Biology of vascular mural cells

The vasculature comprises a conduit of interconnected tubes of different diameters that are optimised for efficient flow distribution to minimise cardiac strain while ensuring the delivery of the appropriate amount of blood to all organs (Mohrman and Heller, 2018). In addition to appropriate baseline flow patterns, mechanisms must exist to allow tissues to increase blood flow rapidly, for instance in muscles during exercise or in the brain when performing cognitive tasks (Joyner and Casey, 2015; Zhu et al., 2022). Our knowledge of how the vasculature meets these demands is still in its infancy. Blood vessels are composed of two main cell types: endothelial cells (ECs), which line the inner lumen of a given blood vessel, and mural cells, which are located on the abluminal side in close contact with ECs. Mural cells can be broadly divided into vascular smooth muscle cells (SMCs) and pericytes, with different subtypes for both present along the vascular network (Armulik et al., 2011; Muhl et al., 2022; Vanlandewijck et al., 2018). There are striking morphological differences among pericyte subtypes (Grant et al., 2019). Furthermore, SMCs on arteries and veins differ greatly in their gene expression patterns (Vanlandewijck et al., 2018), suggesting functional differences among SMCs. However, the relevance of this and the mechanisms establishing the observed differences have remained mysterious. It is also unknown how mural cells might interact with ECs during vessel contraction and relaxation. Both cell types are located next to each other and share the same basement membrane (in the case of pericytes on capillaries); contracting one type of cell would require the other to follow suit. However, whether this occurs and, if so, through which mechanisms has not been investigated. It is also unclear whether different tubular architectures, such as unicellular versus multicellular tubes, present in capillaries and larger veins and arteries contribute to differences in blood flow regulation exerted by mural cells (Lammert and Axnick, 2012).

Work over recent years has furthermore highlighted the importance of mural cells in controlling the growth of new blood vessels independently of their role in regulating blood vessel constriction (Armulik et al., 2005). In addition to these functions of mural cells in controlling EC biology, ECs also control mural cell biology. In this Review, I discuss the mechanisms through which vascular mural cells and ECs interact and how these interactions ensure proper vascular morphogenesis and blood vessel diameters. I also review the relevance of mural cell biology to human disease.

Mural cells play important roles in the regulation of vascular function. Studies in cultured ECs have suggested that during the angiogenesis phase of vascular development, pericytes modulate endothelial sprouting (Chang et al., 2013) and cell cycle progression (Orlidge and D'Amore, 1987) (Fig. 1A). These features depend on their contractility (Durham et al., 2014). In the mouse retina, pericytes are recruited to the angiogenic front through EC expression of Pdgfb (Gerhardt et al., 2003). Mechanistically, pericytes control Vegfa bioavailability through expression of Vegfr1 (Flt1), which acts as a sink for Vegfa (Eilken et al., 2017). In the absence of pericytes, Vegfa levels are increased, leading to impaired angiogenic sprouting and vascular patterning. Other studies suggest that pericyte-supplied angiopoietin 1 regulates angiogenic sprouting (Uemura et al., 2002) and overexpression of angiopoietin 1 is able to rescue the retinal patterning defects that resulted from pericyte depletion. Angiopoietin signalling has also been implicated in pericyte survival and recruitment (Cai et al., 2008). Further studies showed that deletion of the angiopoietin receptor Tie2 in pericytes enhances tumour angiogenesis and tumour growth (Teichert et al., 2017). Thus, proper angiogenic sprouting relies on the interplay between ECs and pericytes. Pericytes can also regulate vascular patterning by promoting selective vessel regression that occurs after the completion of angiogenesis (Simonavicius et al., 2012).

In addition to their function during angiogenesis, mural cells control individual features of organ-specific vascular beds. For instance, pericytes are instrumental during the establishment and maintenance of the blood–brain barrier (BBB) (Langen et al., 2019; Zhao et al., 2015) (Fig. 1B). The neurovascular unit consists of ECs, pericytes and astrocyte end-feet embedded in a common basement membrane (Iadecola, 2017; Schaeffer and Iadecola, 2021). Pericytes control BBB permeability by regulation of EC transcytosis, polarisation of astrocyte endfeet (Armulik et al., 2010) and prevention of the accumulation of serum proteins in the brain (Bell et al., 2010). They further control the formation of tight junctions between BBB ECs (Daneman et al., 2010). Pericytes can also control EC gene expression patterns (Mae et al., 2021). Together, vascular mural cells act in concert with ECs during the establishment of organ-specific vascular beds.

After completion of vascular growth, vascular mural cells help establish correct blood flow patterns in the brain, also referred to as neurovascular coupling (Kaplan et al., 2020; Zhu et al., 2022) (Fig. 1C). This allows for the appropriate amount of blood flow to the brain depending on neuronal demand, known as functional hyperaemia (Hall et al., 2014; Petzold and Murthy, 2011). SMCs have long been recognised as contractile cells that control vascular tone during this process. The contractile state of a given mural cell population has obvious implications for blood flow regulation (Hall et al., 2014; Hartmann et al., 2021; Hill et al., 2015; Peppiatt et al., 2006; Yemisci et al., 2009). An ongoing debate concerns the contractile state of pericytes (Kaplan et al., 2020), which might be resolved by adopting the same nomenclature used for ensheathing, mesh and thin-strand pericytes (Attwell et al., 2016) (see below). However, there is currently no consensus on such nomenclature in the field, as recent reviews refer to different cell populations as ‘capillary pericytes’ (Attwell et al., 2016; Hartmann et al., 2022; Holm et al., 2018).

During neurovascular coupling, neurons and astrocytes release vasoactive substances that act on vascular mural cells, causing their relaxation or contraction, thereby controlling vascular diameters and blood flow to specific brain regions (Attwell et al., 2010). One important molecule in this process is nitric oxide (NO) (Hoiland et al., 2020; Hosford and Gourine, 2019; Stobart et al., 2013). Further studies have implicated NO signalling in arteriolar dilation, whereas capillary dilation relies on astrocyte-produced arachidonic acid (Mishra et al., 2016). These results highlight differential requirements in terms of SMC or pericyte function in blood vessel diameter control in the brain. It will be important in subsequent research to address the question of how these requirements might interface with reported differences in gene expression patterns and contractile properties between SMCs and pericytes.

In skeletal muscle, vasodilation and vasoconstriction are coordinated among vessel branches (Rowell, 1993). This integration is mainly carried out by the sympathetic nervous system, which ensures an appropriate blood supply to active muscles through local vasodilation while preventing a decline in peripheral vascular resistance and blood pressure by means of vasoconstriction (Just et al., 2016). One molecule implicated in this function is NO released from either muscle or endothelial cells (Thomas and Segal, 2004; Thomas and Victor, 1998). Sympathetic nerve terminals also release noradrenalin, neuropeptide Y and adenosine triphosphate, which act on SMCs and lead to vasoconstriction. Some of the vasoactive molecules and their receptors on SMCs differ between the central nervous system and the peripheral circulation (Koep et al., 2022). This emphasises organ-specific adaptations of vascular mural cells in relation to their specific functions in controlling blood flow. It will be important in the future to elucidate how these activities are integrated at the level of the organism.

Another type of specialised mural cells constitutes pre-capillary sphincters, located at the transition between the penetrating arteriole and the first-order capillary (Fig. 1C). Although initial observations identified pre-capillary sphincters on the vasculature of the mesentery, conflicting results existed for other vascular beds (Altura, 1971; Harris and Longnecker, 1971; Sakai and Hosoyamada, 2013). More recently, pre-capillary sphincters have also been described associated with brain vasculature in mice (Grubb et al., 2020), where they control capillary perfusion (Zambach et al., 2021). Capillary sphincter mural cells display a distinct morphology; however, as with other mural cell types, no specific gene expression patterns have been reported (Grubb et al., 2020). These findings underscore the importance of vascular mural cells for blood flow control, with several distinct mural cell types potentially exerting unique functions during this process.

Mural cells also stabilise blood vessels (Armulik et al., 2005, 2011) (Fig. 1D). This function requires integrin signalling between mural cells and the basement membrane (Abraham et al., 2008). In zebrafish, interactions between ECs and mural cells sustain the dorsal aorta, relying on establishing an appropriate basement membrane (Stratman et al., 2017). Without an appropriate basement membrane, the elasticity of the dorsal aorta is increased, ultimately resulting in increased vessel diameter. Pericytes can also stabilise tumour vasculature, thereby enhancing tumour growth (Furuhashi et al., 2004). Therefore, mural cells play important roles in controlling blood vessel functionality through a diverse array of cellular functions.

One understudied aspect of mural cell biology pertains to the interactions between ECs and mural cells during blood vessel diameter control. It is not clear how mural cells change their shapes during blood vessel dilation or constriction and how these changes might impinge on EC shapes. Previous studies have shown that mural cells align along EC junctions (Ando et al., 2016). However, blood vessels within the capillary vasculature show distinct architectures with some segments consisting of unicellular tubes (Lammert and Axnick, 2012). Whether mural cells align differently along unicellular or multicellular blood vessels and whether this might affect mural cell–EC interactions remains to be determined (Fig. 2). It is also unclear whether ECs play a structural role in blood vessel diameter control apart from releasing vasoactive substances, such as NO. One recent study found that the tubular architecture of brain capillaries correlates with their diameters (Sargent et al., 2023 preprint). Thus, blood vessel diameter control likely relies on an interplay between ECs and mural cells of distinct shapes and sizes. The elucidation of the mechanisms controlling this interplay will be of great interest.

One key signalling pathway influencing mural cell development is platelet-derived growth factor B (Pdgfb) signalling (Andrae et al., 2008), which comprises four ligands (Pdgfa-d) and two receptors (Pdgfrα and Pdgfrβ) (Fig. 3). Analysis of mouse mutants for either the Pdgfb ligand or its receptor Pdgfrβ demonstrated that both are required for mural cell development during embryogenesis (Hellstrom et al., 1999; Leveen et al., 1994; Lindahl et al., 1997; Soriano, 1994). The initial specification of mural cells is not affected in these mutants, but vessel recruitment and proliferation are compromised. Mutant animals exhibit dilated blood vessels, changes in endothelial junctional protein distribution and cell shape, and signs of increased transendothelial permeability (Hellstrom et al., 2001). Similar phenotypes have since been reported for pdgfrb and pdgfb mutants in zebrafish (Ando et al., 2016, 2021b). These studies revealed a lack of pericytes on central arteries in the brain, although mural cells readily invest blood vessels at the cerebral base, albeit at lower numbers (Ando et al., 2021b). In addition, blood vessels are markedly enlarged in later-stage pdgfb signalling mutant embryos. Mural cell development in the fin of juvenile zebrafish also requires Pdgfrβ signalling, as does their recruitment during tissue regeneration (Leonard et al., 2022). Work in fish has further revealed the importance of Cxcr4 signalling for controlling the expression of the Pdgfb ligand in arteries, leading to increased SMC recruitment (Stratman et al., 2020). By contrast, in veins, the shear stress-responsive transcription factor Klf2a represses cxcr4a and pdgfb expression, reducing venous SMC recruitment. Together, these results indicate an evolutionary conservation for Pdgfrβ signalling in mural cell biology between fish and mouse (Ando et al., 2021a). They also implicate Pdgfrβ signalling in the generation of both pericytes and SMCs. However, it remains unclear whether this is achieved by influencing the numbers of an earlier progenitor cell or regulating mural cell differentiation at later developmental stages.

Earlier work suggested a role of transforming growth factor beta (Tgfβ) in regulating Pdgfb ligand expression (Daniel et al., 1987; Makela et al., 1987), presumably through Smad2/3 transcription factors (Taylor and Khachigian, 2000). However, we still lack an understanding of the spatial regulation of Pdgfb expression. For instance, during angiogenic sprouting, Pdgfb transcripts can be mainly found in tip cells at the angiogenic front (Gerhardt et al., 2003). However, Smad2/3 activation is absent from these cells and can instead be found in trailing stalk cells (Aspalter et al., 2015) that do not express Pdgfb. How these findings can be reconciled remains to be determined.

Tgfβ signalling can also directly influence SMC differentiation and homeostasis (Sinha et al., 2014; Wang et al., 2015) (Fig. 3). In embryonic stem cells, Tgfb1 contributes to SMC development (Sinha et al., 2004) and further work has identified deltaEF1 (Zeb1) as a transcription factor mediating Tgfβ signalling during SMC differentiation (Nishimura et al., 2006). Studies in mice have shown that endothelium-specific deletion of Tgfbr1 (also known as Alk5) or Tgfbr2 in ECs results in a failure of SMC differentiation (Carvalho et al., 2004). Exogenously supplied Tgfb1 ligand can rescue this phenotype, suggesting that Tgfβ signalling in ECs is necessary for enhancing the availability of Tgfβ ligands. A follow up study has also established a requirement for Tgfbr2 in SM22 (Tagln)-expressing SMCs (Carvalho et al., 2007; Frutkin et al., 2006). Interestingly, however, Smad2 phosphorylation could still be detected in global Tgfbr1 knockout mutant mice, which permits SMC differentiation. Carvalho and colleagues attributed this finding to an upregulation of the Alk4 (Acvr1b) receptor in the Tgfrb1 mutants. Thus, Tgfbr2 appears to be able to interact with different Tgfbr1 molecules during SMC differentiation.

Notch signalling also influences mural cell biology (O'Hare and Arboleda-Velasquez, 2022) (Fig. 3). In mice, mutations in the Notch ligand jagged 1 cause a failure in SMC development (High et al., 2008), and mutations in Notch3 convert arterial SMCs into those normally found on veins, affecting blood flow regulation in the brain of mutant animals (Domenga et al., 2004). Such animals also show a reduction in mural cell numbers within the retina (Liu et al., 2010). Similar findings have also been reported in zebrafish, where Notch3 positively regulates brain pericyte proliferation (Wang et al., 2014) and mural cell specification (Ando et al., 2019). Mechanistically, Notch3 can lead to the upregulation of Pdgfrβ in SMCs, thereby influencing mural cell biology (Jin et al., 2008). Work from my laboratory has previously shown that, in the zebrafish fin, proximally located acta2-expressing mural cells do not show activation of the Notch pathway (Leonard et al., 2022). However, mural cells in distal locations exhibited upregulated Notch signalling. Therefore, regional differences exist in terms of signalling pathway activation within mural cell populations, potentially contributing to their functional differentiation.

Mural cells can arise from a surprisingly varied pool of progenitor cells depending on the anatomical location and developmental time point examined (Fig. 4). In mice, neural crest-derived cells contribute mural cells to the retina, the choroid plexus (Trost et al., 2013) and the thymus (Foster et al., 2008; Muller et al., 2008). Transplantation studies in chick embryos have revealed a similar capacity of neural crest and other neuroectodermal cells (Korn et al., 2002). The mesothelium, a squamous epithelial cell layer of mesodermal origin gives rise to mural cells in the lung, liver and gut (Armulik et al., 2011; Asahina et al., 2011; Que et al., 2008; Wilm et al., 2005). The major aorta has vascular mural cells from various origins; depending on the location along the aorta, the mural cells investing this vessel have four different developmental origins, including the secondary heart field, neural crest, splanchnic mesoderm and somites (Armulik et al., 2011; Majesky, 2007; Sinha et al., 2014; Sinha and Santoro, 2018). Even within these areas, SMCs of inner and outer regions of the media can be derived from distinct embryonic sources (Sawada et al., 2017).

Despite our insights into the embryonic sources of early mural cell progenitors, much less is known about their differentiation in situ. Recent studies have shown that, in embryonic zebrafish and the regenerating fin vasculature, Col1a2-expressing fibroblasts can give rise to pericytes (Leonard et al., 2022; Rajan et al., 2020). However, it has not been established whether fibroblasts contribute significantly to the pericyte lineage in other developmental or regeneration settings.

Lineage-tracing studies in the heart have shown that epicardial cells can give rise to coronary SMCs (Dettman et al., 1998; Mikawa and Gourdie, 1996; Zhou et al., 2008). Epicardial cells first differentiate into pericytes present on the coronary microvasculature. Subsequently, pericytes can give rise to SMCs populating coronary arteries (Volz et al., 2015). Endocardial cells can also contribute to the SMC lineage of the heart through a pericyte intermediate (Chen et al., 2016). Although single-cell sequencing data (Muhl et al., 2020) support this trajectory, it remains to be determined whether pericytes are a source for SMCs in other organs.

Several studies have examined precursor cells capable of giving rise to endothelial and SMC lineages, such as meso-angioblasts originating from the dorsal aorta (Minasi et al., 2002). Other studies reported the transdifferentiation of ECs into SMCs through endothelial-to-mesenchymal transition (Arciniegas et al., 2000; DeRuiter et al., 1997). Another study showed that Tie1-expressing precursor cells can contribute to the SMC lineage (Chang et al., 2012). Thus, ECs can give rise to mural cells, and, within the mural cell population, pericytes can transdifferentiate into SMCs. It will be of interest to investigate whether these lineage relationships are maintained in settings of tissue regeneration.

The varied mural cell ontogeny might be a prerequisite for obtaining their distinct functions in different organs. As detailed above, mural cells in the brain interact with astrocytes and neurons in establishing the BBB (Langen et al., 2019; Zhao et al., 2015), In the heart, mural cells contribute to the myocardial barrier, where they form a syncytium with cardiac ECs, connected through gap junctions (Avolio and Madeddu, 2016; Juchem et al., 2010; Murray et al., 2017; Nees et al., 2013). The kidney contains two kinds of specialised vascular mural cells: glomerular mesangial cells and podocytes (Schell et al., 2014; Stefanska et al., 2013). These cells work in concert with ECs to constitute the glomerular functional unit. In addition to filtering urine, they also secrete the vasoactive protease renin, which works in blood pressure control (Berg et al., 2013; Shaw et al., 2018; Stefanska et al., 2016). In the liver, hepatic stellate cells are specialised storage compartments for vitamin A metabolites (Hellerbrand, 2013; Kamm and McCommis, 2022). They further interact with liver ECs and regulate sinusoidal blood flow. During liver regeneration, hepatic stellate cells become activated and differentiate into fibrogenic myofibroblasts (Tsuchida and Friedman, 2017). Together, these findings highlight the varied roles of vascular mural cells. They further illustrate the need for a better understanding of the similarities and differences among mural cell populations forming from distinct precursor populations during early embryogenesis stages and how they might impinge on mural cell function.

There is an ongoing debate whether pre-existing pericytes can function as mesenchymal stem cells. (Birbrair et al., 2017; Morikawa et al., 2019). Work in several organs in mice has shown limited potential for pericytes and SMCs to contribute to other cell populations during aging or pathological challenges (Guimaraes-Camboa et al., 2017). Similar findings have been obtained in the regenerating zebrafish fin, where pre-existing vascular mural cells do not contribute to newly forming mural cells or other tissues (Leonard et al., 2022). Other studies, however, have shown that pericytes readily contribute to axonal scar tissue, both in mice and in zebrafish (Dias et al., 2018; Goritz et al., 2011; Tsata et al., 2021). It will be important in the future to determine the cause for the observed differences among organs and experimental systems. Together, depending on their origin and later tissue localisation, distinct signalling pathways control the differentiation of mural cells. These signalling pathways might be reactivated during tissue regeneration, stimulating the differentiation of new mural cells from mesenchymal stem cell precursors. Furthermore, a mural cell's developmental trajectory might be a prerequisite for its separate function within a target organ, for example in maintaining the BBB.

Work over recent years has provided fundamental new insights into the phenotypic diversity of mural cell populations (Hartmann et al., 2015) (Fig. 5). Detailed analysis of mural cell morphologies in the mouse brain has revealed distinct morphologies of SMCs and pericytes depending on the branch order of the vasculature (Dore-Duffy and Cleary, 2011; Grant et al., 2019; Hartmann et al., 2015). These studies coined the terms ‘ensheathing pericyte’ (also called ‘smooth muscle-pericyte hybrid’; Hartmann et al., 2015), ‘mesh pericyte’ and ‘thin-strand pericyte’ (Fig. 2A). Except for differences in alpha smooth muscle actin (α-SMA) expression, which can be detected in ensheathing but not in mesh and thin strand pericytes, no differences in expression patterns have been reported (Pfeiffer et al., 2021). Notably, a recent publication has shown that the detection of α-SMA expression depends on the tissue fixation method (Alarcon-Martinez et al., 2018). Thus, all pericytes (as well as SMCs) express α-SMA, albeit at different levels. This debate highlights the need for unique molecular markers that would allow identification of differently shaped types of vascular mural cells based on gene expression patterns.

Single-cell sequencing studies have provided a detailed characterisation of differences in gene expression among organ-specific pericyte and SMC populations (Muhl et al., 2022; Vanlandewijck et al., 2018). These studies uncovered that mural cells exist in two distinct subclasses. In one subclass, pericytes exist in a continuum with venous SMCs that is characterised by progressive loss of pericyte markers and acquisition of venous SMC markers. In the other subclass, arteriolar SMCs are in a continuum with artery SMCs, with a gradual acquisition of artery SMC markers. Interestingly, the transcriptional similarity of mural cells within these subclasses (pericyte, venous SMC, arteriolar SMC and arterial SMC) does not follow the distribution of mural cells along the arteriovenous axis (arterial SMC, arteriolar SMC, pericyte, venous SMC). However, how these differences in gene expression patterns might instruct discrete mural cell shapes remains elusive.

So far, no unique molecular markers have been identified to allow different pericyte subtypes to be distinguished (Table 1). Early studies examining differences in gene expression patterns between mural cell populations detected a gradual decrease in α-SMA expression in pericytes along the capillary vascular network, as discussed above (Nehls and Drenckhahn, 1991). Two other markers frequently used to identify pericytes are the proteoglycan neuron-glial antigen 2 (NG2; Cspg4) (Ozerdem et al., 2001) and desmin (Nehls et al., 1992). However, both are also readily expressed by SMCs. Further studies have used cDNA microarrays to identify differentially expressed genes in Pdgfb or Pdgfrb knockout mice, which almost completely lack pericytes (Bondjers et al., 2006, 2003). These studies revealed mural cell-specific expression of regulator of G protein signaling 5 (Rgs5), which has also been reported in a separate study (Cho et al., 2003), in addition to potassium inwardly rectifying channel subfamily J member 8 (Kcnj8), ATP-binding cassette subfamily C member 9 (Abcc9) and delta homologue 1 (Dlk1). Mural cell expression of Kcnj8 and Abcc9 differs between organs. In the brain, both are expressed in pericytes but are absent from arteriolar SMCs (Vanlandewijck et al., 2018; Zeisel et al., 2018). In other organs, however, arteriolar SMCs also express Kcnj8 and Abcc9 (Aziz et al., 2014; Cui et al., 2002; Foster and Coetzee, 2016).

Subsequent studies have used double-transgenic animals expressing NG2-DsRed (Zhu et al., 2008) and Pdgfrβ-eGFP to fluorescently label pericytes, identify pericyte-specific genes and follow the expression of these genes over time (He et al., 2016; Jung et al., 2018). However, the markers identified through this approach are also expressed in some SMCs. Work in zebrafish has recently identified a conserved gene signature shared between pericytes present in adult mouse or embryonic zebrafish (Shih et al., 2021). The authors have further identified ndufa4l2a as a marker driving pericyte-specific transgene expression in zebrafish. Ndufa4l2a is also highly expressed in pericytes in mice and humans (Mesa-Ciller et al., 2023) and the authors of this study detected Ndufa4l2a expression in SMCs, with decreasing expression along the arteriole-arterial SMC axis. Together, these discoveries highlight the difficulty of assigning specific molecular markers to either the SMC or the pericyte population, let alone the ability to distinguish between individual pericyte populations. Therefore, it remains to be established whether the vastly different cellular shapes of distinct pericyte and SMC populations associate with unique gene expression profiles or whether they merely correlate with differences in gene expression levels, as has been suggested for α-SMA.

Regional differences in mural cell phenotypes might also contribute to the aetiology of diseases involving mural cells. In atherosclerosis, SMCs derived from distinct embryonic sources might contribute differently to atherosclerotic plaques (Haimovici and Maier, 1964, 1971; Majesky, 2007; Sinha and Santoro, 2018). SMC contribution differs depending on the stage of atherosclerosis. It is generally correlated with a phenotypic SMC switch from contractile to synthetic (Zhang et al., 2021). In the initiation stage, diffuse intimal thickenings in arteries contain SMCs that have more synthetic organelles. This is followed by lipid deposition and SMC-mediated changes in the extracellular matrix (Basatemur et al., 2019). SMCs in these lesions can also undergo apoptosis, resulting in microcalcifications (Clarke et al., 2008). Subsequently, macrophages populate these lesions. Gene expression studies have suggested that a subset of these macrophages is SMC derived (Andreeva et al., 1997). These macrophages fuel atherosclerosis progression through lipid ingestion, thereby becoming foam cells, which ultimately induce the formation of a fibrous cap that is rich in SMC-derived cells (Bentzon et al., 2007, 2006; Yu et al., 2011). Together, these studies underscore the important contribution of SMCs to all stages of atherosclerosis development (Grootaert and Bennett, 2021). Future studies will aim to understand the molecular pathways involved in SMC phenotypic switching and how to manipulate SMC phenotypes in a beneficial manner.

Stroke causes an ischemic insult in a brain region, most often as the result of an occluded blood vessel (Campbell et al., 2019). This leads to changes in ATP availability and ion imbalances, ultimately causing death of neurons and accessory brain cells (Hossmann, 2006; Lipton, 1999). Importantly, the ischemic core region of a stroke is surrounded by a hypo-perfused penumbra that can be salvaged through rapid reperfusion (Astrup et al., 1981). Current therapies, therefore, aim to restore flow to the penumbra region (National Institute of Neurological and Stroke rt-PA Stroke Study Group, 1995). Recent studies have revealed that pericytes play a major role in controlling blood flow patterns in infarcted brain regions (Hall et al., 2014). In ischemic settings, pericytes constrict brain capillaries and die in rigor, exacerbating blood flow perturbations. Ablating pericytes from the adult brain using diphtheria toxin causes similar neuronal death as a result of tissue swelling caused by BBB breakdown (Nikolakopoulou et al., 2019). Together, these studies show that pericytes are important players in determining the severity of stroke outcomes and therefore potential targets for therapeutic interventions.

Mutations in NOTCH3 cause CADASIL (cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy), which most strongly affects SMCs within the central nervous system (Ferrante et al., 2019; Joutel, 2011; Joutel et al., 1996; Mizuno et al., 2020). CADASIL is characterised by loss of SMCs, a thickening of the arterial wall and the deposition of granular osmiophilic material. At the molecular level, mutations in NOTCH3 cause an abnormal vascular accumulation of the extracellular domain of NOTCH3. Of note, disease onset occurs only at around 30 years of age, with ischemic strokes manifesting at around 50 years old (Joutel, 2011). The reasons for this late disease onset are still not clear. Studies involving CADASIL mouse models have shown that mutant Notch3 can either retain signalling function or exhibit increased signalling (Baron-Menguy et al., 2017; Monet et al., 2007). Thus, how disease manifestation relates to Notch3 function or to the accumulation of granular osmiophilic material over time remains to be established. A possible mechanism involves changes in extracellular matrix composition because Notch3 mutants present with an accumulation of Timp3 (tissue inhibitor of metalloproteinases-3) (Kast et al., 2014; Monet-Lepretre et al., 2013) and Timp3 reduction can ameliorate disease burden in mouse models (Capone et al., 2016). Interestingly, changes in Timp3 have also recently been reported in mice lacking Tgfbr1 in pericytes. In this setting, Timp3 is downregulated, resulting in reduced pericyte coverage and blood vessel dilation, ultimately leading to germinal matrix haemorrhage-intraventricular haemorrhage (Dave et al., 2018). Thus, changes in extracellular matrix composition appear to be an important driver of vascular pathologies involving mural cells, as is also proposed for BBB maintenance (see above).

Work investigating spontaneous deep intracerebral haemorrhage (ICH) has revealed that an alteration of the extracellular matrix can directly influence mural cell numbers and contractile properties (Schlunk and Greenberg, 2015). Further work has revealed SMC degeneration on larger brain arteries in a collagen mouse model of ICH, causing vessel rupture and bleeding (Ratelade et al., 2018). Similarly degenerated SMCs have been found in humans with ICH (Masawa et al., 1994; Sutherland and Auer, 2006; Takebayashi and Kaneko, 1983). Of note, the effect on mural cells differs between larger diameter arterioles and transitional mural cells situated on pre-capillary blood vessels. On the latter vessel segments in collagen mutants, mural cells increase in number and contractility (Ratelade et al., 2020). This, in turn, increases blood pressure in arterioles with sparse SMC coverage, collectively precipitating ICH. The authors also show that upregulation of Notch3 signalling causes the observed hypercontractility in pre-capillary mural cells. Together, these findings reveal complex interactions between the extracellular matrix and mural cells that can differentially affect mural cell properties, such as their contractility and proliferation. Further studies are necessary to clarify the contribution of these effects on human pathology.

We are just beginning to understand mural cell biology and how it relates to vascular function and disease. Although initial advances towards identifying specific mural cell morphologies and gene expression patterns have been made, we clearly lack an understanding of the transcriptional programmes that might differentiate mural cells present on distinct segments of the vasculature and displaying vastly different morphologies. We also do not understand how these morphologies impinge on mural cell function and how disease-causing mutations might interfere with these properties. More work is also needed to reveal the complex interactions between ECs and mural cells, both during vascular morphogenesis and later in regulation of vascular physiology. Ultimately, these insights will help to develop strategies for promoting vascular regeneration, both within the body and for improving tissue engineering approaches aiming at rebuilding human organs.

I thank Christine Mau, Zeenat Diwan and Jia Kang for critical reading of the manuscript.