Cell subpopulation characteristics of human ascending aorta
Five Stanford type A ATAD samples and 4 normal ascending aortic samples were digested to get single cell suspension. Upon quality control and normalization, 39525 cells were used for further analysis. After identification of cell clusters, the population characteristics of 9 samples were shown in Supplementary file I. In this study, 14 clusters were primarily obtained (Supplementary Figure IA). Upon examination of conserved genes in each cluster, 8 cell types were identified, including VSMCs, ECs, 5 clusters of FBs, macrophages, 2 clusters of monocytes, T lymphocytes, mast cells and 2 clusters of neutrophils. The proportion of each cluster between control and ATAD group was shown in Supplementary Figure IB. The marker genes of these clusters were shown in Supplementary Figure IC-D.
In 3 non-immune cells, most VSMCs highly expressed CALD1, but the traditional contractile marker MYH1115 and synthetic marker MYH1016, 17 exhibited a separated expression pattern, indicating the existence of contractile and synthetic phenotype of VSMCs. However, in those immune cells, a small proportion of neutrophils showed higher levels of S100A12 and CD177, implying activation of these neutrophils. Though all monocytes highly expressed CD163, their markers exhibited a splitted trend, which characterized by high levels of SERPINB2 and EREG in monocytes 1 as well as overexpression of MT1G in monocytes 2. These results revealed the heterogeneity in each cell type, which prompted us to explore the subpopulation composition of these cell types.
Heterogeneous subpopulations of VSMCs in ascending aorta
We got 8 subpopulations upon re-clustering VSMCs (Figure 1A). The composition of subpopulations in each sample was shown in Supplementary Figure IIA. The proportion of each subpopulation in ATAD and control group was shown in Figure 1B.
VSMCs 1 was identified as synthetic VSMCs for its higher expressions of complement activation, ECM and apoptotic genes including CXCL12, CFH, VCAN, MYH10 and IGFBP418, which also expressed growth factors such as BMP4, TGFA, NRG1, FGF9 and GDF5, whereby mediating cell-cell signaling, cell proliferation and differentiation (Figure 1C-E, Supplementary Figure IIB). The genes distinctly expressed in VSMCs 1 played roles in ECM and collagen metabolism, cell adhesion, antigen-processing and interferon response, which were consistent with the functions of synthetic VSMCs (Supplementary Figure IIC-D). It also exhibited increased type I IFN response, ROS pathway and oxidative phosphorylation to regulate inflammation, oxidative stress and enhanced energy metabolism (Figure 1F). Further analysis found its involvement in ECM modulation and moderate expression of collagen and cell cycle genes19 (Figure 1G-H). Notably, we found specific expression of STEAP4 in VSMCs 1 (Figure 1E), a protein mainly expressed on plasma membrane20. IHC and IF showed higher expressions of MYH10, STEAP4, CXCL12 in the same area of ATAD media, but the expression of contractile VSMCs marker MYLK was reversed. IF verified the expression of STEAP4 and CXCL12 in MYH10+, but not MYLK+ VSMCs of aortic media in both ATAD and control group (Figure 1J-K).
We defined VSMCs 3 as stressed VSMCs for the expressions of stress-associated genes including HSPA1B, ATF3 and SOCS3 (Figure 1C, 1E, Supplementary Figure IIB). Genes uniquely expressed in VSMCs 3 modulated sarcomere organization, cardiac muscle cell apoptosis and signal transduction (Supplementary Figure IIC, F). QuSage analysis revealed its moderate enrichment in vascular contraction and activation of TNF-α and notch signaling pathways (Figure 1F-H), indicating its response to inflammation and reserved function of contraction.
VSMCs 6 might be proliferating VSMCs based on the expressions of cell proliferation and growth factor response genes such as APOLD121, ADAMTS4 and NR4A322, with higher levels of growth factors including FGFR3, FGF18, FGF19, FGF5 and ARTN to regulate cell proliferation, differentiation and survival (Figure 1C-E, Supplementary Figure IIB). Its distinctly expressed genes regulated gene transcription and cell cycle (Supplementary IIC, IIH). We also found its activation of PI3K-Akt-mTOR, wnt-β-catenin, TNF-α, notch and inflammatory response signaling pathways, with the function of VSMCs differentiation and higher levels of cell cycle and VSMC contraction genes, which rendered the differentiation and proliferation activity of VSMCs 6 (Figure 1F-H).
VSMCs 2, 4, 5 were identified as contractile VSMCs for their higher expression of RGS5 (Figure 1C, 1E), a gene involved in arteriogenesis23. Notably, VSMCs 2 expressed growth factors including PTN, AREG, PSPN and OGN to improve cell survival, with distinct functions in cell death and actin filament capping (Figure 1D, Supplementary Figure IIC, IIF), which showed relatively higher enrichment of elastic fiber assembly and VSMCs contraction gene-sets (Figure 1G-H). VSMCs 4 and 5 both expressed cell adhesion and apoptotic genes including CLMP24 and EGLN325, with the expressions of growth factors such as GDFs, BMP5, FGFR2 and TGFB1 whereby modulating cell growth and development (Figure 1C-E, Supplementary Figure IIB). They also exhibited similar functions in glycolytic process, protein metabolism and apoptosis (Supplementary Figure IIC, IIG). Qusage analysis showed their enrichment of glycolysis and VSMCs contraction, implicating the alteration of energy metabolism in these subpopulations (Figure 1F, 1H). Furthermore, VSMCs 5 showed relatively higher expressions of COL8A1 and MFAP5 (Figure 1C, 1E, Supplementary IIB) as well as involvement in ATAD, activation of hedgehog signaling pathway and enhanced collagen synthesis (Figure 1F-H), implying it might be unmature contractile VSMCs.
VSMCs 7 was defined as monocyte-like VSMCs for its expressions of monocyte markers CD93 and THBD (Figure 1C, 1E, Supplementary Figure IIB), which lost the function of vascular contraction but showed enhanced gene transcription and glycolysis (Figure 1F, 1H, Supplementary Figure IIC, III). VSMCs 8 highly expressed metallothionein superfamily genes including MT1G and MT1M (Figure 1C, 1E, Supplementary Figure IIB), which distinctly regulated cell response to metal ion (Supplementary Figure IIC, IIJ).
Differential expressed genes (DEGs) of VSMCs between ATAD and control group were shown in Figure 1K. The results showed most subpopulations of VSMCs in ATAD group highly expressed genes involved in ECM organization, metal ion response, glycolysis and hypoxia, while exhibited lower levels of genes relating to cell adhesion and muscle contraction (Figure 1L).
We noticed that most subpopulations of VSMCs in ATAD group exhibited higher proportions except for VSMCs 3 (stressed) and 6 (proliferating) (Figure 1B). Though the augmentation of synthetic VSMCs in ATAD has been widely accepted, the higher proportion of contractile VSMCs was not consistent with previous studies.
Differential phenotypes of FBs in ascending aorta
Nine subpopulations were obtained after re-clustering 5 clusters of FBs (Figure 2A). The composition of subpopulations in each sample and proportion of each subpopulation between ATAD and control group were shown in Supplementary Figure IIIA and Figure 2B, respectively.
FBs 7 highly expressed glycolytic genes including ENO126 and PKM as well as THBD, (Figure 2C-D, Supplementary IIIB), with high levels of growth factors including EPGN, IL11 and NRP2 to support cell proliferation, migration and cardiovascular function (Figure 2E). It also distinctly expressed genes relating to cellular component movement and angiogenesis (Supplementary Figure IIIH). Qusage analysis demonstrated its functions in activation of PI3K-Akt-mTOR, DNA repair, oxidative phosphorylation and glycolysis, implicating the elevated requirement to energy (Figure 2G). We also found its function in VSMC contraction and relevance to aortic atherosclerotic lesion (Figure 2F, 2H), which implied its potential roles in cell proliferation, differentiation and pathogenesis of aortic dissection.
FBs 8 was defined as synthetic VSMCs-like FBs for its higher expressions of STEAP4 and CXCL12, with elevated levels of growth factors containing GDF7, GDF9, TGFA and NRG1 whereby promoting cell proliferation and differentiation (Figure 2C-E). Furthermore, it displayed unique functions in ECM organization, cell adhesion and blood vessel remodeling (Supplementary Figure IIII). Qusage analysis revealed its roles in VSMCs differentiation and moderate activation of notch signaling pathway (Figure 2F-G). These characteristics were in consistent with the functions of synthetic VSMCs, implying that FBs might differentiate into synthetic VSMCs.
FBs 1, 4, 5 and 6 were identified as collagen-synthetic FBs for their expressions of collagen genes including COL1A2, COL3A1, COL1A1 and COL14A1 (Figure 2C-D, Supplementary Figure IIIB). Though these FBs exhibited differential expressions of growth factors, most of them played roles in cell growth and survival (Figure 2E). FBs 1 and 6 showed analogical functions in cell proliferation and migration, but FBs 4 exhibited functions in ECM organization, disassembly and collagen metabolic process (Supplementary Figure IIIC-D, IIIF). Interestingly, FBs 5 showed enhanced level of immune-associated gene C7 and NGFR, an indicator of phenotype switching27 (Figure 2C-D, Supplementary Figure IIIB). It also distinctly regulated ribosome biogenesis (Supplementary Figure IIIC, IIIG), indicating its enhanced translational activity. Qusage analysis further revealed the activation of IL6-JAK-STAT3 signaling pathway and response to IFN-γ as well as enrichment of cell-matrix adhesion, collagen fibril organization and collagen synthesis for FBs 5 (Figure 2F-H). These results hinted us FBs 5 might be involved in phenotype switching to maintain aortic homeostasis. FBs 6 specifically expressed ALDH1A3, a gene relating to various metabolic processes, cell proliferation and regulating the expression of ECM proteins28 (Figure 2C-D). Upon Qusage analysis, FBs 6 regulated cell migration, angiogenesis and TGF-β signaling, with moderate enrichment of cell adhesion, migration and relevance to abnormal aortic arch morphology, aortic dissection and aneurysm (Figure 2F-G).
FBs 2 and 3 might be stressed FBs for their overexpressions of HSPA1B, SOCS3, JUN and JUNB (Figure 2C-D, Supplementary Figure IIIB). FBs 2 and 3 uniquely regulated transcription (Supplementary Figure IIIC, IIIE). Qusage analysis revealed their roles in apoptosis and activation of TNF-α, IFN α and γ response signaling pathways (Figure 2G).
Though FBs 9 expressed several collagen genes (Figure 2C-D), it showed no specific markers. Moreover, it functioned in neutrophil chemotaxis and inflammatory response (Supplementary Figure IIIC, IIIJ), which might be involved in the process of neutrophil infiltration.
Differential expressed genes (DEGs) of FBs between ATAD and control group were shown in Figure 2I. The results showed most subpopulations of FBs in ATAD group highly expressed genes of transcriptional and translational processes, while exhibited lower levels of genes regulating collagen and ECM organization (Figure 2J).
As the major cell type of adventitia, most FBs exhibited higher proportions in control group (Figure 2B). Nevertheless, FBs 7 and 8 (synthetic VSMCs-like FBs) were 2 dominant subpopulations in ATAD group (Figure 2B), which might be associated with the development of ATAD and prompt us to study their relationship with synthetic VSMCs.
Unique subpopulations of ECs in ascending aorta
ECs was re-clustered and identified 9 heterogenous subpopulations (Figure 3A). The characteristics of populations in each sample and proportion of each population between ATAD and control group were shown in Supplementary Figure IVA and Figure 3B, respectively.
ECs 1, 3, 4 and 5 were identified as canonical ECs for their expressions of tissue development and vascular adhesion genes including POSTN and SELE (Figure 3C-D, Supplementary Figure IVB). ECs 1 and 3 highly expressed growth factors involved in endothelial growth and survival and angiogenesis such as ARTN, VEGFD, GDNF, HGF, GDF10 and BDNF (Figure 3E). Moreover, ECs 4 showed higher levels of FGF7 and NRG2 whereby regulating wound healing and response to stimulus, while ECs 5 expressed PDGFB, NTF3 and IL6 to participate endothelial survival and inflammation (Figure 3E). Furthermore, ECs 1, 3 and 4 showed similar functions in IFN-γ and cytokine-mediated signaling pathways and defense response to virus (Supplementary Figure IVC-D). Notably, ECs 3 distinctly regulated cell response to stimulus (Supplementary Figure IVC, IVF). ECs 4 also modulated immune response, antigen processing and presentation (Supplementary Figure IVC, IVG). Unexpectedly, ECs 5 did not display special function. Qusage analysis revealed significant activation of IL6-JAK-STAT3, hedgehog, PI3K-Akt-mTOR and TNF-α signaling pathways in ECs 1 and 3 (Figure 3F) to regulate inflammation and immune response, angiogenesis, proliferation and differentiation.
ECs 2 was identified as angiogenic ECs for its expressions of ECs migration, vascular modulation and development genes including SEMA3G29, IGFBP330 and HEY131, with expressions of growth factors including FGF1, FGF2, HBEGF and CXCL1 to support angiogenesis and neutrophil chemotaxis (Figure 3C-E, Supplementary Figure IVB). Further analysis unveiled its distinct functions in angiogenesis, vasculogenesis, monocyte and eosinophil chemotaxis (Supplementary Figure IVC, IVE). Qusage analysis displayed the activation of wnt-β-catenin signaling pathway (Figure 3F), which might be favorable to cell migration and promoted angiogenesis.
We deduced ECs 7 might be remodeling ECs for its expressions of ECM organization and tissue remodeling genes including OMD, EFEMP132 and MGP, which also partly expressed IGFBP3, with ubiquitous expressions of tissue remodeling growth factors such as BMP4-6 and OGN (Figure 3C-E, Supplementary Figure IVB). Moreover, ECs 7 functioned in oxidation-reduction process, IL-5, IL-12, IFN-γ production and cell differentiation (Supplementary Figure IVC, IVH). Qusage analysis revealed its activation of ROS, TGF-β and K-Ras signaling pathways (Figure 3F), which might be favorable to ECM wound healing.
ECs 8 might be lymphatic-like ECs for its expressions of lymphatic formation and chemokine genes including CCL21, LYVE1 and IGFBP5, with higher expressions of growth factors such as TGFB3, TGFB1, TGFA and LIF to support lymphangiogenesis (Figure 3C-E, Supplementary Figure IVB). It also regulated deacetylation of several proteins, PI3K activity and lymphangiogenesis (Supplementary Figure IVC, IVI). Moreover, ECs 8 showed higher activity in notch signaling pathway and multiple metabolic processes (Figure 3F) to regulate cell fate, proliferation and differentiation.
ECs 6 expressed canonical ECs marker SELE, but similar to stressed populations in VSMCs and FBs, it exhibited higher levels of SOCS3, HSPA1A, HSPA1B, which regulated transcription, cell growth and death (Figure 3C-D, 4G, Supplementary Figure IVB-C, IVG). ECs 9 showed highly expressed lymphatic-like ECs marker IGFBP5 (Supplementary Figure IVB), with activation of wnt-β-catenin signaling pathway (Figure 3F). It also functioned in negative regulation of ECs migration and angiogenesis (Supplementary Figure IVJ). These results indicated its potential origin from lymphatic-like ECs.
Differential expressed genes (DEGs) of ECs between ATAD and control group were shown in Figure 3G. The results showed most subpopulations of ECs in ATAD group highly expressed genes involved in transcriptional and translational processes, while exhibited lower levels of genes relating to immune response, antigen processing and presentation (Figure 3H).
Subpopulations of infiltrated neutrophils in ascending aorta
Eight subpopulations were identified upon re-clustering neutrophils (Figure 4A). For the differences between myeloid and peripheral neutrophils, we analyzed the markers and functions of neutrophils in ascending aorta according to the study conducted by Xie et al33. Neu 8 exhibited higher enrichment of G0, G1, G2, GM, G3 and G4 markers as well as functions in neutrophil activation, degranulation and ROS production, implying it might be a mixture of myeloid-derived neutrophils (Figure 4B-C). On the contrary, Neu 4 and Neu 7 might be mature peripheral neutrophils for their higher enrichment of G4 and G5b markers as well as functions in neutrophil aging, maturation, activation and degranulation, but Neu 1-3, 5 and 6 showed no enrichment in these markers and functions (Figure 4B-C). The composition of subpopulations in each sample was shown in Supplementary Figure VA. The proportions of Neu 8, 4 and 7 exhibited higher levels in control group, but Neu 1-3, 5 and 6 were dominant subpopulations in ATAD group (Figure 4D), implying Neu8, 4 and 7 might be the main subpopulations in physiological condition.
As previously described, Neu 8 expressed G2 and G3 neutrophil markers including LTF and CAMP as well as G4 neutrophil marker MMP8, with high levels of cytokines including CCL13, GPI, IL18 and AIMP1 to chemoattract monocyte and lymphocyte and induce leukocyte migration, angiogenesis and inflammation (Figure 4E-F, 5H, Supplementary Figure VB). It also functioned in mitochondrial DNA replication and translation, indicating its enhanced proliferation activity (Supplementary Figure VC, VL). Moreover, Neu 8 displayed enhanced activities in oxidative phosphorylation, notch and mTORC1 signaling pathways (Figure 4G).
Neu 4 and 7 exhibited similar markers including calcium-dependent signal transduction, neutrophil activity regulation and transmigration genes S100A12, S100A6 and CD177 with lower levels for the latter (Figure 4E-F, Supplementary Figure VB). Neu 4 highly expressed cytokines including IL16, IL27, CXCL13 and IL6R to promote activation of T cells, production of IFNG, migration of B cells and immune response, while Neu 7 showed higher levels of IL24, IL7, IL15 and CCL23 to regulate apoptosis, T cell and NK cell chemotaxis and B cell development (Figure 4H). Both of them exhibited similar functions in chemotaxis, glycolysis and innate immune response (Supplementary Figure VC, VJ). Moreover, Neu 4 regulated endocytosis, T cell tolerance induction, cell migration and TLR signaling pathway, while Neu 7 functioned in platelet activation, immune response and PI3K activity (Supplementary Figure VC, VH, VK). Qusage analysis unveiled enhanced activities in IFN-α response, hypoxia, ROS, PI3K-Akt-mTOR and IL2-STAT5 signaling pathways for Neu 4 and 7 with lower for the latter (Figure 4G).
Neu 1, 2, 5 and 6 showed higher expressions of IL1B and CXCL8, which also regulated cell death, lipid and protein metabolism (Figure 4E-F, Supplementary Figure VB-D). Notably, Neu 1 expressed cytokines such as CCL19 and CCL22 to chemoattract NK cells, T cells and monocytes, whereas Neu 2 showed higher levels of CCL3 and IL12B to regulate inflammation and NK cell activation (Figure 4H). Neu 5 expressed more chemokines represented by CCL11, CCL14, CCL21, CCL8, CXCL6 and CXCL9, implying its chemotactic activities for other immune cells, but Neu 6 expressed several lymphocyte, basophil and eosinophil chemotactic cytokines including CCL20, CC26 and CD70 (Figure 4H). Furthermore, Neu 1 distinctly functioned in macrophage-derived foam cell differentiation, glycoprotein and polyamine metabolism (Supplementary Figure VC, VE). Neu 2 and 6 regulated macrophage activation, chronic inflammation and integrin signaling pathway (Supplementary Figure VC, VF). Qusage analysis further showed the moderate enrichment of Neu 6 in coagulation, angiogenesis and hedgehog signaling pathway (Figure 4G).
Neu 3 was identified as stressed neutrophils for it lost markers of mature neutrophil but gained stress-related genes including EGR1, FOS and JUNB (Figure 4E-F, Supplementary Figure VB), which uniquely functioned in innate immune response, necroptosis, mRNA processing and type I IFN signaling pathway (Supplementary Figure VG). Interestingly, it showed mild enrichment of markers in G5a and G5b neutrophils (Figure 4B). These results elucidated its potential derivation from mature neutrophils.
Differential expressed genes (DEGs) of neutrophil between ATAD and control group were shown in Figure 4I. The results showed most subpopulations of neutrophil in ATAD group highly expressed genes relating to immune and inflammatory response, while exhibited lower levels of genes associated with cell migration and innate immune response (Figure 4J).
Monocytes/macrophages subpopulations in ascending aorta
Nine clusters of monocytes and 7 clusters of macrophages were identified upon re-clustering (Figure 5A, 5G). The composition of subpopulations for monocytes and macrophages in each sample were shown in Supplementary Figure VIA and Supplementary Figure VIIA. All subpopulations of monocytes were dominant in ATAD group (Figure 5B).
Most cells of Mono 1-6 and 8 highly expressed immune-related genes including CCL20, IL1B and IL1RN (Figure 5C-D, Supplementary Figure VIB). Moreover, Mono 1 and 8 showed higher levels of CXCL1, TNFAIP634, IL1A and F3, with expressions of several cytokines represented by GDF3, GDF6 and IL36B for Mono 1 as well as XCL2, IL24 and CXCL6 for Mono8, indicating their roles in neutrophil chemotaxis, inflammatory response and apoptosis (Supplementary Figure VIB, Figure 5E). Qusage analysis unveiled significant activation for Mono 1 and mild activation for Mono 8 of TNF-α and IL-6-JAK-STAT3 signaling pathways (Figure 5F). Mono 2-4 highly expressed TNIP3, ACSL4 and SMOX, with higher levels of cytokines including members of CCL, CXCL and interleukin family, playing roles in apoptosis, ferroptosis, inflammation and chemotaxis (Figure 5C-E, Supplementary Figure VIB). Qusage analysis revealed their enrichment in angiogenesis, coagulation, oxidative phosphorylation and PI3K-Akt-mTOR activation (Figure 5F). Mono 5 and 6 showed no specific markers, but expressed members of interleukin and CCL family, while Mono 5 exhibited similar gene-sets activity with Mono 2-4 (Figure 5E-F). Furthermore, Mono 1, 2 and 4 showed similar functions in cell migration, IFN-γ response and T cell activation (Supplementary Figure VIC-D). Mono 3 and 8 played roles in protein modification, TLR signaling pathway and T cell activity, with distinct functions of mono 3 in apoptosis, coagulation and platelet activation (Supplementary Figure VIC, VIE-F). Mono 5 also regulated protein modification and TLRs activity (Supplementary Figure VIC, VIG). Though Mono 7 and 9 showed higher expression of MT2A, only Mono 7 highly expressed stress-related genes HSPA1A and HSPA1B with functions in response to stress and protein modification (Figure 5C-D, Supplementary Figure VIB-C, VIH), which both functioned in response to metal ion and mineral absorption (Supplementary Figure VII).
Mφ 1 and 3 were identified as monocyte-like macrophages for their differential expressions of monocyte markers. Mφ 1 expressed cell migration and apoptotic genes including S100A4 and EMP1, with high levels of several cytokines such as CCL13, CCL19 and CCL28 to recruit monocytes and lymphocytes, which regulated ECs function and protein modification (Figure 5I-K, Supplementary Figure VIIC-D). However, Mφ 3 expressed cell-cell adhesion and inflammation genes THBS1 and CCL20, with higher levels of CCL17, IFNG and CXCL11 to chemoattract lymphocytes, exerting effects on antigen processing and presentation as well as T cell proliferation and apoptosis (Figure 5I-K, Supplementary Figure VIIB-C, VIIE). Qusage analysis revealed similar activation of angiogenesis, IL6-JAK-STAT3 and PI3K-Akt-mTOR signaling pathways for Mφ 1 and 3 (Figure 5L). We defined Mφ 2 as M2 macrophage for its higher expression of CXCR435 with expressions of several cytokines including XCL1, CXCL15 and IL5 to induce immune cell infiltration and inflammation, which regulated apoptosis, antigen processing and presentation, TLR and cytokine-mediated signaling pathways (Figure 5I, 6K, Supplementary Figure VIIB-C, VIIF). Qusage analysis displayed its similar enrichment to Mφ 1 and 3 (Figure 5L). Mφ 4 was identified as stressed macrophage for its expressions of HSPA1B, FOS and JUN, which regulated transcriptional processes (Figure 5I-J, Supplementary Figure VIIB-C, VIIH). Mφ 5 might be M1 macrophage for its high levels of IL1B, PTGS2 and EREG, with the expressions of multiple cytokines including members of CCL, interleukin and CSF family, which also modulated apoptosis, adaptive immune response, inflammatory response, TLR and TNF signaling pathways (Figure 5I-K, Supplementary Figure VIIB-C, VIIG). Qusage analysis showed its significant enrichment in IFN-γ response, TNF-α and TGF-β and wnt-β-catenin signaling pathways (Figure 5L). Mφ 6 and 7 did not showed specific markers and functions, which might under an unknown condition. Most subpopulations of macrophage were dominant in control group, while Mφ 3 exhibited higher proportion in ATAD group, implying its derivation from monocyte (Figure 5H).
Differential expressed genes (DEGs) of monocytes and macrophages between ATAD and control group were shown in Figure 5M-N. The results showed most subpopulations of monocytes in ATAD group highly expressed genes involved in transcriptional and translational processes, while exhibited lower levels of genes relating to immune response, antigen processing and presentation (Figure 5M). Most sub-clusters of macrophages showed higher levels of genes regulating inflammatory response, coagulation and multiple metabolic processes, while exhibited lower levels of genes involved in transcriptional and translational processes and antigen processing and presentation (Figure 5N).
Synthetic VSMCs-derived CXCL12 mediated chemotaxis of neutrophils and transdifferentiation of FBs
Previous results demonstrated that VSMCs specifically expressed chemokine CXCL12, the ligand of CXCR4 and ACKR3, which propelled us to examine their expressions in different cell types. We found most FBs highly expressed ACKR3 except for FBs 8, and most neutrophils and T cells highly expressed CXCR4. Alternatively, ECs, monocytes and macrophages hardly expressed CXCL12, CXCR4 and ACKR3 (Figure 6A-B). We hypothesized synthetic VSMCs-derived CXCL12 might exert regulatory effects on neutrophils, T cells and FBs. Upon re-clustering previously identified T cell, we first discriminated NK cell and T cell (Supplementary Figure VIII). Then we re-clustered other T cells and identified 5 subpopulations including CD8-TEM, CD8-TEFF, CD4, naïve and stressed T cells with their specific markers (Supplementary Figure IXA, IXC-D). However, all subpopulations of T cells in ATAD group exhibited lower or similar proportion compared with control group (Supplementary Figure IXB). As acute lesion in ATAD, we predicted cell-cell communication when neutrophils and FBs were selected as the origin of receptor. Most subpopulations of neutrophils interacted with VSMCs 1 via CXCL12-CXCR4 in a mild intensity, while all subpopulations of FBs did not exhibit the interaction with VSMCs 1 via CXCL12-ACKR3 (Supplementary Figure X, Supplementary Figure XI). Nevertheless, IHC revealed the expressions of ACKR3 and CXCR4 in FBs of adventitia and neutrophils of media, respectively (Figure 6C), implying the interaction between VSMCs 1 and FBs could not be excluded via CXCL12-ACKR3. The predicted downstream pathways of CXCL12-CXCR4 contained JAK-STAT, ERK1/2, PI3K-Akt and PLC-PKC signaling pathways and their potential targets (Figure 6D), whereby regulating cytokine production, chemotaxis, ROS production, cell differentiation, migration and apoptosis. Unfortunately, related pathways and targets could not be predicted upon combination of CXCL12 with ACKR3.
To detect whether FBs 8 derived from other subpopulations of FBs, we predicted the differentiation trajectory of FBs via CytoTRACE and found FBs 7 was the initiate of cell differentiation trajectory, followed by collagen synthetic FBs 5, 6 and 1 (Figure 6E). FBs 8, the synthetic VSMCs-like FBs, was the terminal state of FBs (Figure 6E). Genes predicted to be correlated with less differentiated and more differentiated FBs were also screened. The genes involved in protein translation and elongation were associated with less differentiated FBs including RPS18, RPS5, RPL18 and RPL29 as well as ENO1 (Figure 6F). However, synthetic VSMCs markers such as CFH and STEAP4 as well as bone development genes including FRZB and OGN showed higher correlation with terminal differentiated cells (Figure 6F). These results implied the higher potential that ENO1+ FBs 7 might differentiate into other subpopulations. Pseudo-time analysis displayed 2 main branches in the cell differentiation trajectory upon selecting FBs 7 as the initiate, with the confluence of FBs 7 at initiate and FBs 8 at terminal (Figure 6G, Supplementary Figure IXE-F). FBs 1, 5 and 6 distributed all over the trajectory, but FBs 2, 3, 4 and 9 populated in 2 terminal branches (Supplementary Figure IXE-F). Gene alteration along with the trajectory showed that synthetic VSMCs markers, ECM-related genes and genes correlated with more differentiated cells including STEAP4, CFH, VCAN, collagens, FRZB and OGN overexpressed after branching to cell fate 1, but the level of ACKR3 decreased at the terminal (Figure 6H-I, Supplementary Figure IXG, IXL). Nevertheless, stress-related and RNA catabolic genes such as FBLN2, HSPA1A, ATF3, EGR1 and HSPA6 overexpressed after branching to cell fate 2 (Supplementary Figure IXH-K, IXM-N). IHC and IF revealed higher expression of ENO1 in DCN+ FBs in adventitia of ATAD group, moreover, IHC and IF also displayed higher proportion of CXCL12+/STEAP4+ cells and expression of STEAP4 in DCN+ FBs in adventitia of ATAD group (Figure 6J-K). Homogeneity analysis among FBs 7, FBs 8, VSMCs 1 and VSMCs 2 also demonstrated similar marker genes between FBs 8 and VSMCs 1 (Figure 6L).
Cell differentiation trajectory of VSMCs and neutrophils
In consideration of the interaction among VSMCs, FBs and neutrophils, we further analyzed the cell differentiation trajectory of VSMCs and neutrophils.
CytoTRACE unveiled that VSMCs 1 was the initiate in the predicted trajectory with higher differentiation potential, followed by VSMCs 6, an intermediate state between synthetic VSMCs and contractile VSMCs, with subsequent 3 clusters of contractile VSMCs in order of VSMCs 5, 2, 4, representing different stages of contractile VSMCs (Figure 7A). VSMCs 8 was the terminal state in differentiation trajectory with the loss of VSMCs markers (Figure 7A). After filtration, we identified genes specifically correlated with less differentiated and more differentiated VSMCs. The expression of CFH, B2M, FN1, EFEMP1, VCAN and IGFBP4 showed more correlation with less differentiated VSMCs, while the expression of MYH11, PLN, MYL9, MYLK and TNS1 were more related with differentiated VSMCs (Figure 7B). Upon VSMCs 1 was selected as the origin of cell differentiation, pseudo-time analysis of VSMCs exhibited 2 cell fates. VSMCs 2 resided all over of the trajectory, but VSMCs 3 and 4 populated in all branches except for the pre-branch. The terminal branches were populated by VSMCs 5 and 7 for cell fate 1 as well as part of VSMCs 6 and VSMCs 8 for cell fate 2 (Figure 7C, Supplementary Figure XIIA-B). After branching, the genes relating to metal ion, response to stimulus and contractile VSMCs markers overexpressed in cell fate 2 such as MYH11, MYL9, ADAMTS4, APOLD1, ATF3, MT1G and THBD (Figure 7D, 7F, Supplementary Figure XIID-E), but the expressions of synthetic VSMCs markers and ECM organization, cell adhesion and migration genes decreased in cell fate 2 including MYH10, RGS5, VCAN, VCAN, OGN and FRZB (Figure 7E-F, Supplementary Figure XIID, XIIF). Moreover, glycolysis, apoptosis and cell adhesion genes overexpressed in cell fate 1 represented by CLMP and EGLN3 (Supplementary Figure XIIC-D, XIIG). This trajectory revealed the differentiation potential of VSMCs 1 and energy metabolism and function alteration of other subpopulations in cell development.
CytoTRACE analysis revealed Neu 8 was the initiate of differentiation trajectory with the highest differentiation potential, nearly followed by Neu 4 and 7, with other neutrophils in order of Neu 6-Neu 1-Neu 2-Neu 5 (Figure 7G). The genes correlated with less differentiated neutrophils represented by S100A8, TMSB4X, S100A4, PFN1 and CD63, which modulated cell differentiation, proliferation and migration, while IL1B, CXCL8 and PTGS2 were significantly correlated with more differentiated neutrophils, playing roles in inflammation, cell migration, apoptosis and vascular permeability (Figure 7H). Pseudo-time analysis revealed 2 terminal cell fates in this differentiation trajectory upon selecting Neu 8 as the initiate. Neu 8 and a fraction of Neu 4 were the only 2 subpopulations that existed in initial branch with higher differentiation potential, while most Neu 4 and other neutrophils distributed all over the trajectory with 2 different cell fates (Figure 7I, Supplementary Figure XIIH-I). Neutrophils located in cell fate 2 showed elevated expressions of genes relating to chemotaxis, inflammatory and immune response such as CCL4, CXCL1, CXCL2, CXCL8 and NLRP3 as well as decreased expression of Neu 8 marker LTF, indicating their roles in pro-inflammation (Figure 7J, 7L, XIIK-L). For cell fate 1, the genes of cellular component movement, immune response and Th1 cell activation exhibited high levels including S100A8, S100A12, S100A6, CD63, TNFRSF1B, IFITM3, IFITM2 and CST7 (Figure 7K-L, Supplementary Figure XIIJ-K, XIIM-N). These results demonstrated the differentiation potential of Neu8 and differential state of other subpopulations.
Co-expression network among VSMCs, FBs and neutrophils
WGCNA was performed to demonstrate the co-expression regulatory network among VSMCs, FBs and neutrophils (Figure 8A-C). Twelve modules (Figure 8D-E) were identified, regulating cell-cell/cell-ECM interaction (mediumpurple 3, paleturquoise, salmon and sienna3), transcription and translation (green, steelbule and turquoise), immune and inflammation (blue and royalblue), muscle contraction (cyan and violet) and calcium-dependent signal (orange).
The distinctly correlated modules splitted neutrophils into 2 different parts including Neu 4, 7 and 8 characterized by expressions of genes in blue and royalblue module as well as Neu 2, 5 and 6 represented by expression of genes in sienna3, steelblue and mediumpurple module. Myeloid-derived and peripheral neutrophils including Neu 4, 7 and 8 highly expressed genes in blue and royalblue modules, which were characterized by involvement of innate immune response and inflammatory response (Figure 8F, Supplementary Figure XIIIA). The core genes such as SCL11A36, CLEC4E, LCP237, SYK38 and ITGAM in blue and royalblue module regulated susceptibility to the intracellular pathogens, TCR-mediated intracellular signal transduction, adherence of neutrophils, inflammatory and immune response.
Neu 2, 5 and 6 showed distinct expression of genes in highly correlated modules including mediumpurple3, sienna3 and steelblue, which played roles in cell-cell signal, adhesion and leukocyte migration (Supplementary Figure XIIIB-D). Upon excluding untitled genes in these modules, we found the core regulators of these modules including RNU1-87P and LINC00676, which needed further studies to illustrate their functions.
All FBs and most VSMCs except for VSMCs 8 highly expressed genes in orange and turquoise module, which were responsible to cell adhesion, calcium-mediated signaling pathway, cGMP metabolism as well as translational and RNA metabolic process (Figure 8G, Supplementary Figure XIIIE). The hub genes of turquoise module including RPL10A, RPS13 and other members of RPL and RPS family, which regulated translation and RNA metabolic process. The core genes in orange module such as THRB39, FRZB, MYH10 and FBLIM1 modulated growth, cell adhesion, cell morphology and cell motility. Moreover, VSMCs 2, 3, 4, 6 and 8 also distinctly expressed genes in green and violet module, which regulated muscle contraction, actin crosslink formation, cell adhesion and protein modification (Figure 8H, Supplementary Figure XIIIF). The key regulators of these modules including EIF2B340, contractile VSMCs markers MYL9, TPM1 and TAGLN and NOTCH341, participating in vascular development and VSMCs differentiation.
VSMCs 1 and FBs 8, with similar characteristics of marker genes, highly expressed genes in cyan, paleturquoise and salmon module, which functioned in ECM-cell signaling, cell adhesion, CCL2 secretion and multiple metabolic processes (Figure 8I, Supplementary Figure XIIIG-H). Upon filtration of untitled genes, PKP142, DCHS2 and COL4A3 were identified as hub genes for these modules, acting as regulators of cell adhesion and ECM organization, which were in accordance with the functions of synthetic VSMCs.
Immediate early genes (IEGs) in subpopulations of different cell types
Tissue dissociation induces expressions of IEGs and influences the accuracy in identification of cell subpopulations. We analyzed the expressions of dissociation-induced IEGs to evaluate the influence of dissociation on these subpopulations identified in our study.
Previously, we found most stressed subpopulations in different cell types conservatively expressed stress-related genes including HSPA1B, SOCS3 and JUN. Upon correlation analysis among all subpopulations on the basis of top 2000 variable genes, overall expression of dissociation-induced IEGs43 for each subpopulation was calculated (Supplementary Figure XIVA). We noticed that FBs 2, FBs 3, FBs 9 and Mφ 4 exhibited higher overall expression of dissociation-induced IEGs, implying these subpopulations were influenced by tissue dissociation, which further intervened the identification of functions for these subpopulations (Supplementary Figure XIVB).