Multiomic Landscape of Extracellular Vesicles in Human Carotid Atherosclerotic Plaque Reveals Endothelial Communication Networks

Abstract

Background: Carotid atherosclerosis is orchestrated by cell-cell communication that drives progression along a clinical continuum (asymptomatic to symptomatic). Extracellular vesicles (EVs) are cell-derived nanoparticles representing a new paradigm in cellular communication. Little is known about their biological cargo, cellular origin/destination, and functional roles in human atherosclerotic plaque.

Methods: EVs were enriched via size exclusion chromatography from human carotid endarterectomy samples dissected into paired plaque and marginal zones (symptomatic n=16, asymptomatic n=13). EV-cargos were assessed via whole transcriptome microRNA-sequencing and mass spectrometry-based proteomics. EV multiomics was integrated with bulk and single-cell RNA-sequencing datasets to predict EV cellular origin and ligand-receptor interactions, and multimodal biological network integration of EV-cargo was completed. EV functional impact was assessed with endothelial angiogenesis assays.

Results: Carotid plaques contained more EVs than adjacent marginal zones, with differential enrichment for EV-microRNAs and EV-proteins in key atherogenic pathways. EV cellular origin analysis suggested that tissue EV signatures originated from endothelial cells, smooth muscle cells, and immune cells. Integrated tissue vesiculomics and single-cell RNA-sequencing indicated complex EV-vascular cell communication that changed with disease progression and plaque vulnerability (ie, symptomatic disease). Plaques from symptomatic patients, but not asymptomatic patients, were characterized by increased involvement of endothelial pathways and more complex ligand-receptor interactions, relative to their marginal zones. Plaque EVs were predicted to mediate communication with endothelial cells. Pathway enrichment analysis delineated an endothelial signature with roles in angiogenesis and neovascularization, well-known indices of plaque instability. This was validated functionally, wherein human carotid symptomatic plaque EVs induced sprouting angiogenesis in comparison to their matched marginal zones.

Conclusions: Our findings indicate that EVs may drive dynamic changes in plaques through EV-vascular cell communication and effector functions that typify vulnerability to rupture, precipitating symptomatic disease. The discovery of endothelial-directed angiogenic processes mediated by EVs creates new therapeutic avenues for atherosclerosis.

Keywords: atherosclerosis; endothelial cells; extracellular vesicles; microRNAs; proteomics.

Figure 1.

Whole tissue proteomics identifies a potential role for extracellular vesicles (EVs) in modulating plaque biology corroborated with increased enrichment of tissue-resident EVs from human atherosclerotic carotid plaques. A, Experimental overview. Carotid endarterectomy specimens (N=64) were collected from the operating room and dissected into plaque and marginal zones for downstream analysis. B and C, Whole tissue proteome analysis (n=5). B, Principal component (PC) analysis of carotid tissue whole proteomes between the plaque zone (PZ) and marginal zone (MZ; donor matched n=5; donor uncorrected in Figure S2C). C, Percent incidence of Gene Ontology (GO; Biological Process and Cellular Component) terms (false discovery rate [FDR] <0.05) containing the word "vesicle" derived from differentially enriched proteins (q<0.05) from plaque and marginal zones (left; n=5). Right, Top 26 (by adjusted [adj] P<0.05) GO Biological Process terms where vesicle-associated terms from plaque regions are highlighted (MZ in Figure S2D). D through G, Analysis of EV-enriched fractions. D, Heat map representing per-fraction proteomics identifies prototypical EV markers enriched in fractions 7 to 10 (left). Protein abundance was Z score normalized by protein (representative sample, n=4; additional heat maps shown in Figure S5A). Western blot analysis depicting EV markers (CD63, CD81, CD9, and Flot1) in EV-enriched fractions (7–10) isolated from human atherosclerotic plaque tissue (right, n=1). E, Cryogenic electron microscopy of EVs isolated from plaque and adjacent marginal zones. Arrows indicate select EV structures as bilayered nanoparticles with dense cores (representative sample, n=4). Scale bar=200 nm. F, Quantification of EV mean diameter by nanoparticle tracking analysis (NTA; n=20). G, Nanoparticle concentration (normalized to tissue weight, NTA dilution factor, and total EV-enriched volume; n=11) binned by particle size from PZ (purple) and MZ (pink). Quantification of nanoparticle concentration across all sizes by NTA (n=11, right). Bar graphs show mean±SEM. Statistical significance was assessed by Fisher exact test (C) and a 2-tailed paired t test (11–20 pairs; F and G). SEC indicates size exclusion chromatography.

Figure 2.

Carotid plaque extracellular vesicle (EV)–vesiculome is zone-specific (plaque vs marginal) with differentially expressed EV-microRNA (miRNA) and proteins associated with proatherogenic functions. A and B, Donor-corrected principal component analysis (PCA) showing EV-miRNA (A) and EV-protein (B) profiles of EVs enriched from carotid plaque zones (PZ, purple) vs marginal zones (MZ, pink; miRNA n=20 pairs; protein n=23 pairs). C, Edge quantification of protein-protein interaction network with all differentially enriched EV-proteins (n=661) in plaque vs marginal zones. Represents full STRING network wherein edges indicate both functional and physical protein interactions with a medium confidence interaction score and nodes represent proteins. D and E, Volcano plot of carotid tissue derived EV-miRNA (D) and EV-protein (E) contents with positive and negative fold change representing miRNAs and proteins enriched in plaque vs marginal zones, respectively (false discovery rate [FDR] step-up/q<0.05). Highlighted in red are miRNAs and proteins known to function in inflammation. F, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis of all differentially expressed (FDR/q<0.05) predicted EV-miRNA targets (mRNA, miRTarBase via MicroRNA Enrichment Turned Network [MIENTURNET]) and EV-proteins. Venn diagram of unique and shared KEGG pathways between predicted EV-miRNA targets and EV-proteins (top left). Bubble plots of KEGG pathways that were significantly (FDR<0.05) modulated by all differentially enriched plaque EV-miRNAs (top right) and EV-proteins (bottom right), with all pathways that were shared by EV-miRNA and EV-proteins shown in bottom left. Labeled are pathways of interest (FDR<0.05). Cancer-, and infection-associated pathways were excluded from analysis. adj indicates adjusted; ECM, extracellular matrix; IL, interleukin; miR, microRNA; NFκB, nuclear factor κB; PC, principal component; TGF, transforming growth factor; TNF, tumor necrosis factor; and VEGF, vascular endothelial growth factor.

Figure 3.

In silico analysis of human atherosclerotic tissue derives extracellular vesicle (EV)–cell of origin and global communication strategies in plaque and marginal zones. A, Pictorial representation of cell of origin using the TISSUES database with all EV-proteins. Top 11 cells by strength metric calculated as a ratio of observed and background gene count after inclusion of cell types only (tissue annotations filtered out; further details in Figure S8A). B, Box plot of global EV-protein signature enrichment among cells from atherosclerotic plaques, informed by integrating all detected EV-proteins (global EV signature) with single-cell RNA-sequencing (scRNA-seq) of carotid atherosclerotic plaques. C, Stacked bar plot visualizing total number of predicted cell-cell communication interactions in plaque (purple) and marginal (pink) zones with CellChat. D, Heatmap depicting EV–cell signaling strength used by all EV-proteins to both plaque and marginal single cells. Strength of EV-derived signaling molecule and cell type shown on right and top bar graph, respectively, and represents communication probability. E, Cell type annotations derived using the Tabula Sapiens database (vascular tissue) by inputting all significant (adjusted P<0.05) predicted mRNA targets of all differentially expressed EV-microRNA (miRNA) in plaque and marginal zones combined (false discovery rate <0.05). F, Box plot of predicted EV-miRNA-mRNA target signature enrichment among cells of the atherosclerotic plaque, informed by integrating differentially detected EV-miRNA-mRNA targets with scRNA-seq of carotid atherosclerotic plaques. DN indicates double negative; NK, natural killer; NKT, natural killer T cell; and VSMC, vascular smooth muscle cell.

Figure 4.

Site-specific (plaque vs marginal zone) differences in extracellular vesicle (EV) cell of origin, recipient cell communication strategies, and directed communication with vascular cells (endothelial cells [ECs], smooth muscle cells [SMCs], macrophages). A and B, Box plot of differential marginal (A) and plaque (B) EV-protein signature enrichment among cells of the atherosclerotic plaque, informed by integrating differentially enriched EV-proteins with single-cell RNA-sequencing (scRNA-seq) of carotid atherosclerotic plaques. C and D, Pictorial representation of cell of origin using the TISSUES database with differentially expressed EV-proteins in marginal (C) and plaque (D) zones. Top 10 cells by strength metric were calculated as a ratio of observed and background gene count after filtering for tissues (further details in Figure S8B). E and F, Heat map depicting differential marginal and plaque EV to cell signaling strength used by differentially enriched EV-proteins in marginal (E) and plaque (F) zones to all atherosclerotic single cells in plaque and marginal regions. Strength of EV-derived signaling molecule and cell type shown on right and top bar graph, respectively. G and H, Chord diagrams representing differentially expressed EV ligands to EC (G) and SMC (H) ligand-receptor pairs to single cells in marginal zones (left) and plaque zones (right). Arrows represent EV-ligand interactions. Individual cell populations are annotated in legend. DN indicates double negative; ICAM, intercellular adhesion molecule; and VSMC, vascular SMC.

Figure 5.

Extracellular vesicle (EV) concentration and greater differential EV-cargo (microRNA [miRNA] and protein) discriminate between symptomatic and asymptomatic atherosclerotic plaques. A, Schematic of experimental design. Carotid endarterectomy specimens from symptomatic and asymptomatic patients were dissected into plaque zone (PZ) and marginal zone (MZ), followed by EV enrichment, quantification, and miRNA transcriptomics and proteomics. B, Hematoxylin and eosin (H&E) and Movat pentachrome stain of asymptomatic (top) and symptomatic (bottom) carotid plaque specimens (patients 1 and 2). Large, calcified plaque with calcified/necrotic center, but with no significant intraplaque hemorrhage (top). Large, calcified plaque with calcified/necrotic center with intraplaque hemorrhage (bottom). H&E staining with intraplaque hemorrhage indicated by green arrowheads (bottom left) and Movat pentachrome staining with intraplaque hemorrhage indicated by cyan arrowheads (bottom right). Scale bar (500 µm) as shown. C and D, Nanoparticle tracking analysis (NTA) of nanoparticle concentration in asymptomatic (C) and symptomatic (D) atherosclerotic PZ vs MZ. E, Assessment of EV size and morphology. NTA of mean diameter of nanoparticles enriched from asymptomatic (left) and symptomatic (middle) carotid plaques. Cryogenic electron microscopy of EVs from asymptomatic (right, top) and symptomatic (right, bottom) carotid plaques (n=1–2 per group). Arrows indicate select EV structures as bilayered nanoparticles with dense cores. Scale bar=200 nm. F and G, Volcano plot of carotid tissue-resident EV-miRNA transcriptome in asymptomatic (F) and symptomatic (G) with positive and negative fold change (FC) representing miRNAs enriched in PZ vs MZ (false discovery rate [FDR] step-up <0.05). H and I, Volcano plot of carotid tissue-resident EV-proteome in asymptomatic (H) and symptomatic (I) with positive and negative FC representing EV-proteins enriched in PZ vs MZ (q<0.05). Bar graphs show mean±SEM. Statistical significance was assessed by nonparametric paired Wilcoxon test (C through E). miR indicates microRNA.

Figure 6.

Differential plaque extracellular vesicle (EV)-secretome highlights key pathways involved in plaque vulnerability. A, Venn diagram depicting shared and unique plaque-derived EV-microRNA (miRNA) in asymptomatic (green) and symptomatic (purple) conditions (false discovery rate [FDR] step-up<0.05, filtered for miRNAs<1000). B, Total numbers of significant Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways (adjusted P<0.05) modulated by the predicted mRNA targets (miRTarBase via MicroRNA Enrichment Turned Network, threshold of minimum interactions 1; FDR<0.05) of differentially expressed plaque EV-miRNAs in asymptomatic (green), symptomatic (purple), and shared (brown) conditions. C, Dot plot of all KEGG pathways derived from unique asymptomatic plaque–derived predicted EV-miRNA mRNA targets (P<0.05). Labeled are significant (adjusted P<0.05) pathways graphed by odds ratio. D, Dot plot of KEGG pathway analysis from unique symptomatic plaque–derived predicted EV-miRNA mRNA targets. Graphed are top 42 significant (adjusted P<0.05) pathways by odds ratio (total 120 pathways with adjusted P<0.05). E, Venn diagram depicting shared and unique plaque-derived EV-proteins in asymptomatic (green) and symptomatic (purple) conditions (q<0.05). F, Total numbers of significant GO and KEGG pathways (adjusted P<0.05) modulated by differentially expressed plaque EV-proteins (q<0.05) in asymptomatic (green), symptomatic (purple), and shared (brown) conditions. G, Bar graph of all significant KEGG pathways (adjusted P<0.05) from unique asymptomatic plaque–derived EV-proteins (adjusted P<0.05). H, Bar graph of all significant KEGG pathways (adjusted P<0.05) from unique symptomatic plaque–derived EV-proteins (adjusted P<0.05). STRING (db Version 12.0) is used to generate protein-protein interaction maps (right) of pathways of interest. Cancer-, and infection-associated pathways were excluded from analysis. ECM indicates extracellular matrix; IL, interleukin; and TNF, tumor necrosis factor.

Figure 7.

Endothelial cells (ECs) are cellular players modulating cell-cell communication through extracellular vesicles (EVs) in vulnerable plaques. A through D, Box plots of symptom and regional EV-protein signature enrichment among cells of the atherosclerotic plaque, informed by integrating differentially expressed EV-proteins in asymptomatic and symptomatic, marginal vs plaque zones with single-cell RNA-sequencing of carotid atherosclerotic plaques. Asymptomatic marginal (A) and plaque (B) zone signature enrichment among cells of the atherosclerotic plaque. Symptomatic marginal (C) and plaque (D) zone signature enrichment among cells of the atherosclerotic plaque. E and F, Heat map depicting differential asymptomatic plaque- (E) and symptomatic plaque- (F) EV–cell signaling strength used by differentially expressed EV-proteins in asymptomatic and symptomatic cohorts with plaque cells. Strength of EV-derived signaling molecule and cell type shown on right and top bar graph, respectively. NKT indicates natural killer cell; and VSMC, vascular smooth muscle cell.

Figure 8.

Overrepresentation analysis identifies angiogenesis and neovascularization as functional effects of vulnerable (symptomatic) plaque extracellular vesicles (EVs) with support from a functional endothelial cell (EC) sprouting assay. A, Chord diagrams representing ligand-receptor pairs between differentially enriched EV-proteins (ligands) to ECs (receptor) in asymptomatic (top) and symptomatic (bottom) plaques. Red arrows represent EV–ligand-to-receptor interactions. Individual cell populations are annotated in legend. B, Filtered inquiry of EC, angiogenesis, and neovascularization specific pathways from Gene Ontology (GO) overrepresentation analysis of differentially expressed predicted EV-microRNA (miRNA)-mRNA targets in asymptomatic and symptomatic groups. Fold enrichment is represented in legend. C, Pairwise comparison of mean fold enrichment of EC-specific pathways in symptomatic vs asymptomatic plaques. D through F, Functional effect of symptomatic marginal zone (MZ) and plaque zone (PZ) EVs on EC angiogenesis. Representative confocal microscopy images of human umbilical vein EC–spheroid angiogenesis after addition of EVs (2–4×10¹⁰ EVs; 24 h), or assay control from 3 independent experiments (scale bar 100 µm; D). Representative grayscale images with annotations for sprout number (E, left) and quantification of sprout number per spheroid (n=7-10 from 3 independent experiments; E, right). Representative grayscale images with tracings representing individual sprout lengths (F, left) and quantification (n=188–331 sprouts from 7 to 10 spheroids from 3 independent experiments). Bar graphs show the mean±SEM. Violin plots show median±interquartile range. Statistical significance was assessed by 1-way ANOVA with Sidak multiple comparisons test (E) and Kruskal-Wallis test with Dunn multiple comparisons correction (F). P values meeting significance (<0.05) are shown in bolded text (E and F).