfused with PBS followed by 4% PFA. The vessels were isolated and embedded in paraffin. Successive 8 μm sections were cut for further experiments. Then, paraffin was removed by using xylene and samples were dehydrated using ethanol of graded concentrations. Antigen retrieval was completed using high pressure cooking in 1 mM EDTA buffer (pH = 9.0). Tissue sections were blocked with 5% BSA and 15% goat serum in 1 × PBS for 1 h at 25°C. Then, the primary antibodies were incubated at 4°C overnight and Alexa Fluor–conjugated secondary antibodies were incubated at 25°C for 2 h. The following primary antibodies were used: α-SMA (Proteintech, Cat. #14,395–1-AP, 1:200); ceramides (Enzo, Cat# 09082115, 1: 50). Alexa Fluor® AffiniPure IgG (H + L) were used as secondary antibodies. Nuclei were stained using DAPI (Sigma, Cat. # D9542). Finally, the images of stained cells were taken on Nikon A1 PLUS. Protein isolation and western blot analysis Cells were washed by ice-cold PBS and lysed with RIPA lysis buffer (Beyotime, Cat. #P0013B) supplemented with phosphatase and protease inhibitors (Roche). Lysates were separated on 10%-12.5% SDS-PAGE gels (Beyotime, Cat. #P0012AC), and then transferred to 0.45 μm polyvinylidene difluoride membranes (Merck Millipore) by Trans-Blot System (Bio-Rad Laboratories). Membranes were then blocked with 5% skimmed milk diluted in TBST for 1h at 25 °C and then were incubated in diluted primary antibodies overnight at 4°C. Subsequently, the membranes were incubated with secondary antibodies conjugated with horseradish peroxidase (HRP) for 1h at 25 °C. Western HRP substrate (Millipore) and imaging system were employed to blot detection. The antibodies used in this study are as followed: α-SMA (Proteintech, Cat. #14,395–1-AP, 1:1000); SM22α (Proteintech, Cat. #, 1:1000); β-Tubulin (Abways, Cat. #AB0057, 1:2000). RNA extraction and quantitative real-time PCR (qRT-PCR) Total RNA was extracted from tissues (carotid artery and abdominal aorta) according to the protocols of Axypre™ Multisource Total RNA Miniprep kit (Axygen). The isolated RNA was then reverse-transcribed by using PrimeScript™ RT Mix with gDNA eraser (Takara). SYBR®Premix Ex Taq™ II (Takara) was used for qRT-PCR by using 7500 system (Applied Biosystems). All the primers are listed in Supplementary Table 2. The relative mRNA expression was analysis using the comparative threshold cycle method (2−ΔΔCt) using β-actin as the endogenous control gene. Transmission electron microscopy study After washing with PBS, tissues (carotid artery and abdominal aorta) were fixed in glutaraldehyde (2.5%) and sodium cacodylate buffer (0.1 M) at 4 ℃ overnight. All the samples were embedded and subjected to ultramicrotomy. Transmission electron microscope images were captured by JEM-1400 (Japan Electron Optics Laboratory) and iTEM 5.2 (EMSIS). Quantification of the intravascular elastic membrane (μm2) were analyzed by Fiji Image J (NIH, Bethesda, MD, United States). Statistics Data are expressed as mean ± s.e.m. Statistical analysis and graphical representation were performed in the GraphPad Prism software (version 9.0.0) or Rstudio (version 4.1). Normality of the data was evaluated using the Shapiro–Wilk test. Two-tailed unpaired Student t tests were used to assess the differences between two groups. The differences between multiple groups were compared using two-way ANOVA, followed by the Tukey’s multiple comparisons test, when normal distribution was satisfied; otherwise, the nonparametric Kruskal–Wallis test was performed for further analysis. A 95% confidence interval was computed for each mean difference. A P-value of < 0.05 was considered statistically significant. Statistical significance was set at P < 0.05; two asterisks (**) for P < 0.01; three asterisks (***) for P < 0.001; and four asterisks (****) for P < 0.0001. Supplementary Information Supplementary file 1. Supplementary file 2. Supplementary file 3. Author contributions JB and YC supervised the project. BL and KZ were major contributors in writing the original manuscript. SJ, XW, and ZG established the animal models. CG, YH, and YL collected samples from humans. BL, KZ, and KC: software, methodology, and qRT-PCR. ZG and XW: data collection, and data validation. All authors contributed to the article and approved the final manuscript. Funding This study was supported by National Natural Science Foundation of China (No. 82171866), Research program of Air Force Medical University of China (2023RCJB09, 2024GJJH01-03, 2023KXKT093); "Clinical Medicine + X" research program of Xijing hospital of China (LHJJ24KQ03). Data availability The data that support the findings of this study are available from the corresponding author upon reasonable request. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Bowen Li, Kaixiang Zhou and Kun Chen contributed equally. Contributor Information Yonghong Liu, Email: liuyhong@fmmu.edu.cn. Yaoming Chang, Email: changym@fmmu.edu.cn. Junxiang Bao, Email: baojunx@fmmu.edu.cn. References Associated Data Supplementary Materials Supplementary file 1. Supplementary file 2. Supplementary file 3. Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request.