Astrocyte regulation of cerebral vascular tone

Abstract

Cerebral blood flow is controlled by two crucial processes, cerebral autoregulation (CA) and neurovascular coupling (NVC) or functional hyperemia. Whereas CA ensures constant blood flow over a wide range of systemic pressures, NVC ensures rapid spatial and temporal increases in cerebral blood flow in response to neuronal activation. The focus of this review is to discuss the cellular mechanisms by which astrocytes contribute to the regulation of vascular tone in terms of their participation in NVC and, to a lesser extent, CA. We discuss evidence for the various signaling modalities by which astrocytic activation leads to vasodilation and vasoconstriction of parenchymal arterioles. Moreover, we provide a rationale for the contribution of astrocytes to pressure-induced increases in vascular tone via the vasoconstrictor 20-HETE (a downstream metabolite of arachidonic acid). Along these lines, we highlight the importance of the transient receptor potential channel of the vanilloid family (TRPV4) as a key molecular determinant in the regulation of vascular tone in cerebral arterioles. Finally, we discuss current advances in the technical tools available to study NVC mechanisms in the brain as it relates to the participation of astrocytes.

this article is part of a collection on Unique Features of Cerebral Circulation. Other articles appearing in this collection, as well as a full archive of all collections, can be found online at http://ajpheart.physiology.org/.

Introduction

Given its limited energy reserves, the brain requires constant perfusion for proper function. This is accomplished through three fundamental processes, cerebral autoregulation (CA), neurovascular coupling (NVC) or functional hyperemia (FH) (88), and endothelium-mediated signaling (8). Whereas autoregulation ensures constant blood flow over a wide range of systemic pressures, FH ensures rapid spatial and temporal increases in cerebral blood flow (CBF) in response to neuronal activation. In addition, endothelial cells (ECs) release both vasodilators [nitric oxide (NO), endothelium-derived hyperpolarizing factor, prostacyclin, and prostaglandin E₂ (PGE₂)] and vasoconstrictors (endothelin-1, thromboxane A₂, and prostaglandin F_2α), which are capable of modulating cerebrovascular tone (8). In the next sections, we will expand on the mechanisms by which astrocytes participate in NVC and, to a lesser extent, CA. The signaling mechanisms arising from neurons, in particular, interneurons, have been reviewed previously (9, 42, 76). These pathways are an important component of the hyperemic response and must not be underestimated, especially given that they constitute the principal mechanism controlling CBF in some brain areas such as the cerebellum (169). Moreover, the intimate anatomical association of cortical neurons and cerebral blood vessels via neuronal-astrocyte-vascular appositions also lends support to the idea that vasoactive peptides released from neurons act directly on the vasculature (2, 22, 26, 50, 153, 159). While neuronal signaling is undoubtedly involved in NVC mechanisms, experimental evidence also demonstrates the involvement of astrocytes in NVC. In this review, we focus on the mechanisms by which astrocytes contribute to the regulation of vascular tone during FH or NVC (21, 148).

The Neurovascular Unit

Blood supply to the brain is conducted by the internal carotid and vertebral arteries (28, 48). The vertebral arteries give rise to the basilar artery, which along with the internal carotid and communicating arteries form the circle of Willis (28). The circle of Willis further divides, giving rise to the anterior, middle, and posterior cerebral arteries that then continue to branch into smaller arteries and arterioles (28). For a detailed description of the neurovascular control of the cerebral arteries that do not penetrate the brain parenchyma, we refer the readers to excellent review articles (15, 73, 77). As arterioles outside the brain branch into pial arterioles that surround the surface of the brain, they then penetrate the brain parenchyma as penetrating or parenchymal arterioles (48) (Fig. 1). Parenchymal arterioles branch into an extensive capillary network with a heterogeneous density distribution that is dependent on the network proximity to neuronal populations (48). Higher capillary density is observed in the gray matter relative to the white matter regions (23). Accordingly, the neurovascular control of the cerebral circulation varies depending on the location and caliber of the vessels. The density of peripheral nerve terminals decreases as they penetrate the brain parenchyma and pass the Virchow-Robin space (25, 29, 31, 77); consequently, parenchymal microvessels are primarily regulated by local interneurons and neuronal terminals from a central origin (intrinsic innervation) such as the basal forebrain, raphe nucleus, and locus coeruleus (73, 77, 85).

Fig. 1.

Hypothetical illustration of the major cellular components comprising the neurovascular unit in the cerebral cortex. As pial arterioles penetrate the brain parenchyma, they gradually lose innervations from extracellular nerves originating from peripheral ganglia. The vascular tone of parenchymal arterioles is chiefly regulated by signals derived from neurons and astrocytes that encase much of the abluminal surface of the arterioles.

In recent years, much attention has been placed on the participation of glial cells, primarily astrocytes, in the regulation of cerebrovascular tone (9, 21, 72, 88, 91). The growing interest to further our understanding of the mechanisms underlying NVC was prompted by the large amount of evidence of the highly sophisticated arrangement of various cells types at the neuronal-glial-vessel interface or the neurovascular unit (NVU). The NVU (Fig. 1) is comprised of vascular cells [ECs, pericytes, and vascular smooth muscle cells (VSMCs)], neuron terminals or varicosities, astrocytes and their specialized end-foot processes, and microglia. The functional role of these various cell types varies throughout development and in disease conditions. For example, astrocytes establish a tight association with ECs during vessel formation/maturation (1, 132, 133) and with neurons during the establishment of new synapses and circuit organization (122, 145). The interaction between these various cell types aids in blood brain barrier development/maintenance (1) and CBF control. This fine-tuned anatomical organization is disrupted in disease conditions such as Alzheimer's disease (69, 86, 172), hypertension (69, 87), and stroke (7, 38, 39, 69) to name a few. Cell-to-cell communication among astrocytes (67) and/or between astrocytes-ECs (104) is supported and propagated by structural components such as gap junctions (60, 141) and anchoring proteins [e.g., integrins (38)]. It has been reported that astrocytic end feet cover about 99% of the abluminal surface of the vessel wall (92, 141). It is not clear, however, whether this extensive end-foot coverage is homogeneous among all vessel subtypes (arterioles, venules, capillaries) and/or throughout different brain regions, and if not, what functional implications varying degrees of end-foot coverage of the vasculature may have on NVC mechanisms. Clearly, the structural and functional organization of the cells comprising the NVU is critical for optimal CBF distribution in the brain.

On a larger scale than that of the NVU, additional functional organization is provided by glial networks (66). Specialized gap junctions connect neighboring astrocytes forming a syncytium (149) that is capable of efficiently modulating the activity of large neuronal populations (66) and, of particular relevance to this review, vascular networks as well. For example, the gap junction protein connexin 43 functionally links astrocytic end feet surrounding parenchymal arterioles to the glia limitans, the thick layer of astrocytic processes surrounding pial vessels and separating it from the underlying neuropil (60, 167). This functional coupling allows local information from these cells to be transmitted upstream to pial arterioles and thus vessels that supply the brain circulation (167), thereby ensuring an efficient increase in CBF during FH. Paisansathan et al. (126) have shown that adenosine and K⁺ mediate this upstream pial arteriole dilation following neuronal activation.

NVC and Astrocytes

An increase in neuronal activity triggers the activation of a number of pathways originating from both neurons and astrocytes that elicit rapid spatiotemporal delivery of glucose and oxygen to working neurons via increased CBF. Studies in the cortex and hippocampus suggest that neuronal activity-induced astrocytic activation occurs primarily through glutamatergic signaling pathways (58, 173). In a seminal study, Carmignoto's group (173) showed that activation of metabotropic glutamate receptors (mGluRs) induced an increase in astrocytic Ca²⁺, resulting in arteriolar vasodilation. The study provided mechanistic evidence that a cyclooxygenase product downstream from arachidonic acid (AA) metabolism (9), likely PGE₂-mediated astrocyte-induced arteriole dilation (173). Additional studies followed that further supported the role of astrocytes in the control of vascular tone via signals released from AA metabolism. Specifically, in vivo data obtained using two-photon laser-scanning microscopy provided evidence for a signal derived from the cyclooxygenase-1 pathway in astrocyte-induced arteriole vasodilation (148). In addition to prostaglandin production (21, 35, 36, 95, 148, 173, 174), increased intracellular Ca²⁺ in astrocytes (3, 33) also leads to the formation and release of other vasoactive signals including NO (21, 27, 37, 94, 105, 119, 134, 164), epoxyeicosatrienoic acids (EETs) (4, 14, 21, 82, 95, 108, 119, 128), glutamate, adenosine, and ATP (6, 21, 95, 101, 137, 141, 154), all which are capable of altering the vascular tone of parenchymal arterioles (55, 61, 173). Table 1 lists some of the reported astrocyte-derived vasoactive signals acting on cerebral vessels.

Table 1.

Vasoactive astrocyte-derived signals altering pial and/or parenchymal arteriole vascular tone

Signals	Vascular Response	Possible Mechanism	References
EETs	VD	BK channel activation on VSMC	4, 14, 21, 82, 95, 108, 119, 128
20-HETE	VC	BK channel inhibition on VSMC	119, 121
Nitric oxide	VD/VC	likely via its interaction with prostanoids	21, 27, 37, 94, 105, 119, 134, 164
Lactate	VD	PGE2 levels	71
PGE2	VD/VC	VD via prostanoid (EP4) receptor	21, 35, 36, 95, 148, 173, 174
		VC via EP1
K+	VD/VC (Conc. dependent)	VD via VSMC Kir channel activation	62, 68
		VC via VSMC depolarization
Ca2+	VD/VC (Conc. dependent)	VD via VSMC BK channel activation	61, 68
		VC via VSMC depolarization
ATP→AMP→adenosine	VD	via A2R in VSMC	6, 21, 95, 101, 137, 141, 154

EETs, epoxyeicosatrienoic acids; VD, vasodilation; VC, vasoconstriction; Conc, concentration; BK, large conductance Ca²⁺-activated and voltage-dependent K⁺ channels; VSMC, vascular smooth muscle cells; PGE₂, prostaglandin E₂; EP, E-prostanoid; K_ir, inwardly rectifying K⁺ channel; A₂R, adenosine type 2 receptor.

Research investigating the effects of astrocyte-derived signals on brain arterioles was promptly complicated by the fact that different laboratories used a variety of technical approaches and parameters, which led to variability in the types of vascular responses observed. One such parameter is the age of the animal used for study. In young or neonatal animals (which are often used for NVC studies), astrocytes may have a different resting dynamic state and arterioles may not be fully differentiated leading to variability in the resting level of tone. To this end, our group showed that the polarity of vascular responses following astrocyte stimulation could be altered by the resting level of arteriole tone (14). Astrocytic stimulation induced constrictions in arterioles with little or no basal tone and dilations in arterioles with 30% or greater basal tone (14). In addition, a recent study by Sun et al. (147) demonstrated that astrocytic expression of mGluR5, a commonly targeted receptor in NVC studies, is developmentally regulated. Thus studies addressing NVC mechanisms using mGluR agonists must take into consideration shifts in the expression/pattern of these receptors in older animals. Another factor that may account for variability in vascular responses is that some of the approaches used to stimulate astrocytes or neurons may have been too strong, thus eliciting nonphysiological responses as may be the case with Ca²⁺ uncaging or electrical field stimulation (68, 121). Finally, experimental conditions such as temperature and oxygen gradients may have a strong impact on the activity of astrocytes as well as the levels of vascular tone, particularly as it relates to in vitro studies. For example, Gordon et al. (71) suggested that the polarity of the vascular response to glia-derived vasoactive signals is coupled to oxygen concentration levels, which modulate the metabolic state of the tissue, specifically extracellular lactate and adenosine concentrations. The authors proposed that when Po₂ levels are low, the increased lactate (following neuronal stimulation) inhibits the activity of PGE₂ transporters, raising extracellular PGE₂ levels and thus favoring vasodilation (24). They also concluded that vasoconstriction occurs under hyperoxic conditions, where lactate levels are low and extracellular PGE₂ availability is decreased (71). These observations were challenged by Lindauer et al. (106) when they demonstrated that CBF responses to electrical forepaw stimulation or cortical spreading depression were independent of O₂ levels in anesthetized rats under conditions of hyperbaric oxygenation. Along these lines, Newman and colleagues (120) showed that O₂ levels altered the polarity of vascular responses in vitro but had little effect in vivo in the retina. Similarly, Metea and Newman (119) attributed the polarity of astrocyte-induced vascular responses to NO availability and its interactions with AA metabolites, such that vasoconstriction is likely favored under conditions of increased NO due to NO-mediated inhibition of cytochrome P-450 (152) and the subsequent decrease in EET formation (119). In agreement with these data, Rancillac et al. (134) demonstrated that glutamate-induced neuronal NO release elicited vasoconstriction of cerebellar microvessels via a prostanoid and endothelin-dependent mechanism.

The contribution of NO signaling to NVC is likely brain region specific and dependent on the intensity of neuronal stimulation. NO has been reported as the primary NVC signal in the cerebellum (168, 169), yet it is a suggested modulatory signal in the cortex (107). Parenchymal arterioles in hippocampal brain slices constricted in response to NO synthase inhibition, suggesting that in this brain region NO has a tonic vasodilatory impact on vascular tone (56). Chisari et al. (27) provided evidence for the involvement of astrocyte-derived NO in NVC using a coculture system in which NO originating from LPS-activated glia induced basilar artery dilation. Furthermore, de Labra et al. (37) suggested that vascular or glial-derived NO was the primary mediator of low frequency-induced vasodilation, whereas neuronal NO was the primary mediator of vasodilation following intense stimulation.

An additional gas messenger molecule, carbon monoxide (CO), has also been shown to participate in glutamate-induced cerebral arteriole dilation (63). CO, derived from vascular cells (63, 102) and astrocytes (103, 127), relaxes VSMCs through its effect on Ca²⁺ sparks and the subsequent activation of large conductance Ca²⁺-activated and voltage-dependent K⁺ channels (BK) (90, 166). As with other glial and neuronal-derived signals (e.g., NO, K⁺, Ca²⁺), the polarity of the vascular response to CO may be modulated by its effect on NO production (89).

Potassium has long been shown to regulate vascular tone (Fig. 2). Activation of endothelial intermediate and small conductance K⁺ channels (IK and SK, respectively) can trigger a hyperpolarizing current from the endothelium to VSMC, via myoendothelial gap junctions (143), resulting in arteriole dilation. Secondary to gap junction activation, K⁺ efflux from IK and SK channels generate a K⁺ cloud (78) that hyperpolarizes neighboring VSMCs via activation of both inwardly rectifying (K_ir) channels and the Na⁺/K⁺ pump (117), further contributing to endothelial-VSMC interactions. The contribution of K_ir channels to VSMC hyperpolarization may be secondary to the activation of other K⁺ conductances as Smith et al. (142) showed blunted vasodilation to pinacidil (ATP-sensitive K⁺ channel activator) in the presence of the K_ir channel blocker Ba²⁺; the authors suggested that K_ir channels participate as electrical amplifiers of responses initiated by EC K⁺ channels. Similarly, we showed that following mGluR-mediated astrocyte activation, there is a significant increase in BK channel activity in astrocytic end-feet processes, leading to an increase in extracellular K⁺ that also hyperpolarizes VSMCs (62). In astrocytes, BK channels are selectively expressed in astrocytic end feet (131), thus providing a potentially high density of K⁺ currents in these structures and eliciting rapid changes in extracellular K⁺ concentrations at the gliovascular interface upon their activation. Experimental evidence suggests that the polarity of vascular responses to K⁺ (vasodilation vs. vasoconstriction) is dependent on the degree of K⁺ efflux from astrocytes. Girouard et al. (68) demonstrated that increasing the concentration of uncaged Ca²⁺ in astrocytes shifted dilations to constrictions and that these responses were mediated by BK channels. The putative mechanisms leading to responses of opposing polarity are the activation of VSMC K_ir channels to modest K⁺ elevations (<20 mM) and closure of VSMC voltage-gated calcium channels resulting in membrane hyperpolarization for dilations (62) and depolarization in response to K⁺ elevations > 20 mM for constrictions (45). The contribution of K⁺ signaling to NVC is further supported by the observation that EETs increase outward K⁺ currents in astrocytes (84). This would suggest that upon the release of AA metabolites, K⁺ signaling is either enhanced or prolonged, thus contributing to hyperemia in the brain. While Higashimori et al. (84) demonstrated that outward K⁺ currents in astrocytes are mediated exclusively by BK and SK channels, Longden et al. (109) showed that IK channels also contribute to K⁺ efflux from astrocytes following electrical field stimulation. Along with the observation that extracellular K⁺ affects the polarity of the vascular response to astrocyte stimulation (68), our group demonstrated the importance of resting vascular tone in the K⁺-induced vascular response (14). We showed a positive correlation between the magnitude of K⁺-induced vasodilation and the level of arteriolar tone before dilation (14). K⁺ signaling studies suggest that, similarly to ECs, astrocytes modulate vascular tone through the efflux of K⁺ from various K⁺ channel subtypes (BK, SK and IK) and that this mechanism can induce both dilation and constriction of parenchymal arterioles in a K⁺-concentration manner.

Fig. 2.

K⁺ signaling mechanisms at the neurovascular unit. Schematic representation of the signaling mechanism by which endothelial cell (EC) and astrocyte signaling induce membrane potential (V_m) hyperpolarization of vascular smooth muscle cells (VSMCs) and arteriole vasodilation. Shear stress or agonist-induced activation of G protein-coupled receptors (GPCR) increases Ca²⁺ in ECs via transient receptor potential (TRP) channel-dependent mechanisms resulting in the activation of intermediate conductance (IK) and small conductance (SK) potassium channels. The hyperpolarization mediated by these channels spreads to the VSMCs through myoendothelial gap junctions (MEGJ). Moreover, the resulting K⁺ efflux from ECs activates VSMC inwardly rectifying K⁺ channel (K_ir) channels and the Na⁺/K⁺ pump inducing V_m hyperpolarization and closure of voltage-dependent calcium channels (VDCC), decreased intracellular Ca²⁺, and VSMC relaxation. On the other hand, glutamate released during neuronal activation binds to GPCRs on astrocytes, triggering an increase in Ca²⁺ that, in turn, activates large conductance Ca²⁺ activated potassium channels (BK) on astrocytic end feet (EF). Astrocytic BK channel activation causes potassium efflux into the perivascular space and, again, VSMC K_ir channel activation followed by relaxation. TRP of the vanilloid family (TRPV4) channel activation has recently been shown to elicit increases in astrocytic EF Ca²⁺ as well. The intracellular Ca²⁺ increase in ECs and astrocytes also results in PLA₂ activation, arachidonic acid (AA) metabolism, and the production/release of various vasoactive substances including epoxyeicosatrienoic acids (EETs), which in turn can modulate BK channels in astrocytes and vascular cells, further contributing to the regulation of vascular tone. TRPC1, transient receptor potential canonical-1.

Relevant to K⁺ signaling and NVC mechanisms is the K⁺ buffering/siphoning mechanism in which astrocytes remove K⁺ from the synapse during neuronal activity and redistribute it to areas with lower K⁺ concentrations such as the gliovascular space to prevent excessive neuronal excitability induced by continued neuronal depolarization (96). However, evidence from the retina suggests that this passive mechanism does not hold true as direct depolarization of Müller cells does not induce K_ir-mediated efflux of K⁺ at the gliovascular space or elicit dilation (118). Nevertheless, alternative mediators, such as other classes of K⁺ channels, may underlie a K⁺ buffering/siphoning mechanism that contributes to NVC. Furthermore, in a recent study, Nedergaard's group (158) showed that neuronal stimulation activates the Na⁺/K⁺ pump. Activation of the Na⁺/K⁺ pump will result in hyperpolarization of the astrocytic membrane, which would then facilitate the influx of K⁺ through K_ir channels on perisynaptic astrocytic processes. Moreover, upstream events (e.g., activation of Na⁺-Ca²⁺ exchanger and/or G protein-coupled receptors) triggering increases in astrocytic Ca²⁺ would activate BK (62) or other potassium channels at the end feet, thereby contributing to vasodilation and the removal of K⁺ from the synapse (K⁺ buffering). Thus we suggest that the K⁺ siphoning hypothesis and its contribution to NVC be revisited, keeping in mind variations in brain regions as well as the contribution of other classes of K⁺ channels at the synapse and the gliovascular space.

CA and Astrocytes

CA constitutes a critical physiological mechanism of cerebral arterioles that ensures constant perfusion throughout a broad range of systemic pressures (50–150 mmHg) (17, 28). It is well established that CA involves at least three major components: myogenic, neurogenic, and metabolic. In the myogenic component of CA, the increased transmural pressure in cerebral arterioles increases vascular resistance through a number of mechanisms, including activation of L-type calcium channels, membrane potential (V_m) depolarization, and increased production of vasoconstrictors (e.g., 20-HETE) downstream from inositol 1,4,5-triphosphate and diacylglycerol activation (81, 151). 20-HETE induces VSMC constriction via the inhibition of BK channels followed by V_m depolarization and opening of calcium channels (80). In the cerebral circulation, the AA metabolite 20-HETE can be produced by VSMCs (65) and the endothelium (124). Work by Faraci et al. (52) demonstrated that autoregulatory responses persist following endothelial injury, further supporting the myogenic component of CA. While much has been discovered concerning the myogenic constriction of cerebral vessels, less is understood on the contribution of the metabolic and neurogenic pathways. Even fewer studies have addressed the potential mechanisms by which cells within the brain parenchyma (e.g., astrocytes) contribute to the regulation of arteriole tone during CA.

Within the intricate structural arrangement of the NVU, VSMCs are situated between two very different cell types, ECs on the luminal side and astrocytes on the abluminal side. The literature suggests that many of the endothelial vasoactive pathways that allow for EC-VSMC interactions (e.g., AA metabolites, NO, EETs, adenosine, K⁺) (53, 54) are present in astrocytes as well, thus potentially contributing to astrocyte-VSMC interactions and the regulation of vascular tone (Fig. 1). Although speculative at this moment, one reason for such a structural arrangement may stem from the need to maintain brain blood volume within physiological levels. Because the brain is encased in a closed cranium with no space to expand, astrocytes may participate in the control of CBF as an additional system to prevent overexpansion of brain volume. They may also participate in the signaling necessary (besides NVC/FH) (148) to maintain a level of vascular tone that allows for optimal perfusion to the needed areas. Thus astrocytes may be constant monitors of both neuronal activity and the degree of basal CBF. While the contributions of astrocytes to NVC and FH-induced increases in CBF are well documented, little is known as to whether astrocytes contribute to resting CBF. The ability of astrocytes to both dilate and constrict arterioles favors this idea since their intimate contact with active neurons places them in an ideal position to regulate local vascular tone. Thus astrocytes may be key contributors to CA mechanisms as well (82). However, critical questions remain unanswered, particularly whether astrocytes, as well as neurons, have the ability to sense autoregulatory-mediated adjustments in perfusion/energy supply and thus participate in bidirectional communication at the NVU.

Studies performed in cannulated cerebral arterioles showed increased flow/pressure-induced vasoconstriction in parenchymal arterioles in the presence or absence of a functional endothelium (19, 20, 70). While a significant portion of this pressure/flow-induced response is mediated by the intrinsic properties of the VSMCs, it is possible that signals released from activated astrocytes also contribute to the constriction. As mentioned earlier, the abluminal surface of parenchymal cerebral vessels is almost completely covered by astrocytic end-feet processes (92, 141); these specialized structures are unique in that they express a distinct subset of channels including BK, aquaporin-4, and the recently characterized transient receptor potential channel of the vanilloid subfamily, TRPV4. Using brain slices, Mulligan and MacVicar (121) showed that astrocytic stimulation via uncaging of Ca²⁺ (in the absence of neuronal stimulation) induced parenchymal arteriole vasoconstriction. This observation provided evidence that in addition to contributing to NVC during increases in neuronal activity, astrocytes may also be capable of constricting arterioles, a response that may be useful to the brain under conditions of increased perfusion pressures. Importantly, the anchoring of some of the channels mentioned is also dependent on the presence of mechanosensory proteins such as integrins (40). The polarized expression of specific proteins and channels in astrocytic end feet is ideal for continuous monitoring and subsequent adjustment of vascular tone. While evidence for this mechanism is lacking, there is indeed evidence that TRPV4 channels, which upon activation cause an increase in intracellular Ca²⁺ (43), are expressed in astrocytic end feet. Thus it is reasonable to suggest that in addition to a myogenic component, signals from perivascular cells may also contribute to increases in vascular tone in response to increases in luminal pressure and/or flow (79, 81, 151).

TRPV4 Channels as Regulators of Vascular Tone

Structurally, the TRPV4 protein is comprised of six transmembrane segments with a cation pore located between segments 5 and 6 (51). Four subunits are required for the formation of a functional TRPV4 channel (51). In addition to forming homotetramers, recent studies suggest that TRPV4 can also heteromerize with transient receptor potential canonical-1 and transient receptor potential polycystin-2 (43, 112). Functionally, these cation channels are moderately selective for Ca²⁺ with a permeability ratio of 5.8–6.9 P_Ca/P_Na (146, 155, 160). TRPV4 channels are activated by both chemical [endocannabinoids (162), AA (162) and 4-α-phorbol esters (5, 160, 163)] and physical stimuli [cell swelling (146), heat (74, 163) and mechanical displacement (125)]. EETs, which are produced by both ECs and astrocytes, are potent endogenous activators of TRPV4 channels (46, 47, 130, 156, 162); cytochrome P-450 epoxygenase metabolism of AA to EETs mediates TRPV4 activation by endocannabinoids and cell swelling (157, 162).

TRPV4 channels are important mediators of vascular tone. VSMC and EC TRPV4 expression has been demonstrated in multiple tissues including aorta and extraaveolar vessels as well as carotid, cerebral, mesenteric, and pulmonary arteries (5, 64, 97, 113, 114, 156, 160, 161, 165, 170). In VSMCs, EET-induced activation of TRPV4 channels initiates VSMC hyperpolarization and vascular relaxation via the subsequent activation of BK channels (46, 47). In ECs, increases in intracellular Ca²⁺ induced by TRPV4 activation initiate both NO and endothelial-derived hyperpolarizing factor-mediated vasodilation (98, 144).

In the brain, TRPV4 expression has been demonstrated in both neuronal and nonneuronal cell types, including astrocytes and microglia in addition to cerebral artery VSMCs and ECs (12, 13, 30, 46, 51, 100, 113, 139). TRPV4 channels expressed in both cerebral artery VSMCs (46) and ECs (113) have been shown to contribute to cerebral arteriole tone. In astrocytes, TPRV4 channel expression is localized primarily to astrocytic end feet (12, 13) that envelop the cerebral vasculature, an ideal location for monitoring and/or regulating vascular tone. The importance of astrocytic TRPV4 channels in modulating vascular tone has been demonstrated in a recent study by Dunn et al. (44) in which they show that astrocytic TRPV4 channel activation enhances vasodilation during NVC. Whereas they show that EETs are capable of inducing TRPV4-mediated increases in astrocytic end-foot Ca²⁺, EETs do not seem to mediate this TRPV4-induced amplification of NVC; instead, the response is induced by the subsequent activation of inositol 1,4,5-trisphosphate receptors (44).

Other indirect evidence also suggests that TRPV4 channels may play a more comprehensive role in monitoring and/or regulating vascular tone. Higashimori et al. (84) demonstrated that the synthetic EET analog 11-nonyloxy-undec-8 (Z)-enoic acid increased both Ca²⁺ oscillation frequency and BK channel currents in astrocytes. Given that EETs are endogenous TRPV4 agonists (46, 47, 130, 156, 162) and astrocytic K⁺ signaling is an important mediator of vascular tone (62), these data support the hypothesis that EET-induced activation of TRPV4 channels and subsequent increases in intracellular Ca²⁺ may contribute to astrocyte K⁺ signaling and the regulation of cerebral vascular tone. Upon their activation, astrocytic TRPV4 channels also engage inositol 1,4,5-trisphosphate receptors to increase Ca²⁺ as recently demonstrated (44), providing an additional mechanism by which TRPV4 channel-induced Ca²⁺ signaling in astrocytes may contribute to the regulation of vascular tone. Furthermore, TRPV4 channel activation is associated with the production of NO (41, 171), which itself can elicit sustained increases in astrocytic Ca²⁺ (11). These data suggest that TRPV4 channels may also survey and/or modulate vascular tone via an NO-dependent signaling mechanism. Given their strategic expression in astrocytic end feet combined with their capacity to respond to a multitude of signals, TRPV4 channels are ideal candidates to mediate a bidirectional communication modality within the NVU.

Tools to Study NVC in the Brain

The role of astrocytes in the control of vascular tone has been addressed with a number of in vivo and in vitro techniques. Both include their own set of advantages and disadvantages. It is reasonable to state that in vivo simultaneous vascular and astrocyte recordings are ideal as the approach provides an intact system including pressurized and perfused vessels with the complete anatomical organization of astrocytes and neurons in their surroundings. Moreover, in vivo techniques have been complemented with advances in the use of genetically engineered animals, adenoviruses, and Ca²⁺ sensors (e.g., GCamP), as well as additional stimulation modalities such as uncaging and optogenetics (59, 115, 138). Although these approaches provide powerful data, confounding limitations lie in the fact that imaging is confined to the first few hundred microns below the surface of the brain, acquisition rates are generally slower, and pharmacological approaches are limited given the need to use higher drug concentrations. In addition, the use of anesthesia must be taken into consideration given that it can have profound effects on constituents of the NVU (10, 150). Nonetheless, in vivo two-photon laser scanning microscopy is among the most sophisticated tools to study astrocyte-vascular interactions.

Among the in vitro approaches, the most commonly used technique is the brain slice model. Slices are an ideal model as all of the constituents of the NVU are intact allowing for the study of neuronal-to-astrocyte-to-vessel communication. Moreover, the fact that virtually any portion of the brain can be sliced allows for the study of NVC in different brain regions which may shed light on unexplored mechanisms that may differ from those in commonly studied areas, namely the cortex and hippocampus. Earlier studies that lacked some of the important steps we know today must be taken into account when studying arterioles that are not pressurized and perfused. To this end, it is clear now that to attain a physiological vascular response, arterioles must have tone. The latter can be achieved by either perfusing the slice with a vasoconstrictor such as the thromboxane A₂ receptor agonist U-46619 (14, 61, 62, 110, 111) or NO synthase inhibitors (57, 173), as previously reported, or by cannulating and pressurizing the arterioles as previously performed in excised cerebral vessels (123). The latter approach consists of introducing a cannula into the open end of the arteriole at the pial surface and perfusing the arteriole (93, 110) to attain physiological levels of luminal pressure and shear stress (19, 151). Given that this approach allows cerebral arterioles to develop myogenic tone, arteriole cannulation in brain slices is an ideal approach for studying the cellular mechanisms underlying NVC in the brain.

In addition to vascular tone, the role and effects of tissue O₂ tension on NVC have been debated and clearly deserve further consideration. Extensive studies were performed on brain slices to optimize the conditions by which the tissue would remain viable for several hours (18, 32, 75). Perfusing brain slices with a bicarbonate buffered artificial cerebrospinal fluid (aCSF) gassed with 95% O₂-5% CO₂ clearly aids in prolonging tissue viability (83, 135). However, while these solutions have led to successful electrophysiological and imaging recordings from both neurons and astrocytes, a closer look at the constituents of the aCSF is needed. For example, glucose concentrations used by most researchers (10–25 mM) exceeds that of the actual CSF (∼2.5 mM) (140). Moreover, the use of 95% O₂ may result in variable levels of neuronal excitability as well as superoxide production and cell death (34, 83). Additionally, tissue Po₂ levels may also have profound effects on vascular responses since it may alter the levels of vasoactive agonists at the gliovascular interface (71). Similarly, high O₂ levels may have direct effects on NO availability (99), also interfering with the vasodilator/vasoconstrictor ratios acting on the arterioles. Thus future considerations must be made as to the need to switch brain slice studies (at least for NVC mechanisms) to lower glucose and O₂ concentrations that approximate physiological levels without compromising the viability of the tissue.

In summary, it is clear that NVC is mediated by a multiplicity of signals and conditions including astrocyte release of vasodilatory signals. However, to better understand the physiology of how astrocytes regulate vascular tone, we must first take into consideration the conditions under which these mechanisms are being studied and the questions elicited by previous studies. Consequently, when studying NVC, one must consider that the predominant vascular response is highly dependent on the resting levels of vascular tone, O₂ gradients, anesthesia, and strength/duration/type of stimulus. Future studies addressing whether astrocytes participate in regulating basal vascular tone will shed light on the role these cells have in CA (82, 136).

In light of the large amount of knowledge gained on the signaling modalities by which astrocytes and neurons control vascular tone, new questions have arisen. For example, what are the functional differences between astrocyte signaling to arterioles, venules, and capillaries? To this end, Peppiatt et al. (129) demonstrated norepinephrine-induced constrictions of pericytes and glutamate-induced dilations providing evidence for the control of blood flow at the level of the capillaries. Thus it is possible to speculate that given the intimate association between astrocytes and pericytes (16, 116), astrocyte-induced pericyte dilation/constriction may also contribute to the regulation in flow at the level of the capillaries. A mechanism of this nature may provide very tight regulation of CBF without the need to involve upstream vessels. The potential for regulatory diversity within the brain microcirculation itself highlights the necessity of continuing research, using a range of advanced methodologies to reveal the intricate cellular signaling modalities involved in NVC and CBF autoregulation.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author (s).

AUTHOR CONTRIBUTIONS

J.A.F. prepared figures; J.A.F. and J.A.I. drafted manuscript; J.A.F. and J.A.I. edited and revised manuscript; J.A.F. and J.A.I. approved final version of manuscript.