Endothelial SK Ca and IK Ca Channels Regulate Brain Parenchymal Arteriolar Diameter and Cortical Cerebral Blood Flow

Animals and Dissections

Animal procedures used in this study were performed in accordance with institutional guidelines and were approved by the Institutional Animal Care and Use Committee of the University of Vermont. Male Sprague—Dawley rats (aged 3 to 4 months) were killed by intraperitoneal injection of sodium pentobarbital (150 mg/kg), followed by rapid decapitation. The brain was removed and placed in cold MOPS (3-(N-morpholino)propanesulfonic acid)-buffered saline. A section of the brain that included the M1 region of the middle cerebral artery (MCA), as it bifurcates from the Circle of Willis and travels along the temporal lobe, was removed. Parenchymal arterioles were dissected by loosening the brain tissue using forceps and by gently pulling on the arachnoid layer around the MCA to remove the arterioles from the brain tissue while still attached to the MCA (Straub et al, 2009). From these arterioles (~2 to 4 mm in length), a nonbranching segment originating at least 0.5 mm beyond the point of origin of the MCA and with little associated brain tissue was selected.

Pressurized Parenchymal Arterioles

Segments of PAs ~40 to 80 μm in diameter were mounted on similar-sized glass cannulas in an arteriograph system (Living Systems Instrumentation, Burlington, VT, USA) and pressurized to 40 mm Hg in prewarmed (35°C to 37°C) artificial cerebral spinal fluid equilibrated with 95% O₂ and 5% CO₂ for at least 45 minutes. Only viable arterioles, defined as those that developed pressure-induced myogenic tone, were used in subsequent experiments. Vessel diameter was continuously monitored using a CCD camera and edge-detection software (IonOptix, Milton, MA, USA). All compounds were applied to the perfusate (artificial cerebral spinal fluid), which was constantly refreshed through the arteriograph chamber. Arteriole viability was tested again at the conclusion of the experiment by depolarizing with 60 mmol/L potassium (K⁺) to induce constriction, after which arterioles were exposed to a Ca²⁺-free solution containing 100 μmol/L papaverine (a phosphodiesterase inhibitor) to induce maximal dilation. Where indicated, arterioles were denuded of the endothelium by passing a large air bubble through the lumen of the vessel; elimination of endothelial function was verified by the lack of a response to bradykinin. Changes in parenchymal arteriolar diameter were calculated as percentage change in diameter from the previous diameter.

Arteriolar Smooth Muscle Membrane-Potential Recordings

Parenchymal arterioles were dissected, mounted, and pressurized on an arteriograph chamber as described above; the Ca²⁺-channel blocker nimodipine (200 nmol/L) was included in the bath solution to minimize arterial wall movement. Microelectrodes were filled with a 0.5 mol/L KCl solution to produce electrode resistances of 200 to 300 Mω, and membrane potential was measured using a WPI Intra 767 amplifier (Sarasota, FL, USA) (Brayden and Nelson, 1992) at a sample-recording rate of 5 kHz. Acceptable membrane-potential recordings were those that showed an abrupt change in potential when the cell was impaled, a stable membrane potential for 30 seconds, and unchanged tip resistance before and after impalement (tip potentials of < 2 mV upon exit from the cell).

Endothelial Cell and Smooth Muscle Isolation and Electrophysiology

The isolation procedures used were modifications of our previously described methods for isolating ECs (Taylor et al, 2003; Ledoux et al, 2008a) and SMCs (Straub et al, 2009). Digestion protocols were optimized to produce a greater number of the desired cell type. For EC isolation, PAs were digested in dissociation solution containing Worthington neutral protease (dispase; 1 mg/mL) and elastase (1 mg/mL) for 60 minutes at 37°C; 0.5 mg/mL of a 70%/30% mixture of type F and type H collagenase (Sigma, St Louis, MO, USA) was included for the final minute. For isolation of SMCs, PAs were digested in dissociation solution containing papain (0.3 mg/mL) and dithioerythritol (1 mg/mL) for 20minutes at 37°C, and then with 1.0mg/mL of a 70%/30% mixture of type F and type H collagenase (Sigma) for 5 minutes at 37°C. Endothelial cell isolation yielded small sheets of ECs and isolated ECs, which had a characteristic round and rough shape. Smooth muscle cell isolation yielded primarily single SMCs, which were identified by their spindle-like shape and contractile behavior. Membrane currents were recorded from freshly isolated cells using the conventional whole-cell configuration of the patch-clamp technique (dialyzed with 3 μmol/L free-Ca²⁺ solution) and a voltage-ramp protocol (–125 to + 45 mV, 200 milliseconds). Currents were sampled at 20 kHz, acquired in a voltage-clamp mode using an Axopatch 200A (MDS Analytical Technologies, Sunnyvale, CA, USA), and analyzed using pClamp Suite software (Axon Instruments, Foster City, CA, USA). The patch-clamp pipette series resistance was ~5 to 8MΩ. The series resistance error was negligible because of the small amplitude of the recorded currents; therefore, no compensation was applied. The capacitance error was also negligible under our conditions, with a time constant (τ) of < 100 microseconds.

In Vivo Cerebral Blood Flow Measurements

Cerebral blood flow responses were assessed using a cranial window preparation, as described previously (Niwa et al, 2000). Male C57Bl/6 mice were anesthetized with an intraperitoneal injection of urethane (750 mg/kg) and α-chloralose (50 mg/kg), and a femoral arterial catheter was inserted for blood gas determination and arterial pressure measurement. Animals were mechanically ventilated for the duration of the experiment, and arterial blood gases were maintained at 32 to 35 mm Hg pCO₂ and 100 to 110mmHg pO₂ (Capstar-100, CWE, Ardmore, PA, USA). A 2 × 2mm² section of the parietal bone and the underlying dura in the region of the somatosensory cortex were removed. The surface of the cerebral cortex within the cranial window was continuously superfused with artificial cerebral spinal fluid (at 37°C, pH 7.4), with or without pharmacological agents. Cortical CBF was measured using a laser-Doppler probe (Perimed, Järfalla, Sweden) positioned just above the cortex and away from the pial arteries.

Immunofluorescence

The MCA and attached PAs, isolated as described above, were fixed in a 4% paraformaldehyde/phosphate-buffered saline (0.01 mol/L phosphate buffer, 0.15 mol/L NaCl) solution for 2 hours, and then permeabilized with 0.02% Triton X and blocked with 3% donkey serum or 2% gelatin for 2 hours at room temperature to prevent nonspecific binding. Middle cerebral artery/PA preparations were then incubated overnight at 4°C with primary antibodies against AQP4 (aquaporin 4) (goat anti-AQP4; Santa Cruz Biotechnology, Santa Cruz, CA, USA; 2.0 μg/mL) or rabbit anti-UCH-L1 (ubiquitin carboxyl-terminal hydrolase L1; Zymed Laboratories, Carlsbad, CA, USA; 2.5 μg/mL diluted in 0.2% gelatin/0.1% Triton/phosphate-buffered saline). After washing with phosphate-buffered saline, Cy5-donkey-anti-goat IgG and Alexa Fluor-555-donkey-anti-rabbit IgG secondary antibodies (Invitrogen, Carlsbad, CA, USA; diluted 1:500 in 0.2% gelatin–0.1% Triton—phosphate-buffered saline) were added, and the MCA/PA preparation was incubated for 2 hours at room temperature. Nuclei were stained with a 0.5 μg/mL solution of the fluorescent nucleic acid dye DAPI (4',6-diamidino-2-phenylindole; Invitrogen). Segments of the MCA containing at least one attached PA were mounted onto glass slides using Aqua Polymount mounting medium (Polysciences Inc., Warrington, PA, USA) and examined at × 10, × 20, and × 40 magnification using an Olympus BX50 Light Microscope (Olympus America Inc., Lake Success, NY, USA) equipped with a QImaging Retiga 2000R digital camera (QImaging, Surry, BC, Canada), and at × 40 and × 63 magnification using a Zeiss LSM 510 META Confocal Microscope (Carl Zeiss MicroImaging Inc., Thornwood, NY, USA). Nonspecific staining was determined in images collected from MCA/PA preparations incubated with secondary antibodies alone, and a threshold was applied to each individual channel to exclude pixels that corresponded to nonspecific staining.

Solutions

The extracellular solution for isolated cell experiments consisted of 134 mmol/L NaCl, 6 mmol/L KCl, 10mmol/L HEPES (pH 7.4), 1 mmol/L MgCl₂, 2 mmol/L CaCl₂, and 10mmol/L glucose. Low-Ca²⁺ dissociation solution contained 55 mmol/L NaCl, 80 mmol/L Na-glutamate, 5.9mmol/L KCl, 2 mmol/L MgCl₂, 0.1 mmol/L CaCl₂, 5 mmol/L glucose, and 10 mmol/L HEPES (pH 7.3). The composition of the pipette solution used for whole-cell experiments was 134 mmol/L KCl, 5.53 mmol/L MgCl₂, 0.207 mmol/L CaCl₂, 5 mmol/L hydroxy-EDTA (ethylene-diaminetetraacetic acid), and 10 mmol/L HEPES, adjusted to pH 7.2 with 17.8 mmol/L KOH; the final K⁺ concentration was 151.8 mmol/L and free-Ca²⁺ and free-Mg²⁺ concentrations were 3 μmol/L and 1mmol/L, respectively. A MOPS-buffered saline solution (3 mmol/L MOPS pH 7.4 at 25°C, 145mmol/L NaCl, 3mmol/L KCl, 1.2mmol/L NaH₂PO₄, 1.17 mmol/L MgSO₄, 2mmol/L CaCl₂, 5mmol/L glucose, 10mg/mL bovine serum albumin) was used for dissection of PAs. Prewarmed artificial cerebral spinal fluid (125mmol/L NaCl, 3mmol/L KCl, 26mmol/L NaHCO₃, 1.25 mmol/L NaH₂PO₄, 1mmol/L MgCl₂, 2mmol/L CaCl₂, 5 mmol/L glucose; 35°C to 37°C) was used for pressurized arteriole and in vivo experiments.

Reagents

Apamin, ChTx, paxilline, and TRAM-34 were purchased from BioMol International (Farmingdale, NY, USA). NS309 was kindly provided by Dr Søren-Peter Olesen (Neuro-search A/S, Ballerup, Denmark). All other compounds used were purchased from Sigma.

Statistics

All data are presented as mean ± s.e.m. Data were analyzed using ANOVA (analysis of variance), followed by Tukey's multiple comparison tests, or paired or unpaired Student's t-test, as appropriate. Differences with P-values < 0.05 were considered statistically significant.

Isolated Parenchymal Arterioles Lack Extrinsic Innervation and are Encased by Astrocytic Processes

Parenchymal arterioles that branched from the MCA were studied (Figure 1). In 25 PAs evaluated from 6 MCAs, the unpressurized diameter ranged from 8 to 53 μm (mean, 25 ± 2.8 μm; median, 26 μm). The nerves were identified by staining for the panneuronal marker UCH-L1 (previously known as neuron cytoplasmic protein 9.5), and astrocytes were identified by staining for AQP4, which is uniquely expressed in astrocytes and is concentrated in astrocytic endfeet (Iadecola and Nedergaard, 2007). Aquaporin4 staining was not detected in the MCA, but intense staining was observed in PAs beginning at a point shortly after branching off the MCA (Figure 1). The MCA exhibited prominent UCH-L1 staining, consistent with extensive innervation. In contrast, little or no innervation was observed on PAs, and any detectable nerve staining was lost after a short distance (< 50 μm) from the branch point with the MCA, consistent with previous reports (Cipolla et al, 2004).

OPEN IN VIEWER

Figure 1

Isolated MCAs with attached parenchymal arterioles. Representative photomicrograph of an isolated MCA and single parenchymal arteriole stained with UCH-L1 (green), aquaporin 4 (red), and DAPI (blue) to detect nerves, astrocytes, and nuclei, respectively, showing that parenchymal arterioles are encased by astrocytic processes and lack extrinsic nerves. There was no detectable astrocyte staining on isolated MCAs and little or no detectable nerve staining on parenchymal arterioles. The image is a two-dimensional representation of a compressed Z-stack series. Scale bar, 50 μm. DAPI, 4',6-diamidino-2-phenylindole; MCA, middle cerebral artery; UCH-L1, ubiquitin carboxyl-terminal hydrolase L1.

Functional KCa Channels Exhibit a Cell Type-Specific Distribution in Endothelial and Smooth Muscle Cells of Parenchymal Arterioles

The electrophysiological properties of parenchymal arteriolar ECs have not been characterized previously. To determine whether ECs in PAs have functional SK_Ca and IK_Ca channels, we measured whole-cell currents in freshly dissociated ECs using the patch-clamp technique. Cells were dialyzed with a 3 μmol/L free-Ca²⁺ solution to maximally stimulate SK_Ca and IK_Ca channels, and current—voltage relationships of whole-cell currents were recorded using a voltage-ramp protocol (–125 to + 45mV, 200 milliseconds). Endothelial cells from PAs had a mean capacitance of 8.0 ± 0.7 pF, consistent with previous reports from mouse aortic ECs (10.6 ± 0.8 pF (Ledoux et al, 2008a)). Positive to the K⁺ equilibrium potential (E_K), membrane current increased and then decreased (inward rectification) with progressive membrane-potential depolarization (Figures 2A and 2B), similar to I—V relationships reported in other types of ECs (Ledoux et al, 2008a). The selective SK_Ca channel blocker apamin reduced the K⁺ current by 37.5% ± 5.6% at + 40mV, and the remainder of the current was inhibited in the presence of the IK_Ca channel blocker ChTx. The SK_Ca and IK_Ca current densities were 19.9 ± 3.5 pA/pF and 28.4 ± 2.8 pA/pF, respectively, at + 40mV (Figure 2C), and reversed near E_K (−80.1 and −81.7mV for SK_Ca and IK_Ca, respectively; E_K = −83.0 mV). The selective BK_Ca channel inhibitor paxilline (Imlach et al, 2008) had no effect (Figure 2B), indicating that BK_Ca channels did not contribute to the observed current.

OPEN IN VIEWER

Figure 2

SK_Ca and IK_Ca currents in ECs, and BK_Ca currents in SMCs from parenchymal arterioles. (A, B) Current-voltage (I—V) relationships of whole-cell currents recorded from an isolated parenchymal EC using a voltage-ramp protocol (–125 to + 45mV, 200 milliseconds) and dialyzed with a 3 μmol/Lfree-Ca²⁺ and 1 mmol/Lfree-Mg²⁺ solution. The characteristic inward rectification must be noted. EC currents were insensitive to the BK_Ca channel blocker paxilline (500nmol/L) (panel B), but were almost completely eliminated by sequential addition of the SK_Ca blocker apamin (300nmol/L) and the IK_Ca blocker ChTX (100nmol/L) (P < 0.05) (panel A). (C) Mean IK_Ca current density (ChTX sensitive, n = 5 cells) and SK_Ca current density (apamin sensitive, n = 3 cells) measured at + 40 mV. (D) Whole-cell current-voltage relationship for a SMC from a parenchymal arteriole recorded under the conditions described above. The typical outward rectification caused by the activation of BK_Ca channels, and the much larger currents at positive membrane potentials compared with EC cells must be noted (~ 7-fold larger at + 40 mV; red trace in panel A). Paxilline almost completely abolished the K⁺ current in SMCs isolated from parenchymal arterioles (P < 0.05). The lower I—V curve illustrates the lack of an effect of both apamin and ChTx on membrane currents in the presence of paxilline. (E) The remaining SMC K⁺ current at + 40 mV was not altered by both SK_Ca and IK_Ca channel blockers (n = 5). BK_Ca, large-conductance, voltage-dependent calcium-sensitive potassium channel; ChTX, charybdotoxin; EC, endothelial cell; IK_Ca, intermediate-conductance calcium-sensitive potassium channel; SK_Ca, small-conductance calcium-sensitive potassium channel; SMC, smooth muscle cell.

Parenchymal arteriolar SMCs exhibited a markedly different current—voltage relationship with the same solutions, and showed very prominent outward rectification (Figure 2D). The outward K⁺ current was greatly reduced (91.5% ± 3.3% at + 40mV) by the BK_Ca channel inhibitor paxilline. In the presence of paxilline, application of the IK_Ca and SK_Ca blockers ChTx and apamin did not affect outward currents (Figures 2D (inset) and 2E). The BK_Ca current density in SMCs was quite substantial (154.4 ± 72.4 pA/pF at + 40mV), particularly in comparison with EC SK_Ca and IK_Ca current densities. These results indicate that expression of functional K_Ca channels is cell-type specific, with BK_Ca channels in SMCs of PAs, and SK_Ca and IK_Ca channels in ECs.

Blockade of Endothelial Cell SKCa and IKCa Channels Constricts Parenchymal Arterioles, whereas Inhibition of BKCa Channels has a Minimal Effect

To elucidate the roles of EC SK_Ca and IK_Ca channels and SMC BK_Ca channels in the regulation of arteriolar function, we determined the effects of SK_Ca, IK_Ca, and BK_Ca blockers on parenchymal arteriolar diameter. Increase in intravascular pressure to physiologic levels (40mmHg) constricted PAs by 26.0% ± 1.2% (n = 38), consistent with previous observations (Cipolla and Bullinger, 2008). Combinations of K_Ca channel blockers were used to isolate the relative role of each channel type in the regulation of PA tone, as shown by representative diameter traces in Figure 3. Application of the SK_Ca channel blocker apamin constricted pressurized arterioles by 15.8% ± 1.9% (Figures 3A and 3B). Paxilline, a BK_Ca channel blocker, caused a small constriction (5.3% ± 2.4%); a similar constriction to paxilline was observed in the presence apamin (Figures 3A and 3B). As paxilline (500 nmol/L) did not affect IK_Ca or SK_Ca currents in isolated ECs, but essentially eliminated BK_Ca currents in SMCs (see Figure 2), it was always applied before the IK_Ca/BK_Ca blocker ChTx to isolate IK_Ca channel effects on parenchymal diameter. In the presence of paxilline, addition of ChTx to block IK_Ca channels constricted PAs by 15.9% ± 2.7%, similar to the constriction caused by apamin alone (Figure 3). The synthetic IK_Ca channel inhibitor, TRAM-34 (1mmol/L), constricted PAs by 21% ± 5% (n = 3), consistent with a previous report (Cipolla et al, 2009). When applied together, SK_Ca, IK_Ca, and BK_Ca channel blockers constricted pressurized (40mmHg) PAs by 30.7% ± 3.9%.

OPEN IN VIEWER

Figure 3

Blockade of SK_Ca and IK_Ca channels causes a greater constriction of pressurized parenchymal arterioles than that of BK_Ca channels. (A) Representative diameter traces of parenchymal arterioles pressurized to 40mmHg. The upper trace illustrates the effects of sequential abluminal application of apamin (300 nmol/L), paxilline (500 nmol/L), and ChTx (100 nmol/L). The lower trace illustrates the effect of paxilline (500 nmol/L) and ChTx (100 nmol/L) in the absence of apamin. (B) Summary data illustrating mean constrictions to apamin alone (‘SK_Ca channel block’), to ChTx in the presence of paxilline to block BK_Ca channels (‘IK_Ca channel block’), to paxilline alone (‘BK_Ca channel block’), and to the combination of apamin, paxilline, and ChTx (‘all K_Ca channel blockers’). Individual inhibition of SK_Ca, IK_Ca, and BK_Ca channels significantly constricted arterioles with myogenic tone (*P < 0.05, n = 15, repeated-measures ANOVA, Tukey's test). ANOVA, analysis of variance; BK_Ca, large-conductance, voltage-dependent calcium-sensitive potassium channel; ChTX, charybdotoxin; IK_Ca, intermediate-conductance calcium-sensitive potassium channel; SK_Ca, small-conductance calcium-sensitive potassium channel.

Activation of Endothelial Cell IKCa Channels Dilates and Hyperpolarizes Parenchymal Arterioles

Blockade of EC SK_Ca and IK_Ca channels caused significant vasoconstriction. Therefore, opening SK_Ca and/or IK_Ca channels should produce vasodilation. To test this hypothesis, we determined the effect of NS309, a selective opener of SK_Ca and IK_Ca channels (Strøbaek et al, 2004), on pressurized arterioles (Figure 4). NS309 (1 μmol/L) caused a large increase in both SK_Ca and IK_Ca channel currents in isolated parenchymal arteriolar ECs (Supplementary Figure 1), consistent with the results of the study by Strøbaek et al (2004). NS309 (1 μmol/L) caused maximal dilation of pressurized (40mm Hg) PAs (n = 11; Figures 4A and 4C). The dilatory effect of NS309 was prevented by preincubation with a combination of apamin and ChTx, or by removing the endothelium (Figure 4B), consistent with NS309 activation of endothelial SK_Ca and IK_Ca channels.

OPEN IN VIEWER

Figure 4

The SK_Ca and IK_Ca channel opener NS309 maximally dilates pressurized (40mmHg) parenchymal arterioles. (A) Representative diameter trace illustrating dilation to NS309 (1 μmol/L) and reversal by apamin (300 nmol/L) and/or ChTX (100 nmol/L). (B) Representative diameter trace illustrating the lack of an effect of NS309 on endothelium-denuded parenchymal arterioles. (C) Summary data showing the effects of NS309 (1 μmol/L) alone (control), and in the presence of L-NNA (50 μmol/L) + indomethacin (10 μM), BaCl₂ (30 μmol/L), paxilline (500 nmol/L), apamin (300 nmol/L) + ChTx (100 nmol/L), and without endothelium. The dilatory response to NS309 was significantly blunted in the presence of SK_Ca and IK_Ca channel blockers and in endothelium-denuded arterioles (*P < 0.01, ANOVA and Tukey's post hoc analysis; n = 3 to 4). (D) Summary data showing the effects of 500 nmol/L NS309 on pressurized parenchymal arterioles; NS309-induced dilation was prevented with IK_Ca channel block (paxilline was used in combination with ChTX to isolate the IK_Ca channel effects), but not with SK_Ca channel block (*P < 0.01, paired t-test; n = 4 to 5). ANOVA, analysis of variance; ChTX, charybdotoxin; EC, endothelial cell; IK_Ca, intermediate-conductance calcium-sensitive potassium channel; L-NNA, L-NG-nitroarginine; NS309, 6,7-dichloro-1H-indole-2,3-dione3-oxime; SK_Ca, small-conductance calcium-sensitive potassium channel.

It is possible that activation of EC SK_Ca and IK_Ca channels could indirectly affect endothelial nitric oxide (NO) synthase (NOS) and cyclooxygenase activity by increasing EC global Ca²⁺ (Ledoux et al, 2008a; Sheng et al, 2009). The combination of L-NG-nitroarginine (L-NNA) (50 μmol/L) and indomethacin (10 μmol/L) constricted PAs by 26.0% ± 3.9% (n = 14), which is similar to the previously reported response to NOS inhibition alone (Cipolla et al, 2009). Inhibition of NOS and cyclooxygenase with L-NNA (50 μmol/L) and indomethacin (10 μmol/L) had no effect on NS309-induced dilation of PAs; the K_ir channel inhibitor BaCl₂ (30 μmol/L) and the BK_Ca channel inhibitor paxilline (500 nmol/L) were similarly ineffective in blocking the dilation to NS309 (Figure 4C). Interestingly, inhibition of IK_Ca channels with ChTx (in the presence of paxilline), prevented the dilation to NS309, whereas inhibition of SK_Ca channels did not (Figure 4D). Collectively, these results indicate that NS309 dilates PAs primarily by activation of EC IK_Ca channels, independent of NOS or cyclooxygenase activity.

A likely mechanism for this EC IK_Ca channel-dependent relaxation of the smooth muscle is through SMC hyperpolarization. To test this possibility, we measured the membrane potential of SMCs in pressurized (40 mm Hg) PAs using microelectrodes. The resting membrane potential was −38.6 ± 0.8mV (n = 10), which is 10 to 15 mV more depolarized than that of the pial arteries at the same pressure (Knot and Nelson, 1998; Albarwani et al, 2003). NS309 (1 μmol/L) caused a 27mV hyperpolarization of SMCs to −65.6 ± 3.6 mV. Blocking BK_Ca or SK_Ca channels in the presence of NS309 did not significantly affect membrane-potential hyperpolarization to NS309 (Figures 5A and 5B). However, blocking IK_Ca channels with ChTx reversed the membrane-potential hyperpolarization to NS309 (Figures 5A and 5B). These results indicate that activation of endothelial IK_Ca channels by NS309 dilates PAs through hyperpolarization of parenchymal arteriolar SMCs.

OPEN IN VIEWER

Figure 5

NS309 acts through IK_Ca channels to hyperpolarize the smooth muscle membrane potential in pressurized (40 mm Hg) parenchymal arterioles. (A) Illustration of smooth muscle membrane potential (resting −35.5 mV) hyperpolarization to NS309 (1 μmol/L) (–66.4 mV), the lack of an effect of apamin (300nmol/L) and paxilline (500nmol/L), and reversal by ChTx (100nmol/L) (−36.9 mV). (B) Summary data showing changes in membrane potential in response to 1 μmol/L NS309. The hyperpolarization to NS309 was not altered by blocking SK_Ca or BK_Ca channels, yet IK_Ca channel block repolarized parenchymal arterioles (*P < 0.01 compared with membrane potential in the presence of NS309; n = 4 arterioles). BK_Ca, large-conductance, voltage-dependent calcium-sensitive potassium channel; ChTX, charybdotoxin; IK_Ca, intermediate-conductance calcium-sensitive potassium channel; NS309, 6,7-dichloro-1H-indole-2,3-dione3-oxime; SK_Ca, small-conductance calcium-sensitive potassium channel.

Contribution of SKCa and IKCa Channels to Resting Cortical Cerebral Blood Flow In Vivo

The results obtained in isolated arterioles indicate a tonic dilatory influence of EC SK_Ca and IK_Ca channel activity on PA tone. To determine whether this is the case in vivo, we assessed the effects of EC SK_Ca and IK_Ca channel modulation on cortical CBF using laser-Doppler flowmetry in a cranial window preparation. Combined inhibition of SK_Ca and IK_Ca channels by superfusing the cortical surface of the brain with apamin (300 nmol/L) and TRAM-34 (10 μmol/L), a specific inhibitor of IK_Ca channels, decreased resting cortical CBF by 15% ± 2% (Figure 6A). The reduction in resting cortical CBF with combined SK_Ca and IK_Ca blockade was similar to that observed with NOS inhibition (14% ± 6%). Inhibition of IK_Ca channels reduced resting cortical CBF by 9% ± 2%. We have previously shown that BK_Ca channel inhibition with paxilline does not affect resting cortical CBF (Girouard et al, 2010). These results indicate that EC SK_Ca and IK_Ca channels contribute to the determination of resting cortical CBF.

OPEN IN VIEWER

Figure 6

Effects of SK_Ca and IK_Ca channel modulation on cortical CBF. (A) Effect on resting cortical CBF. Combined SK_Ca and IK_Ca block (300nmol/L apamin + 10 μmol/L TRAM-34) reduced resting CBF by 15% ± 2% (*P < 0.05, n = 11), compared with 14% ± 6% with 1 mmol/L L-NNA (*P < 0.05, n = 5). Inhibition of IK_Ca only (TRAM-34) accounted for about half of the combined effect of apamin + TRAM-34 (9% ± 2%, n = 5, P < 0.05 compared with resting CBF). (B) Effect of the SK_Ca/IK_Ca activator NS309 on cortical CBF under control conditions and in the presence of various pharmacological inhibitors. Cranial superfusion of NS309 (10 μmol/L) increased cortical CBF by 40% ± 3% (*P < 0.05 versus baseline CBF, n = 15). Combined SK_Ca and IK_Ca inhibition (apamin + TRAM-34) and inhibition of IK_Ca only significantly attenuated the hyperemic response to NS309 (^†P < 0.05 versus control; n = 7 and n = 6, respectively), whereas NOS blockade with L-NNA had no effect on the NS309 response (n = 5). (C) Representative laser-Doppler traces of cortical CBF responses to cranial window superfusion of the SK_Ca and IK_Ca channel opener NS309. Cortical superfusion of NS309 began at 0minutes. Pharmacological inhibitors were administered by cortical superfusion 30 minutes before the addition of NS309. CBF, cerebral blood flow; IK_Ca, intermediate-conductance calcium-sensitive potassium channel; L-NNA, L-NG-nitroarginine; NOS, nitric oxide synthase; NS309, 6,7-dichloro-1H-indole-2,3-dione3-oxime; SK_Ca, small-conductance calcium-sensitive potassium channel; TRAM-34, 1-[(2-chlorophenyl)diphenylmethyl]-1H-pyrazole.

Activation of SK_Ca and/or IK_Ca channels by cortical superfusion of NS309 (10 μmol/L) produced a robust 40% ± 3% increase in cortical CBF that exhibited a rapid onset (< 1 minute after administration) (Figures 6B and 6C, Control). This NS309-induced cortical hyperemia was largely prevented by apamin and TRAM-34 (9% ± 3% in CBF to NS309). In agreement with its effects on isolated PAs (see Figure 4), IK_Ca channel block attenuated the NS309-induced increase in CBF to a similar extent as combined SK_Ca and IK_Ca block (Figure 6). Inhibition of NOS (1 mmol/L L-NNA) had no effect on the CBF response to NS309, suggesting that activation of SK_Ca/IK_Ca channels with NS309 does not produce cortical hyperemia through increased NO synthesis. These results suggest that the CBF response to NS309 is primarily IK_Ca mediated, which is consistent with observations in isolated PAs.

The Unique Environment of Parenchymal Arterioles

Previous studies have shown that cortical arterioles are encased in astrocytic processes and lack extrinsic nerves (Iadecola and Nedergaard, 2007). This environment is unique compared with that of other vascular beds. We showed that isolated PAs lack extrinsic innervation, consistent with previously published data (Cipolla et al, 2004). In addition, AQP4 staining surrounded isolated PAs, indicating that astrocytic processes tightly adhere to the isolated arterioles. These data also indicate that the loss of innervation and association with astrocytes occur very near (< 50 μm) the point at which a PA branches from a pial artery on the brain surface.

KCa Channels in the Vascular Wall

The BK_Ca channels have been identified in virtually all types of the smooth muscle (Ledoux et al, 2006). In this study, we show that SMCs, but not ECs, from PAs also have functional BK_Ca channels (Figure 2D). Although there are conflicting reports regarding the expression of BK_Ca channels in ECs (see Sandow and Grayson (2009) and references therein), our data from parenchymal arteriolar ECs support the consensus view that functional BK_Ca channels are absent in the healthy, native endothelium (Gauthier et al, 2002; Sandow and Grayson, 2009).

The responses of arteries to SK_Ca and IK_Ca channel blockers suggest that these channels are likely present in the endothelium of most vascular beds. However, there are relatively few direct measurements of SK_Ca and IK_Ca channel currents in freshly dissociated ECs, and to date, all such measurements have been obtained using ECs from larger caliber arteries (Table 1) (Kohler et al, 2001; Burnham et al, 2002; Bychkov et al, 2002; Taylor et al, 2003; Si et al, 2006; Ledoux et al, 2008a; Braähler et al, 2009). In this study, we provide the first measurements of SK_Ca and IK_Ca channel currents from arteriolar ECs (Figure 2). The most notable difference between SK_Ca and IK_Ca currents in parenchymal arteriolar ECs and those measured in ECs from larger peripheral arteries is the current density. The SK_Ca and IK_Ca current densities in parenchymal arteriolar ECs are ~1.6- to 6.7-fold higher than those measured in ECs from other vascular beds (Table 1), suggesting a greater influence of these channels on adjacent SMCs in these arterioles. To our knowledge, functional SK_Ca and IK_Ca channels (i.e., current measurements) have not been detected in vascular SMCs from healthy arteries, although immunostaining for IK_Ca channels has been observed in pial artery SMCs (McNeish et al, 2006), and vascular injury has been shown to induce IK_Ca channel expression in proliferating SMCs (Grgic et al, 2009).

The BK_Ca channel current density recorded in SMCs (dialyzed with 3 μmol/L intracellular Ca²⁺) from PAs was 154pA/pF at + 40mV and accounted for 92% of the outward current, values similar to those reported previously (Gerzanich et al, 2001). The remaining current in SMCs was insensitive to ChTx and apamin, and likely represents current through voltage-dependent K⁺ channels (K _V 1.2/1.5) (Straub et al, 2009).

Regulation of Parenchymal Arteriole Tone by Endothelial SKCa and IKCa Channels

In the pial and peripheral arteries, endothelium-dependent hyperpolarization to endothelial agonists—the EDHF response—is prevented by blockers of SK_Ca and IK_Ca channels. Although in these arterial preparations, SK_Ca and IK_Ca blockers have little effect on arterial tone and SMC membrane potential in the absence of agonists, the defining feature of this EDHF phenomenon is its dependence on SK_Ca and IK_Ca channels (Andresen et al, 2006; Edwards et al, 2010). Our results indicate that EC SK_Ca and IK_Ca channels, in the absence of exogenous agonists, have a significant tonic role in opposing pressure-induced constriction in PAs, independent of NOS or cyclooxygenase activity. Thus, in PAs, apparently unlike many vascular beds, EDHF has a role in regulating myogenic tone, possibly through direct electrical coupling of the single endothelial and smooth muscle layers in these vessels.

NS309, a potent opener of SK_Ca and IK_Ca channels (Strøbaek et al, 2004), caused a profound arteriolar smooth muscle hyperpolarization, vasodilation, and an increase in cortical CBF (Figures 4, 5, and 6). Interestingly, blockers of IK_Ca channels, but not of SK_Ca channels, largely prevented NS309-induced hyperpolarization, dilation, and hyperemia. This observation is perplexing in light of the clear effects of both SK_Ca and IK_Ca channel blockers on membrane potential, myogenic tone, and CBF under basal (unstimulated) conditions, and the similarity in the potency of NS309 towards exogenously expressed SK_Ca (EC₅₀ = 30nmol/L) and IK_Ca (EC₅₀ = 10nmol/L) channels in transfected HEK 293 cells (Strøbaek et al, 2004). Consistent with the data obtained from the study by Strøbaek et al (2004), we found that NS309 activates both SK_Ca and IK_Ca currents in isolated ECs. Similar to the effects of NS309, Marrelli et al (2003) and You et al (1999) have reported that agonist-induced EDHF responses in the pial arteries and PAs depend on IK_Ca channels, and not on SK_Ca channels. This may reflect clustering of IK_Ca channels in endothelial projections through the internal elastic lamina that abut the smooth muscle (Sandow et al, 2006; Dora et al, 2008; Ledoux et al, 2008b). Given the differential spatial organization of SK_Ca and IK_Ca channels, it is possible that SK_Ca channels are maximally activated by local Ca²⁺ but IK_Ca channels are not. Therefore, blockade of either channel would cause constriction, but activation by NS309 would primarily act through IK_Ca channels. The resolution of this intriguing observation, although beyond the scope of current study, may provide insights into the unique and functionally divergent roles of SK_Ca and IK_Ca channels in ECs.

NS309 caused a profound increase in cortical CBF, which was largely prevented by IK_Ca channel inhibition (Figure 6), consistent with the effects on isolated, pressurized PAs (Figure 4) and the prominent role of IK_Ca channels in EDHF-mediated dilations of cerebral arteries (Marrelli et al, 2003). It is likely that this hyperemic response reflects activation of endothelial IK_Ca channels because current evidence suggests that IK_Ca channels are not expressed in central nervous system neurons or astrocytes (Pedarzani and Stocker, 2008). The SK_Ca channels in neurons could potentially be affected by cortical superfusion of apamin (SK_Ca channel blocker) or NS309 (SK_Ca/IK_Ca channel activator), and indirectly alter CBF by modulating neuronal activity. However, inhibition of SK channel activity in neurons would be predicted to increase neuronal excitability by depolarizing the membrane potential, producing an increase in CBF (neurovascular coupling). Yet, with cortical superfusion of the SK_Ca and IK_Ca blockers, apamin and TRAM-34, we observed a reduction in CBF. Conversely, activation of neuronal SK_Ca channels with NS309 would be expected to hyperpolarize neurons and reduce neuronal metabolism, which would (if anything) lead to a reduction in CBF. In stark contrast to this predicted outcome, we found that cortical superfusion of NS309 produced robust hyperemia. Thus, it is unlikely that modulation of SK_Ca channel activity in neurons contributes to the observed effects of cortical superfusion of apamin and TRAM-34, or NS309, on CBF.

Role of the BKCa Channel in Parenchymal Arterioles

In contrast to the effects of SK_Ca and IK_Ca channel blockers on PA tone and CBF, BK_Ca channel block had only a modest effect on parenchymal arteriolar tone in our studies (< 5%). We have previously reported that BK_Ca channel block also causes only a modest constriction of PAs in brain slices and has little effect on resting cortical CBF (Girouard et al, 2010). This is in contrast with pressurized pial arteries, where blockade of BK_Ca channels causes a constriction of ~30% (Brayden and Nelson, 1992; Nelson et al, 1995). It should be noted that effects of K channel blockers on myogenic tone similar to those described in this study for PAs have been reported in pial arteries (rat MCA) (Marrelli et al, 2003). However, differences in experimental methodology used in this previous study, including the use of a nonpaired experimental design and differences in methods used to calculate changes in diameter, make direct comparisons difficult. In vivo, blockers of BK_Ca channels have been shown to cause a small (3%), but significant, constriction of pial arterioles (Taguchi et al, 1995), and produce a larger effect only at high systemic blood pressures or in the presence of an exogenous vasoconstrictor (Paterno et al, 2000). In the pial arteries, BK_Ca channels are activated by micromolar intracellular Ca²⁺ delivered by local calcium release events through ryanodine receptors (Ca²⁺ sparks) in the sarcoplasmic reticulum (Nelson et al, 1995). Our results indicate that the density of BK_Ca currents in SMCs of PAs is substantial; thus, it is unlikely that the diminished contribution of the BK_Ca channel to parenchymal arteriolar tone is attributable to a reduction in functional BK_Ca channels. Although Ca²⁺ dynamics in SMCs of PAs have not been studied in pressurized PAs, it is tempting to speculate that the diminished role of BK_Ca channels in the regulation of arteriolar tone may reflect a lack of Ca²⁺ sparks in arteriolar myocytes. Alternatively, a loss of the regulatory b1 subunit of the BK_Ca channels, which enhances the apparent Ca²⁺ and voltage sensitivity of the pore-forming α-subunit (Brenner et al, 2000), might account for a muted BK_Ca involvement in regulating arteriolar tone. Regardless of the underlying mechanism, the loss of BK_Ca channel functionality in intact PAs may contribute to the enhanced pressure-induced constriction of these arterioles.

In conclusion, our data are consistent with a model in which PAs depolarize and constrict more to intravascular pressure because SMC BK_Ca channels do not oppose vasoconstriction, an effect that is only partially compensated by the elevated hyperpolarizing and dilating influence of EC SK_Ca and IK_Ca channels. Our results indicate that modulation of SK_Ca and IK_Ca channels in the endothelium of PAs has a profound effect on cortical blood flow. Therefore, dysfunction of EC SK_Ca and IK_Ca channels could compromise intra-CBF, whereas activation of IK_Ca channels could enhance CBF, and thus may represent a therapeutic target to ameliorate reduced CBF that accompanies a number of cerebrovascular disorders.