Single-cell profiling reveals age-associated immunity in atherosclerosis

Abstract

Aims: Aging is a dominant driver of atherosclerosis and induces a series of immunological alterations, called immunosenescence. Given the demographic shift towards elderly, elucidating the unknown impact of aging on the immunological landscape in atherosclerosis is highly relevant. While the young Western diet-fed Ldlr-deficient (Ldlr-/-) mouse is a widely used model to study atherosclerosis, it does not reflect the gradual plaque progression in the context of an aging immune system as occurs in humans.

Methods and results: Here, we show that aging promotes advanced atherosclerosis in chow diet-fed Ldlr-/- mice, with increased incidence of calcification and cholesterol crystals. We observed systemic immunosenescence, including myeloid skewing and T-cells with more extreme effector phenotypes. Using a combination of single-cell RNA-sequencing and flow cytometry on aortic leucocytes of young vs. aged Ldlr-/- mice, we show age-related shifts in expression of genes involved in atherogenic processes, such as cellular activation and cytokine production. We identified age-associated cells with pro-inflammatory features, including GzmK+CD8+ T-cells and previously in atherosclerosis undefined CD11b+CD11c+T-bet+ age-associated B-cells (ABCs). ABCs of Ldlr-/- mice showed high expression of genes involved in plasma cell differentiation, co-stimulation, and antigen presentation. In vitro studies supported that ABCs are highly potent antigen-presenting cells. In cardiovascular disease patients, we confirmed the presence of these age-associated T- and B-cells in atherosclerotic plaques and blood.

Conclusions: Collectively, we are the first to provide comprehensive profiling of aged immunity in atherosclerotic mice and reveal the emergence of age-associated T- and B-cells in the atherosclerotic aorta. Further research into age-associated immunity may contribute to novel diagnostic and therapeutic tools to combat cardiovascular disease.

Keywords: Aging; Atherosclerosis; Cardiovascular disease; Immunology; Immunosenescence.

Graphical Abstract

Figure 1

Aging promotes immunosenescence and atherosclerosis in Ldlr^−/− mice. (A) Experimental setup: Ldlr^−/− mice aged 3 months (green bars) or 20 months (violet bars) were fed with a standard chow diet (white circles) or a Western diet (grey circles) for 10 weeks. (B) Using flow cytometry, percentages (% from live) of circulating myeloid cells (CD11b⁺), neutrophils (CD11b⁺Ly6C^intLy6G⁺), inflammatory monocytes (CD11b⁺Ly6C^hiLy6G⁻), and (C) circulating CD4⁺ and CD8⁺ T-cells were determined (% from live). (D) Splenic naïve (T_N: CD44⁻CD62L⁺), effector (T_EFF: CD44⁻CD62L⁻), central memory (T_CM: CD44⁺CD62L⁺), and effector memory (T_EM: CD44⁺CD62L⁻) T-cells were quantified as a percentage of CD4⁺ and CD8⁺ T-cells and plotted in pie charts. (E) Intracellular cytokine production of interferon-gamma (IFN-γ), interleukin (IL)-4, IL-10, and IL-17 were measured as percentage (mean) of splenic CD4⁺ and CD8⁺ T-cells after 4 h of stimulation with PMA and ionomycin. Colour scale is normalized for each cytokine. (F) Circulating CD19⁺ B-cells were determined with flow cytometry. (G) Total and oxidized LDL (oxLDL)-specific IgM titres were measured in the serum. (H) Total serum cholesterol levels at sacrifice were measured. (I) Cross sections of the aortic root were stained for lipid and collagen content, and (J) atherosclerotic lesion volume was quantified. (K) Collagen content and (L) necrotic cores were quantified as percentage of lesion area. (M) Cholesterol crystallization in atherosclerotic lesions was categorized on a scale of 0 (no cholesterol crystallization) to 3, and presence of calcification (purple) or no calcification (grey) was presented as percentage of group. (N) Macrophage content (MOMA-2) was measured as percentage of lesion area. Data are from n = 12–15 mice per group. Statistical significance was tested by one-way ANOVA. Mean ± S.e.m. plotted. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 2

Integrated scRNA-seq analysis reveals age-associated leucocyte alterations in atherosclerotic mouse aortas. (A) Workflow of scRNA-seq on aortic CD45⁺ cells of chow diet-fed young Ldlr^−/− mice (young CD, n = 9), or Western diet-fed (10 weeks) young Ldlr^−/− mice (young WD, n = 29) and chow diet-fed aged Ldlr^−/− mice (old CD, n = 12). UMAP visualization of clustered aortic leucocytes grouped by (B) sample or (C) immune cell clusters. (D) Stacked diagram showing the relative proportions of major immune cell subtypes within CD45⁺ cells of Ldlr^−/− aortas. NK, natural killer; pDC, plasmacytoid dendritic cell.

Figure 3

Identification of age-associated T-cell populations and gene signatures in atherosclerotic aortas from Ldlr^−/− mice. (A) Cd3e⁺ clusters were extracted from the principal clustering and reclustered, after which the T-cell clusters were identified. (B) UMAP plots and stacked diagrams visualizing the identified T-cell subclusters in young CD, young, WD and old CD aortas, in which Gzmk⁺CD8⁺ T-cells are encircled in the dashed red shape. (C) Dot plot showing the average expression of immune cell cluster-defining markers for each cluster. (D) Heatmap of hierarchically clustered top 25 variable genes across T-cell subclusters. (E) Volcano plot of the differentially expressed genes (DEGs) in the Gzmk⁺CD8⁺ T-cell cluster compared to other CD8⁺ T-cells in clusters 2, 3, 5, and 6. (F) Top canonical pathways of the Gzmk⁺CD8⁺ T-cell cluster compared to CD8⁺ T-cells in clusters 2, 3, 5, and 6. (G) Heatmap showing average expression of biological process-associated genes in T-cell clusters of young CD, young WD, and old CD Ldlr^−/− aortas. (H) Using flow cytometry, absolute numbers of Ly6C⁻CD44⁺Tox⁺PD-1⁺ CD8⁺ T-cells (GzmK⁺CD8⁺ T-cells) were measured in aortas of young and aged Ldlr^−/− mice (n = 11–15). Gating strategy is shown in Supplementary material online, Figure S5B. Statistical significance was tested by one-way ANOVA. Mean ± S.e.m. plotted. *P < 0.05, ****P < 0.0001. DP, double positive; SP, single positive; Tregs, regulatory T-cells; NKT, natural killer T; MC, mast cells.

Figure 4

Characterization of age-associated B-cells in atherosclerotic aortas of Ldlr^−/− mice. (A) Cd79a⁺ clusters were extracted from the principal clustering and reclustered, after which the B-cells clusters were identified. (B) UMAP plots and stacked diagrams visualizing the identified B-cell subclusters, in which ABCs are encircled in the dashed blue shape. (C) Dot plot showing the average expression of immune cell cluster-defining markers for each cluster. (D) Heatmap of hierarchically clustered top 25 variable genes across B-cell subclusters. (E) Volcano plot of the differentially expressed genes (DEGs) in the age-associated B-cells (ABCs), excluded of plasma cells (PCs), compared to other B-cells in the B-cell subclustering. (F) Top canonical pathways of the ABCs. (G) Heatmap showing average expression of biological process-associated genes in B-cell clusters of young CD, young WD, and old CD Ldlr^−/− aortas. (H) ABCs and follicular (FO) B-cells were tested for their capability to present OVA323 peptide antigen to CD4⁺ OTII T-cells and induce T-cell activation (CD69⁺) or (I) proliferation (n = 5). (J) Absolute numbers of CD19⁺CD11b⁺CD11c⁺ ABCs and representative plot of associated protein expression of CD11c and T-bet within the ABCs in aortas of Ldlr^−/− mice (n = 12–15). Gating strategy is shown in Supplementary material online, Figure S5A. Statistical significance was tested by two-tailed paired t-test (FO B-cells vs. ABCs) or one-way ANOVA (three groups). Mean ± S.e.m. plotted. *P < 0.05, **P < 0.01, ****P < 0.0001.

Figure 5

Integrated analysis of myeloid cells reveals age-induced phenotype alterations in macrophage subpopulations in atherosclerotic aortas. (A) Cd68⁺ and Itgam⁺ clusters were extracted from the principal clustering and reclustered, after which the myeloid clusters were identified. (B) UMAP plots and stacked diagrams visualizing the identified myeloid subclusters. (C) Dot plot showing the average expression of immune cell cluster-defining markers for each cluster. (D) Dendrogram heatmap based on the 25 most differentially expressed genes from all macrophage clusters. (E) Heatmap showing average expression of biological process-associated genes in myeloid cell clusters of young CD, young, WD and old CD Ldlr^−/− aortas. (F and G) Percentage of M1- or M2-like bone marrow-derived macrophages (BMDMs) of young and aged Ldlr^−/− mice (n = 5), which have taken up apoptotic cells (F) or BODIPY-labelled cholesteryl lipids (G), was measured by flow cytometry, of which the gating strategy is shown in Supplementary material online, Figure S7C. Statistical significance was tested by one-way ANOVA. Mean ± S.e.m. plotted. ***P < 0.001, ****P < 0.0001.

Figure 6

Flow cytometry analysis confirms expansion of age-associated T- and B-cells in human atherosclerotic plaques. (A) Single-cell RNA-sequencing analysis of human atherosclerotic plaques (n = 18), in which gene expression of GzmK⁺CD8 T-cell-associated markers GZMK, EOMES, TOX, PDCD1, and LAG3 are shown. (B) Representative plot and quantification of GZMK⁺TOX⁺ as percentage of CD8⁺ T-cells as measured in human atherosclerotic plaques and corresponding blood samples (n = 15) with flow cytometry. (C) Representative plot and quantification of CD11b⁺CD11c⁺ ABCs as percentage of B-cells, and expression of T-bet within ABCs measured in human atherosclerotic plaques and corresponding blood samples (n = 9) with flow cytometry. Gating strategies are shown in Supplementary material online, Figure S9. Statistical significance was tested by two-tailed paired t-test. Mean ± S.e.m. plotted. **P < 0.01, ***P < 0.001, ****P < 0.0001.