Single-cell multimodal profiling of atherosclerosis identifies CD200 as a cell surface lineage marker of vascular smooth muscle cells and their derived cells

Atherosclerotic cardiovascular disease is the leading cause of death worldwide. Vascular smooth muscle cells (VSMCs) play a central role in atherosclerosis due to their capability to phenotypically transition into either a protective or harmful state. However, the ability to identify and trace VSMCs and their progeny is limited due to the lack of well-defined VSMC surface markers¹. Investigations into VSMCs must utilize lineage-tracing mouse models, which are time-consuming, challenging to generate, and not feasible in humans. Here, we employed CITE-seq to phenotypically characterize 119 cell surface proteins in mouse atherosclerosis. We also performed CITE-seq in human atherosclerosis and found that CD200 is highly expressed and specific for VSMCs that persists even with phenotypic modulation.

To characterize cell types at different stages of atherosclerosis, we used male LDLr−/−, ROSA26LSL- ZsGreen1/+, Myh11-CreERT2 mice, which permanently induces ZsGreen1 expression in VSMCs following tamoxifen administration. At four different durations of western diet feeding (0, 8, 16, and 26 weeks), ZsGreen1⁺ and ZsGreen1⁻ cells were submitted for CITE-seq profiling (Fig [A]). 13 cell clusters were identified, including macrophages, T cells, endothelial cells, VSMCs, and fibroblasts. Based on ZsGreen1⁺, VSMC-derived cells comprised the VSMC 1, VSMC 2, and modulated VSMC populations (Fig [B]). Our analysis identified CD200 as a highly expressed and specific marker of VSMCs and VSMC-derived cells (Fig [C] and Fig [D]). There was also a marked lower and less specific expression on fibroblast and endothelial cells. Additionally, the expression of CD200 was consistent and maintained throughout different durations of atherosclerosis up to and including 26 weeks (Fig [E]).

Figure.

Identification and validation of CD200 as a cell lineage marker in mouse and human atherosclerosis. (A) UMAP visualization of multimodal integration of all CITE-seq data (n=3 per time point) identifying 13 distinct cell populations that was analyzed with Seurat v4.3.0. (B) UMAP depiction of ZsGreen status. (C) Feature plot showing the expression of CD200 protein across all cell clusters. (D) Violin plot quantifying the expression of CD200 in each cell cluster. (E) Violin plot quantifying the expression of CD200 at each time point of western diet feeding. (F) Flow cytometry gating strategy to assess the percentage of CD200+ cells that are ZsGreen+. (G) Quantification of the percent of CD200+ cells that are ZsGreen+. Values are shown as mean±SD, n=6. (H) Relative mRNA expression of Acta2, Cnn1, Tagln, Dcn, and Lum in FACS-sorted CD200+ and CD200− cells. Values are shown as mean±SD, n=4. (I) UMAP visualization of human carotid atherosclerosis CITE-seq data with expression of CD200 across all cells, n=6. (J) Violin plot comparing the expression of CD200 in asymptomatic (n=2) and symptomatic (n=4) carotid plaques. (K) Immunohistochemistry staining of human coronary artery sections. CD200 (green), α-SMA (red), and DAPI (blue). (L) Immunohistochemistry staining of human coronary artery sections. CD200 (green), CD31 (red), and DAPI (blue)

To validate our CITE-seq findings, we developed a flow cytometry panel to determine if CD200 can discriminate between VSMC and fibroblasts and used this panel to analyze cells from the ZsGreen1^+/− mice (Fig [F]). CD45⁺ leukocytes and CD31⁺ endothelial cells were excluded, so that the remaining cells included VSMCs and fibroblasts. We examined the expression of CD200, identifying two populations of CD200⁺ and CD200⁻ cells. Finally, we inspected the proportion of each population by ZsGreen1 status. The CD200⁺ cells were over 95% ZsGreen1⁺, whereas the CD200⁻ cells were only 5% ZsGreen1⁺ (Fig [G]). For further validation, we sorted CD45⁻CD31⁻CD200⁺ and CD45⁻CD31⁻CD200⁻ cells from C57BL6J mice and assessed the expression of VSMC- and fibroblast-specific genes (Fig [H]). The sorted CD200⁺ cells displayed a 10–15-fold increase in the VSMC-specific genes Acta2, Cnn1, and Tagln, and a greater than 90% reduction of fibroblast-specific genes Dcn and Lum. Although modulated VSMCs are suggested to express CD45 and CD31, our analysis (data not shown) and recent studies show that phenotypically modulated VSMCs do not express these proteins². Therefore, these data support that CD45⁻, CD31⁻, CD200⁺ cells can identify VSMC and VSMC-derived cells extremely effectively and that the expression of CD200 is maintained on VSMC-derived cells throughout atherosclerosis. In support of clinical relevance and translational utility, we corroborated these findings in human atherosclerosis with CITE-seq data from a previously reported carotid endarterectomy dataset and observed a similar pattern (Fig [I])³, and there was an increase in expression within symptomatic plaques (Fig [J]). We also performed immunohistochemistry on human coronary arteries to localize CD200-expressing cells. We stained these tissues with α-SMA (VSMCs), and CD200 (Fig [K]). There was a high degree of colocalization in the media with a lack of staining within the adventitia. Within the neointima, there was strong CD200 staining; however, there was less colocalization, likely because as VSMCs phenotypically modulate, they lose the expression of contractile VSMC proteins, including ACTA2⁴. Additionally, there was a lack of colocalization with CD31 (Fig[L]).

Here, our CITE-seq analysis revealed that CD200 is highly expressed on VSMCs, and this expression persists in their derived cells during the progression of atherosclerosis. Although CD200 is known to be produced by epithelial, hematopoietic, and endothelial cells⁵, ours is the first to report that CD200 is a surface marker of VSMCs and their modulated progeny. Based on our mouse data, leukocytes and endothelial cells are the only other major cell types of the vasculature that express CD200. Therefore, combining CD200 positivity with CD31 and CD45 negativity allows for the isolation of VSMCs.

Our discovery that CD200 is a persistent VSMC surface marker, even during phenotypic switching, abrogates the need to include a lineage-tracing reporter and facilitates investigations of VSMC and VSMC-derived cells in atherosclerosis, including in humans where VSMC lineage-tracing has limited feasibility. In conclusion, by identifying a cell surface marker of VSMCs, multiple techniques, including flow cytometry, can now be employed to characterize VSMCs and their derived cells in atherosclerosis. Mouse CITE-seq data are available under NCBI Gene Expression Omnibus database accession number GSE246779. All other data are available from the corresponding author upon request. The Institutional Animal Care and Use Committee of Columbia University approved all procedures (AABQ5576), and all procedures were in accordance with NIH guidelines. The Institutional Review Board of Columbia University approved all human protocols under IRB AAAJ2765 and AAAR6796 with written informed consent obtained from all individuals.

Nonstandard Abbreviations and Acronyms

VSMCs

Vascular Smooth Muscle Cells

CITE-seq

Cellular Indexing of Transcriptomes and Epitopes Sequencing

CD200

Cluster of Differentiation 200

LDLr

Low Density Lipoprotein receptor

α-SMA

alpha Smooth Muscle Actin

IRB

Institutional Review Board