Heterogeneous subpopulations of adventitial progenitor cells regulate vascular homeostasis and pathological vascular remodelling

Abstract

Cardiovascular diseases are characterized by chronic vascular dysfunction and provoke pathological remodelling events, such as neointima formation, atherosclerotic lesion development, and adventitial fibrosis. While lineage-tracing studies have shown that phenotypically modulated smooth muscle cells (SMCs) are the major cellular component of neointimal lesions, the cellular origins and microenvironmental signalling mechanisms that underlie remodelling along the adventitial vascular layer are not fully understood. However, a growing body of evidence supports a unique population of adventitial lineage-restricted progenitor cells expressing the stem cell marker, stem cell antigen-1 (Sca1; AdvSca1 cells) as important effectors of adventitial remodelling and suggests that they are at least partially responsible for subsequent pathological changes that occur in the media and intima. AdvSca1 cells are being studied in murine models of atherosclerosis, perivascular fibrosis, and neointima formation in response to acute vascular injury. Depending on the experimental conditions, AdvSca1 cells exhibit the capacity to differentiate into SMCs, endothelial cells, chondrocytes, adipocytes, and pro-remodelling cells, such as myofibroblasts and macrophages. These data indicate that AdvSca1 cells may be a targetable cell population to influence the outcomes of pathologic vascular remodelling. Important questions remain regarding the origins of AdvSca1 cells and the essential signalling mechanisms and microenvironmental factors that regulate both maintenance of their stem-like, progenitor phenotype and their differentiation into lineage-specified cell types. Adding complexity to the story, recent data indicate that the collective population of adventitial progenitor cells is likely composed of several smaller, lineage-restricted subpopulations, which are not fully defined by their transcriptomic profile and differentiation capabilities. The aim of this review is to outline the heterogeneity of Sca1+ adventitial progenitor cells, summarize their role in vascular homeostasis and remodelling, and comment on their translational relevance in humans.

Open in new tabDownload slide

1. Introduction

The tunica adventitia was traditionally depicted as an inert layer composed of connective tissue and vasa vasorum.¹ However, it is now clear that the adventitia is home to a diverse cell population ranging from resident progenitor cells, adipocytes, fibroblasts, dendritic cells, macrophages, lymphocytes, and the vascular neural network all scaffolded by extracellular matrix, perivascular fat, and vasa vasorum.^2–5 Robust adventitial expansion is consistently observed in various models of vascular disease, and despite its relatively distant location from the luminal interface, disease-induced adventitial remodelling is hypothesized to exert subsequent pathological remodelling on the tunica media and intima.^6–9 The field’s traditional understanding of vascular remodelling is being refined and is complemented by the ‘outside-in hypothesis’, a more recent model that depicts the adventitia as a biological processing centre that can regulate the effector functions of medial and intimal cells and suggests that certain adventitial cell populations are poised to integrate environmental cues and initiate mechanisms that promote vascular remodelling.^6,7 Certainly, it is well understood and acknowledged that vascular pathology can be initiated ‘inside-out’, which involves initial endothelial dysfunction and expression of surface adhesion molecules and leading to monocyte recruitment and subsequent inflammation within the subendothelial space resulting in medial remodelling and neointima development.^10–15 However, applications of in vivo cell fate-mapping systems and single-cell sequencing technology have led to new discoveries about adventitia-mediated vascular remodelling and have identified adventitial biology as an important point of focus for the field.¹⁶

While the response to vascular injury is likely initiated on both sides of the vascular wall (intima and adventitia), the impacts of crosstalk between adventitial and medial/intimal cells have not been adequately addressed. To what extent each side contributes to atherogenesis, inflammation, vascular fibrosis, and acute vascular remodelling remains unclear. What is clear is that the adventitia is highly involved in vascular homeostasis and focused studies are warranted to understand the adventitial response to disease states.^1,2,6

The adventitia contains multipotent and heterogeneous populations of progenitor cells, which express the glycosyl phosphatidylinositol-anchored cell surface protein stem cell antigen-1 (Sca1) hereto defined as AdvSca1 cells, which may be critical to the concept of ‘outside-in’ vascular remodelling. AdvSca1 cells reside along the media: adventitia interface and are maintained in a restricted niche of sonic hedgehog (Shh) signalling.^2,17,18 Over the past several years, many groups have demonstrated that AdvSca1 cells can differentiate into smooth muscle cells (SMCs), endothelial cells (ECs), myofibroblasts, macrophage-like cells, chondrocytes, and adipocytes under specified conditions.^1,19–24 These cells can become activated and exhibit migratory phenotypes after vascular insult and are involved in the development of vascular fibrosis and inflammation in hyperlipidaemic and atherogenic models.^25–27 AdvSca1 cells are crucial players involved in vascular remodelling and homeostasis. Diphtheria toxin ablation of AdvSca1 cells ameliorated vascular fibrosis in the kidney and cardiac tissue and improved left ventricular function in models of organ fibrosis.²⁶ Specific ablation of adventitial progenitor cells decreased vascular calcification in a model of chronic kidney disease.²⁷ Finally, deletion of AdvSca1 progenitor populations resulted in decreased SMCs and a poorly healed medial layer in model of severe transmural vascular injury.²⁸ These data support that AdvSca1 cells are important for regulating both vascular homeostasis and disease. Typical markers used to identify these progenitor cells include Sca1, CD34, c-Kit, PDGFRα, Klf4, and Gli1.^{2,18,26,29–31} While less abundant than Sca1, c-Kit has been shown by several investigators to be expressed by subsets of adventitial progenitor cells suggesting that c-Kit is also an important lineage marker for vascular progenitor cells.^18,31 Consistent with their stem-like, progenitor phenotype, AdvSca1 cells typically do not express CD31 and αSMA, markers of mature ECs and SMCs, respectively.³² AdvSca1 cells do not originate from the bone marrow or from neural crest cells.^18,20 Sca1 is a marker only of the adventitial progenitor population in mice. As discussed below, a similar cell population of adventitial progenitor cells in humans exists but the functions and, as Sca1 is not expressed in humans, consensus markers are less well defined (Table 1).

It should be pointed out that while Sca1 is frequently used as a marker to identify adventitial progenitor cells, its function in regulating a stemness phenotype is poorly understood. However, Sca1-null mice exhibit increased stem/progenitor cell differentiation and reduced self-renewal potential leading to exhaustion of the stem/progenitor cell pool, which supports not only a role for Sca1 as a reliable marker, but also a functional role for Sca1.³³ In addition, reliance on cell surface markers alone limits a complete understanding of subtle differences that divide AdvSca1 cells into smaller, refined subpopulations, each of which is likely unique in its genetic profile, differentiation tendencies, and overall functional significance. Importantly, as differentiation into specific cell types is associated with loss of these AdvSca1 markers, a major gap in the field includes the lack of distinct lineage-mapping systems to faithfully track these cells and define their function during disease progression.

1.1 Advsca1 progenitor cells are heterogeneous

The multipotency of AdvSca1 cells was observed in early ex vivo experiments. When primary AdvSca1 cells were placed into serum-containing medium, 50% differentiated into SMCs while roughly 25% retained Sca1 and exhibited self-renewal proliferation.^2,20 BMP2 stimulation induced a small percentage (≤5%) of AdvSca1 cells to display osteogenic activity while insulin, isobutylmethylxanthine, and dexamethasone stimulation resulted in adipocyte differentiation.^2,20 While these experiments are valuable, they do not address whether AdvSca1 cells represent a homogeneous population of progenitor cells with a wide-range of differentiation capacities or whether AdvSca1 cells are comprised of several, smaller lineage-restricted subpopulations that have less diverse cell lineage propensities. As discussed below, the current literature supports the latter hypothesis. The future challenges ahead are to delineate AdvSca1 subpopulations, to define their cellular behaviours in the context of adventitial remodelling, and to understand their contributions to vascular remodelling.

Our group characterized a subpopulation of AdvSca1 cells that are derived from mature SMCs (defined by us as AdvSca1-SM cells).¹⁹ Using a SMC-reporter lineage-tracing system, we were the first to describe the potential of mature SMCs to migrate outward into the adventitia, where they lose all characteristics of SMCs, acquire Sca1 surface expression, and display a stem cell phenotype that is dependent upon induction of the transcription factor, Klf4. This process of SMC reprogramming into AdvSca1-SM cells occurs under normal, physiological conditions, but also under pathological settings. In response to acute wire-induced femoral injury or carotid artery ligation, there is a significant expansion of AdvSca1-SM cells, which is associated with substantial adventitial remodelling. Using a comprehensive immunophenotyping panel, we characterized two clusters of AdvSca1 cells: the previously mentioned AdvSca1-SM subpopulation and a separate, non-SMC-derived population termed AdvSca1-MA cells. Depending on the vessel, AdvSca1-SM cells account for 35–65% of the total AdvSca1 cell population (∼65% of total in combined aortic arch plus carotid artery, ∼40% of total in the descending aorta, ∼35% of total in femoral arteries). Of the AdvSca1-SM cell population, >98% are CD45− and nearly all express CD140a/PDGFRα or CD140b/PDGFRβ, which are associated with fibroblast or mesenchymal cell differentiation.^34,35 In contrast, the AdvSca1-MA cells are suspected to represent several smaller subpopulations as 31% of this population expressed CD45, implicating this subtype as part of a previously described subtype of Sca1+ myeloid-like progenitor cells.^{23,24,32,36,37} On the other hand, 69% of AdvSca1-MA cells were CD45− and expressed Bst1, suggesting their commitment to EC differentiation, which is consistent with studies that demonstrated Sca1+ adventitial cells with endothelial propensity.³⁸ Interestingly, a study by Dobnikar et al.³⁹ identified a rare population of SMCs that express Sca1 in healthy mouse aorta. Only 1% of medial SMCs stained positive for Sca1 underlying their rarity. These Sca1+ SMCs exhibited a broad range of contractile genes, such as Myh11, Acta2, and SM22, indicating these cells were not uniformly mature contractile SMCs. Instead, Dobnikar et al. discovered that these Sca1+ SMCs were enriched for ontology terms of the dedifferentiated SMC phenotype including cell migration, proliferation, and extracellular matrix production. The Sca1+ SMCs were also enriched for genes associated with activation of signalling pathways (PI3K, TGF-β, and chemokines), reflecting a responsive cell state. The Sca1+ SMC population described by Dobnikar et al.³⁹ likely correlates with the AdvSca1-SM population, which reflects SMCs being reprogrammed to a more dedifferentiated, stem-like state. In addition, the existence of medial Sca1+ SMCs could complicate the interpretation of lineage-tracing studies that conclude an adventitial progenitor cell origin for SMCs.

As alluded to above, the adventitia also contains a subpopulation of Sca1+ macrophage-progenitor cells. Two seminal studies by Psaltis et al.^23,24 describe a definitive population of Sca1+/CD45+ cells that resides in the adventitia throughout the full length of the aorta. Primary culture of aortic cells in methylcellulose media generated a broad spectrum of colony forming units including a significant proportion of macrophage-specific colony forming units (CFU-M).²³ The CFU-M colonies expressed Sca1 and CD45 and further flow cytometric analysis revealed that Ly6C, a marker used to define circulating monocytes, and macrophage colony stimulating factor receptor were enriched in Sca1+/CD45+ populations. In subsequent studies,²⁴ RNA microarray identified >500 differentially expressed genes in Sca1+/CD45+ populations as compared to Sca1+/CD45− (SMC progenitors) and Sca1−/CD45+ (resident leucocytes) populations including granulocyte colony stimulating factor receptor (Csf2) and the chemokine receptor CX3CR1. Compared to wild-type mice, Sca1+/CD45+ populations are increased in ApoE- and LDL-R-null mice indicating that the adventitial macrophage progenitor populations can be activated in the setting of hyperlipidaemia and atherosclerosis. Adoptive transfer studies demonstrated that the aortic-macrophage progenitors are not significantly replenished or maintained by circulating bone marrow-derived cells. Donor GFP+ cells constituted only 1% of CFU-M cells generated from aortic Sca1+/CD45+ progenitor cells. To further support the hypothesis that resident vascular macrophages can be derived from local progenitor cells, Psaltis et al. treated mice with liposomal clodronate, which induces depletion of circulating monocytes. Despite marked reductions in circulating monocyte numbers, clodronate was in fact associated with a slight but significant increase in the numbers of CFU-M colonies from the aortic wall. In addition, this group showed that a rare vessel-derived haematopoietic stem cell population is capable of self-renewal and low-level reconstitution of leucocytes.²⁴ This evidence supports a subpopulation of adventitial Sca1+/CD45+ progenitor cells that can give rise to macrophages in the vascular wall independent from bone marrow contribution.

Ultimately, these findings support the existence of multiple, distinct subtypes that make up the global AdvSca1 progenitor cell population, which raises an important question: Do the origins of AdvSca1 subpopulations affect their differentiation capabilities and responses to vascular perturbation? In development, ancestral stem/progenitor cells that give rise to haematopoietic blood cells also play a part in the formation of blood vessels.⁴⁰ Others have showed Sca1+/CD45+ cells that lack expression of Tie-2, CD31, and CD144, but formed CD31+ colonies under endothelial differentiation conditions in vitro.⁴¹ Sca1+/CD45+ cells from donor mice generated new vessels in ischaemic hindlimbs of mice and improved perfusion 50% as compared to sham.⁴¹ These surface markers and differentiation tendencies show striking similarities to our identified AdvSca1-MA population. While these studies do not define the origins of these cells, they provide insight into how some of these AdvSca1-MA cells are derived.

Additional data supporting the heterogeneity among AdvSca1 cells comes from a single-cell RNA sequencing (scRNA-Seq) study of mouse aortic adventitial cells. The adventitia layers of wild-type and ApoE-null mice fed a standard chow diet were dispersed and analysed using scRNA-Seq. Unbiased cell cluster analysis identified immune cell clusters of macrophages, T-cells, and B-cells as well as rare neuron and lymphatic cell types.⁴² ApoE-null mice exhibited increased numbers of adventitial cells compared to wild-type mice. The study further analysed the non-immune adventitial cell clusters that included ECs, SMCs, and four mesenchymal populations (Mesen I-IV). Notably, the transcriptome of Mesen II cells expressed Sca1, CD34, and other stemness factors, suggesting their similarity to previously defined AdvSca1 cells. Using this same technology, Gu et al.⁴³ also characterized mesenchymal stem cells (MSCs) that reside in perivascular adipose tissue (PVAT), a tissue that is physically approximated to the adventitia and is proposed to impact vascular inflammation and endothelial dysfunction. The investigators sorted a rare population of Sca1+/CD29+/CD34+ cells from pooled PVAT of murine thoracic aortae. These cells were negative for CD45, CD11b, and Plin1, supporting they are not of haematopoietic or adipocyte lineage, and also negative for SMC markers (Cnn1, Myh11, and Acta2). Cluster analysis revealed two distinct populations of Sca1+ cells. Cluster A cells had endothelial potential, expressed CD31 and CD36, and GSEA pathway analysis revealed pathways for VEGF and PPAR signalling. Cluster B demonstrated stromal and SMC potential, expressed Tgfβr2, and GSEA pathway analysis revealed ontology signatures related to PDGF signalling, insulin-like growth factor signalling, and TGF-β signalling, all of which are associated with SMC and mesenchymal differentiation.⁴³ Together these studies showcase the power of scRNA-Seq as a valuable tool to identify signalling pathways and mechanisms that can further our understanding of AdvSca1 subpopulations.^44,45 However, a limitation of scRNA-Seq technology is that the technique captures cell phenotypes and differentiation trajectories on a single time point, which cannot exclude that heterogeneity is an interchangeable and dynamic phenotype. Therefore, clonal analysis remains fundamental to establish the differentiation tendencies of a single cell.

The current data indicate that the full population of AdvSca1 cells is remarkably heterogeneic indicating that more narrowly defined subpopulations of Sca1+ cells arising from different origins collectively make up the global population of AdvSca1 cells. It is likely that certain subpopulations of AdvSca1 cells may be more committed to mesenchymal phenotypes, such as SMCs, fibroblasts, and myofibroblasts in disease states^19,26,30 while other populations may be more committed to myeloid or endothelial-like cells to support innate immune cell differentiation or vasa vasorum expansion, respectively.^23,24,38,41 It is possible the various subpopulations may mutually cooperate to support adventitial remodelling and impart physiologically significant effects on all layers of the vessel wall (Figure 1). Table  1 summarizes studies of murine adventitial progenitor cells.

1.2 Signalling pathways and factors involved in AdvSca1 regulation

Some of the underlying signalling pathways that regulate AdvSca1 differentiation, migration, and response to injury have been characterized. Early studies demonstrate that Shh signalling is required to maintain the population of Sca1+ adventitial cells in the aortic root. Shh signalling was first detected at embryonic day 15.5 (e15.5), reached peak levels between postnatal days 1–10, and then became restricted in adult mice to the media-adventitia interface. Shh-deficient mice exhibited a dramatic decrease in Sca1+ cells thus highlighting the importance of this signalling niche.²⁰ The transcription factor Klf4 is required to maintain the stemness phenotype in AdvSca1 cells. Klf4 is highly expressed by AdvSca1-SM cells and is required for in vivo SMC reprogramming to the AdvSca1-SM cell subpopulation.¹⁹ Klf4 silencing in vitro was associated with increased numbers of adventitial cells exhibiting αSMA contractile filaments suggesting that loss of Klf4 accelerates differentiation of AdvSca1 cells towards the SMC fate.¹⁹ Cre-mediated knockdown of Klf4 selectively in AdvSca1 cells resulted in decreased Sca1 expression but surprisingly promoted a spontaneous increase in AdvSca1-SM-derived cells expressing a myofibroblast phenotype as compared to wild-type animals. Compared to wild-type, carotid arteries from mice deficient of Klf4 specifically in AdvSca1-SM cells exhibited increased adventitia-to-media ratio and increased adventitial collagen deposition.³⁰ Taken together, these results demonstrate that Klf4 is essential for the maintenance of AdvSca1-SM progenitor cell phenotype. Others have shown that AdvSca1 differentiation into ECs requires the transcription factor ETV2.²² AdvSca1 cells transduced with ETV2 demonstrated increased expression of endothelial genes (VE-cadherin and Tie2) and down-regulation of SMC genes. Wire-injured femoral arteries grafted with ETV2-transduced progenitor cells demonstrated reduced neointimal hyperplasia as compared to controls supporting ETV2 as an important transcription factor in EC differentiation and vascular repair.²² In the setting of hyperlipidaemia, Sca1+ adventitial cells up-regulate the microRNA miR-29b. Overexpression of miR-29b led to increase signalling related to migration and up-regulation of matrix metalloproteinase-9 and sirtuin 1.²⁵ Finally, TGF-β signalling is known to be involved in AdvSca1 differentiation into myofibroblasts. Song et al.⁴⁶ report that stimulation of AdvSca1 cells with TGF-β1 induced Smad2/3 phosphorylation within 30 min, and AKT phosphorylation was observed 8 h after stimulation. These intracellular signalling changes were associated with expression of myofibroblast markers including αSMA, vimentin, and S100A4 supporting that TGF-β signalling is important for myofibroblast differentiation of AdvSca1 cells. Thus, further identification of pathways and factors involved inAdvSca1 regulation and function could lead to therapeutically relevant approaches to manipulate specific pathways in vivo to alter AdvSca1 cell function and contributions to various pathological settings.

1.3 Adventitial progenitor cell function in health and disease

1.3.1 Physiologic functions of AdvSca1 cells in growth and development

The function of AdvSca1 cells has been examined in models of vascular injury, hypertension, and other vascular diseases, but there is also data to support their physiologic functions. In development, AdvSca1 cells were identified to first populate the adventitia from E15.5 to E18.5, well after the full complement of medial SMCs has been established and layering has stopped.²⁰ Typically, mice grow tremendously in the first 6 weeks of life and this rapid growth requires expansion of the vascular system.^47,48 The aorta must elongate, increase its calibre, and thicken in proportion to the growing organism. The physiologic functions of AdvSca1 cells are supported by numerous studies demonstrating these cells can differentiate into ECs, pericytes, and SMCs to contribute to vessel growth and maturation.^49–51 Considering the clustering of these resident progenitor cells near the media-adventitia border and their differentiation to SMCs in vitro, it is likely that AdvSca1 cells represent a local reservoir of cells that can proliferate and differentiate into cells that contribute to the growing blood vessel. Additional stem/progenitor niche signalling pathways, including Wnt/β-catenin and BMP signalling, have been implicated in maintaining the stem-like phenotype of AdvSca1 cells. Our group employed unbiased bulk RNAseq approaches and found Hedgehog, Wnt/β-catenin, BMP, and Klf4 signalling were overrepresented in distinct subpopulations of AdvSca1 cells isolated from uninjured vessels.³⁰ Acute carotid ligation-induced dramatic down-regulation of these signalling pathways indicating differentiation of the progenitor cells and loss of the basal, stem-like phenotype. These data support the concept that the adventitial microenvironment sustains progenitor niche signalling pathways that are important for maintaining the stem-like phenotype of AdvSca1 cells under physiologic conditions.

1.3.2 Distinct populations of AdvSca1 cells contribute to pathological vascular remodelling and fibrosis

While evidence shows AdvSca1 cells demonstrate proliferation under basal conditions and may function to support physiologic processes, such as self-renewal and SMC cell turnover,²⁰ these cells can become pathologic effectors in disease settings. To assess the activity of vascular progenitor cells in the setting of pathologic organ fibrosis, Kramann et al.²⁶ utilized a lineage-mapping system exploiting a Gli1 promoter-Cre^ERT driver to track adventitial progenitor cells. The choice to use Gli1 as a progenitor-specific marker was based on a previous study that identified a population of Gli1+ periarterial MSCs from mouse incisor that express typical stem-related markers, including Sca1.⁵² These periarterial MSCs exhibited trilineage differentiation, and after incisor injury, Gli1+ cells became activated and increased secretion of dentine, a protein involved in enamel repair.⁵² Using a Gli1-Cre^ERT-mTomato reporter mouse, Kramann et al. identified Tomato+ (Gli1+) cells in the adventitia of many large arteries and tissues, including the aorta, carotid arteries, lung, kidney, and liver parenchyma. These cells highly co-expressed Sca1 and CD29 and were negative for CD31 and CD45. Induction of cardiac, lung, and liver fibrosis revealed Gli1+ progenitor cells from the vasculature of these organs expanded in the adventitia, differentiated into myofibroblasts, and were spatially correlated with increased perivascular collagen deposition. Interestingly, Gli1 is overrepresented in the SMC-derived AdvSca1-SM population as compared to the AdvSca1-MA population and can be used as a marker to distinguish AdvSca1-SM from AdvSca1-MA cells;^19,30 thus Kramann’s findings using a Gli1 reporter system likely relate specifically to the behaviour of AdvSca1-SM cells.

The findings from Kramann et al. make up only part of the evidence supporting AdvSca1 involvement in pathological fibrosis. Our group used a similar lineage-mapping system as Kramann et al. and showed that Gli1+ AdvSca1 cells from the carotid artery expand after acute ligation-induced injury, lose expression of Sca1 and Klf4 and other stem cell markers, but gain expression of myofibroblast markers, such as periostin, to contribute to pathological adventitial fibrosis.³⁰ In separate experiments studying wound repair, wound-induced differentiation of Sca1+ cells resulted in more αSMA+ myofibroblasts in cutaneous vascular beds.⁵³ A separate study showed that Sca1+/CD31− adventitial cells significantly up-regulated ECM proteins (Col1a1, Col3a1, Col5a1, and fibronectin) after angiotensin II administration as compared to Sca1−/CD31+ cells and Sca1−/CD31− cells.⁵⁴ This strong body of evidence supports the concept that specific AdvSca1 cell populations demonstrate preferential tendencies towards mesenchymal differentiation and contribute to pathological vascular fibrosis.

1.3.3 Direct involvement of AdvSca1 cells in acute injury-induced neointima formation?

Neointimal hyperplasia primarily involves mature SMC proliferation, migration, and phenotypic modulation from a quiescent, contractile cell into a more synthetic phenotype. This is a complex process regulated by transcription factors, miRNAs, and paracrine signalling.^55,56 Bone marrow transplant studies, SMC-lineage-tracing systems, and single-cell sequencing have been instrumental to establish that medial SMCs are the predominant source of neointimal cells and highlight the phenotypic diversity of SMCs.^57–59 It is important that we recognize and appreciate the years of strong evidence positing that the activation and expansion of medial SMCs is the predominant mediator of neointimal remodelling.

Given AdvSca1 progenitor cells can give rise to SMCs, it will be interesting to clarify the extent to which AdvSca1-derived SMCs contribute to neointimal hyperplasia. The answer to this question remains unclear as the data from several laboratories have been variable. Using separate lineage-mapping systems to track AdvSca1 progenitor cells, two independent groups employed a very similar protocol to induce neointima formation and to track reporter-positive cells within the neointima. Tang et al.²⁸ developed a Sca1-Cre^ERT; Pdgfrα-Dre^ERT-mTomato dual recombinase fate-mapping system to track a select population of adventitial Sca1+PDGFRα+ progenitor cells while Kramann et al.²⁷ used the Gli1-Cre^ERT-mTomato system described earlier. In both experiments, femoral artery samples were analysed 4 weeks after wire injury, and the results were notably disparate. On the one hand, the Tang et al. found the femoral neointimal lesions were completely devoid of mTom+ cells that expressed SMMHC, a marker of a mature SMCs. However, Tang et al. also investigated Sca1+PDGFRα+ neointimal involvement in an arterial anastomosis model that involves full transmural injury to the arterial wall and loss of local SMCs followed by regeneration of the medial layer and healing of the anastomosis site. In this model of severe vascular injury, 13% of SMCs in the neointima were derived from Sca1+PDGFRα+ cells. This small population of progenitor-derived SMCs expressed CNN1 and SM22 but not SMMHC suggesting these were immature SMCs.

On the other hand, Kramann identified a significant number of mTom+ progenitor cells in the neointima using their Gli1 reporter system and the majority of these mTom+ cells expressed αSMA and CNN1. While both of these studies were consistent in that the majority of neointimal cells originated from SMCs, these studies offer opposing conclusions regarding the contribution of progenitor-derived SMCs cells to neointima formation. These differences can potentially be explained by the different lineage tracking systems used and, therefore, it is feasible that each group was tracking slightly different cell populations. Further, the results suggest that progenitor contribution to the neointima is partially dependent on the severity of vascular injury. Additional considerations regarding the disparate findings in these studies include the limitations of Cre/Lox mediated fate-mapping systems. The Cre-driving gene (e.g. Gli1, Sca1) may exhibit non-uniform levels of expression in different cell types. Therefore cells with the lowest expression of Cre in this heterogeneous population may not be sufficiently labelled with the reporter leading to underestimation of the cell population being studied.⁶⁰ Our recent findings using a Gli1-Cre^ERT-YFP reporter system demonstrated that AdvSca1-SM cells contribute predominantly to adventitial remodelling and fibrosis after acute carotid ligation injury.³⁰ While some AdvSca1-SM cells migrated into the vascular media and adopted a mature SMC phenotype, very few contributed to neointima formation. These findings are entirely consistent with our previous work using a Myh11-Cre^ERT lineage-mapping system to label mature SMCs, which definitively showed that mature SMCs are the dominant cell type contributing to neointima formation.^58,61 As AdvSca1-SM cells most likely represent the AdvSca1 cells reported in Kramann’s study, contribution to neointima formation is quite likely contextual and due to the method and extent of vascular injury.

Collectively, these studies offer early insights into the involvement of AdvSca1 cells in neointimal formation and thus provide clues to how AdvSca1 cells fit into the outside-in hypothesis and potentially regulate the tunica media and intima remodelling. While AdvSca1 cells demonstrate differentiation into mesenchymal remodelling cells, such as SMCs and myofibroblasts, other subpopulations of Sca1+ adventitial cells demonstrate endothelial and inflammatory tendencies, which speak to the diversity of the entire AdvSca1 population (Figure 2).

1.4 The role of progenitor cells in experimental atherosclerosis

Atherosclerosis is a complex inflammatory condition and the major driver of cardiovascular disease including clinically acute events, such as myocardial infarction or stroke.^62,63 Over 610 000 people in the USA die every year from cardiovascular-related illnesses and 735 000 people suffer from myocardial infarctions every year.⁶⁴ Clinical management of atherosclerosis targets lowering cholesterol and lipid levels with statins and other lipid-lowering therapies combined with primary prevention, such as maintaining a healthy diet and smoking cessation.⁶⁵ Atherosclerosis is also an inflammatory disease and many people remain resistant to intensive lipid lowering due to increased inflammatory risk. Consistent with this, the IL-1β inhibitor, canakinumab, reduced myocardial infarctions, strokes, and death in patients with atherosclerosis in a landmark clinical trial with over 10 000 subjects, underlying the importance of inflammation in atherosclerosis progression.^66–68 Adventitial inflammation is linked to atherosclerosis progression and plaque severity;⁶⁹ therefore, the cells residing in this layer may show promise as a novel target to treat atherosclerosis since statins fail to fully resolve cardiovascular disease risk and anti-inflammatory treatments carry increased risk of infection and sepsis.^70,71 Because adventitial remodelling is associated with plaque advancement, understanding the behaviour of AdvSca1 cells in the setting of atherosclerosis is pertinent.

1.4.1 Adventitial progenitors and vasa vasorum

The vasa vasorum is a network of microvessels found in the adventitia that function to provide oxygen and nutrients for the cells of the vascular wall.⁷² Advanced atherosclerosis is dependent on vasa vasorum expansion and infiltration into the developing plaque. Independent groups showed that antiangiogenic peptides r-PAI1 and angiostatin decreased vasa vasorum density and plaque size in atherosclerotic mice,^73,74 and AdvSca1 cells may act as a local cellular source for this expansion.⁷⁵ In human studies, there is evidence to support a population of CD34+/CD90+, progenitor-like populations that associate with vasa vasorum and clinical specimens from patients with diffuse coronary artery disease who underwent coronary endarterectomy demonstrated enhanced vasa vasorum associated with lymphocytic infiltrates of CD79+ B and CD3+ T cells.^76,77 Mice do not exhibit vasa vasorum expansion as robustly compared to humans; long-term timepoints are required to observe neovascularization in atherogenic mouse models. Nonetheless a collection of studies substantiates the correlation between the extent of the vasa vasorum with the severity of the atherosclerotic lesion in hypercholesterolemic mouse models.^{73,74,78–81} An important study by Toledo-Flores et al.⁴¹ provides evidence supporting a subpopulation of AdvSca1 cells that participate in vasa vasorum expansion. A population of adventitial Sca1+/CD45+ progenitor cells was shown to differentiate into ECs and demonstrate vasculogenic properties in vitro and in an atherogenic mouse model. CD31 and Tie-2, strong markers of ECs, were not expressed at baseline by Sca1+/CD45 progenitor cells, but after ApoE-null mice were fed high-fat diet (HFD) for 24 weeks, the progenitor cells began to express these strong endothelial markers and displayed vasculogenic gene expression. Cultured CD45+/Sca1+ cells in defined media formed CD31+ EC colonies and Matrigel assays demonstrated CD45+/Sca1+ cells forming interconnected vascular cord-like structures consisting of ECs and occasional macrophages. CD45+/Sca1+ cells from wild-type mice injected into the adventitia of carotid arteries of ApoE-null mice on atherogenic diet produced donor-derived functional adventitial vasa vasorum.⁴¹ Therefore, the existing data support the involvement of at least one subpopulation of AdvSca1 cells with vasa vasorum expansion and plaque progression.

1.4.2 Adventitial progenitors and atherosclerotic plaque progression

Phenotypically modulated SMCs give rise to the majority of the cells that form the fibrous cap in atherosclerosis.^56,82,83 Rainbow lineage-mapping murine models demonstrated that clonally expanded populations SMCs migrate and form the fibrous cap and subsequently invade into the plaque core.⁸⁴ However, the current data are limited to support AdvSca1-derived cell contribution to plaque cells. One study by Kramann et al.²⁷ tracked AdvSca1 cells using the Gli1-Cre^ERT-mTomato lineage system in ApoE-null mice that also underwent 5/6 nephrectomy and were fed a HFD for 10–16 weeks to develop severe atherosclerosis. Gli1+ cells were not observed in the plaque after 10 weeks, but the aortic lesions at the 16-weeks timepoint revealed a small number of Gli1 (mTom+) cells in intima atheromas. Interestingly, the few Gli1+ cells detected in the plaque were invariably adjacent to CD68+ macrophages. Kramann did not assess atherosclerotic burden in mice with ablated Gli1+ cells, precluding a full assessment of the significance of AdvSca1 progenitors to atherosclerosis progression and lesion development.

1.4.3 Adventitial progenitor gene expression in atherosclerosis

Given AdvSca1 cells can differentiate into specific cell types depending on the experimental context, defining their changes in gene expression in the setting of atherosclerosis can help elucidate their behaviours in this disease. Single-cell RNA-Seq approaches have been used to identify leucocyte populations in atherosclerotic lesions including heterogeneic populations of macrophages with stem-like qualities, B-cells, T-cells, and natural killer cells,^85–88 however there are limited examples of single-cell RNAseq studies that delineate progenitor cell populations. A notable study by Kokkinopoulos et al.²⁵ applied single-cell RNAseq in mouse atherosclerotic lesions and identified adventitial Sca1+ populations that display altered signalling pathways related to cell migration and matrix protein degradation. AdvSca1 cells isolated from ApoE-null mice exhibited enhanced migration behaviour in transwell assays as compared to control AdvSca1 cells from wild-type mice.²⁵ Because Sca1 was the sole marker used to define progenitor cells in this study, it is likely that they captured multiple populations of these progenitor cells. The scRNAseq study of adventitial cells performed by Gu et al.⁴² (described above) clarified four mesenchymal subpopulations (Mesen I-IV) implicated in atherosclerosis. Limitations of this study were the lower proportion of cells from WT mice compared to ApoE-null mice and the small numbers of adventitial EC and SMC that were captured and sequenced due to the low abundance of vasa vasorum in mouse aorta. Ligand–receptor analysis revealed proinflammatory interactions between Mesen II Sca1+ cells and immune cells that were increased in the adventitia of ApoE-null mice compared to wild-type mice. These data underscored the dynamic changes of multiple cell types within the outer adventitia that can occur during the early stages of atherosclerosis.

The field’s understanding of AdvSca1 behaviour in atherosclerosis is limited, but existing evidence suggests these cells become activated and support vasa vasorum expansion and potentially migrate into the plaque and contribute to the plaque microenvironment. More data are needed to clarify the contributions of AdvSca1 cells in atherosclerosis as it is still not clear whether AdvSca1 cells contribute to the fibrous cap or if they participate in inflammatory signalling found in the plaque core. In summary of the experimental data discussed above, activated AdvSca1 cells can differentiate into various effector cell types and contribute to vascular fibrosis, inflammation, neovascularization, and medial wall repair (Figure 2).

1.5 Adventitial progenitor cells in human vessels

1.5.1 Adventitial progenitors in human aorta and small blood vessels

Given the evidence for AdvSca1 cells in animal models that allow lineage tracing, it is essential to identify vascular progenitor cells in human blood vessels, determine whether the human vascular progenitors share or diverge from murine counterparts, and understand their relevance to human vascular biology and disease. Currently, our understanding of adventitial vascular progenitors in human vessels is limited and challenged by the lack of consensus markers to isolate and define progenitor populations. Sca1 (Ly6a) is an important marker for progenitor cells in murine vessels, but the Sca1 gene was evolutionarily deleted between mouse and rat speciation and thus is not present in humans.³³ The CD34 receptor is highly expressed on a population of progenitor cells from both mouse and human vessels that reside in the adventitia along the media-adventitia border and in a niche around adventitia blood vessels. CD34 is co-expressed with Sca1 in murine AdvSca1 progenitor cells. However, CD34 is not specific for resident adventitial progenitors as it is expressed on mature EC and haematopoietic stem/progenitor cells. Thus, human adventitial CD34+ progenitors are often distinguished by the absence of mature EC markers, including CD31, vWF, and VE-cadherin.

In general, the differentiation and lineage capacities of human adventitial progenitors have been demonstrated using robust flow cytometry panels and under various ex vivo conditions that promote differentiation towards SMCs, pericytes, ECs, adipocytes, or chondrocyte lineages.^89,90 It is currently unclear whether human vascular progenitors uniformly have multi-lineage capacities in vivo or, like the mouse, consist of heterogeneous progenitor subpopulations that are predisposed for specific lineages and phenotypes.

Early studies showed the foetal aorta contains CD34+/CD31− cells in the outer adventitia.⁹¹ In an aortic explant assay, the foetal aorta radiated capillary-like structures comprised of CD34+/CD31+ and vWF+ cells. Primary CD34+/CD31− progenitors were embedded in Matrigel and developed tube-like structures in Matrigel after 7–10 days. In control assays, mature CD31+ cells formed tubes within a day, suggesting the CD34+/CD31− cells required differentiation to obtain the mature EC phenotype and other cells of the capillary-like structure. More recently, Zengin et al.⁹² showed human internal thoracic arteries (ITAs) contain CD34+ cells that reside along the media/adventitia border and do not express CD31 and endoglin (CD105). When adenovirus expressing GFP was applied to the outer zone of ITA rings where CD34+ cells typically reside, sprouting GFP+ cells were observed in the capillary structures and acquired mature endothelial markers. Another study showed human ITAs contain adventitial CD34+cKit+ cells that have high Ki67 staining denoting a proliferative phenotype and acquire endothelial markers under VEGF stimulation.⁹³ Finally, injections of human foetal aorta-derived adventitial CD34+CD133+ cells into ischaemic hindlimbs of SCID mice improved limb perfusion and salvage compared to mature EC or progenitor conditioned media controls.⁹⁴

Some of the above studies are limited because the CD34+ adventitia cells were not highly purified or labelled to trace daughter cells. Hence, the proportions of mature ECs, SMCs, or other cell types that are derived from adventitial CD34+ progenitors cannot be estimated. Many early studies differentiated CD34+ adventitial cells under angiogenic conditions; however, adventitial CD34+ cells can also differentiate to SMC, adipocytes, pericytes, fibroblasts, and chondrocytes under different in vitro conditions.^51,90 In addition to aorta and ITA, CD34+ adventitial cells have been studied in human pulmonary arteries,⁹⁵ saphenous veins,^96,97 and renal arteries,⁹⁸ but not in human coronary arteries. However, our preliminary data support the existence AdvSca1-SM-like cells in human coronary arteries (our unpublished data). A study of cells surrounding vasa vasorum in human adult thoracic aorta also identified a population of ‘supra-vasa’ cells that express CD34+ and CD90+, but lack CD146 and αSMA (Table 2).⁹⁹

1.5.2 Heterogeneous adventitial progenitors

An important study by Michelis et al.⁷⁷ identified a small subpopulation of adventitial cells in human ITA and aorta that met the classical criterion for MSCs, including functional colony formation, tube-formation, and differentiation into adipocytes, osteoblasts, and chondrocytes. This population of adventitial cells expressed CD90 (Thy1) and co-expressed other MSC markers, including CD44, PDGFR, CD73, CD105; however, CD44 and PDGFRα also labelled αSMA+ medial cells and other cell types. The CD90+ cells also expressed CD34, as previously reported. Injections of the CD90+ cells into ischaemic limbs of SCID mice promoted angiogenesis and limb reperfusion. RNA-sequence analysis of CD90+ cells purified from surgical samples of healthy aorta, diseased aorta (aneurysm), and ITA identified over 900 genes that were differentially expressed in diseased compared to healthy aorta and enriched in multiple gene ontology (GO) terms that represent wound healing, inflammation, extracellular matrix, vessel formation, and atherosclerosis. Since human ITA contained CD90+ progenitors and is resistant to atherosclerosis, the sequencing data for CD90+ cells of diseased aorta were also compared to CD90+ cells of healthy ITA. The combined results identified 115 genes that shared concordant expression in the diseased vs. healthy aorta and the diseased aorta vs. ITA comparisons.⁷⁷ This gene set may provide important insights for the molecular regulation of CD90 progenitors in vascular disease.

This genetic analysis of a rare CD90+ subpopulation underscores the likelihood that resident vascular progenitors include other subpopulations, similar to the emerging model of murine AdvSca1 vascular progenitors that include multipotent progenitors that express SMC genetic tracers, and progenitors that are restricted to EC and/or haematopoietic lineage capacities.^19,32,38 Current efforts to resolve human progenitors are confounded by inconsistencies in the consensus markers and nomenclature of these cells across multiple studies. scRNA-Seq has a strong potential to resolve the genetic signatures of progenitor clusters or subpopulations residing in the adventitia without selecting cells based on the expression of only a few cell surface markers. For translational significance, the genetic signature of human adventitial progenitors can be aligned with mouse AdvSca1 progenitors expressing lineage marks to identify homologous human-mouse counterparts. These data will likely yield important insights for the developmental origins and functions of human vascular progenitors in healthy and diseased blood vessels.

1.5.3 Functions of adventitial progenitors in development and human disease

The potential roles for adventitial progenitor cells in vascular growth, repair, and in diseases such as vascular inflammation, fibrosis, hypertension, vascular aging, and atherosclerosis have been detailed above in discussions related to murine adventitial progenitor cells, but data from human samples are more limited. Although acute vascular injury mouse studies have found moderate numbers of genetically marked adventitial progenitor cells in the neointima,^27,30 human vessels have distinct features within each vessel wall layer that may differentially impact the number of neointimal cells that are derived from adventitial vascular progenitors in human vascular disease. Human vessels at birth have collections of αSMA+ cells that reside between the endothelium and internal elastin membrane.¹⁰⁰ These cells may be activated by endothelial-dependent shear, inflammatory signals, and/or retained subendothelial lipoproteins and subsequently expand to contribute to the neointima. Thus, the proportion of adventitia-derived progenitor cells in human neointima may be fewer than mouse vessels that lack early non-EC intimal cells.

To invade the intima, adventitial cell cells must transmigrate across the media and migrate through lamellar structures of compacted SMCs, elastin, and matrix molecules, which have physical and biochemical properties that support the barrier function and immune privilege property of the media.³ Fenestrations facilitate transmural solute transport and cell migration; however, the media structure of large animal blood vessels is more extensive compared to the mouse aorta.¹⁰¹ Additional conditions, such as chronic inflammation and the actions of metalloproteinases may be required to disrupt the medial components, which could facilitate adventitial progenitor cell entry to the intima.

An obvious question is how adventitial remodelling impacts the development and progression of intimal atherosclerotic plaques. Interesting observations from optical coherence tomography of human coronary arteries show that the adventitia undergoes dramatic remodelling during the early stages of atherosclerosis, well before significant intimal lesions have developed.¹⁰² Flow wire measurements of coronary endothelial dysfunction temporally correlate with dysfunction of adventitia vasa vasorum. Thus, the adventitia becomes dysfunctional in the early stages of atherosclerosis and the altered permeability and vasomotor regulation may affect the exchange of solutes and immune cell trafficking, which triggers adventitial remodelling in early coronary artery disease.¹⁰² Interestingly, adventitial progenitors are abundant in the niche surrounding adventitial microvessels, which position these cells to sense immune cell signalling, hemodynamic forces, and diffusible solutes in the adventitia.³²

Importantly, human arteries contain an abundant native network of adventitial vasa vasorum compared to the mouse aorta.⁸¹ In addition, adventitial neovascularization often invades the media to vascularize the intima in the setting of plaque progression, which is a source for intraplaque haemorrhage in human atherosclerosis.¹⁰³ Thus, penetrating adventitial-derived vessels may also create conduits for adventitial progenitor cells to enter the intima and contribute to the developing plaque. Alternatively, adventitial progenitor cells adjacent to vasa vasorum may contribute to plaque neovascularization by incorporating into angiogenic neovessels that invade the plaque.

1.5.4 Summary of AdvSca1 cells in humans

The dramatic remodelling of the adventitia and the contributions of human vascular progenitor cells to vascular repair and disease merits further investigation. Future studies of adventitial progenitor cells will resolve heterogeneous subpopulations in human blood vessels, most likely via single-cell sequencing methods. Human tissues have technical limitations because unknown in vivo conditions may have previously influenced adventitial progenitor subpopulations in presumed ‘healthy vessels’. Similarly, human blood vessel samples with vascular disease may include plaques with heterogeneous cell compositions and exposure to clinical risk factors and medications. The analysis of vascular progenitors will require blood vessel samples from many individuals to represent the broader genetic diversity of human cells. Although somatic mutations in mitochondria may track daughter cells derived from human adventitial progenitor cells,¹⁰⁴ it will be important to establish correlations with animal models of atherosclerosis to sort out the contributions of these cells in human disease in order to develop rationales for therapeutic intervention.

2. Conclusion

The field has made notable progress in developing the outside-in hypothesis and demonstrating the tunica adventitia as a major vascular wall structure that is at least partially responsible for supporting physiologic functions and regulating pathological vascular remodelling across all three layers of the arterial wall. Vascular adventitial progenitor cells are consistently shown to be involved in adventitial remodelling and are likely to be key cells involved in integrating environmental cues and orchestrating changes to the vasculature in pathological settings. Many studies demonstrate their bioactive behaviour across various models of disease including vascular fibrosis, acute injury, and atherosclerosis. The most salient challenges and opportunities ahead will be to unravel the complex heterogeneity of adventitial progenitor cells to establish a consistent nomenclature and distinct functional characterization of AdvSca1 subpopulations. Our current markers to label progenitor cells are not ideal as many of these markers are present on other cell populations, such as mature ECs, SMCs, and haematopoietic-derived cells. Modern technologies, such as single-cell sequencing and big data approaches (transcriptomics, metabolomics, and epigenomics) will become important to push the field forward, resolve AdvSca1 subpopulations, and define their contributions to normal vessel homeostasis vs. disease progression. The second major opportunity for the field is to reconcile cell markers in the mouse and human. To justify clinical exploration of AdvSca1 cells, it is important to translate the data obtained from animal models to human progenitor cells, which we and others propose exist and will support the development of novel therapeutics. The most exciting opportunity ahead is to target vascular progenitor cells for therapeutic benefit. Specifically, it is conceivable to manipulate the differentiation tendencies of AdvSca1 progenitors and preferentially drive their differentiation trajectory through signalling pathway intervention, epigenetic remodelling, and altering the microenvironment. For example, on the one hand, AdvSca1-SM cells demonstrate strong myofibroblast differentiation tendencies in disease states and significantly contribute to vascular fibrosis. On the other hand, AdvSca1-SM cells can also differentiate into mature SMCs and may be involved in SMC replenishment and repair of the vessel wall. Developing methods to leverage AdvSca1-SM differentiation towards SMCs rather than pathologic myofibroblasts could provide novel approaches to treat patients with cardiovascular diseases at risk for developing irreversible vascular fibrosis. Adventitial progenitor cells show great potential as a paradigm-shifting target to expand treatment options for cardiovascular disease.

Authors’ contributions

A.J.J., K.S.M., and M.C.M.W.-E. wrote the review manuscript. S.L., K.A.S., A.M.D., M.F.M., R.A.N., and M.W.M. edited the manuscript.

Conflict of interest: none declared.

Funding

The work was supported by grants R01 HL121877 and R01 HL123616 from the National Heart, Lung, and Blood Institute, NIH, Chernowitz Foundation Research Grant, and School of Medicine Transformational Research Consortium for Fibrosis Research and Translation Pilot Award to M.C.M.W.-E., grant 20PRE35200015 from the American Heart Association to A.J.J., grant 18POST34030397 from the American Heart Association to S.L. and grant 1F31HL147393 from the National Heart, Lung, and Blood Institute, NIH to K.A.S.

References