Role of epoxyeicosatrienoic acids as autocrine metabolites in glutamate-mediated K⁺ signaling in perivascular astrocytes

Abstract

Epoxyeicosatrienoic acids (EETs), synthesized and released by astrocytes in response to glutamate, are known to play a pivotal role in neurovascular coupling. In vascular smooth muscle cells (VSMC), EETs activate large-conductance, Ca²⁺-activated K⁺ (BK) channels resulting in hyperpolarization and vasodilation. However, the functional role and mechanism of action for glial-derived EETs are still to be determined. In this study, we evaluated the effect of the synthetic EET analog 11-nonyloxy-undec-8(Z)-enoic acid (NUD-GA) on outward K⁺ currents mediated by calcium-activated K⁺ channels. Addition of NUD-GA significantly increased intracellular Ca²⁺ and outward K⁺ currents in perivascular astrocytes. NUD-GA-induced currents were significantly inhibited by BK channel blockers paxilline and tetraethylammonium (TEA) (23.4 ± 2.4%; P < 0.0005). Similarly, NUD-GA-induced currents were also significantly inhibited in the presence of the small-conductance Ca²⁺-activated K⁺ channel inhibitor apamin along with a combination of blockers against glutamate receptors (12.8 ± 2.70%; P < 0.05). No changes in outward currents were observed in the presence of the channel blocker for intermediate-conductance K⁺ channels TRAM-34. Blockade of the endogenous production of EETs with N-methylsulfonyl-6-(2-propargyloxyphenyl)hexanamide (MS-PPOH) significantly blunted (dl)-1-aminocyclopentane-trans-1,3-dicarboxylic acid (t-ACPD)-induced outward K⁺ currents (P < 0.05; n = 6). Both NUD-GA and t-ACPD significantly increased BK channel single open probability; the later was blocked following MS-PPOH incubation. Our data supports the idea that EETs are potent K⁺ channel modulators in cortical perivascular astrocytes and further suggest that these metabolites may participate in NVC by modulating the levels of K⁺ released at the gliovascular space.

in recent years, significant progress has been made in the understanding of the signaling mechanisms underlying neurovascular coupling (NVC) in the brain (22, 26). Epoxyeicosatrienoic acids (EETs), metabolites of arachidonic acid (AA), are among the various proposed signaling candidates contributing to NVC (3, 42). In addition to their potent vasodilatory action (19, 24), EETs have been associated with a number of physiological functions such as angiogenesis (46, 53) and anti-inflammation (30, 43). Both in vitro and in vivo data have validated the importance of these metabolites in cerebral blood flow (CBF) regulation (32, 33, 39). In addition, changes in EET signaling have been observed in disease conditions such as hypertension and stroke (14, 38, 40, 51, 54, 55). It has been reported that as a compensatory mechanism, soluble epoxide hydrolase (sEH), the enzyme that converts EETs to their corresponding diols, is increased in hypertensive animals (1, 14, 52). For example, inhibition of sEH significantly lowered arterial blood pressure in angiotensin II-induced hypertensive animals (25). In spontaneously hypertensive stroke-prone rats, similar treatment significantly reduced cerebral infarct size following middle cerebral artery occlusion (12, 40). These findings highlight the importance of a detailed study addressing the mechanism of action for EETs at the neurovascular unit, the site where multiple target cells (neuronal terminals, astrocytes, and vascular cells) interact to optimize CBF to the brain.

The current hypothesis in the field suggests that EETs, released by activated astrocytes (2), diffuse to vascular smooth muscle cells (VSMC) where they induce vasodilation (23, 26). Specifically, glutamate-induced activation of metabotropic glutamate receptors (mGluR) in astrocytes results in an increase in intracellular Ca²⁺ triggering AA mobilization and CYP 2C11 activation, which leads to the formation and release of all four EETs regioisomers (5,6-, 8,9-, 11,12-, and 14,15-EETs) (2, 4, 9, 29).

The signaling mechanism by which astrocyte-derived EETs induced vasodilation is thought to occur through the activation of the large-conductance, Ca²⁺-activated K⁺ channel (BK) in VSMC (13, 27). In addition to the effects on VSMC, several lines of evidence support the idea that EETs may also act in an autocrine manner as seen by a significant increase in intracellular Ca²⁺ in cortical astrocytes (7) and the activation of BK channel currents in cultured hippocampal astrocytes (50). The rapid Ca²⁺ response, combined with the notion that astrocytic endfeet are enriched with BK channels (36), raises the possibility that EETs, in addition to their VSMC effects, may also contribute to K⁺ signaling in cortical astrocytes (17).

In a previous study, we showed that K⁺ signaling from astrocytic endfeet, via BK channel activation, induced vasodilation during NVC (17). Moreover, 11,12-EET-induced vasodilations were blocked in the presence of the BK channel blocker tetraethylammonium (TEA) (7). In cultured hippocampal astrocytes, Gebremedhin et al. (20) showed that inhibition of endogenous EETs formation with miconazole blunted the glutamate-induced increase in BK single channel open probability following mGluR activation (20). Based on these findings, we hypothesized that endogenously produced EETs may act in an autocrine manner, modulating BK channels expressed at the very same site where these metabolites are released, namely the gliovascular space. In this way, following neuronal activation astrocyte-derived EETs would sustain astrocytic K⁺ efflux, thus prolonging vasodilation.

Using patch-clamp electrophysiological recordings and calcium imaging from cortical perivascular astrocytes, we characterize the effects of the EETs analog 11-nonyloxy-undec-8(Z)-enoic acid (NUD-GA) on intracellular calcium responses and Ca²⁺-sensitive K⁺ currents.

METHODS

Brain slice preparation.

Cortical brain slices were prepared from juvenile (P23-31) Wistar rats following protocols approved by the Office of Animal Care Management at the Medical College of Georgia. The cortex was rapidly removed and cut into 300-μm thick coronal slices using a vibratome (Leica VT 1200S, Leica Microsystems, Wetzlar, Germany) in cold artificial cerebrospinal fluid (aCSF) (in mM): 3 KCl, 125 NaCl, 1 MgCl₂, 26 NaHCO₃, 1.25 NaH₂PO₄, 10 glucose, 2 CaCl₂, and 400 μM l-ascorbic acid, with osmolarity at 300–305 mosM, equilibrated with 95% O₂-5% CO₂. Slices were then incubated at room temperature (RT) in aCSF equilibrated with 95% O₂-5% CO₂ (pH 7.4) until needed.

Electrophysiology.

Cortical astrocytes in close proximity (<50 μm) to a parenchymal arteriole were identified using a ×63 water-immersion objective (numerical aperature 1.0). Whole cell patch-current recordings were obtained using an Axopatch 200B amplifier (Axon instruments, Foster City, CA). Patch pipettes were made from thin-walled borosilicate glass (outer diameter 1.5 mm, internal diameter 0.86 mm; Sutter instrument BF150-86-7.5) and pulled (P-97 puller Sutter Instruments, Novato, CA) to resistances between 5 and 7 mΩ. The internal solution for whole cell recordings consisted of (in mM) 130 K⁺ gluconate, 10 HEPES, 0.2 EGTA, 10 KCl, 0.9 MgCl₂, 4 Mg₂ATP, 0.3 Na₂GTP, and 20 phosphocreatine; pH adjusted to 7.2 with KOH. For intracellular labeling, 5% Lucifer yellow potassium salt (Invitrogen, Carlsbad, CA) was added to the pipette solution. The osmolarity of the intracellular solutions was 291–295 mosM. Current signals were filtered at 1 kHz low-pass filter and digitized with a Digidata 1320 board (Axon instrument). pClamp 9.2 (Axon instrument) was used for data acquisition and storage. Single-channel K⁺ currents were recorded using the cell-attached configuration from an astrocytic endfoot attached to a cerebral arteriole. The pipette solution for cell-attached recordings consisted of (in mM) 140 KCl, 1 MgCl₂, and 10 HEPES; pH adjusted at 7.2 with KOH and osmolarity at 280 mosM. Single channel activity was measured at +50 mV. Mean open state probability (NPo) was determined as previously described (11, 49). Single channel conductance was calculated from a linear slope fitted regression of the current-voltage (I-V) relationship obtained from channel current amplitudes at potentials between −60 and +60 mV. Perforated patch recordings were used to measure astrocytic endfeet whole cell currents for N-methylsulfonyl-6-(2-propargyloxyphenyl)hexanamide (MS-PPOH) experiments. The perforated pipette solution consisted of (in mM) 30 KCl, 110 K⁺ aspartate, 10 NaCl, 1 MgCl₂, and 10 HEPES adjusted pH to 7.2 with NaOH. 500 μg/ml amphotericin-B was added to the solution. Lucifer yellow (5%) was added to assure membrane integrity as shown by the lack of intracellular staining (37). All experiments were conducted at 33–35°C. The liquid junctional potential (+6.5 mV) was previously determined experimentally and described in Sonner et al. (41).

Confocal immunohistochemical imaging.

For glial fibrillary acidic protein (GFAP) immunostaining of recorded cells, cortical slices containing Lucifer yellow (5%) filled cells were fixed in 4% paraformaldehyde overnight. For colocalization studies of astrocytic endfeet and BK channel expression, paraformaldehyde-fixed, frozen brain sections were permeabilized and blocked in PBS containing 0.3% Triton X-100 and 10% horse serum for 1 h. Slices were incubated overnight at 4°C with mouse anti-GFAP (1:10,000; Chemicon) and rabbit-anti-BK α-subunit (1:100; Chemicon) antibodies. GFAP signal was detected with donkey anti-mouse CY3 (Jackson ImmunoResearch). Anti-BK channel antibody was labeled with anti-Rb biotinylated antibody followed by Streptavidin-FITC (Jackson ImmunoResearch). Images of Lucifer yellow-filled astrocytes or GFAP and BK channel-stained astrocytes were captured by Zeiss LSM510 confocal microscope with ×40 and ×63 magnification objectives using the Zeiss LSM510 program (Carl Zeiss, Thornwood, NY).

Calcium imaging.

Ca²⁺ imaging was performed using the Andor Technology Revolution system (iXON EMCCD camera with the Yokogawa CSU 10, confocal scanning unit). This unit was attached to a Zeiss microscope (Axioscope 2FS). Briefly, cortical slices were incubated at room temperature in aCSF containing 10 μM Fluo-4 AM and pluronic acid (2.5 μg/ml). After a 2-h incubation period, slices were placed in aCSF at room temperature until needed. At the time of the experiment, a slice was transferred to a perfusion chamber on the microscope and held with a nylon grid and continuously superfused with aCSF at 33–35°C. Using this loading protocol, we are able to visualize Ca²⁺ transients in astrocytes. In accordance with previous reports (34), neurons do not load sufficiently under these conditions for Ca²⁺ detection. Parenchymal microvessels were visualized with a ×63 water-dipping objective (numerical aperature 1.0). Fluorescence images were obtained using a krypton-argon laser at 488 nm and emitted light at >495. Images were acquired at 2 image/s for 5–8 min.

Drugs and chemicals.

The synthetic EET analog, the glycine amide of NUD-GA, and 14,15-epoxyeicosa-5(Z)-enoic acid (14,15-EEZE) were synthesized by Dr. J. R. Falck. All chemicals were obtained from Sigma Chemicals (St. Louis MO) otherwise noted. mGluR agonist (dl)-1-aminocyclopentane-trans-1,3-dicarboxylic acid (t-ACPD), metabotropic glutamate receptor5 (mGluR5) antagonists 2-methyl-6-phenylethynylpyridine hydrochloride (MPEP), mGluR1a receptor antagonist (S)-(+)-α-amino-4-carboxy-2-methylbenzeneacetic acid (LY-367385), CYP epoxygenase inhibitor MS-PPOH, and 1,2-bis(2-aminophenoxyl)ethane-N,N,N,N,-tetraacetic acid tetra (acetoxymethyl ester) (BAPTA-AM) were purchased from Tocris Bioscience (Ellisville, MS). Isolection-GS-IB₄-Alexa Fluor-568 conjugate was purchased from Invitrogen (Carlsbad, CA). Tetrodotoxin (TTX) was obtained from Ascent Scientific LLC (Princeton, NJ).

Data analysis.

Analysis for electrophysiological recordings was performed using Clampfit 9.2, pClamp version 9.2 (Axon instrument) and GraphPad Prism software 4 (Graphpad software, La Jolla, CA). Wilcoxon paired t-test was used to analyze current changes (in pA/pF) between control and each drug treatment. A one-way analysis of variance was used to compare current changes (pA/pF) between control and drug-treated groups presented in Fig. 3E. Ca²⁺ imaging was analyzed using Andor IQ 1.6 software. Fractional fluorescence (F/F_o) was determined by dividing the fluorescence intensity (F) within a region of interest (ROI) by a baseline fluorescence value (F_o) determined from ∼50 images showing no activity. The frequency of Ca²⁺ oscillations was determined within a ROI (10 × 10 pixels or 2.5 × 2.5 μm) on a cell exhibiting Ca²⁺ oscillations. The number of peaks over a given time was automatically detected from oscillations crossing a set threshold value (>1.30 F/F₀). All values are expressed as means ± SE. Statistical significance was tested at 95% (P < 0.05) confidence level and denoted with asterisks, when the P < 0.05 (*), P < 0.01 (**), or P < 0.001 (***).

Fig. 3.

Calcium and electrophysiological responses of cortical astrocytes to the EET agonist 11-nonyloxy-undec-8(Z)-enoic acid (NUD-GA). A: NUD-GA-induced calcium transients in cortical astrocytes, representative traces and fluorescence images are shown. B: summary data of NUD-GA-induced changes in averaged peak F/F₀ (***P < 0.0001, n = 11) and calcium oscillation frequency (**P < 0.01, n = 11). C and D: representative NUD-GA-induced outward voltage-dependent and linear membrane currents in response to a ramp protocol from −120 to +80 mV in perivascular astrocytes, respectively. The corresponding subtracted currents for C and D are shown below. E: summary data of peak currents (pA/pF) at +80 mV under control conditions, in the presence of NUD-GA, in the presence of 14,15-epoxyeicosa-5(Z)-enoic acid (14,15-EEZE) and in the presence of 14,15-EEZE and NUD-GA (*P < 0.005, n = 7). F: summary data of peak currents (pA/pF) at +80 mV in the presence and absence of the calcium chelator BAPTA (**P < 0.05, n = 5). Calibration bar, 10 μm.

RESULTS

Properties of cortical perivascular astrocytes.

Only astrocytes in the vicinity of a cortical arteriole were studied; these cells were identified by their lack of action potential firing and their distinct morphology, namely small round somata and, in some cases, endfeet extending to the nearest arteriole. Figure 1A shows a differential interference contrast (DIC) image of a perivascular astrocyte loaded with Lucifer yellow. After a successful seal, astrocytes were subjected to a voltage ramp protocol (from −120 mV to +80 mV, in 400 ms, Fig. 1B, inset). In whole cell configuration, mean resting membrane potential (V_m) was −72.4 ± 1.9 mV and input resistance 74.6 ± 3.7 mΩ (n = 81). Figure 1, B and C, shows representative traces of the two I-V response patterns obtained from cortical perivascular astrocytes. Of 81 recorded astrocytes, 60.5% exhibited a voltage-dependent outwardly rectifying pattern (Fig. 1B), whereas 39.5% showed a linear I-V response (Fig. 1C). These response patterns are comparable to those previously reported (8, 10, 56). The averaged reversal potential for the overall whole cell current was −77.01 ± 0.52 mV, near the expected E_K+.

Fig. 1.

Representative whole cell current profiles from cortical perivascular astrocytes. A: differential interference contrast (DIC) image of a Lucifer yellow-filled astrocyte in direct contact to an intraparenchymal arteriole. Calibration bar, 10 μm. Representative voltage-dependent (B) and linear membrane currents (C) in response to a ramp protocol from −120 to +80 mV.

Functional implication of CYP 2C11 activity in EETs-induced outward K⁺ currents in cortical astrocytes.

The vasodilatory action of EETs has been attributed to their hyperpolarizing effect on VSMC, resulting from the activation of BK channels in these cells (13, 19, 23, 26, 27). Likewise, EETs also induce BK channel activation in cultured hippocampal astrocytes (50). Since there is evidence that astrocytes synthesize EETs (3), and this synthesis is enhanced by glutamate (2), we hypothesized that endogenously produced EETs may act in an autocrine manner, modulating BK channels expressed at the very same site where these metabolites are released, namely, the gliovascular space. In this way, astrocyte-derived EETs would sustain astrocytic K⁺ efflux following neuronal activation, thus prolonging vasodilation.

To test our hypothesis, brain slices were preincubated with MS-PPOH (20 μM for 30 min), a selective CYP epoxygenase substrate inhibitor (45), followed by stimulation of astrocytes with the mGluR agonist t-ACPD (100 μM). We speculated that following astrocytic mGluR activation, K⁺ currents would be substantially attenuated due to the lack of EETs-related autoregulatory effects. As expected, in the absence of MS-PPOH application, t-ACPD (100 μM) elicited a significant increase in outward currents (64 ± 29.8%; P < 0.05; n = 6), as measured at +80 mV (Fig. 2A). On the other hand, in slices preincubated with MS-PPOH, t-ACPD induced a nonsignificant increase of only 4.4 ± 2.6% in K⁺ outward currents (at +80 mV; n = 5; Fig. 2B). These data provide, for the first time, strong evidence that most of the K⁺ current (∼96%) elicited by mGluR activation in perivascular astrocytes in situ depends on endogenous EET synthesis.

Fig. 2.

Effect of endogenous epoxyeicosatrienoic acids (EETs) on outward currents induced by activation of metabotropic glutamate receptors (mGluR) in perivascular astrocytes. (dl)-1-Aminocyclopentane-trans-1,3-dicarboxylic acid (t-ACPD)-induced (100 μM) outward currents in the absence (A) and presence (B) of the CYP epoxygenase substrate inhibitor N-methylsulfonyl-6-(2-propargyloxyphenyl)hexanamide (MS-PPOH) (20 μM) (solid black line: control, solid gray line: t-ACPD, dashed line: subtracted current) are shown. C: summary data of averaged peak currents (pA/pF) at V_m = 80 mV induced by t-ACPD in the presence and absence of MS-PPOH and the large-conductance, Ca²⁺-activated K⁺ (BK) channel blocker paxilline. D: representative traces of single BK channel activation induced by t-ACPD in the presence and absence of MS-PPOH. E: changes in single channel open probability (NPo) induced by t-ACPD in the presence and absence of MS-PPOH (*P < 0.05, n = 4).

To assess the participation of BK channels in mGluR-induced outward currents, experiments were repeated in the presence of the BK channel blocker paxilline (2 μM). Under these conditions, t-ACPD-induced currents were significantly attenuated (by 87.6%) compared with control responses (+80 mV; P < 0.05; n = 5). Importantly, in slices preincubated with MS-PPOH, t-ACPD-induced outward currents were partially restored in the presence of NUD-GA (Fig. 2C). Finally, the activity of single BK channels at the astrocytic endfoot in response to t-ACPD was assessed using the cell attached configuration. After t-ACPD exposure, a ∼16-fold increase in single channel NPo (from 0.0185 ± 0.003 to 0.301 ± 0.033; P < 0.005, n = 4) was observed (Fig. 2, D and E). Similarly, BK single channel recordings showed that mGluR-mediated increases in NPo were significantly reduced following MS-PPOH incubation (from 0.0277 ± 0.004 to 0.149 ± 0.025). Altogether, these studies support BK channels as mediators of EETs-induced outward currents.

EET analog NUD-GA elicits calcium-dependent K⁺ currents in perivascular astrocytes.

Given the importance of K⁺ channels, we investigated in detail the effects of EETs on outward K⁺ currents. We previously showed that 11,12-EET triggers potent Ca²⁺ responses in cortical astrocytes (7). To verify that similar responses could be induced by the EET analog used in the present study, brain slices were perfused with NUD-GA (1 μM). As shown in Fig. 3, A and B (see online supplementary movie at the AJP-Cell Physiology website), NUD-GA significantly increased both the frequency of astrocytic Ca²⁺ oscillations, from 0.013 ± 0.003 to 0.030 ± 0.005 Hz (P < 0.005; n = 11 slices) as well as the peak averaged fluorescence (F/F₀) from 1.05 ± 0.02 to 2.26 ± 0.13 (P < 0.0001; n = 11 slices).

Next, we determined whether the robust Ca²⁺ response induced by NUD-GA could lead to the activation of Ca²⁺-sensitive K⁺ channels. After a 5-min exposure, NUD-GA (1 μM) significantly increased astrocytic outward K⁺ currents by 37.5 ± 11.9% at +80 mV (P < 0.001, n = 10) (Fig. 3, C and D). The reversal potential under these conditions was not altered. No differences were observed when comparing the magnitude of the NUD-GA-induced response between astrocytes displaying voltage-dependent or linear ramp responses. Thus data from these two populations were pooled. In a subgroup of cells, we tested NUD-GA specificity by measuring astrocytic outward K⁺ currents in the presence or absence of the specific EETs antagonist 14,15-EEZE. In the presence of 14,15-EEZE (10 μM), NUD-GA-induced increases in outward currents at 80 mV were significantly decreased from 29.72 ± 4.5% to 12.54 ± 2.5% (n = 7, P < 0.005) (Fig. 3E).

To confirm the Ca²⁺ dependence of the observed responses, two sets of experiments were conducted. First, brain slices were incubated with the Ca²⁺ chelator BAPTA-AM (10 μM) for 30 min and the electrophysiological response to NUD-GA was determined. In a second set of experiments, BAPTA (10 mM) was added to the pipette solution and allowed to dialyze and equilibrate into the cell for 10 min before exposure to NUD-GA. Under both experimental conditions, the stimulatory effect of NUD-GA on K⁺ currents was abolished, indicating that a raise in intracellular Ca²⁺ mediates such effect (Fig. 3F). We next proceeded to study the identity of the K⁺ channels responsible for the outward currents elicited by NUD-GA.

NUD-GA-induced changes to BK channel currents in cortical perivascular astrocytes.

Examination of the I-V plots shown in Fig. 3, C and D, revealed a more complex pattern than that expected from the activation of a single channel type, suggesting that multiple K⁺ conductances may be activated in astrocytes upon exposure to the EET analog NUD-GA. Whereas the BK channel has been shown to be the effector of outward K⁺ currents induced by EETs in cultured hippocampal astrocytes and VSMCs (19, 50) and is thus a main candidate to mediate NUD-GA-induced currents in our preparation, other potential contributors to this current are the intermediate- (IK) and small-conductance (SK), Ca²⁺-sensitive K⁺ channels IK and SK, respectively. Therefore, we employed pharmacological inhibitors to evaluate the contribution of each of these channels to the NUD-GA-sensitive K⁺ current.

We first determined the contribution of BK channels to whole cell outward currents in the absence of the EETs agonist. To this end, ramp protocols were run in the presence and absence of BK channel blockers paxilline (2 μM) and TEA (1 mM). In the presence of both inhibitors, outward K⁺ currents were significantly reduced by 23.4 ± 2.4% at +80 mV (P < 0.0005, n = 11) (Fig. 4A). A similar attenuation was observed in the presence of the specific BK channel inhibitor paxilline alone (23.1 ± 2.6%; P < 0.05, n = 6) (Fig. 4C).

Fig. 4.

Contribution of BK channels to NUD-GA-induced outward currents. A: representative current profiles of perivascular astrocytes in the presence and absence of the BK channel blocker paxilline (2 μM) and tetraethylammonium (TEA, 1 mM) in response to a ramp protocol from −120 to +80 mV (P < 0.001, n = 11). Subtracted currents are shown below. B: representative NUD-GA-induced outward K⁺ currents from an astrocyte in the presence of the BK channel blocker paxilline and TEA. Corresponding subtract currents are shown below. C: summary data of peak currents at +80 mV in the presence (*P < 0.05, n = 7) and absence of BK channel blockers.**P < 0.01; ***P < 0.001 D: left, immunolabeling against BK channels (green) and the astrocyte marker glial fibrillary acidic protein (GFAP) (red) (scale bar = 10 μm); right, representative single channel activity from an astrocytic endfoot in the absence and presence of NUD-GA, O, open state and C, closed state. E: left, averaged current-voltage relationship for BK channels expressed in astrocytic endfeet; right, NUD-GA-induced increased in NPo (**P < 0.01, n = 4).

We next determined the contribution of BK channels to NUD-GA-induced K⁺ outward currents. In the presence of both TEA and paxilline, NUD-GA still elicited a significant increase in outward currents (11.5 ± 1.8%; P < 0.05; n = 7), which represents, however, a decrease of 69.3% compared with responses to NUD-GA alone (compare Fig. 3C with 4B). These data indicate that about 70% of the NUD-GA-sensitive current in perivascular astrocytes can be attributed to activation of BK channels.

To further confirm and to assess the effects of NUD-GA on the properties of single BK channels, we next performed cell-attached recordings on astrocytic endfeet, the site where these channels were previously shown to be highly concentrated (36). In agreement with previous studies, we found a robust BK channel immunoreactivity in clusters around parenchymal microvessels and which colocalized with the astrocyte-specific marker GFAP (Fig. 4D). The EETs analog NUD-GA caused a 12-fold increase in single channel NPo from 0.0231 ± 0.041 to 0.284 ± 0.028 (P < 0.05; n = 4) (Fig. 4, D and E). The calculated slope conductance for these channels was 171 pS. This value is consistent with the range of single channel conductance previously shown for BK channels in cultured astrocytes (16, 39) (Fig. 4E). These data further confirm the involvement of BK channels in NUD-GA-induced K⁺ currents in perivascular astrocytic endfeet, giving strength to the idea that EETs modulate astrocyte K⁺ efflux, being a likely mechanism affecting vascular tone in the mammalian brain (7, 15).

NUD-GA-induced changes to IK channel currents in cortical perivascular astrocytes.

The incomplete NUD-GA-induced current block observed upon BK channel inhibition suggests that other K⁺ channels are also activated upon exposure to this EET analog. To evaluate the potential contribution of IK channels on NUD-GA-induced K⁺ currents, whole cell recordings were performed in the presence of the specific IK channel blocker TRAM-34 (1 μM). TRAM-34 did not alter the I–V relationship of cortical astrocytes either in control conditions (no agonist; Fig. 5A) or upon combined exposure with NUD-GA (Fig. 5B; n = 6). In addition, in the presence of TRAM-34, NUD-GA induced similar changes in current amplitude at all different voltages at (+80 mV; 22.52 ± 3.7%) suggesting that cortical astrocytes do not possess functional IK channels or that these channels are not activated by the EET analog NUD-GA (Fig. 5B).

Fig. 5.

Lack of contribution of intermediate conductance (IK) channels to NUD-GA-induced outward currents. A: representative ramp profile of a perivascular astrocyte in the presence and absence of the IK channel blocker TRAM-34 (1 μM). B: representative ramp profile of a perivascular astrocyte in the presence and absence of NUD-GA in the presense of TRAM-34.

NUD-GA-induced changes to SK channel currents in cortical perivascular astrocytes.

Finally, we used the specific SK channel blocker apamin to determine whether SK channels, reportedly expressed in astrocytes from the supraoptic nucleus of the hypothalamus (5), contribute to the increase in outward K⁺ currents induced by NUD-GA. Unexpectedly, in the presence of the SK channel blocker apamin (300 nM), control currents were significantly increased (18.4 ± 5.6% at +80 mV; P < 0.05; n = 7) (Fig. 6A). We reasoned this response could be attributed to the inhibition of neuronal SK channels leading to the release of glutamate and subsequent astrocytic activation (31), as it would be possible in the slice preparation. Under these conditions, the concomitant rise in astrocytic Ca²⁺ expected to result from the activation of mGluR (35) may explain the rise in outward currents in the presence of apamin. To test this possibility, the intracellular Ca²⁺ response to apamin was measured in cortical astrocytes. As shown in Fig. 6, D and E, apamin (300 nM) significantly increased both the frequency (0.009 ± 0.006 to 0.03 ± 0.008; P < 0.05, n = 7 slices) and the amplitude (F/F₀) (1 ± 0.002 to 3.66 ± 0.5; P < 0.01, n = 7 slices) of astrocytic intracellular Ca²⁺ oscillations. These data confirm that SK channel blockade in the slice preparation induces astrocytic Ca²⁺ elevations. Attesting to the involvement of neuronal activity in this response, the apamin-mediated increases in astrocytic Ca²⁺ oscillation frequency and amplitude were abolished upon combined inhibition of Na⁺ channels (0.5 μM TTX) and glutamate receptors: ionotropic (1 mM kynurenic acid) and group I mGluR (50 μM LY-367385 and 100 μM MPEP) (n = 9, data not shown).

Fig. 6.

Contribution of small conductance (SK) channels to NUD-GA-induced outward currents. A: representative ramp profiles of perivascular astrocytes in the presence and absence of the SK channel blocker apamin (300 nM) (P < 0.05, n = 7), subtracted currents are shown below. B: representative ramp profiles of perivascular astrocytes in the presence of tetradotoxin (TTX) (0.5 μM) and glutamate receptor blockers kyneurenic acid (1 mM), LY-367385 (50 μM), and 2-methyl-6-phenylethynylpyridine hydrochloride (MPEP, 100 μM) with or without apamin. C: NUD-GA-induced outward currents in the presence of apamin and combined inhibitors against ionotropic and metabotropic GluR. D: representative calcium trace showing apamin-induced calcium oscillations in astrocytes. E: summary data of apamin-induced changes in averaged peak F/F₀ and calcium oscillation frequency from cortical astrocytes (**P < 0.01, n = 7). F: summary data of averaged peak currents at +80 mV in the presence and absence of NUD-GA, combined inhibitors against ionotropic and metabotropic GluR and the SK channel blocker apamin. *P < 0.05.

Thus we applied the above combination of inhibitors to elucidate the effects of apamin on both control (no agonist) and NUD-GA-sensitive K⁺ currents in astrocytes. After application of voltage-gated Na⁺ channel and glutamate blockers, baseline I-V ramp responses were unaffected. In the presence of apamin, outward currents were 12.8 ± 2.7% smaller than those measured in the presence of the blockers alone (Fig. 6B; +80 mV; P < 0.05; n = 7). When both apamin and the blocker cocktail were present, NUD-GA exposure induced an increase in outward currents of 17.6 ± 4.7% (+80 mV; P < 0.05; n = 6) (Fig. 6C), which represents a ∼53% reduction with respect to the response to NUD-GA alone. These data suggest that SK channels also participate in NUD-GA-induced K⁺ currents.

DISCUSSION

Although a contribution of EETs to the regulation of vascular reactivity and CBF (32, 33, 39) has been supported by several studies, the cellular mediators, the specific mechanisms involved and overall physiological relevance remain incompletely understood. Among the questions that deserve further attention are the endogenous site(s) of EET production, the signals that mediate its production and/or release, and the molecular mechanisms by which changes in vascular tone are induced.

A physiologically relevant pathway leading to cerebrovascular vasodilation during NVC comprises the release of glutamate by neurons and subsequent activation of mGluR in astrocytes (16, 57). The rise in intracellular Ca²⁺ following mGluR activation has been linked to a number of vasoactive compounds, potential mediators of the hyperemic response (22, 26, 39). To this end, we showed that a BK channel-dependent K⁺ efflux from astrocytic endfeet contributes to the vasodilatory response associated with neuronal stimulation (17). Moreover, we also showed that exogenously applied 11,12-EET elicits a rapid increase in intracellular Ca²⁺ in cortical astrocytes in situ (7). These findings raise the possibility that EETs, rather than, or in addition to, diffusing to VSMC, may act in an autocrine manner to potentiate K⁺ efflux from astrocytic endfeet, thus contributing to astrocyte-induced vascular responses (17). In this study, we tested the hypothesis that endogenous EET production in astrocytes contributes to the increase in K⁺ efflux that follows activation of mGluR in astrocytes. In support of our hypothesis, our results show that inhibition of endogenous production of EETs with MS-PPOH abolished whole cell mGluR-induced outward K⁺ currents in astrocytes and attenuated the open probability of single BK channels at their endfeet. Concurrent with our previous observations that BK channel inhibition blocked EETs-mediated vasodilatory responses of parenchymal arterioles (7), the present data supports a role for EETs and K⁺ signaling in glutamate-mediated NVC.

Since astrocytes synthesize (3) and respond to EETs with an increase in K⁺ conductance (50), we employed the synthetic EET analog NUD-GA in the present study to identify specific K⁺ channels activated by EETs in perivascular astrocytes in situ and potentially involved with NVC. We show that perivascular astrocytes respond to the EET analog NUD-GA with a raise in intracellular Ca²⁺, which leads to a significant increase in Ca²⁺-dependent outward K⁺ currents mainly through the activation of BK and SK channels. The Ca²⁺ dependence of NUD-GA-induced K⁺ currents was confirmed by the complete current inhibition observed upon either bath application or cell dialysis with the Ca²⁺ chelator BAPTA. Of note, the increase in Ca²⁺ oscillatory frequency and amplitude elicited by NUD-GA was comparable to that observed previously by us using a similar dose of 11,12-EET (5). Calcium responses induced by EETs have been attributed to the putative activation of TRPV4 channels of the vanilloid family (13), which are targets for these metabolites (47). In astrocytes, TRPV4 channels are preferentially expressed on perivascular processes (6), which is consistent with a preferential action of EETs on glial endfeet.

The present data show that BK channel activation, assessed upon its blockade with TEA or paxilline, predictably accounts for the largest portion of the NUD-GA-induced K⁺ current (∼70%). To further assess this effect, single BK channel activity was measured in native astrocytic endfeet, the site where these channels are densely expressed (36). In response to NUD-GA, we observed a ∼12-fold increase in BK channel open probability, with a concurrent single channel conductance of 171 pS, similar to that previously reported (50). Hence, the strategic expression of BK channels in astrocytic endfeet (36) and their prominent activation by EETs (18) support the idea that EET-induced cell Ca²⁺ increase and BK channel activation are coordinated events that contribute importantly to the control of vascular tone in the brain (7, 13, 17).

Altogether the data suggest that EETs are potent modulators of BK and SK channel activity in astrocytes and likely function as key contributors of the glutamate-mGluR-K⁺ signaling pathway during NVC. In this respect, it is conceivable that after the increase in Ca²⁺ resulting from mGluR activation, EETs production secondary to intracellular AA release may serve to amplify or prolong the raise in Ca²⁺, leading to a more robust or sustained change in vascular tone. As mentioned before, an open question pertains to the possible release of astrocytic EETs into the gliovascular space (23, 26), where they may directly exert an increase in Ca²⁺ in VSMCs leading to BK channel activation and vasodilation (13, 19).

We propose EETs contribute to NVC by increasing intracellular Ca²⁺ in astrocytes and altering the activity of K⁺ channels. Through these mechanisms, EETs may modulate the levels of K⁺ being released at endfeet processess and may further affect the type of vascular response elicited (vasodilation vs. vasoconstrictions). In this sense, we previously showed that vascular responses of penetrating arterioles in brain slices to exogenously applied EETs were dependent on the level of tone of the arteriole (7). Moreover, although commonly deemed as a vasodilator, the amount of K⁺ released by astrocytes (which correlated with the amount of intracellular Ca²⁺ increased in astrocytes) was recently shown to differentially induce vasodilatory or vasoconstricting responses, regardless of the level of tone of the arteriole, thus highlighting the importance and complexity of K⁺ signaling in NVC (21).

While not directly related to the data presented in this study, in a recent study, Marowsky et al. (28) showed expression of sEH (the enzyme that converts EETs to their corresponding dihydroxyeicosatrienoic acids) primarily in astrocytes and particularly in processes outlining the vasculature (endfeet). Thus the strategic expression of sEH in astrocytic endfeet (28) may help modulate K⁺ levels at the gliovascular space.

In summary, in this study we showed that the EET analog NUD-GA increased intracellular calcium in perivascular astrocytes and significantly increased BK and SK channel currents. In addition, blockade of the endogenous formation of EETs blunted outward K⁺ currents induced by the activation of mGluR receptors. Thus, through modulation of astrocyte K⁺ channels, EETs stand as critical players in glutamate-related NVC. A current hypothesis about NVC states that the rise in calcium in astrocytes initiates the metabolism of AA, which leads to the formation of metabolites such as prostaglandins (44, 57) and EETs (3), and that these metabolites diffuse to VSMCs inducing vasodilation (22, 23). Our results expand on previous studies in hippocampal cultured astrocytes (50) and suggest an additional pathway by which EETs may also participate in NVC by acting directly on astrocytic endfeet and in turn modulating Ca²⁺ levels and K⁺ currents in an autocrine manner. Given the limitation of the brain slice preparation, lack of intraluminal pressure/flow, the present study focused primarily on signaling events in astrocytes with the exclusion of potential factors arising from the vasculature itself. Future studies on EETs-mediated mechanisms arising from the circulation as it occurs during changes in shear stress should not be ruled out since they may also significantly influence K⁺ signaling in astrocytes. This form of communication may be the basis for bidirectional flow of information at the neurovascular unit (18).

GRANTS

This work was funded by grants from National Heart, Lung, and Blood Institute (R01HL-089067) and National Institutes of Health (GM-32178) and by the Robert A. Welch Foundation (GL-625910).

DISCLOSURES

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