Direct Deviations in Astrocyte Free Ca²⁺ Concentration Control Multiple Arteriole Tone States

Abstract

Astrocytes elicit bidirectional control of microvascular diameter in acutely isolated brain slices through vasoconstriction and vasodilation pathways that can be differentially recruited via the free Ca²⁺ concentration in endfeet and/or the metabolic status of the tissue. However, the Ca²⁺-level hypothesis has not been tested using direct manipulation. To overcome this, we used Ca²⁺-clamp whole-cell patching of peri-arteriole astrocytes to change astrocyte-free Ca²⁺ to different concentrations and examined the vascular response. We discovered that clamping Ca²⁺ at the approximate resting value (100 nM) had no impact on arteriole diameter in a pre-constricted arteriole. However, a moderate elevation to 250 nM elicited sustained vasodilation that was blocked by the COX-1 antagonist SC-560 (500 nM). The vasodilation to 250 nM Ca²⁺ was sensitive to the metabolic state, as it converted to vasoconstriction when oxygen tension was dramatically elevated. In normal oxygen, clamping astrocyte Ca²⁺ well above the resting level (750 nM) produced sustained vasoconstriction, which converted to vasodilation in the 20-HETE blocker HET0016 (1 μM). This response was fully blocked by the addition of SC-560 (500 nM), showing that 20-HETE-induced vasoconstriction dominated the dilatory action of COX-1. These data demonstrate that direct changes in astrocyte free Ca²⁺ can control multiple arteriole tone states through different mediators.

1. Introduction

Ca²⁺-dependent signaling in astrocytic endfeet contributes to the regulation of arteriole diameter [1,2,3]. While different transmitters and G-protein-coupled receptors drive distinct cell pathways that cause different vascular responses, it is notable that within the neurovascular unit, even the same transmitter and receptor system can elicit vasoconstriction or vasodilation depending on the physiological context or the current state of the microenvironment. These include the level of free Ca²⁺ in astrocyte endfeet [4], the degree of arteriole tone [5,6], and the metabolic status of the tissue [7,8]. These different scenarios may not be completely mechanistically separate, as both increases in arteriole tone and low oxygen levels elevate astrocyte free Ca²⁺ [6,9,10]. In acute brain slices, when using afferent stimulation or direct Ca²⁺ uncaging in astrocytes, a small elevation in free Ca²⁺ above baseline yielded vasodilation of penetrating arterioles in the neocortex, whereas a larger elevation in free Ca²⁺ caused vasoconstriction [4]. A similar concept was proposed to explain differential vessel responses to seizure activity in a 4-AP model of epilepsy in vivo [11] and the observation of vasoconstriction during spreading depression in vivo where the intracellular Ca²⁺ level became high [12]. The [Ca²⁺]_i crossover point to see these different responses was determined to be ~500 nM in astrocytes by calculating the concentration based on the observed ∆F/F and by determining the max ∆F/F at the end of each experiment with a Ca²⁺ ionophore [4]. An alternative method for testing this hypothesis is to directly ‘clamp’ astrocyte free Ca²⁺ at different concentrations via a patch pipette, though this method has not yet been performed to either support or refute the hypothesis.

Direct Ca²⁺ clamping of astrocytes also allows one to explore how sustained changes in astrocyte-free Ca²⁺ impact vascular tone. For example, it is unclear if arteriole diameter tracks long-lasting changes in astrocyte Ca²⁺ or if any evoked vascular response is transient in nature. This is important because most investigations have focused on brief vascular responses to astrocyte Ca²⁺ transients. However, astrocytes have been more recently implicated in the regulation of ongoing arteriole tone, independent of neural activity. This may occur through the steady-state Ca²⁺ concentration in astrocyte/Müller cell endfeet [13,14], relatively small elevations in endfoot Ca²⁺ evoked by static or rhythmic arteriole constrictions [6], and/or elevations in endfoot Ca²⁺ in response to increases in intraluminal pressure [9]. Similar tonic effects on astrocyte Ca²⁺ and arteriole diameter have also been observed after plasticity-inducing trains of afferent stimulation in neocortical slices [15]. Nevertheless, it remains unclear how directly maintaining astrocyte free Ca²⁺ at specific concentrations affects the arteriole response. We tested the hypothesis that arteriole tone tracks sustained changes to astrocyte free Ca²⁺ in a bidirectional manner depending on the absolute Ca²⁺ concentration.

2. Materials and Methods

All procedures abided by Canadian standards for animal research and were approved by the Animal Care and Use Committee of the University of Calgary (protocol AC19-0109). Male Sprague Dawley rats (P23-30, Charles River, Wilmington, MA, USA) received an intravenous injection of fluorescein isothiocyanate–dextran (FITC–dextran) (Sigma-Aldrich, St. Louis, MO, USA, 2000 KDa; 15 mg in 0.4 mL lactated ringers) to visualize the vasculature. Acute coronal slices of the sensory–motor cortex were cut with a vibratome (Leica VT1200S) while submerged in a slicing solution containing (in mM): N-methyl d-glucamine (119.9), KCl (2.5), NaHCO₃ (25), CaCl₂–2H₂O (1.0), MgCl₂–6H₂O (6.9), NaH₂PO₄–H₂O (1.4), and glucose (20). Brain slices then recovered in artificial cerebrospinal fluid (ACSF) continuously bubbled with carbogen (95% O₂ and 5% CO₂) at 34 °C for 45 min. ACSF contained (in mM): NaCl (126), KCl (2.5), NaHCO₃ (25), CaCl₂ (1.5), MgCl₂ (1.2), NaH₂PO₄ (1.25), and glucose (10). Slices were bulk-loaded with Rhod-2/AM (Biotium Inc.) at 15 µM (0.2% DMSO; 0.006% pluronic acid; 0.0002% Cremophor EL (Sigma-Aldrich)) for 45 min in carbogen-bubbled ACSF. Slices were transferred to ACSF continuously bubbled with physiological levels of oxygen (30%) and were provided 45 min to equilibrate. In a subset of experiments, slices were maintained in a high oxygen condition (carbogen). The arteriole pre-constrictor U-46619 (Cayman Chemical, Ann Arbor, MI, USA) (100 nM) was added to the bath for all experiments. Fluorescence imaging was performed with a custom two-photon laser-scanning microscope [16] equipped with a 40x W/1.0 NA Zeiss objective lens and a Chameleon Ultra Ti:Sapph laser (Coherent, Santa Clara, CA, USA). Imaging was performed at 22 °C, and slices were superfused with ACSF (aerated with 30% or 95% O₂) at ~2 mL/min.

Astrocytes were identified by bright Rhod2/AM uptake, endfeet opposing the vasculature, and input resistance (10–20 mΩ). Astrocytes that were ~30–60 microns from the coronal surface of the slice and less than ~60 microns from a penetrating arteriole were targeted for patch clamp. A giga-ohm seal was maintained for ~10–20 min prior to imaging to allow the effects of the patching process on arteriole tone and Ca²⁺ signal to subside. We recorded 100 s of stable baseline prior to going whole-cell and allowing the internal solution (visualized by Alexa 488; 200 µM) to diffuse throughout the astrocyte network. Astrocyte Ca²⁺ was clamped at different target concentrations by patching with internal solutions containing different ratios of BAPTA and free Ca²⁺. The base internal solution contained (in mM): potassium gluconate (68), KCl (8), sodium gluconate (8), MgCl₂ (2), HEPES (10), potassium ATP (4), sodium GTP (0.3), and potassium BAPTA (10). To clamp astrocyte Ca²⁺ at different concentrations, the CaCl₂ concentration was adjusted in each internal solution. The 100, 250, and 750 nM Ca²⁺ clamp solutions contained 0.043, 0.067, and 0.086 mM CaCl₂, respectively, as derived with the Maxchelator Calculator [17]. Once in whole-cell configuration, astrocytes were voltage clamped at −80 mV. A stable holding current and input resistance were monitored throughout the experiment, and the rate and extent of Alexa-488 filling in endfeet were quantified to ensure comparable results across groups. ImageJ (National Institutes of Health, Bethesda, MD, USA) and Prism 7 software (GraphPad, La Jolla, CA, USA) were used to analyze the data.

Rhod-2/AM intensity over time was analyzed to measure relative changes in astrocyte endfoot Ca²⁺ signal as ΔF/F = ((F₁ − F₀)/F₀) × 100. Lumen area was measured to assess changes in arteriole diameter using particle analysis in ImageJ. The baseline Rhod-2/AM signal and arteriole diameter were compared to peak values, the latter of which were determined by examining the average trace data to find the time in which the peak occurred. This time point was then used to extract a measurement from each of the individual experiments for statistical comparison. p < 0.05 was deemed statistically significant. Paired student’s t-tests were used to assess changes within groups (comparing baseline to peak). Unpaired t-tests were used to compare the 250 nM Ca²⁺ clamp in 30–95% O₂ conditions. A one-way ANOVA was used to compare peak means across groups with Tukey’s multiple comparisons.

3. Results

3.1. Clamping Astrocyte Ca²⁺ to Different Elevated Levels Elicits Opposite Arteriole Tone States

Using a whole-cell patch technique designed to buffer free Ca²⁺ to a particular target concentration using BAPTA plus CaCl₂ [18], we tested the hypothesis that clamping astrocyte Ca²⁺ near resting physiological levels would have little impact on existing arteriole tone. Quantitative Ca²⁺ imaging studies have reported resting astrocyte Ca²⁺ levels of ~80–130 nM [19,20,21], so we clamped astrocyte Ca²⁺ at 100 nM. Neocortical brain slices were bulk-loaded with a Rhod-2/AM Ca²⁺ indicator, arterioles were pre-constricted with U-46619 (100 nM), and the tissue was equilibrated in physiological levels of O₂ (30%) before the start of the astrocyte patching experiment. After sealing onto an astrocyte soma, on-cell configuration was maintained for ~15 min before going whole-cell. The time course and extent of internal solution diffusion into the patched astrocyte, its adjoining endfoot, and additional spread throughout the astrocyte network were visualized with Alexa-488 (200 µM) [13]. Following the arrival of the Alexa-488 signal (along with BAPTA-base Ca²⁺-clamp solution) into astrocyte endfeet, we observed no significant change in arteriole diameter (2.2 ± 1.0%, n = 5, p > 0.05; Figure 1A–D). To test whether moderately elevating astrocyte Ca²⁺ produced vasodilation, we patched astrocytes to deliver a BAPTA-containing internal solution designed to clamp intracellular Ca²⁺ at 250 nM, which was above an approximate resting value of 100 nM [21]. In response to this manipulation, we observed a sustained vasodilation (12.6 ± 2.7%, n = 8, p < 0.01; Figure 1E,G,H) that lasted the duration of the recording (~8 min). We observed a significant increase in relative endfoot Ca²⁺ signal (ΔF/F_Rhod-2 = 37.0 ± 15.1%, p < 0.05; Figure 1F–H), thus confirming that the 250 nM Ca²⁺ solution elevated Ca²⁺ from baseline concentration. We then tested whether the vasodilation evoked by 250 nM free Ca²⁺ in astrocytes was sensitive to the metabolic status of the tissue, as dictated by oxygen availability. Similar to previous reports [7,8], increasing the O₂ to supra-physiological levels (95%) switched the vasodilation to vasoconstriction (−7.4 ± 3.0%, n = 7, p < 0.01; Figure 1N,S) despite the same increase in Ca²⁺ within astrocytes (p = 0.3; Figure 1Q,S).

Next, to examine the effect of a high increase in endfoot Ca²⁺ on arteriole tone at physiological levels of O₂, we clamped astrocyte Ca²⁺ at 750 nM. The arrival of Alexa-488 to the endfeet corresponded with a long-lasting arteriole constriction (−8.4 ± 1.6%, n = 8, p < 0.01; Figure 1I,K,L) and a rapid, sustained increase in relative Ca²⁺ signal (ΔF/F_Rhod-2 = 92.6 ± 36.3%, p < 0.05; Figure 1J–L). We measured the tau of the monophasic exponential arrival of the Alexa-488 signal in endfeet to control for differences in the time course of internal solution delivery between experimental conditions [13]. There was no significant difference in tau values between the various Ca²⁺ clamp groups (p > 0.05; Figure 1O,R). Consequently, the observed differences in both vascular tone and relative endfoot Ca²⁺ cannot be attributed to the differential kinetics of internal solution diffusion through the astrocyte network.

3.2. Bidirectional Changes in Arteriole Diameter Evoked by Elevations in Astrocyte Ca²⁺ Are Mediated by COX-1 and 20-HETE

We previously demonstrated that lowering astrocyte Ca²⁺ below the resting level with intracellular BAPTA next to a pre-constricted arteriole induced vasoconstriction. This effect was eliminated by the COX-1 blocker SC-560 and astrocyte Ptgs1 knockdown [6]. This suggested that the resting endfoot Ca²⁺ level (in part determined by arteriole tone itself) produced the tonic release of vasodilators via COX-1 activity. To determine whether 250 nM Ca²⁺-clamp-evoked vasodilation was mediated by the further activation of COX-1, we repeated this experiment in the presence of SC-560 (500 nM) (Figure 2A–D). Pre-treatment of brain slices with this COX-1 antagonist blocked 250 nM Ca²⁺ clamp-evoked vasodilation (−1.3 ± 1.6%, n = 5, p = 0.007), despite no differences in either the relative Ca²⁺ increase (p = 0.2) of the rate of endfoot filling (p = 0.9) (Figure 2D).

High astrocyte Ca²⁺-induced vasoconstriction has been previously linked to arachidonic acid conversion to the potent vasoconstrictor 20-HETE [2,22], ATP release and P2Y1 receptor activation [23], and BK channel opening with K⁺ efflux [4]. Given that our Ca²⁺ clamp approach at 250 nM was generating AA due to COX-1 recruitment, we tested whether 750 nM astrocyte Ca²⁺ resulted in enough AA to drive the 20-HETE constriction pathway to dominate vasodilatory prostaglandins. Indeed, the presence of the CYP450 blocker HET0016 (1 µM), used to block the synthesis of 20-HETE, reversed the vascular response to 750 nM astrocyte Ca²⁺ from a constriction (−8.4 ± 1.6, n = 8) to a dilation (14.3 ± 4.3%, n = 5, p = 0.009) (Figure 2E–G). To test the idea that this reversal in diameter response in HET0016 was due to an intact vasodilatory COX-1 pathway in high intracellular Ca²⁺, we pre-treated slices with HET0016 plus SC560. In this cocktail, 750 nM astrocyte Ca²⁺ elicited no significant changes in vascular tone (2.9 ± 1.0%, n = 5, p < 0.0001) (Figure 2G–J).

4. Discussion

Here, we found that two different and direct elevations from the resting astrocyte Ca²⁺ concentration controlled opposite arteriole tone states, whereas clamping astrocyte Ca²⁺ at the presumed resting level caused no change in diameter. Specifically, we observed that clamping astrocyte Ca²⁺ at a moderate level above resting (250 nM) elicited vasodilation that was eliminated by the COX-1 antagonist SC-560 (500 nM). A high Ca²⁺ concentration (750 nM) produced arteriole constriction that was reversed to dilation by the 20-HETE blocker HET0016 (1 μM) and fully eliminated by a cocktail of SC-560 (500 nM) and HET0016 (1 μM). Though previous investigations have predominantly examined vascular responses to astrocyte Ca²⁺ transients, our data showed that arteriole tone tracks sustained changes to astrocyte free Ca²⁺.

A previous study using calculation-based [Ca²⁺]_i estimation in astrocytes placed the crossover point for astrocytes evoking vasodilation to vasoconstriction at ~500 nM [4]. This corresponds well with our data, in which clear opposite arteriole responses were observed when the Ca²⁺ concentration straddled this value at 250 and 750 nM using patch pipettes to directly control free Ca²⁺. The previous work describing opposing arteriole responses to elevations of free Ca²⁺ to different levels provided mechanistic data in support of BK channels in endfeet [4]. While our work supports AA metabolites in these patch-evoked effects, BK channels were not explored, though they may have contributed to our described effects via the PGE2-mediated opening of BK channels [24]. Alternatively, prostaglandin release and K⁺ efflux from endfeet may cooperatively act on vascular smooth muscle membrane potential.

This vascular polarity-switching phenomenon occurs in other conditions. Previously, the metabolic state of brain tissue determined the polarity of vascular responses to Ca²⁺ uncaging events in astrocytes [7] or neural activity [8]. Consistent with this, we found that clamping astrocyte Ca²⁺ at 250 nM elicited arteriole dilation in 30% O₂ and vasoconstriction was produced in supra-physiological O₂ (95%). The effector molecules released by physiological levels of O₂ may be lactate and adenosine, which help promote vasodilation and block vasoconstriction, respectively [7]. Additionally, large differences in arteriole tone induced by U-46619 permit opposite arteriole responses to the same mGluR agonist, t-ACPD [5]. When strongly constricted, t-ACPD produces vasodilation, and when the vessel is much more relaxed, t-ACPD generates vasoconstriction. However, in our study, we used a consistent and constant degree of tone (~20%) elicited by U-46619 (100 nM) across experiments; thus, differences in tone were unlikely to be a major factor, though this may have contributed to within-group variability.

Our data provide support for the idea that 20-HETE-mediated vasoconstriction dominates prostaglandin-mediated vasodilation pathways. Similar trends, in which constriction overpowers active dilation pathways were reported for oxygen and for K⁺. For O₂, astrocyte-mediated vasoconstriction dominated dilation in the high O₂ condition unless adenosine and lactate were provided to the extracellular milieu [7]. For K⁺, ongoing K⁺-mediated dilation can be overcome by vasoconstriction when a certain threshold concentration of K⁺ was reached [4]. Such a switch to an entirely different vessel response is readily explained by the sudden opening of L-type Ca²⁺ channels in vascular smooth muscle as threshold membrane potential is reached, either through the action of different mediators (PGE2 vs. 20-HETE) or when external K⁺ becomes high enough to drive intracellular depolarization rather than hyperpolarization [1,2,25,26,27].

An important limitation of our work was the use of acute brain slices, maintained at room temperature and containing arterioles that lack blood flow and perfusion pressure. Though an unrealistic preparation, our data, which show vasodilation at modest levels of astrocyte free Ca²⁺ and vasoconstriction at higher levels, are consistent with in vivo studies [11,12,25], one of which employed quantitative Ca²⁺ measurements with fluorescence lifetime imaging microscopy (FLIM) [11]. Another limitation was our inability to confirm the [Ca²⁺]_i-clamp achieved in endfeet. Though we directly patched astrocytes abutting arterioles, both the concentration of BAPTA and free Ca²⁺ in the Ca²⁺-clamp solution was somewhat lower in the endfeet compared to the patch pipette and significantly lower in neighboring astrocytes as measured by FLIM [21]. Assuming equal reductions in both [BAPTA] and [Ca²⁺] as distance increases from the pipette within the patched astrocyte, the clamp should still occur at the same intended Ca²⁺ concentration but the overall buffering capacity of the solution will be less (i.e., less able to deal with increases or decreases away from this equilibrium). Future work could employ quantitative Ca²⁺ measurements to understand the exact changes, but the opposite vascular responses we measured under our different patch conditions indicated clear differences in the achieved Ca²⁺ concentrations.

The role of astrocytes in brief, stimulus-evoked, functional hyperemia is debated [28,29,30,31,32,33,34,35]. However, astrocytes express COX-1 [36,37] and downstream microsomal prostaglandin synthase [38], and this pathway may contribute to the regulation of arteriole diameter [39,40]. However, several groups have reported no role of COX-1 in functional hyperemia [36,41,42]. Clarifying this, recent work demonstrated that at least one role of COX-1 in astrocytes is to help constrain the degree vasoconstriction through an arteriole–endfoot feedback mechanism [6]. In vivo, astrocytic Ptgs1 knockdown experiments revealed that COX-1 helped constrain the power of rhythmic oscillations in arteriole diameter over a wide range of very low frequencies [6]. Additional or complementary roles for astrocytic-COX-1 could include mediating CO₂-induced vasodilation [38] or mediating capillary dilation to afferent neural activity [37]. At this point, it remains unclear which of these roles is relevant for our 250 nM Ca²⁺-recruitment of COX-1. The recruitment of 20-HETE at high levels of Ca²⁺ may be important for neuromodulators such as noradrenaline that evoke large Ca²⁺ elevations in astrocytes and vasoconstriction through this pathway [2] or during pathological events such as spreading depression when astrocyte Ca²⁺ is aberrantly elevated [43,44]. Nevertheless, we have demonstrated that these pathways can be constantly engaged in a steady-state manner depending on the set concentration of astrocyte Ca²⁺. Additionally, our data further support the idea that the level of endfoot Ca²⁺ helps dictate the polarity of the arteriole response, that such responses are sensitive to the metabolic micro-environment, and that constrictions readily dominate active dilation pathways.