Single-cell profiling reveals inflammatory polarization of human carotid versus femoral plaque leukocytes

Abstract

Femoral atherosclerotic plaques are less inflammatory than carotid plaques histologically, but limited cell-level data exist regarding comparative immune landscapes and polarization at these sites. We investigated intraplaque leukocyte phenotypes and transcriptional polarization in 49 patients undergoing femoral (n = 23) or carotid (n = 26) endarterectomy using single-cell RNA-Seq (scRNA-Seq; n = 13), flow cytometry (n = 24), and IHC (n = 12). Comparative scRNA-Seq of CD45+-selected leukocytes from femoral (n = 9; 35,265 cells) and carotid (n = 4; 30,655 cells) plaque revealed distinct transcriptional profiles. Inflammatory foam cell-like macrophages and monocytes comprised higher proportions of myeloid cells in carotid plaques, whereas noninflammatory foam cell-like macrophages and LYVE1-overexpressing macrophages comprised higher proportions of myeloid cells in femoral plaque (P < 0.001 for all). A significant comparative excess of CCR2+ macrophages in carotid versus plaque was observed by flow cytometry in a separate validation cohort. B cells were more prevalent and exhibited a comparatively antiinflammatory profile in femoral plaque, whereas cytotoxic CD8+ T cells were more prevalent in carotid plaque. In conclusion, human femoral plaques exhibit distinct macrophage phenotypic and transcriptional profiles as well as diminished CD8+ T cell populations compared with human carotid plaques.

Keywords: Atherosclerosis; Cardiology; Inflammation.

Figure 1. Canonical correlation analysis (CCA) clustering of CD45⁺ selected cells derived from scRNA-Seq of femoral (n = 9; 35,265 cells) and carotid (n = 4; 30,655 cells) atherosclerotic plaque reveals distinct immune cell types and populations.

(A) CCA clustering and UMAP visualization of all femoral (n = 9) and carotid (n = 4) plaque samples for which scRNA-Seq was performed, including UMAP for overall samples colored by 14 cell types. (B) Dot plot of top marker genes per cluster. (C) Cell type identities were validated based on marker gene expression.

Figure 2. Quantitative comparison between femoral (n = 9; 35,265 cells) and carotid (n = 4; 30,655 cells) atherosclerotic plaque from CCA clustering of CD45⁺ selected cells.

(A) UMAP visualization of separated carotid and femoral samples. (B) Corresponding table and the stacked bar graph of the logistic regression comparing cell proportions in carotid versus femoral plaque (log OR, 95% CI expressed). (C) Volcano plot of highly expressed genes for carotid and femoral plaque samples.

Figure 3. Myeloid reclustering of scRNA-Seq data reveals inflammatory foam cell–like macrophage and monocyte predominance in carotid plaque (n = 4; 8,941 myeloid cells) and comparative antiinflammatory and resident-like macrophage biases in femoral plaque (n = 9; 4,461 myeloid cells).

(A) CCA Re-Clustering and UMAP visualization of carotid and femoral myeloid cell samples, including overlay by vascular bed revealed 8 distinct populations. (B) Dot plot of top marker genes per cluster. (C and D) Top marker genes per cluster are shown by violin plot, and macrophages with high APOE and APOC1 expression suggestive of foam cell features consisted of 2 distinct clusters, labeled inflammatory and noninflammatory based on differential gene expression.

Figure 4. Distinct macrophage gene expression and phenotypic profiles confirmed by scRNA-Seq (8,941 carotid myeloid cells, 4,461 femoral myeloid cells) and in a validation cohort by flow cytometry.

(A) Cell clusters as a proportion of myeloid cells were compared for carotid versus femoral samples and are displayed in UMAP visualization of separated carotid and femoral samples. (B) Stacked bar graph and corresponding table of the logistic regression comparing cell proportions in carotid versus femoral plaque (log OR, 95% CI expressed; *P < 0.0017, between carotid and femoral plaques that were significant at Bonferroni-corrected value. (C) Volcano plot of highly expressed genes for carotid versus femoral plaque overall. (D–F) Flow cytometry of digested plaque macrophages, identified as CD11b⁺CD14⁺CD64⁺HLA-DR^hi live cells and further distinguished based on CCR2 expression, revealed a significantly higher proportion of carotid plaque macrophages expressing CCR2 than carotid plaque macrophages (representative plot in E, comparison in F, left plot). Classical (CD14⁺⁺CD16^–) monocytes from blood also expressed CCR2 more highly in patients undergoing carotid versus femoral endarterectomy (F, right plot).

Figure 5. Differential gene expression analyses of femoral versus carotid myeloid clusters reveal comparative homeostatic, inflammation-regulating biases in femoral plaque.

(A–D) Volcano plots of differential gene expression and gene ontology (GO) analyses of biological processes in femoral versus carotid plaque for inflammatory foam cell–like macrophages (A), monocytes (B), mast cells (C), and DCs (D) depict differential gene expression for carotid and femoral plaques.

Figure 6. Lymphoid reclustering of scRNA-Seq data reveals T cell predominance and cytotoxicity-associated gene expression in highly prevalent T cell clusters.

(A) Visualization of carotid and femoral lymphoid cells revealed 9 distinct populations. (B) Highly expressed genes are visualized with the dot plot. (C) Top marker genes per cluster are shown by violin plot. (D) Related gene set enrichment shown by the dot plot.

Figure 7. Carotid plaque exhibits comparative cytotoxic CD8⁺ T cell bias, whereas B cells are more prevalent in femoral plaque.

(A and B) UMAP visualization of separated carotid and femoral samples with stacked bar graph and corresponding table of the logistic regression comparing cell proportions in carotid versus femoral plaque (log OR, 95% CI expressed). *P < 0.0017, differences between carotid and femoral plaques that were significant at Bonferroni-corrected value. (C) Flow cytometry of T cells from paired plaque and blood samples revealed comparative excess in CD8⁺ T cells (as a proportion of overall T cells) in plaque and blood from carotid endarterectomy patients (**P < 0.05 using 2-tailed t test). (D and E) In situ determination and numbers of B and T cells per high-powered field in intraplaque leukocyte aggregates. Magnification, 4× (left) and 40× (middle and right).

Figure 8. Differential gene expression analyses of femoral versus carotid lymphoid clusters demonstrate comparative inflammation-regulating bias in femoral plaque.

(A–H) Volcano plots of differential gene expression in femoral versus carotid plaque for IL7R⁺ CD4⁺ Effector T cells (A), GZMK⁺CD8⁺ T cells (B), GZMH⁺CD8⁺ T cells (C), NK cells (D), CXCL8⁺ T cells (E), CD79A⁺ B cells type 2 (F), TIGIT⁺ CD4⁺ Tregs (G), and plasma cells (H). CD79A⁺ B cells type 1 were not included due to the insufficient cell number in carotid plaques.