Fate and State of Vascular Smooth Muscle Cells in Atherosclerosis

Abstract

Vascular smooth muscle cells (VSMCs) have long been associated with phenotypic modulation/plasticity or dedifferentiation. Innovative technologies in cell lineage tracing, single cell RNA-sequencing, and human genomics have been integrated to gain unprecedented insights into the molecular reprogramming of VSMCs to other cell phenotypes in experimental and clinical atherosclerosis. The current thinking is an apparently small subset of contractile VSMCs undergoes a fate switch to transitional, multi-potential cells that can adopt plaque de-stabilizing (inflammation, ossification) or plaque-stabilizing (collagen matrix deposition) cell states. Several candidate mediators of such VSMC fate and state changes are coming to light with intriguing implications for understanding coronary artery disease risk and the development of new treatment modalities. Here, we briefly summarize some technical and conceptual advancements derived from two publications in Circulation and another in Nature Medicine that, collectively, illuminate new research directions to further explore the role of VSMCs in atherosclerotic disease.

The concept of vascular smooth muscle cells (VSMCs) existing in a “transitional stage between smooth muscle and foam cells” of a human atheroma was first postulated in 1961 with the aid of an electron microscope.¹ This proposition gained further support with the development of molecular probes to identify cell-restricted genes or their encoded proteins in human atherosclerosis² and in cultured VSMCs undergoing transdifferentiation from a contractile VSMC state to a lipid-laden macrophage-like state.³ The emergence of lineage tracing reporters in mouse models of atherosclerosis strengthened the view of VSMC-to-macrophage transdifferentiation^{4, 5} and corroborated the decades-old theory⁶ of clonal expansion of a few VSMC-derived cells in the evolving atheroma.^{4, 7, 8} VSMC lineage tracing combined with single-cell RNA-sequencing (scRNA-seq) revealed transcriptomic heterogeneity of VSMCs in atherosclerosis,⁹ and Cre-loxP mouse models continually demonstrate an important role for VSMC-associated genes in atherogenesis.^{5, 10} These studies highlight the importance of innovative technologies in establishing phenotypic modulation of VSMCs from a contractile state to one resembling other cell types. Yet, while there is no doubt as to phenotypic plasticity of VSMCs in atherosclerosis,¹¹ controversy lingers over the precise fate of VSMCs and their phenotypic state(s) in this chronic inflammatory disease.^{12, 13}

Two recent papers in Circulation ^{14, 15} provide compelling proof for the 60-year-old idea of a “transitional stage”¹ between contractile VSMCs of the media and VSMC-derived cell types that contribute to the complex milieu of an atherosclerotic lesion (athero-lesion). Armed with modern innovations in VSMC lineage tracing, scRNA-seq of mouse and human atherosclerotic vessels, and human genomics, both studies demonstrate a multi-potential fate of dedifferentiated VSMCs to cell types postulated to either stabilize or destabilize the evolving athero-lesion. The findings cement the importance of VSMC phenotypic modulation in experimental and clinical atherosclerosis and provide a trove of new research questions to explore. Here, we highlight the major findings of these reports and the integrative technologies used to conclude the presence of a multipotent VSMC-derived cell state in experimental and clinical atherosclerosis.

VSMC-Derived Transitional Cells in Atherosclerosis

Alencar et al¹⁴ utilized Apoe null mice carrying Myh11-CreER^T2 with an inducible, fluorescently-tagged reporter that, upon tamoxifen (Tmx) administration, indelibly marks VSMCs in their native medial compartment of the vessel wall. Following a standard atherogenic diet, micro-dissected athero-lesions of the brachiocephalic artery (BCA), considered by some to be a surrogate for human coronary plaques,¹⁶ underwent cell sorting and scRNA-seq. The results disclosed seven different clusters of cells comprising previously contractile VSMCs. Further analysis of the top 100 genes in each cluster uncovered distinct molecular signatures, including inflammatory, extracellular matrix producing, and osteogenic states. Next, the authors repeated the same study in VSMC-restricted Krüppel-like factor 4 (Klf4) null mice and found a significant reduction in Galectin 3 (Lgals3) positive cells giving rise to cells of an osteogenic state and increases in VSMC-tagged cells exhibiting elevated expression of VSMC contractile markers, possibly plaque-stabilizing cells similar to the “fibromyocyte” of Wirka et al.¹⁷ These results are consistent with previous work showing reduced Lgals3+ VSMCs and lesion burden in VSMC Klf4 knockout mice ⁵ suggesting KLF4 acts as a molecular fate-switch, promoting the dedifferentiation of contractile VSMCs. KLF4 represses Myocardin,^{18, 19} “a component of a molecular switch for smooth muscle differentiation,”²⁰ thereby affording a transcriptional mechanism for the fate switching of contractile VSMCs to a dedifferentiated transitional state.

Alencar et al¹⁴ also made use of a novel dual reporter mouse for unambiguous tracking of cell states emanating from dedifferentiated VSMCs in athero-lesions. The results of these elegant studies showed >60% of VSMCs in athero-lesions activate Lgals3 transcription. These so-called “pioneer cells” generate inflammatory, extracellular matrix, and osteogenic cell states in the athero-lesion. However, contrary to earlier findings,⁵ few formerly contractile VSMCs went on to a macrophage state despite the presence of Lgals3, typically thought of as a marker for macrophages. Thus, under an atherogenic regimen that promotes endoplasmic reticulum stress and activates unfolded protein response pathways,²¹ contractile VSMCs dedifferentiate to a transitional pioneer cell, fated for non-macrophage cellular states (Figure, left). Importantly, scRNA-seq of human carotid endarterectomy specimens revealed similar clusters of VSMC-derived pioneer cells that may give rise to an osteogenic (plaque destabilizing) cell state.¹⁴

Figure.

Diverse VSMC-derived cell states in atherosclerosis. The LGALS3+ transitional cell of the intima is derived from dedifferentiated VSMC, perhaps in a deterministic manner. These cells represent pioneer cells of Alencar et al¹⁴ or Stem/Endothelial/Macrophage (SEM) cells of Pan et al¹⁵ and likely are precursors to plaque-stabilizing fibromyocytes of Wirka et al.¹⁷ Depending on micro-environmental signaling cues (e.g., retinoid signaling) and expression level of key transcription factors (e.g., KLF4), transitional cells may differentiate into a number of cells types that lead to fibrous cap formation (e.g., ACTA2+/PHACTR1+ cells) and plaque stability or expansion of the athero-core (osteogenic and inflammatory cells) leading to plaque instability (red zone). Unresolved questions are indicated. See text for more details.

Meanwhile, Pan et al¹⁵ discovered a similar transitional cell in a temporal scRNA-seq analysis of lineage-marked VSMCs of atherosclerotic aorta in low density lipoprotein receptor (Ldlr) null mice. This transitional cell state increased progressively over the course of the atherogenic diet, giving rise to three macrophage-like cell clusters and one fibrochondrocyte cluster (Figure, right). Similar to Alencar et al,¹⁴ each of these clusters showed attenuated expression of VSMC contractile genes indicating a dedifferentiated VSMC state. Because a previous report, using the Apoe null model of atherosclerosis and a different lineage reporter, failed to show evidence for a macrophage-like state in lineage-tagged VSMCs of athero-lesions,¹⁷ Pan et al¹⁵ repeated their temporal study in the Apoe null background; the results revealed similar clusters of de-differentiated VSMCs exhibiting a macrophage-like state. The transitional cell state of Pan et al¹⁵ displayed expression of markers for stem cell (Ly6a), endothelial cell (Vcam), and monocyte/macrophage (Ly6c1) and so they named these cells SEM cells. The SEM cells were, like the pioneer cells of Alencar et al,¹⁴ Lgals3+ and multipotent (Figure, right).¹⁵ Consistent with results of Alencar et al,¹⁴ scRNA-seq data from human carotid and coronary artery lesions yielded a transcriptionally related SEM cell type indicating conservation of this transitional state independent of embryological origin of VSMCs.¹⁵

While the findings of Pan et al¹⁵ suggesting a VSMC-derived macrophage cell state align with previous in vivo lineage tracing studies,^{4, 5, 8, 22, 23} they differ from those of both Alencar et al¹⁴ and Wirka et al¹⁷ who showed little evidence of VSMC-derived macrophages in athero-lesions. Aside from variations in lineage reporter, dosing of Tmx, and vascular tissues analyzed (Table), there are caveats that must be considered such as the need for rigorous analysis of multiple markers of the macrophage lineage and combined lipid staining with lineage marker or cell state marker.¹⁷ Although VSMC-derived foam cell formation occurs in experimental models^{17, 24} and, as originally proposed by Geer et al,¹ in human atherosclerosis,²² the molecular state of these cells remains controversial (Figure). Nevertheless, both studies in Circulation ^{14, 15} and the report of Wirka et al,¹⁷ provide incontrovertible molecular proof for complex VSMC transitions to plaque cell types in both experimental and human atherosclerosis (Figure). The relative importance of each VSMC-derived cellular state in the pathogenesis of atherosclerosis demands further probing given the fact that a number of patients with significant coronary artery disease (CAD) and its sequelae exhibit normal lipid profiles.²⁵

Table.

Comparative methods and results of recent studies analyzing VSMC fate and state in atherosclerosis

Methods	Alencar et al14	Pan et al15	Wirka et al17
Athero-lesion model	10 or 18 week HFDa in Apoe−/− mice	8, 16, 26 week HFD in Ldlr−/− and Apoe−/− mice	8 and 16 week HFD in Apoe−/− mice
Lineage reporter(s)b	SF-eYFP and SR-F-tdTomato-S-GFP	SF-ZsGreen1	SF-tdTomato
Cre driver mousec	Myh11-CreERT2 Myh11-Cre(Dre)ERT2 Lgals3-SR-IRES-Cre	Myh11-CreERT2	Myh11-CreERT2
Tamoxifen schedule	0.1 mg x 10 days, IP	2 day diet	0.2mg/g x 2 days, PO
Mouse tissue for scRNA-seqd	Micro-dissected lesions of BCA	Whole BCA and ascending-thoracic aorta	Whole aortic root and ascending arch
Human scRNA-seqd	Carotid endarterectomy	Carotid/coronary	Coronary
scRNA-seq clustering	UMAP, 14 clusters	UMAP, 12 clusters	tSNE, 12 clusters
De-differentiated VSMC	Pioneer cell	SEM cell	Modulated SMC
Intimal VSMC-derived cell states	Chondro-osteogenic; ECM-rich (FM?); inflammatory	Monocyte/macrophage; fibrochondrocyte (FM?)	Fibromyocyte (FM)
Protein analysis	CyTOF/IFM	metaVIPER/IFM	CITE-seq/IFM
ChIP-seq	KLF4/OCT4	na	TCF21
CAD GWAS	88 KLF4/OCT4 TFBS near 64/163 CAD GWAS	226 RA targets near 37 CAD GWAS	Neg TCF21 exp with 36 CAD SNPs
VSMC KO	Oct4, Klf4	na	Tcf21

^aAbbreviations: HFD, high fat diet; SF, stop flox; SR, stop rox; F, flox; S, stop; LCA, left carotid artery; BCA, brachiocephalic artery; UMAP, uniform manifold approximation and projection; tSNE, t-distributed stochastic neighbor embedding; SEM, stem cell, endothelial cell, monocyte/macrophage; CyTOF, cytometry by time of flight; VIPER, virtual inference of protein activity by enriched regulon; IFM, immunofluorescence microscopy; CITE-seq, cellular indexing of transcriptomes and epitopes by sequencing; ChIP-seq, chromatin immunoprecipitation sequencing; CAD, coronary artery disease; GWAS, genome wide association study; TFBS, transcription factor binding sites; RA, retinoic acid; Neg, negative; exp, expression; SNPs, single nucleotide polymorphisms; VSMC, vascular smooth muscle cell; KO, knockout; IP, intraperitoneal; PO, per os (oral gavage); na, not applicable.

^bLineage reporters were activated by Myh11-CreER^T2, except for the dual lineage reporter of Alencar et al; this reporter (abbreviated here as SR-F-tdTomato-S-GFP) was sequentially activated by a

^cMyh11-DreER^T2 and Lgals3-SR-IRES-Cre.

^dEmbryological origin of each mouse and human vascular tissue for scRNA-seq analysis is neural crest (ascending aorta, arch, LCA, and BCA); secondary heart field (aortic root); splanchnic mesoderm (thoracic aorta), and proepicardial organ (coronary).

Distinct Regulatory Mechanisms Govern the Fate of VSMC-Derived Transitional Cells in Atherosclerosis

Alencar et al¹⁴ performed bulk RNA-seq and ChIP-seq for KLF4 and OCT4 binding events in BCA lesions from high-fat diet fed Apoe null mice with Klf4 or Oct4 knocked out in VSMCs. Putative KLF4 target genes, reduced in the background of Klf4 null VSMCs, were elevated in Oct4 null VSMCs (e.g., inflammatory genes) whereas KLF4 targets elevated in Klf4 null VSMCs showed attenuated expression in Oct4 null VSMCs (e.g., myogenic genes). These opposing programs of gene expression are consistent with lesion pathology in the context of Klf4 null VSMCs (reduced athero-lesion) versus Oct4 null VSMCs (increased athero-lesion).^{5, 26} Remarkably, 88 of the KLF4 and OCT4 targets are the closest annotated gene in nearly 40% of genome wide association study (GWAS) loci implicated in CAD. As the ChIP-seq experiments were done in mice, it is unclear whether similarly positioned KLF4 and OCT4 binding sites exist in human genomes and if binding has any consequence for altered expression of CAD GWAS targets. Notably, the KLF4 gene is itself a CAD GWAS risk allele with the putative variant (rs944172) located ~265 kb upstream of the KLF4 transcription start site. The nature of this variant is unclear; however, it overlaps a weak AP1 binding site with some evidence for JUND binding (UCSC Genome Browser data). The recent Nobel Prize winning technology related to clustered regularly interspaced short palindromic repeats (CRISPR) genome editing²⁷ will enable functional definition of KLF4 and OCT 4 binding sites in proximity to CAD risk alleles.

Pan et al¹⁵ employed a bioinformatics tool to gain insight into potential “master regulators” of the transitional cell. Meta Virtual Inference of Protein activity by Enriched Regulon (metaVIPER)²⁸ uses differential gene expression between two conditions (VSMC and SEM states) to infer potential protein activities that explain differential transcript profiles. metaVIPER disclosed cellular retinoic acid binding protein 2 as a putative master regulator of the VSMC-SEM transition. Retinoids, such as all-trans retinoic acid (atRA), impact VSMC phenotype by inhibiting VSMC growth, migration, and lesion burden in experimental models of vascular disease, presumably through nuclear receptor-mediated effects on VSMC gene expression.²⁹ Interestingly, one such nuclear receptor, RARB, was reduced in human athero-lesions and atRA inhibited the VSMC-SEM transition in vitro and in vivo (Figure, right).¹⁵ Moreover, atRA reduced intimal area in experimental athero-lesions and elevated the proportion of VSMC-tagged cells destined for the fibrous cap. Pan et al¹⁵ then turned to integrative genomics and found a number of down-regulated retinoid-response genes in unstable human athero-lesions. Of notable interest was the finding that some retinoid-responsive genes co-localized with CAD-associated risk variants. However, as with Alencar et al¹⁴, these associations will require follow up studies, including formal demonstration of retinoid receptor binding sites around CAD risk alleles and variant modeling/correcting using CRISPR or prime editing³⁰ in human induced pluripotent stem cells. In addition, conserved regulatory variants in CAD GWAS risk alleles or retinoid receptor binding sites in mouse models of atherosclerosis could be genome edited as shown for other transcription factor binding sites.^{31, 32}

In the study of Wirka et al, VSMC-restricted knockout of transcription factor 21 (Tcf21) attenuated VSMC phenotypic modulation and reduced the number of fibromyocytes at the fibrous cap.¹⁷ Consistent with the mouse data, expression of human TCF21, another CAD GWAS risk allele, correlated with VSMC phenotypic modulation in human athero-lesions as well as reduced CAD risk.¹⁷ Interestingly, the lead CAD GWAS variant (rs12190287) disrupts a microRNA binding site in the 3’ untranslated region of TCF21.³³

Questions and Future Directions

The two recent studies in Circulation,^{14, 15} and an earlier one in Nature Medicine,¹⁷ integrated state-of-the-art tools in genetics and genomics to formally demonstrate VSMC fate switching in experimental and clinical atherosclerosis. All three studies, while subtly different in design and conclusion (Figure and Table), support a transitional cell type derived from previously contractile VSMCs and possessing a unique molecular signature. An additional study found a VSMC-derived transitional mesenchymal stem cell population in a model of human aneurysm formation.²³ Together, these exciting findings engender new questions.

First, are a limited number of VSMCs pre-determined to migrate from the tunica media and fate switch to the pioneer or SEM cell state in the evolving intima or is the process stochastic? A recent Circulation paper suggests the former, as higher expressing NOTCH3 VSMCs appear to leave the media and clonally expand in pulmonary intimal lesions.³⁴ Further characterization of the transcriptome, epigenome, and proteome of medial VSMCs will assist in distinguishing between a deterministic or stochastic basis for the migration and expansion of VSMC-derived transitional cells in vascular disease.^{7, 35-37} It is perhaps ironic that with all the controversy over whether or not stem cells reside in the medial layer of conduit arteries,^38-40 the most abundant cell type in the vessel wall (VSMC) possesses the inherent potential to acquire so many different cell fates in vivo.^{4, 5, 14, 15, 17, 24, 41}

Second, do pioneer/SEM cells clonally expand in athero-lesions? Although no formal proliferation assays were conducted in any of the three studies,^{14, 15, 17} previous lineage tracing experiments ^{7, 42} and a stem cell antigen 1 (Sca1) signature of pioneer/SEM cells ^{14, 15} would support their clonal expansion. In this context, a recent report concluded little contribution of baseline vascular Sca1 positive cells in the formation of athero-lesions.⁴³ However, it remains to be formally shown whether the acquisition of a Sca1 positive state, as in VSMC-derived transitional cells,^{14, 15, 17} triggers their expansion and subsequent differentiation to other intimal cell types. A confetti reporter mouse study supports clonal expansion of VSMC-derived mesenchymal stem cells in a model of aneurysm formation.²³ Perhaps the pioneer/SEM state represents re-expression of an earlier transitional state passed through during development of VSMCs in the embryo. It will be important to determine whether human VSMCs express genes with functions similar to Sca1 and if such factors drive the clonal expansion of human VSMCs.

Third, what is the complete molecular signature of VSMC-derived cell types in athero-lesions, especially with respect to long noncoding RNAs which, increasingly, play a role in atherogenesis?⁴⁴ Addressing this question will be challenging given the constraints with current scRNA-seq library construction (polyadenylated RNA only) and the low abundance of most long noncoding RNAs. However, we can safely anticipate advances in next generation sequencing that will increase the depth of unbiased sequencing at the individual cell level.

Fourth, what are the signaling events in the local micro-environment that promote one SEM/pioneer cell-derived state over another? Put another way, what are the underlying mechanisms for Klf4¹⁴ and Tcf21¹⁷ transcription or altered retinoid signaling¹⁵ under vascular pathological conditions? Insight into this potential "Achilles heel" could spawn the development of therapeutics to mitigate the number of plaque-destabilizing cells or enhance the number of plaque-stabilizing cells. In this context, conditions of reduced TGFβ signaling and elevated cholesterol in the vessel wall resulted in massive Klf4 induction, suggesting these micro-environmental perturbations are permissive for VSMC dedifferentiation and the emergence of a stem cell fate.²³

Finally, are previously contractile VSMCs licensed to engulf lipid and release inflammatory mediators with a monocyte/macrophage-like phenotype in athero-lesions? Despite decades of research, this question remains to be firmly established.

From a clinical standpoint, one can imagine harnessing advances in imaging and nanotechnology to probe lesions for the presence and possible ablation of plaque destabilizing cells. Alternatively, it may be feasible to target the transitional cells directly to bias their fate towards plaque stabilizing cell types such as the fibromyocyte. If advances in biomedical innovations over the last 60 years are any indication, the next 60 years will surely unfold with unanticipated discoveries for precision-guided control of VSMC fate switching and cellular states in atherosclerosis.

Non-standard Abbreviations and Acronyms

atRA

all-trans retinoic acid

BCA

Brachiocephalic artery

CAD

Coronary artery disease

ChIP-seq

Chromatin immunoprecipitation sequencing

CRISPR

Clustered regularly interspaced short palindromic repeats

GWAS

Genome wide association study

KLF4

Krüppel like factor 4

OCT4

Octamer-binding transcription factor 4

Sca1

Stem cell antigen 1

scRNA-seq

Single cell RNA-sequencing

SEM

Stem cell, endothelial cell, monocyte/macrophage

TCF21

Transcription factor 21

Tmx

Tamoxifen

t-SNE

t-distributed stochastic neighbor embedding

UMAP

Uniform manifold approximation and projection

VIPER

Virtual inference of protein activity by enriched regulon

VSMC(s)

Vascular smooth muscle cell(s)