1. Construction of Rat Single-Cell Atlases by scRNA-Seq
Although scRNA-seq has extensively profiled vascular changes in cardiovascular and metabolic diseases, a detailed single-cell analysis of arterial adaptations in hindlimb-suspended rats remains absent. Addressing this gap, we conducted a systematic scRNA-seq analysis of over 22,000 individual cells derived from the carotid and femoral arteries of rats that underwent 3 months of hindlimb unloading, with 1G-exposed rats serving as controls (Fig. 1A). Prior to clustering, we excluded cells with gene counts exceeding 6000 or falling below 500. Additionally, cells exhibiting over 40,000 UMIs or more than 20% mitochondrial gene content were filtered out. After quality control, 6742, 2983, 4487, 2463 cells were acquired in Con-CA (carotid artery in contol group), Con-FA (femoral artery in contol group), HU-CA (carotid artery in hindlimb unloading group) and HU-FA (femoral artery in hindlimb unloading group) datasets (Fig. 1B), which correspond to the median of 2260.5, 2064, 1909 and 1825 genes per cell, respectively.
Individual clusters were labeled for cell type with known marker genes. In total, we identified 23 cell clusters (Fig. 1C) that could be divided into 10 major cell types (Fig. 1D). Integrated datasets of above four groups displayed satisfactory alignment. Known markers used to divide cell types were shown in Additional file 1: Table. S1. Vascular cell types represented in the distinct clusters included ECs (cluster 5, 7, 11, 13, 20), VSMCs (cluster 0, 3, 9, 10, 14, 15), fibroblasts (cluster 1, 2, 4, 6, 12), macrophages (cluster 8), B cells (cluster 21), T cells (cluster 18), dendritic cells (cluster 17), neutrophil (cluster 19), neuron (cluster 22), and stem cells (cluster 16). The expression of marker genes across the 10 cell clusters were shown in Additional file 1: Fig. S1.
We then examined the distribution of major cell types within each group (Fig. 1E, F, Additional file 1: Table. S2). Notably, the proportion of SMCs was elevated in the CA relative to the FA. Interestingly, a reduction in SMC fraction was observed in HU-CA and HU-FA compared to Con (Fig. 1F). This finding contradicts prior studies suggesting that simulated weightlessness stimulates the proliferation of CA SMCs in rats[5]. Additionally, HU-CA demonstrated a higher percentage of ECs compared with Con-CA. Moreover, the HU-FA group exhibited increased percentage of immune cells—comprising B cells, T cells, dendritic cells, and neutrophils—compared to Con-FA.
Then, we presented top 5 highest differential expression genes relative to all other cells in each cluster (Fig. 1G). For example, the specific markers of ECs are Mall, Cldn5, Vwf, Mmrn2, Gja5. Vwf is the canonical ECs gene. Cldn5, encoding a tight junction protein, and Mmrn2, encoding an extracellular matrix protein known as Multimerin 2, were also highlighted. The markers of SMCs are Acta1 and potassium channel subfamily K member 3 (Kcnk3). Fibroblast marker is Scara5, a class A scavenger receptor with the clearance function of foreign and endogenous material[9]. Here we also showed the number of upregulated genes with average fold change > 1.28, P value < 0.05 for each cell type compared to all other cells (Fig. 1H). Consistent with previous findings, monocytes including dendritic cells and macrophages have more differentially expressed genes than other cell clusters[10].
2. Artery-specific differentially expressed genes between Con and HU group in ECs and distinct gene expression profiles for EC sub-clusters
The report about the differentially expressed genes in SMCs, ECs, fibroblasts and some immune cells from CA and FA under microgravity is lacking so far due to the challenge in isolating each vascular cell type. After sequencing of flow-sorted vascular cells, datasets from different cell types were aggregated in this study.
We first focused on the differentially expressed genes in ECs between Con and HU (Fig. 2A, B). For ECM remodeling, genes encoding collagen such as col4a1, col4a2 were downregulated in both HU-CA and HU-FA. For secretory function of ECs, the increase of Vcam1 and Nos3 suggested enhanced vasodilatation and adhesion of leukocytes in HU-CA ECs. Downregulation of Edn1 which encodes ET-1 may give rise to decreased vasoconstriction and basal vascular tone in HU-CA. The protein levels of eNOS (Fig. 2C), which was coded by Nos3, and ET-1 (Fig. 2D) were verified by immunofluorescence. Mmrn1, which enhances platelet adhesion and thrombus formation[11], and Vwf were upregulated in HU-CA. Ccn1, essential for atherogenesis, induced by oscillatory shear stress and inhibited by laminar shear stress[12], was upregulated in HU-FA but downregulated in HU-CA, highlighting the differential shear stress impacts on the CA and FA. KEGG pathway analysis revealed an upregulation of genes encoding ribosomal proteins, such as Rpl37, Rpl36, Rpl35, in ECs from both HU-CA and HU-FA when compared to their Con counterparts (Fig. 2E, F).
Sub-clustering analysis of ECs revealed 6 sub-clusters. Sub-cluster 0 showed high expression of Isl1 (Fig. 2G), a marker of embryonic stem cell-derived progenitors[13], indicating the potential of self-renewal. The sub-cluster also featured increased expression of Bmp6, a gene implicated in osteogenic differentiation of ECs during vascular calcification[14]. This sub-cluster showed enrichment for KEGG terms related to protein processing in endoplasmic reticulum (Fig. 2H), which was confirmed by GSVA analysis result that sub-cluster 0 was involved in protein secretion and unfolded protein response (Fig. 2I). Compared to FA, more ECs from CA were in sub-cluster 0 (Fig. 2J). Contrary to sub-cluster 0, sub-cluster 1 demonstrated a more enrichment of cell number in FA (Fig. 2H). Lipe and Gpihbp1, involved in lipolytic process[15], were two markers for sub-cluster 1. Sub-cluster 1 specifically exhibited enrichment for the “Rap1 signaling pathway” and its downstream “PI3K/AKT/mTOR signaling pathway” (Fig. 2K), which are known to play roles in cell adhesion, migration, and angiogenesis. Consistent with this, Cxcl12, another marker of sub-cluster 1, is recognized for its involvement in endothelial migration[16]. For EC sub-cluster 2, marker genes Rpl30 and Rps25 were crucial for ribosome assembly and protein translation. Consistently, GO enrichment analysis indicated significant enrichment for the terms “Translation” and “Structural constituent of ribosome” in EC sub-cluster 2 (Additional file 1: Fig. S2). The EC sub-cluster 3 expressed Lyve1, a lymphatic endothelial marker, suggesting this population was lymphatic endothelium. The definition of sub-cluster 3 was further confirmed by the elevated expression of Ccl21, a chemokine produced by lymphatic endothelial cells. Notably, there were more lymphatic endothelial cells in FA than that in CA (Fig. 2J). Cnn1, a canonical marker of SMCs and a late indicator of EndoMT, is a hallmark of EC sub-cluster 4[17]. EndoMT which is integral to the pathogenesis of conditions such as cardiac fibrosis[18] and atherosclerosis[19], is a process whereby endothelial cells undergo phenotype changes towards to mesenchymal cells such as SMCs and fibroblasts. The sub-cluster 4 also expresses the myogenic marker Acta2 which encodes α-SMA and Pln which encodes phospholamban, a regulator of ER Ca2+ homeostasis in angiogenesis[20], suggesting a mesenchymal cell phenotype due to EndoMT.
To delve deeper, we employed pseudotime analysis to simulate the EndoMT transition of EC sub-clusters into fibroblasts and SMCs. As shown in Figure. 2L, a subset of EC sub-cluster 4 appeared to be positioned between the ECs and SMCs in biological evolution. Besides, sub-cluster 5 of ECs was located at the intermediate stage of differentiation between fibroblasts and endothelial cells. Then 3 different cell states were identified: state 1, 2, 3. The ECs were primarily distributed in states 1, located in the initial state of the cell differentiation trajectory. The fibroblasts and SMCs were predominantly distributed in states 2 and 3, representing the final stage of the differentiation trajectory. Subsequently, genes with significant changes over pseudotime were clustered, followed by KEGG enrichment analysis for each cluster (Figure. 2M). Cluster 1 gene set manifested protein processing in endoplasmic reticulum. Cluster 2 gene set exhibited enrichment of vascular smooth muscle contraction. Additionally, we used gene clusters 1 and 2 as target genes for TF prediction and found Egr2, Spi1 and Nrf1 may play important roles in the differentiation process of ECs to fibroblasts, and Klf15, Nr5a2 and Zfx may be crucial for the transition of ECs into SMCs (Additional file 1: Fig. S3). We also compared the pseudotime trajectories between CA and FA and found that the process of EC sub-cluster 4 differentiating into SMCs occured only in the CA while transition of ECs into fibroblasts occurred in CA and FA (Figure. 2N). Besides, EC sub-cluster 4 from HU-CA exclusively exhibited differentiation into SMCs compared to Con-CA, as shown in Figure. 2O. Tail suspension simulated weightlessness might also promote transition of EC sub-cluster 4 to fibroblasts in FA (Additional file 1: Fig. S4).
(A, B) Volcano plot of differentially expressed genes (average log fold change > 0.23, average fold change > 1.28, P value < 0.05) in HU-CA ECs (A) and HU-FA ECs (B) in comparison with Con counterparts. (C, D) Immunostaining of eNOS (C) and ET-1 (D) in Con and HU-CA. Scale bar, 40 µm. (E, F) KEGG analysis of sub-differentially expressed genes in HU-CA ECs (E) and HU-FA ECs (F) in comparison with Con counterparts. (G) Heatmap showing EC sub-cluster identity and the top 5 marker genes of each EC sub-cluster. (H) KEGG analysis of specific genes in EC sub-cluster 0. (I) GSVA analysis of EC sub-cluster-specific genes. (J) Percentage of each sub-cluster of ECs. (K) KEGG analysis of specific genes in EC sub-cluster 1. (L) The trajectory distribution of EC sub-clusters, SMCs and fibroblasts and 3 states in differentiation trajectories. (M) Differentially expressed genes across pseudotime which were divided into cluster 1 and 2 and the represented biological pathways from KEGG analysis of gene cluster 1 and 2. (N) Pseudotime trajectories of EC sub-cluster 4 and 5 in CA + FA, CA and FA. (O) Pseudotime trajectories of EC sub-cluster 4 in Con-CA and HU-CA. Scale bar, 40 µm.
3. Artery-specific differentially expressed genes between Con and HU group in SMCs and distinct gene expression profiles for SMC sub-clusters
In our study, SMCs exhibited heightened transcription of Acta2 in both HU-CA and HU-FA, signifying an increase in α-SMA levels (Fig. 3A, B). This could explain the observed increase in carotid artery intima-media thickness in tail-suspended rats[5], despite a reduced SMC cell count in HU-CA compared to Con-CA. Our research also confirmed the increased intima-media thickness in HU-CA (Fig. 3C). The selective upregulation of the ryanodine receptor (Trdn), the voltage-dependent L-type calcium channel (Cacna1d) and Pln, all of which are integral to ER Ca2+ homeostasis[20], indicated a hypercontractile state in HU-CA SMCs. The protein level of PLN was verified by immunostaining (Fig. 3D). Regarding the ECM, HU-CA SMCs showed a decrease in genes encoding elastin (Eln), collagens (Col1a1, Col4a1, Col6a1), laminins (Lamc1, Lamb1, Lama2), thrombospondins (Thbs2, Thbs3), and fibronectin (Fn1) when compared to Con-CA. Notably, genes encoding components of the ribosome, such as Rpl37 and Rpl41, were upregulated in HU-CA SMCs relative to Con-CA, whereas a downregulation of genes Rps2 and Rps27a was noted in HU-FA compared to Con-FA (Fig. 3B, Additional file 1: Fig. S5). Additionally, the expression level of genes (Spp1, Msx2, Bmp2) encoding vascular calcification markers (OPN, MSX2, BMP2) were elevated in HU-FA compared to Con-FA, indicating the tendency of calcification of FA but not CA. Collectively, these results highlighted the heterogeneity between the forebody and hindbody arteries of hindlimb-unloaded rats, likely due to differential transmural pressures across the CA and FA as a result of hindlimb unloading[21].
Sub-clustering analysis of SMCs revealed 6 sub-clusters (Fig. 3E). Sub-cluster 0 was characterized by elevated levels of Col4a6, Acta2, and glutamine synthetase 2 (Gls2), a regulator of glutaminolysis and ferroptosis. The enriched GO term “organonitrogen compound metabolic process” highlighted this sub-cluster’s metabolic role (Fig. 3F). Sub-cluster 1 was marked by the presence of Sacsin, encoding a chaperone involved in intermediate filament assembly[22] and the ryanodine receptor 2 (Ryr2), crucial for calcium release and SMC contraction. Sub-cluster 2 featured the gene for vitronectin (Vtn), an extracellular matrix glycoprotein involved in cell attachment, adhesion, and migration through integrin receptors[23], and Higd1b, which maintains mitochondrial integrity under hypoxia[24]. This sub-cluster was more prevalent in the FA than in the CA (Fig. 3G). Sub-cluster 3 was defined by Ndufa4, a component of the mitochondrial respiratory chain essential for mitochondrial function and energy metabolism[25], and Rps20, which encodes a ribosomal protein. The sub-cluster’s contribution to ribosome construction was indicated by the GO terms “structure constituent of ribosome” (Fig. 3H). The markers of sub-cluster 4 are oncofetal gene Gpc3 and SMCs progenitor marker Pi16[26]. According to GSVA (Fig. 3I), sub-cluster 4 showed enrichment of “Hedgehog signaling pathway” which is essential in embryogenesis and a key regulator of stem cell[27, 28]. High expression of ECM-related genes in this sub cluster, demonstrated by GO enrichment analysis (Fig. 3J), heralded that sub-cluster 4 was major contributor to the surrounding microenvironmental vascular matrix. The marker of sub-cluster 5 Esco2 is involved in tethering between sister chromatids during DNA double-strand break repair[29]. Another marker Gas2l3 is responsible for linking actin and microtubule filaments in interphase of cells and plays a pivotal role in mitosis and cytokinesis[30]. Consistently, GSVA revealed that “G2M checkpoint”, “mitotic spindle signaling pathway” were enriched in SMCs sub-cluster 5. The enrichment of GO term “cell cycle” (Fig. 3K) and the enrichment of KEGG terms “DNA replication” and “spliceosome” (Fig. 3L) further confirmed the role of this sub-cluster in cell cycle. Notably, this sub-cluster had a higher percentage of G2-phase cells and a lower percentage of non-cycling cells compared to others (Fig. 3M). The percentage of sub-cluster 4 and 5 in CA and FA was lower after hindlimb unloading (Fig. 3G), suggesting that hindlimb unloading may inhibit the cell cycle of SMCs.
(A, B) Volcano plot of differentially expressed genes (average log fold change > 0.23, average fold change > 1.28, P value < 0.05) in HU-CA SMCs (A) and HU-FA SMCs (B) in comparison with Con counterparts. (C) Representative photomicrograph of CA and FA of rats in both Con and HU group showing histo-morphological changes in medial thickness. (D) Immunostaining of α-SMA and PLN in Con and HU-CA. (E) Heatmap of top 5 marker genes of each SMC sub-cluster. (F) Go analysis of specific genes in SMC sub-cluster 0. (G) Percentage of each sub-cluster of SMCs in each group. (H) Go analysis of specific genes in SMC sub-cluster 3. (I) GSVA analysis of SMC sub-cluster-specific genes. (J) Go analysis of specific genes in SMC sub-cluster 4. (K) Percentage of cells at different stages of cell cycle in each SMC sub-cluster. (L, M) Go analysis (L) and KEGG analysis (M) of specific genes in SMC sub-cluster 5. Scale bar, 40 µm. *P < 0.05
4. Artery-specific differentially expressed genes between Con and HU group in fibroblasts and distinct gene expression profiles for fibroblast sub-clusters
The outer adventitia is predominantly composed of collagen and fibroblasts. Fibroblasts showed the greatest number of both downregulated and upregulated genes in HU group compared to Con counterparts across all clusters (Fig. 4A, B), highlighting their high plasticity, which is consistent with previous findings that adventitial plays a pivotal role in vascular remodeling in response to hyperlipidemia or balloon injury[31, 32]. As for ECM, an upregulation of Serpine1 and Serpine2 (Fig. 4C), which encode PAI-1 and PN-1 respectively, indicated heightened antifibrinolytic activity[33] and suggested a limitation in matrix breakdown in HU-CA. The upregulation of Nox4 and Txnip (Fig. 4C, D) may give rise to increased oxidative stress and NLRP3 inflammasome activation[34] in HU-CA and HU-FA. The marker of fibroblast Pdgfra (Fig. 4C, D) which drives fibrosis and proliferation of fibroblast was upregulated in both HU-CA and HU-FA. The protein level of TXNIP (Figure. 4E), PDGFRα and SERPINE2 (Fig. 4F) was verified by immunofluorescence. The upregulation of progenitor marker Pi16 (Fig. 4C, D) in HU-CA and HU-FA suggested the increased progenitor group of fibroblasts[26]. Phenotypic transformation of adventitial fibroblasts, which is characterized by elevated expression of Acta2 (Fig. 4C, D), is important in the process of vascular diseases[35]. This upregulation of Acta2 suggested a shift of fibroblasts towards a myofibroblast phenotype in HU-CA and HU-FA. GSVA analysis revealed an enrichment of myogenesis-related genes in HU-CA (Additional file 1: Fig. S6A), hinting at fibroblast involvement in the modulation of intima-media thickness during hindlimb unloading. It is worth noting that, following hindlimb unloading exposure, differentially expressed genes were obviously enriched in metabolic process in fibroblast of CA and FA after microgravity according to GO terms analysis (Fig. 4G, H).
Sub-clustering analysis of fibroblasts revealed 5 sub-clusters (Fig. 4I). Col1a1 and the glutamate transporter Slc1a3 were markers for sub-cluster 0. The markers of sub-cluster 1 Mup5 was involved in glucose and lipid metabolism[36]. The enrichment of GO terms (Fig. 4J) and GSVA analysis (Additional file 1: Fig. S6B) confirmed this sub-cluster’s role in glycolysis and mitochondrial oxidative phosphorylation. The marker of fibroblast sub-cluster 2 included Alpl, an osteogenic marker which is involved in vascular calcification, and F2rl1 which encoded PAR2, a modulator in vascular inflammation and atherogenesis[37]. Another marker of sub-cluster 2 was Adamts8, a secreted proteinases binding to extracellular matrix involved in ECM remodeling. Sub-cluster 3 maybe progenitor of fibroblast as demonstrated by the expression of Procr, a signature gene providing stem cell signal in various type of cells[38]. Another marker Efhd1 was involved in mitochondrial morphology and energy metabolism in Ca2+ dependent and independent manner[39]. The marker of sub-cluster 4 Spp1 is involved in the process of atherosclerotic plaque formation[40]. GSVA analysis shows an enrichment of genes associated with angiogenesis (Additional file 1: Fig. S6B).
(A, B) Dot plot showing the expression of selected differentially expressed genes in HU-CA fibroblasts (A) and HU-FA fibroblasts (B) in comparison with Con counterparts. (C, D) Immunostaining of PDGFRα, TXNIP(C) and SERPINE2 (D) in Con and HU-CA. (E, F) GO analysis of differentially expressed genes in HU-CA ECs (E) and HU-FA ECs (F) in comparison with Con counterparts. (G) Heatmap of top 5 marker genes of each fibroblast sub-cluster. (H) Go analysis of specific genes in fibroblast sub-cluster 1. (I) GSVA analysis of specific genes in each fibroblast sub-cluster. Scale bar, 40 µm.
5. Artery-specific differentially expressed genes between Con and HU group in macrophages and differentially expressed genes shared across several cell types
Macrophages are crucial for angiogenesis[41] and the formation of atherosclerosis plaques[42]. The binding of CXCR4 with chemokine CXCL12 activates of mTOR and NF-κB signaling pathway, which are known to promote cellular growth and proliferation[43]. The upregulation of Cxcr4 in HU-CA and HU-FA macrophages (Fig. 5A) and the upregulation of Cxcl12 in HU-CA fibroblasts suggested that the interaction of fibroblast CXCL12 with macrophage CXCR4 might be enhanced. The upregulated protein expression of CXCR4 in HU-CA was verified through immunofluorescence (Fig. 5B). KEGG enrichment analysis revealed that upregulated genes (Jund, Nfkbia, Hsp90ab1) were mainly enriched in IL-17 signalling in HU-CA and HU-FA, and downregulated genes (Cyb, Atp6, Cox2, Nd1, Nd4l, Nd5, Nd6) were mainly enriched in oxidative phosphorylation in HU-FA (Fig. 5A, D, E).
Then we scrutinized genes with varying expression levels across multiple cell types to elucidate general vascular remodeling processes (Fig. 5F, J). Notably, several TFs were found to be dysregulated. Klf4, which is involved in the mechanotransduction mechanism and up-regulates many atheroprotective genes[44], was upregulated in HU-CA ECs and fibroblasts as well as in four HU-FA cell types. The upregulation of Cebpb and Cebpd, which encode the transcription factor C/EBPs, was observed across ECs, SMCs and fibroblasts in HU-CA and HU-FA, suggesting the possibility of overall inflammatory state[45]. This was further supported by increased levels of Fosb and Jun, which encode components of the proinflammatory AP1 transcription factor, in several cell types in both HU-CA and HU-FA. Additionally, the exclusive downregulation of ECM–related genes was observed, consistent with previous report that ECM related genes were downregulated in CA of mice[46]. Specifically, Lox, Eln and Dpt were downregulated across HU-CA and HU-FA endothelial cells, SMCs, and fibroblasts, indicating impaired vascular integrity and elasticity. The downregulation of col1a1, col3a1, and Sparc, which are involved in collagen binding and maturation, suggests compromised basal lamina assembly and cellular adhesion[47] in CA and FA after hindlimb unloading. Protein level confirmation was obtained for the upregulation of KLF4, FOSB, CEBPB and the downregulation of SPARC, DPT in HU-CA. (Fig. 5H, I, J, K, L).
-
Dot plot showing the expression of selected differentially expressed genes in HU-CA macrophages. (B) Immunostaining of CD68 and CXCR4 in Con and HU-CA. (C) Volcano plot of differentially expressed genes (average log fold change > 0.23, average fold change > 1.28, P value < 0.05) in HU-FA macrophages in comparison with Con counterparts. (D, E) KEGG analysis of differentially expressed genes in HU-CA macrophages (D) and HU-FA macrophages (E) in comparison with Con counterparts. (F, G) Differential gene expression analysis showing the selected up- and down-regulated genes across several cell types in HU-CA (F) and HU-FA (G) compared to their Con counterparts. (H-L) Immunostaining of CEPBPB (H), DPT (I), FOSB (J), SPARC (K), KLF4 (L) in Con and HU-CA. Scale bar, 40 µm.
6. Cellular communication changes in FA and CA after hindlimb unloading in rats
Intercellular communication between heterogenous clusters within artery was revealed leveraging CellChatDB database. Among HU-related changes in ligand-receptor pairs in CA and FA, fibroblasts and SMCs played a crucial role in orchestrating intercellular communication changes, as supported by the more pronounced changes in the number of interactions and interaction strength (Fig. 6A, B).
To elucidate changes in cell-to-cell communication in the CA due to HU, we utilized CellChat analysis to identify key signaling pathways. This analysis underscored the significant role of fibroblasts and SMCs in cellular communication. Specifically, SMCs received signals from Spp1, Laminin, Collagen, Fn1, and Ptn, while fibroblasts were receptive to signals from Mif, Pdgf, and Wnt (Fig. 6C). In terms of signal emission, SMCs released signals including Spp1, Notch, and Ephb, whereas fibroblasts secreted Cd99, Fn1, Sema3, and Igf (Fig. 6D). Then we compared interaction strength between HU-CA and Con-CA (Fig. 6E). For ECM-related ligand-receptor pairs, Col4a1_Sdc4, Col4a4_Sdc4, Fn1_Itga8/Itgb1, Fn1_Sdc4, Lamb2_Itga7/Itgb1 were increased between ECs and SMCs. However, the interactions of ligands including lama4, lama5, lamb1 and lamc1 and receptors including Itga7/Itgb1 and Dag1 within SMCs of CA were decreased after HU. In our study, we also assessed the secreted signaling of specific ligand-receptor pairs. Notably, the anti-inflammatory interaction between Pros1 and Axl was diminished in the communication between fibroblasts and SMCs after HU in CA[48]. Midkine (Mdk), a heparin-binding growth factor involved in cell growth, survival, migration, and angiogenesis[49], exhibited reduced interactions from fibroblasts with its transmembrane receptors Lrp1, Ncl, Sdc2, and Sdc4 on SMCs in HU-CA compared to Con-CA. These ligand-receptor interactions were similarly attenuated within fibroblasts in HU-CA. Furthermore, arterial endothelial Ackr3, known to mediate endothelium-immune cell adhesion[50], showed an increased interaction with Mif from SMCs in HU-CA versus Con-CA. The interaction of Mif_Ackr3 pair was also enhanced within ECs. Additionally, the Pdgfa_Pdgfrb signaling pair was likewise augmented within SMCs.
HU-induced alterations in cellular communication within the FA were also characterized. ECs received signals from Vegf, Cdh5, and Jam, while SMCs were responsive to Spp1, Laminin, and Sema3 (Additional file 1: Fig. S7). Post-HU analysis of the FA revealed changes in the interaction strength of specific ligand-receptor pairs (Fig. 6F). Notably, the Sema3c_Plxnd1 pair, which suppresses angiogenesis[51], exhibited reduced signaling between fibroblasts and SMCs in HU-FA compared to Con-FA. In contrast to the CA, the anti-inflammatory Pros1_Axl interaction was enhanced in fibroblast-SMC and fibroblast-fibroblast communications following HU in the FA. The secreted signaling pair Mif_Ackr3 was strengthened in SMCs-fibroblasts and fibroblasts-fibroblasts cell crosstalk. ECM-related pairs including Col1a1_Cd44, Col6a1_ Itga1/Itgb1, Col6a1_Cd44 were diminished in fibroblast-SMC crosstalk. Unlike in the CA, ECM pairs such as Col1a1_Itga1/Itgb1, Col4a1_Itga1/Itgb1, and Col4a1_Cd44 were also reduced in EC-SMC communications after HU in the FA. Furthermore, cell-cell contact interactions, including F11r_Jam3, Jag2_Notch1, Jag2_Notch2, and Jam2_Jam3, were attenuated in EC-SMC crosstalk in HU-FA relative to Con-FA.
(A, B) Circle plot showing the differential number of interactions and differential interaction strength among different artery cell types in HU-CA (A) and HU-FA (B) in comparison with Con counterparts. The red line represented increased number and strength of interaction. The blue line represented the opposite trend. The line thickness is proportional to the number of ligands-receptors pairs and interaction strength. (C, D) Heatmap of incoming signaling patterns (C) and outgoing signaling patterns (D) of different cell types identified by CellChat analysis in Con-CA and HU-CA. (E, F) Heatmap revealing top predicted differential expressed ligand-receptor pairs between different cell types in HU-CA (E) and HU-FA (F) compared to their counterparts.