The Ion Channel and GPCR Toolkit of Brain Capillary Pericytes

Abstract

Brain pericytes reside on the abluminal surface of capillaries, and their processes cover ~90% of the length of the capillary bed. These cells were first described almost 150 years ago (Eberth, 1871; Rouget, 1873) and have been the subject of intense experimental scrutiny in recent years, but their physiological roles remain uncertain and little is known of the complement of signaling elements that they employ to carry out their functions. In this review, we synthesize functional data with single-cell RNAseq screens to explore the ion channel and G protein-coupled receptor (GPCR) toolkit of mesh and thin-strand pericytes of the brain, with the aim of providing a framework for deeper explorations of the molecular mechanisms that govern pericyte physiology. We argue that their complement of channels and receptors ideally positions capillary pericytes to play a central role in adapting blood flow to meet the challenge of satisfying neuronal energy requirements from deep within the capillary bed, by enabling dynamic regulation of their membrane potential to influence the electrical output of the cell. In particular, we outline how genetic and functional evidence suggest an important role for G_s-coupled GPCRs and ATP-sensitive potassium (K_ATP) channels in this context. We put forth a predictive model for long-range hyperpolarizing electrical signaling from pericytes to upstream arterioles, and detail the TRP and Ca²⁺ channels and G_q, G_i/o, and G_12/13 signaling processes that counterbalance this. We underscore critical questions that need to be addressed to further advance our understanding of the signaling topology of capillary pericytes, and how this contributes to their physiological roles and their dysfunction in disease.

Keywords: GPCRs (G protein coupled receptors); KATP channels; brain metabolism; cerebral blood flow (CBF); ion channels; neurovascular coupling (NVC); pericytes.

Figure 1

Overview of gene qualification process for pericyte ion channels and GPCRs and other genes of interest. An initial filter of 1 average count/cell was applied to exclude genes with extremely low expression. (A) Heatmap of expression of the remaining genes throughout the neurovascular unit. A small subset of these genes were highly enriched in pericytes (top left), while many showed higher expression in other cell types. To filter out potential contamination, genes that were expressed in <3% of pericytes, and were absent from the PER3 cluster of Zeisel et al. (2018) were excluded. (B) Relationship between pericyte-specificity of expression and fraction of pericytes expressing each gene considered. Genes represented by green circles were excluded according to the above criteria. (C) High resolution view of genes with a <0.1 expression ratio in pericytes, that were expressed in fewer than 10% of pericytes, corresponding to the bottom left corner in (B). Genes represented by green circles were excluded from further consideration as potential contamination.

Figure 2

An overview of brain angioarchitecture. (A) Cross-section of one brain hemisphere illustrating macroscopic vascular architecture. The carotid artery joins the circle of Willis at the base of the brain, then gives rise to major pial arteries which course over the brain surface, from which multiple penetrating arterioles arise and dive into the tissue. (B) Close up view of the components of the vascular network approximating the area in the boxed region in A showing the interconnected organization of pial arteries, penetrating arterioles, the dense capillary network, and venules. The vessel labeling system we use takes the penetrating arteriole as the 0-order vessel and primary reference point, and vessels are numbered sequentially with regard to this. Vessel number automatically increases each time a vessel branches and thus, after vessel n branches, the daughter branches—regardless of diameter or orientation—are labeled vessel n + 1. (C) Illustration approximating the boxed region in (B), showing the cellular elements that make up the arteriolar side of the brain vasculature. Arteries and arterioles consist of SMCs surrounding ECs, which are in direct contact with the blood. The first 3–4 vessels emanating from the penetrating arteriole are a transitional zone and are covered with contractile mural cells that are positive for α-SMA and can change diameter abruptly. Immediately after the α-actin terminus are capillaries covered by mesh pericytes, following which are capillaries where thin-strand pericytes reside. The cross-section at right shows a section through an artery/arteriole and illustrates the presence of the internal elastic lamina (IEL) which separates ECs and SMCs. Occasional fenestrations dot the IEL, through which ECs and SMCs make direct contact via myoendothelial projections (MEPs, circular inset). These are sites of gap junctions (GJs) permitting chemical and electrical cell-cell communication.

Figure 3

Cytoarchitecture and microenvironment of pericytes. (A) Mural cells with a ‘bump-on-a-log’ cell body, with multiple contractile processes that almost completely encase the underlying vessel. 6,000x, rat mammary gland vasculature. Reproduced with permission from Fujiwara and Uehara (1984). (B) A 4,400x magnification scanning electron micrograph of a putative mesh pericyte of the rat mammary gland. Multiple sparse processes enwrap the underlying capillary. Reproduced with permission from Fujiwara and Uehara (1984). (C) A thin-strand pericyte atop a rat retinal capillary, extending fine processes away from the ovoid cell body. Adapted with permission from Sakagami et al. (1999). Scale bar: 10 μm. (D) Illustration of a thin-strand pericyte. The bulk of the volume of the cell body is occupied by the nucleus. The pericyte is prevented from making direct contact with the underlying EC by the basement membrane, shown in the SEM at bottom left, reproduced with permission from Carlson (1989). Multiple small fenestrations are seen in this structure, allowing for pericyte and endothelial projections to make direct contact with one another, forming so-called ‘peg-socket junctions’ which are also sites of gap junction formation. At bottom right electron micrographs depicting a peg-socket junction (left) and a pericyte-endothelial gap junction (right) are shown, reproduced with permission from Díaz-Flores et al. (2009) and Carlson (1989). Abbreviations in micrographs: EC, endothelial cell; N, nerve; P, pericyte.

Figure 4

Overview of CNS pericyte ion channel and GPCR expression. (A) Relative abundance of mRNA for all ion channel subunits meeting our inclusion criteria. The size of each segment represents the relative expression of the underlying gene. Channels are clustered on the basis of the ion species that the corresponding functional channel conducts (denoted by shading of the same color) and are then grouped by family/subfamily. K⁺ channels are the predominant ion channel class due to extremely high expression of Kcnj8 which forms the pore of vascular K_ATP channels. The non-selective TRP channels are the next highest expressed, followed by Ca²⁺ channels, Cl⁻ channels, and lower expression of other channels. (B) Relative expression of pericyte GPCRs. Here, receptors are organized by ligand sensitivity or class. (C) Expression of the K_ATP channel genes Kcnj8 and Abcc9 throughout the brain vasculature. Pericytes express both genes at much higher levels than arterial SMCs or ECs. However, venous SMCs also express high levels of K_ATP channel-forming genes.

Figure 5

Structural topology of K⁺ channels expressed by pericytes. (A) Vascular K_ATP channels are octamers consisting of four 17-transmembrane SUR2 subunits associated with four 2-transmembrane pore-forming K_ir6.1 subunits. (B) K_ir2.2 channels consist of homo or heteromeric assemblies of four 2-transmembrane subunits. (C) K_v channels are composed of four 6-transmembrane alpha subunits with a positively charged voltage sensor at S4 which transduces changes in V_m into conformational alterations. (D) K_2P channels are tetramers of two-pore domain four-transmembrane subunits. (E) K_Na channels have a 6-transmembrane structure that lacks a voltage sensor, with multiple regulatory sites in the long intracellular COOH-terminus including two RCK domains, an ATP binding site, and a PDZ domain. (F) K_Ca2.3 channels consist of four 6-transmembrane domains which lack a voltage-sensor at S4. The COOH-terminus of each is associated with a calmodulin monomer, which imparts Ca²⁺ sensitivity to the channel.

Figure 6

Predicted capillary pericyte-EC interactions to control local blood flow. Neuronal activity drives the release of K⁺ and G_sPCR agonists. Top inset: These are predicted to engage pericyte K_ir2.2 and their cognate GPCRs, respectively. G_sPCR activity activates K_ATP channels, the hyperpolarization by which may feed forward to evoke further K_ir2.2 activity (a sufficient fall in ATP:ADP would also engage K_ATP channels). The hyperpolarization generated by these channels may then be passed via gap junctions to cECs (bottom right inset) or possibly to adjacent pericytes, though direct pericyte-pericyte gap junctions have not been observed to date. In cECs, the incoming hyperpolarization will engage K_ir2.1 channels to amplify hyperpolarization to a sufficient level to pass to adjacent cECs and pericytes. Hyperpolarization-mediated activation of K_ir2.1 and K_ir2.2 in these cells will rapidly regenerate the current so that it can be passed to the next cell, and so on upstream to the arteriole. Upon arrival at the arteriole and its first few offshoots, hyperpolarization will be passed via GJs at MEPs to SMCs and to contractile mural cells, which will close VDCCs, leading to a fall in intracellular Ca²⁺, relaxation of their actin-myosin contractile machinery, vasodilation, and an increase in blood flow.

Figure 7

Structural overview of the TRP families expressed in CNS capillary pericytes, adapted with permission from Clapham (2003). All TRP channels share a common and typical 6-transmembrane structure with profoundly varying intracellular N- and C-terminal domains, the major features of which are illustrated. CC, coiled-coil domain.

Figure 8

Structural topology of Ca²⁺ channels expressed by pericytes. (A) The general structure of Ca_v channels consists of a single 24-transmembrane α subunit which is a repeat of a 6-transmembrane motif with an embedded voltage sensor connected by intracellular loops. This is accompanied by associated β, γ, and α2δ subunits. (B) IP₃Rs consist of a tetrameric assembly of 6- transmembrane subunits with a large N-terminal domain that contains the IP₃ binding site.

Figure 9

Potential G_s- and G_i/o-coupled GPCR–ion channel interactions in capillary pericytes. G_sPCR activation promotes (green) adenylate cyclase (AC) activity, whereas G_i/oPCR activation inhibits (red) AC. AC in turn generates cAMP from ATP, which stimulates PKA activity. PKA interacts with a broad range of ion channels. In pericytes, its activity is expected to couple to plasma membrane K⁺ and VDCC activity, with mixed effects on TRP channel activity. K⁺ channel hyperpolarization will oppose VDCC activity and thus the overall effect of G_s stimulation is membrane hyperpolarization.

Figure 10

Potential G_qPCR-ion channel interactions in capillary pericytes. G_qPCR activation engages PLC, leading to the hydrolysis of PIP₂ into IP₃ and DAG. IP₃ evokes Ca²⁺ release from the ER via resident IP₃Rs, which may engage CaCCs and K_Ca channels. DAG stimulates PKC which has mixed effects on the TRP channels expressed by pericytes, promotes VDCC activity, and inhibits K_ATP, K_ir, and K_v channels. The net effect of engagement of G_qPCRs is thus membrane depolarization and intracellular Ca²⁺ elevation.