The brain is one of the most energy-demanding organs in the body. Despite only accounting for 2% of the body’s mass, over 20% of cardiac output is distributed to the brain to meet oxygen and glucose requirements at rest (Xing et al., 2017). To deliver blood flow to neuronal and glial cells, the mammalian brain has evolved the neurovascular unit (NVU), a complex unit of cells that connects the brain parenchyma to the cerebral vasculature. The NVU consists of brain parenchymal cells including excitatory neurons, inhibitory interneurons, astrocytes, and microglia interacting with vascular cells including pericytes (on capillaries), VSMC (on arterioles and arteries), and endothelial cells (EC) (McConnell et al., 2017; Figure 2). The NVU plays important roles in maintaining brain function, particularly the regulation of cerebral blood flow (CBF) and the formation of the blood-brain barrier (BBB). Pericytes are central to NVU function as they are located at the interface between the brain parenchyma and the blood vessels, and so can act as chemical sensors to enable communication between the two groupings of cells.

Endothelial cells and pericytes communicate by juxtacrine and paracrine signaling through mediators such as TGF-β, angiopoietin 1/2, vascular endothelial growth factor (VEGF), and PDGF-BB (Sweeney et al., 2016). These factors are important in directing vessel development and maturation and formation of the BBB, as well as pericyte survival and contractility (Armulik et al., 2011). The importance of pericyte-EC interactions can be seen in pericyte depletion animal models, such as platelet-derived growth factor receptor beta (PDGFR-β) knockdown mice, that display abnormal vascular growth and extensive EC apoptosis (Hellstrom et al., 2001). Pericytes communicate with each other, allowing the propagation of signals up or down the capillary, which can assist in the control of blood flow through the capillary network (Peppiatt et al., 2006). Astrocytes and pericytes can interact to alter BBB permeability through astrocyte-secreted apolipoprotein binding to the low density lipoprotein receptor-related protein 1 receptor (LRP-1) on pericytes (Ma et al., 2018), and modulate CBF through astrocyte secretion of prostaglandin E₂ binding to pericyte EP4 receptors (Mishra et al., 2016). Neurons may interact with pericytes through the release of neurotransmitters to induce pericyte relaxation (glutamate) or contraction (noradrenaline) thus altering vascular tone (Peppiatt et al., 2006; Hall et al., 2014). It is also possible that pericytes may influence neuronal function through the release of neurotrophic factors (Shimizu et al., 2012). Overall, the interplay between pericytes and other cell types of the NVU is complex but enables precise control over CBF and BBB permeability.

Ischemic stroke involves the disruption of blood supply to the brain and can cause extensive neuronal death. While reopening of the blocked artery has been the predominant focus of ischemic stroke therapy, there has been limited attention given to the restoration of capillary blood flow. Multiple studies have shown that following a stroke, pericytes constrict capillaries and die in rigor causing the capillary to be clamped shut leading to a long-lasting restriction of blood flow (no-reflow). These changes to pericytes post-stroke appear to be driven by an influx of intracellular calcium and oxidative stress (Yemisci et al., 2009; Hall et al., 2014). In addition, ischemia leads to the release of MMP9 by pericytes, which interrupts the tight junctions between EC and the binding of astrocyte endfeet to the vascular wall, increasing BBB permeability (Underly et al., 2017). Therefore, pericytes have been identified as an attractive therapeutic target to prevent ongoing microvascular-mediated restriction of CBF and opening of the BBB following stroke.

Alzheimer’s disease (AD), a prevalent neurodegenerative disorder, is characterized by the accumulation of amyloid beta (Aβ) to form plaques and hyperphosphorylated tau to form neurofibrillary tangles. However, it is becoming increasingly recognized that vascular dysfunction through restriction of CBF and impairment in the BBB could contribute to and even precede AD onset and progression. It is hypothesized that Aβ may cause pericyte death, leading to reduced CBF and loss of BBB integrity, which may in turn accelerate neuronal degeneration and Aβ plaque build-up (Winkler et al., 2014). In vitro studies have confirmed Aβ can cause pericyte toxicity (Wilhelmus et al., 2007) and both rodent and human studies have shown that pericyte loss occurs in AD (Sagare et al., 2013; Sengillo et al., 2013). Pericytes can endocytose perivascular debris, giving them a vital role in the clearance of molecules such as Aβ (Ma et al., 2018) and may have protective effects in AD. A recent publication demonstrates that soluble PDGFR-β is a cerebrospinal fluid biomarker of pericyte injury during early cognitive impairment and correlates with BBB disruption independent of Aβ and tau (Nation et al., 2019). Therefore, identifying ways to prevent pericyte loss in AD is important given that cerebrovascular changes such as BBB dysfunction and reduced cerebrovascular density can occur early in the disease. A recent review that focusses on BBB physiology more comprehensively details pericyte changes in AD (Sweeney et al., 2019).

Tumor development relies on fast and efficient vascularisation to match growth, but the abnormal branching, disorganized networks, uneven basement membranes, and lack of pericytes favors non-productive angiogenesis and metastasis (Barlow et al., 2012). Therefore, delivery of pericyte therapy to the tumor can allow vascular normalization, preventing metastasis and improving the efficacy of chemo- or radio-therapy (Meng et al., 2015). However, many tumor types rely on angiogenesis for growth, for which pericytes are critical, and so the inhibition of pericyte-induced angiogenesis may also be a viable therapeutic target to prevent tumor growth. This has been shown with imatinib, which blocks the phosphorylation of PDGFR-β in pericytes, preventing pericyte survival and proliferation, thus reducing tumor growth (Ruan et al., 2013). Therefore, pericytes appear to be a double-edged sword in cancer and any future therapy options will require detailed understanding of their biology and function in each individual cancer environment (Meng et al., 2015).

Given their heterogeneity in structure and function, an important future direction of pericyte research is the classification of pericyte sub-classes. Sequencing technology has advanced rapidly with researchers now examining the transcriptome, proteome, metabolome, and secretome of cells. A recent study used single-cell RNA sequencing to transcriptionally profile the principal cell types of the brain vasculature (Vanlandewijck et al., 2018). Results showed gradual arteriovenous transcriptional zonation in EC, while no such zonation or subtyping was seen within pericytes. Transmembrane transporter activity was associated with genes that were overexpressed in brain pericytes when compared with lung pericytes, providing evidence for organotypic specialization of pericytes, and suggesting that brain pericytes are directly involved in molecular transport at the BBB. Though this technique has the potential to identify pericyte subtypes, results of this particular study did not find this same zonation profile within brain pericytes. This hints that signaling from nearby cells may give rise to the different pericyte morphology that is observed along the arteriovenous axis.

While pericytes were initially identified based on their distinct cellular morphology, other techniques are available to identify pericytes in vitro and ex vivo using immunochemistry. Several antigen markers exist for pericytes including PDGFR-β, neural/glial antigen 2 (NG2), α-SMA, desmin, and aminopeptidase N (CD13), but all of these are also expressed by other cells, such as oligodendrocyte precursor cells and VSMC (Armulik et al., 2011). To confirm pericyte identity through their juxtaposition with capillaries, pericyte markers can be co-localized with vascular markers such as fluorescently conjugated lectins (e.g., FITC-tomato lectin) or EC markers (e.g., CD31) (Robertson et al., 2015).

Pericytes possess the receptors and intracellular signaling machinery to respond to neurotransmitters and vasoactive mediators including noradrenaline, angiotensin-II, endothelin-1, adenosine, and EC-derived nitric oxide (Hamilton et al., 2010). In vitro pericyte preparations can be used to examine the molecular biology and signaling pathways controlling pericyte function. Primary pericyte cell culture using cells derived from both human and rodent tissue has shown that pericytes can contract and relax in response to vasoactive mediators (Neuhaus et al., 2017), as well as secrete cytokines relevant to neuroinflammatory responses (Rustenhoven et al., 2017). To study how pericytes interact with other neurovascular cells, specific in vitro models of the neurovasculature are being developed. When human brain-derived pericytes and astrocytes were cultured with human induced pluripotent stem cell-derived EC, the cultures formed vascular networks in a fibrin gel (Campisi et al., 2018). These co-cultures can be used to model the BBB and examine interactions between cells as well as transportation and diffusion dynamics of specific molecules. More recently, it was shown that brain-specific mesoderm and neural crest pericyte-like cells could be derived from human induced pluripotent stem cells (Faal et al., 2019; Stebbins et al., 2019). This introduces the possibility of culturing patient-specific pericytes, which would be a valuable tool to study how genetic disease states affect pericyte function. In combination, these techniques could enable the generation of human patient-specific isogenic 3D microfluidic systems, allowing the modeling of human neurovascular function in specific disease states to help develop therapeutics with high specificity (Stebbins et al., 2019). Another approach using ex vivo brain slices with the cerebrovascular architecture intact can be used to reveal molecular pathways that control pericyte functions on capillaries (Hall et al., 2014; Mishra et al., 2016).

Transgenic mice are available with fluorescently labeled pericytes (e.g., NG2-DsRed, PDGFR-β-tdTomato), and are commonly used for imaging of pericytes in vivo using two-photon microscopy (Hall et al., 2014; Hartmann et al., 2015). Pericyte location on capillaries can be visualized using vascular tracers such as FITC-dextran, but a disadvantage of both NG2- and PDGFR-β reporters is that they label all mural cells which includes both pericytes and VSMC (Jung et al., 2017). Fluorescent Nissl dye Neurotrace 500/525 has been shown to preferentially bind to pericytes in vivo allowing the imaging of pericytes without the need for transgenic animals (Damisah et al., 2017). Mouse models of pericyte depletion have been developed (e.g., PDGFRβ^+/–) which can lead to impaired BBB functionality (Armulik et al., 2010). In addition, taking advantage of the Cre-Lox systems, genes can specifically be knocked-in or -out within pericytes to alter their function, for example LRP-1 (Ma et al., 2018), but caution must be noted that other cell types depending on the Cre-reporter line will also be affected (e.g., VSMC and oligodendrocyte precursor cells). Optogenetic techniques, in conjunction with 2-photon imaging, have been shown to stimulate pericytes to generate a state of vessel constriction, eliciting pericyte-driven capillary diameter changes (Kisler et al., 2017a). However, this has not been consistent in all studies (Hill et al., 2015), possibly due to differences in Cre drivers used or the model and/or light source used.

In Hall et al. (2014), an important paper was published stating that pericytes can regulate CBF. Controversially, the next year a contradictory paper was published stating that VSMC, not pericytes regulated CBF (Hill et al., 2015). This disagreement appears to have arisen due to confusion about the nomenclature of pericytes (Attwell et al., 2016) and has revealed the need for a universal pericyte classification system. Several researchers have shown, using high-resolution imaging techniques, three distinct pericyte morphologies: ensheathing pericytes which cover much of the vessel surface at the arteriole-capillary junction and are thought to be contractile; mesh pericytes on the capillary and post-capillary venule which have short processes that are longitudinal and wrap around the vessel; and finally, thin-strand pericytes in the middle part of the capillary which display the traditional bump-on-a-log morphology with thin processes running along the vessel (Figure 1B; Attwell et al., 2016; Berthiaume et al., 2018). However, while single-cell sequencing could be used to show differential gene expression profiles of pericytes depending on their location along the vascular tree, a recent study showed no RNA expression changes within pericytes, and in fact pericytes had a substantially different RNA expression profile to arteriolar VSMC (Vanlandewijck et al., 2018).

The ability to activate or inhibit pericyte function has enormous potential to help treat the underlying cellular pathology in disease. However, specific drug delivery to pericytes is difficult considering pericytes share molecular pathways and surface receptors with other cell types. In addition, drug delivery to cerebral tissue is difficult, as the BBB allows only selective passage of cells and small molecules. While strategies to overcome this issue, such as viral vectors, exosomes, and nanoparticles, have been trialed, their use to target pericytes in the brain has been limited (Dong, 2018). One recent study demonstrated successful peptide-conjugated nanoparticle delivery of the anticancer drug docetaxel to peripheral pericytes in rats with neuroblastoma (Guan et al., 2014), while peptide-conjugated liposomal nanoparticles have also been used to effectively target pericytes and deliver the anticancer drug doxorubicin in mice with lung melanoma (Loi et al., 2010). Nanoparticle formulation with pericyte-binding peptide sequences is possible (Kang and Shin, 2016) but it has not yet been trialed in the brain. The delivery of pericyte-targeted drugs through nanoparticles could provide a novel therapeutic avenue for the treatment of many neurovascular diseases.

Pericytes are essential to tissue repair processes; they direct angiogenesis and therefore the supply of blood to repairing tissue, as well as facilitate inflammatory responses and clear toxic substances (Ribatti et al., 2011; Jansson et al., 2014; Rustenhoven et al., 2017). It has been suggested that administration of pericytes to brain regions where injury has occurred could improve the repair process (Cheng et al., 2018). This concept has recently been examined where the implantation of pericytes into the brains of an AD mouse model improved blood flow and reduced the magnitude of Aβ pathology (Tachibana et al., 2018). The recent developments in induced human pluripotent stem cell-derived pericytes provides the possibility to create autologous pericyte cellular therapies for neurodegenerative diseases that are specific to individual patients, averting the possibility of complications such as immune rejection of cells (Faal et al., 2019; Stebbins et al., 2019). Pericytes themselves can also behave like stem cells and therefore may be reprogrammed to differentiate into other neurovascular cells such as neurons and EC (Nakagomi et al., 2015), which could alleviate neurovascular dysfunction in disease. However, caution over the administration method of pericytes, similar to stem cell therapy, must be considered when contemplating pericyte therapy for disease.

A multi-functional vascular cell that is embedded within the capillary wall and is important in the regulation of CBF, maintenance of the BBB, vascular development and stability, and neuroinflammation. The dysfunction of pericytes in diseases such as stroke and AD has recently been highlighted.

A grouping of cells that regulate the interaction between the brain parenchyma and the blood vessels supplying it. The NVU is made up of neurons, astrocytes, microglia, pericytes, VSMC, and EC.

A semi-permeable membrane separating the blood from the cerebrospinal fluid, and constituting a barrier to the passage of cells, particles, and large molecules. The BBB is formed primarily by non-fenestrated endothelial cells connected by tight junctions, and is supported by pericytes and astrocyte endfeet.

The flow of blood throughout the brain is vital for the maintenance of brain function, as blood delivers the crucial energy substrates oxygen and glucose required for neuronal activity. CBF is tightly controlled at the molecular and cellular levels by the NVU.

Diseases of the brain that can be caused by an acute injury, e.g., stroke, or a chronic degenerative condition, e.g., Alzheimer’s disease.