Acidosis dilates brain parenchymal arterioles by conversion of calcium waves to sparks to activate BK channels

Abstract

Rationale

Acidosis is a powerful vasodilator signal in the brain circulation. However, the mechanisms by which this response occurs are not well understood, particularly in the cerebral microcirculation. One important mechanism to dilate cerebral (pial) arteries is by activation of large-conductance, calcium-sensitive potassium (BK_Ca) channels by local Ca²⁺ signals (Ca²⁺ sparks) through ryanodine receptors (RyRs). However, the role of this pathway in the brain microcirculation is not known.

Objective

The objectives of this study were to determine the mechanism by which acidosis dilates brain parenchymal arterioles (PAs) and to elucidate the roles of RyRs and BK_Ca channels in this response.

Methods and Results

Internal diameter and vascular smooth muscle cell (VSMC) Ca²⁺ signals were measured in isolated pressurized murine PAs, using imaging techniques. In physiological pH (7.4), VSMCs exhibited primarily RyR-dependent Ca²⁺ waves. Reducing external pH from 7.4 to 7.0 in both normocapnic and hypercapnic conditions decreased Ca²⁺ wave activity, and dramatically increased Ca²⁺ spark activity. Acidic pH caused a dilation of PAs which was inhibited by about 60% by BK_Ca channel or RyR blockers, in a non-additive manner. Similarly, dilator responses to acidosis were reduced by nearly 60% in arterioles from BK_Ca channel knockout mice. Dilations induced by acidic pH were unaltered by inhibitors of K_ATP channels or nitric oxide synthase.

Conclusions

These results support the novel concept that acidification, by converting Ca²⁺ waves to sparks, leads to the activation of BK_Ca channels to induce dilation of cerebral parenchymal arterioles.

Introduction

Brain parenchymal arterioles (PAs) constitute a unique vascular bed. In contrast to pial arteries on the surface of the brain, PAs do not receive extrinsic innervation and are enveloped by astrocytic processes called endfeet over nearly their entire basolateral surface ^{1, 2}. PAs account for approximately 40% of total cerebral vascular resistance ³. Thus regulation of their diameter is a determinant process for appropriate perfusion of brain tissue ² and for neurovascular coupling ⁴. Compared with pial arteries, PAs depolarize and constrict to lower levels of intravascular pressure, a physiologically appropriate response given that PAs experience lower pressure levels in situ ^{5, 6, 7, 8}. One plausible explanation is that negative feedback elements which limit depolarization and constriction to pressure in pial arteries are lacking, or less active in parenchymal arterioles. In large diameter pial arteries, activation of smooth muscle large-conductance potassium channels (BK_Ca) by local Ca²⁺-release events (Ca²⁺ sparks) through ryanodine receptors (RyRs) opposes pressure-induced depolarization and constriction ^{9, 10}. In contrast, BK_Ca blockers have little effect on the diameter of cerebral arterioles ^{11, 12} even though functional BK_Ca channels are present ⁷.

Regulation of cerebral blood flow (CBF) by changes in pH has been recognized as a critical homeostatic mechanism for more than a century ¹³. It is well established that acidic pH increases CBF by relaxing vascular smooth muscle cells (VSMC) ^{14, 15, 16}. Protons are thought to act through multiple mechanisms including the inhibition of Ca²⁺ entry through voltage-dependent Ca²⁺ channels (VDCCs) ¹⁷, and activation of K_ATP channels ¹⁸ and NO pathways ¹⁹. In addition, protons can decrease the open probability (P₀) of RyRs ²⁰, which will modify the Ca²⁺-release profile of the sarcoplasmic reticulum (SR) and influence contractile function in VSMCs. The goals of the present study were to determine the mechanism by which normocapnic and hypercapnic acidosis (pH 7.0) causes arteriolar dilation and the roles of Ca²⁺ signaling and BK_Ca channels in this process.

Methods

Pressurized parenchymal arterioles (PAs)

Animal procedures used in this study were in accordance with institutional guidelines and approved by the Institutional Animal Care and Use committee of the University of Vermont. Male mice were euthanized by intraperitoneal injection of sodium pentobarbital (100 mg/kg) followed by rapid decapitation. The brain was removed and placed into 4°C MOPS-buffered saline. PAs (10–40 μm in diameter) were dissected from the middle cerebral artery territory. Pre-capillary arteriolar segments were then mounted and pressurized in an arteriograph system (Living Systems Instrumentation, Inc.) containing artificial CSF solution, and inner diameter or Ca²⁺ release events were measured as described in the supplementary methods.

Solutions

The composition of artificial cerebrospinal fluid (aCSF) was (in mmol/L): 125 NaCl, 3 KCl, 26 NaHCO₃, 1.25 NaH₂PO₄, 1 MgCl₂, 4 D-glucose, 2 CaCl₂; pH was 7.4 when aerated with 5% CO₂ and 7.0 when aerated with 15% CO₂ (hypercapnia). In normocapnic conditions, acidic pH was obtained by decreasing NaHCO₃ concentration (in mmol/L): 23 for pH 7.3, 19 for pH 7.2, 15 for pH 7.1, 11 for pH 7.0, 8 for pH 6.9. NaCl concentration was increased in order to keep [Na⁺]_o constant at 152.25 mmol/L.

Drugs

Paxilline was purchased from A.G. Scientific, Inc. Ryanodine and U46619 were purchased from Enzo Life Sciences. NS11021 was kindly donated by Dr. Søren-Peter Olesen, Neurosearch A/S, Ballerup, Denmark. Iberiotoxin (IBTX) was purchased from Peptides International. All other chemicals were purchased from Sigma-Aldrich.

Data Analysis and Statistics

Changes in arterial diameter were calculated as % change from baseline: [(change in diameter)/initial diameter] or calculated as % of maximal dilation: (change in diameter)/(maximal diameter − initial diameter). Ca²⁺ image experiments were analyzed with custom software (SparkAn) created by Dr Adrian D. Bonev in our laboratory. Fractional fluorescence (F/F_o) was determined by dividing the fluorescence intensity (F) within a region of interest by a mean ₂₊ fluorescence value (F_o) determined from 10 images before stimulation and without any Ca events. Data are expressed as means ± s.e.m. Differences between two means were determined using Student t test. Statistical significance was tested at 95% (P<0.05) confidence level.

Results

Intact parenchymal arterioles display little functional BK_Ca or Ca²⁺ spark activity at normal pH

Elevation of intravascular pressure to 40 mm Hg constricted mouse parenchymal arterioles by 40%, which is similar to the high level of myogenic tone previously observed in rat parenchymal arterioles (PAs) ^{21, 8, 22, 7}. The mean diameter after development of myogenic tone was 19.1 ± 1.5 μm and the mean maximal diameter was 31.9 ± 2.1 μm (n = 31 PAs). Inhibition of BK_Ca channels by paxilline (1 μmol/L) or RyRs by ryanodine (10 μm/L) caused very small constrictions (3.3 ± 0.1%, n=6 and 3.9 ± 1.4%, n=11 respectively) (Figure 1A, 1B, and 1D). Responses to the thromboxane analog U46619 attested to the robust constrictor capability of the arterioles (Figure 1A). Smooth muscle cells in myocytes from PAs have a BK_Ca channel current density comparable to pial arteries ⁷. To test the functionality of the BK_Ca channel we used the agonist NS11021 (3 μmol/L) ²³, which induced a robust, iberiotoxin-sensitive dilation of intact PAs (Figure 1C and 1D). Caffeine was used to test for the presence of functional RyRs. Caffeine, by opening simultaneously all RyRs, typically triggers a large increase in cytosolic Ca²⁺, leading to vasoconstriction ²⁴. Caffeine, in a concentration-dependent manner, caused vasoconstriction, which was prevented by the RyR inhibitor ryanodine, supporting the presence of functional RyRs (Figure 1E and 1F). These results indicate that BK_Ca and RyRs in parenchymal arteriolar smooth muscle cells have little influence on vessel diameter in the absence of exogenous activators.

Figure 1.

Mouse parenchymal arterioles express functional BK_Ca channels and RyRs but the myogenic response is not counterbalanced by Ca²⁺ spark driven BK_Ca currents under basal conditions. Typical recordings of the internal diameter of pressurized parenchymal arterioles (40 mm Hg) during the perfusion of the BK_Ca blocker paxilline 1 μmol/L (A), RyR blocker ryanodine 10 μmol/L (B), or BK_Ca channel agonist NS11021 3 μmol/L (C). (D) Summary data showing change in luminal diameter expressed as percentage change in baseline diameter with the number of experiment indicated in parentheses. *P < 0.05 (E) typical recordings of the internal diameter showing the vasoconstriction induced by noncumulative addition of caffeine in papaverine 100μmol/L in absence (left) or presence (right) of ryanodine 10 μmol/L. (F) summary data showing change in luminal diameter expressed as percentage of baseline diameter. (n = 5).

Intracellular Ca²⁺ signaling in VSMCs has a central role, particularly in regard of the activation of BK_Ca channels. To examine spontaneous Ca²⁺ signals, intact pressurized PAs were loaded with a fluorescent Ca²⁺ dye (Fluo-4). We found that 78% of the VSMCs exhibited asynchronous Ca²⁺ waves propagating through the cytoplasm over a duration of about 10 s, and very few localized Ca²⁺ events (Figure 2A, 2B and 2C). Figure 2A illustrates the difference in spontaneous Ca²⁺ signaling between pial arteries and parenchymal arterioles, with similar levels of tone. In pial arteries (proximal middle cerebral artery) about 80% of the cells exhibited typical brief (<0.5 s) and localized (<20 μm²) Ca²⁺ events (Ca²⁺ sparks), whereas less than 5% of the cells displayed propagated Ca²⁺ waves (Figure 2C). In PAs only 12% of cells exhibited Ca²⁺ sparks.

Figure 2.

Spontaneous Ca²⁺ signals in intact cerebral arteries. (A) Representative recordings showing Ca²⁺ sparks in pressurized (80 mm Hg) mouse pial artery (middle cerebral artery, ~100 μm diameter) and Ca²⁺ waves in pressurized (40 mm Hg) mouse parenchymal arterioles (~ 15 μm diameter). Squares on the photographs are regions of interest, where F/F0 is measured. Spontaneous Ca²⁺ signals (traces below photographs) were detected as F/F₀ from regions of interest marked as colored squares in the top panels. (B) Progression of a Ca²⁺ wave from one end of the cell to the other (bracketed cell from a parenchymal arteriole). Images are color coded as indicated by the color bar. Images were acquired every 35.9 ms (for display, every twentieth image is shown). (C) Summary data of the percent of cells exhibiting Ca²⁺ sparks and Ca²⁺ waves in pial arteries and parenchymal arterioles.

Depletion of Ca²⁺ stores by application of the sarcoplasmic reticulum Ca²⁺- ATPase inhibitor cyclopiazonic acid (30 μmol/L) abolished spontaneous Ca²⁺ waves in all smooth muscle cells of PAs after 3 minutes (n=13 cells). Treatment with ryanodine (10 μmol/L) for 30 minutes, had a similar effect (n=22 cells). To further investigate the role of the RyRs in the generation and propagation of the Ca²⁺ waves, we have tested the effect of the RyR inhibitor tetracaine. In pressurized PAs, perfusion with tetracaine (100 μM) blocked Ca²⁺ waves in 99% of the cells (n=20 cells). Ca²⁺ sparks were not observed in PAs exposed to ryanodine or tetracaine. These results indicate that RyRs mediate both Ca²⁺ waves and sparks in PAs.

Acidic pH converts Ca²⁺ waves to Ca²⁺ sparks

The preceding data indicate that Ca²⁺-spark activity and BK_Ca channel function in PA myocytes are minimal under normal conditions. We next tested the hypothesis that acidosis, a powerful physiological vasodilator signal, relaxes VSMCs in cerebral parenchymal arterioles by enhancing Ca²⁺ spark activity. To test this hypothesis, pH was decreased from 7.4 to 7.0 by decreasing the bicarbonate concentration in the aCSF. This intervention led to a sizeable decrease in Ca²⁺ wave activity (only 9% of cells exhibited waves at pH 7.0, data not shown), accompanied by a robust increase in Ca²⁺ spark activity (Figure 3). Acidic pH induced an increase in both spark frequency and the percentage of cells with Ca²⁺ sparks (Figure 3B). Since this increase in Ca²⁺ spark activity should induce a parallel increase in BK_Ca channel activity ⁹, we tested the hypothesis that acidosis induces an acute dilator response of intracerebral arterioles through Ca²⁺ spark-induced activation of BK_Ca channels.

Figure 3.

Acidic pH reshapes the intracellular Ca²⁺ dynamic from Ca²⁺ waves to Ca²⁺ sparks. (A) Spontaneous Ca²⁺ signals recorded from the same regions of interest of a pressurized parenchymal arteriole at pH = 7. 4 and pH = 7.0. (B) Compiled data showing Ca²⁺ spark frequency (sparks/cell/s, black bars) and the percentage of cells presenting sparks (% of cells with spark, grey bars) at different pHs. At pH 7.2, both the percentage of cells with Ca²⁺ sparks and spark frequency became significantly different from values observed at pH 7.4 (P < 0.05).

BK_Ca or RyR blockers inhibit acidosis-induced dilation

PAs responded to both normocapnic and hypercapnic extraluminal acidosis with a fast, stable, and reproducible vasodilation of about 70% (Figure 4A, 4C and 4D). Ryanodine (10 μmol/L) inhibited 65.5% of the acidic pH-induced dilation in normocapnic conditions and 60.1% in hypercapnic conditions (Figure 4A, 4C and 4D). Paxilline 1 μmol/L induced a similar decrease (60.2%) in the normocapnic acidosis-induced dilation (Figure 4B and 4D). In hypercapnic conditions, ryanodine and paxilline together did not inhibit the dilations beyond that observed during paxilline alone (not shown) or ryanodine alone (Figure 4B and 4D). These results indicate that the Ca²⁺ sparks observed at pH 7.0 indeed activate BK_Ca channels and contribute to the vasodilator response. They also suggest that BK_Ca channels are not directly activated by protons or CO₂ because RyR and BK_Ca blockers have non-additive effects.

Figure 4.

Normocapnic and hypercapnic acidosis dilate pressurized parenchymal arterioles by activating Ca²⁺ spark-driven BK_Ca channel activity. (A) Typical recording of the internal diameter of a pressurized parenchymal arteriole (40 mm Hg) during the perfusion of acidic aCSF under normocapnic conditions, without and with the RyR blocker ryanodine (10 μmol/L). (B) Effect of BK_Ca channel blocker paxilline (1 μmol/L) on acidic pH-induced dilation under normocapnic conditions. (C) Effect of RyR blocker ryanodine (10 μmol/L) and BK_Ca channel blocker paxilline (1 μmol/L) on acidic pH-induced dilation under hypercapnic conditions. (D) Summary data showing vasodilator responses related to initial level of myogenic tone (0%) and fully relaxed diameter (100%; 0 Ca²⁺, papaverine 100 μmol/L), with the number of experiments indicated in parentheses. *P < 0.05

Acidosis-induced dilation depends on BK_Ca channel expression

To further examine the role of BK_Ca channels in acidosis-induced dilation, we used a genetic knockout approach with mice lacking the pore-forming α subunit of the BK_Ca channel (Slo^−/−/Kcnma1^−/−) ²⁵. In the absence of the functional channel, the dilation of intact PAs induced by the agonist NS11021 (1 μmol/L) was reduced to the one observed in wild-type mice in presence of paxilline (1 μm/L) (Figure 5A and 5B). The BK_Ca channel blocker paxilline did not affect tone, or the residual dilations to pH or NS11021 of arterioles from Kcnma1^−/− mice. Moreover the dilations to reduced pH, in both normocapnia and hypercapnia conditions, were substantially attenuated in arterioles from the Kcnma1^−/− mice (Figure 5A and 5B). Indeed, the percent reduction in the response to acidosis in BK_Ca knockout versus wild-type arteries was very close (57.5 % in hypercapnia and 55.2 % in normocapnia) to that observed in normal arteries treated with ryanodine or paxilline versus control. These results support the important role of BK_Ca channels in acidic pH-induced vasodilation.

Figure 5.

Hypercapnic-, normocapnic- and NS11021-induced dilations are decreased in Slo-KO mouse. (A) Typical recordings of the internal diameter of pressurized parenchymal arterioles (40 mm Hg) from a wild type mouse (Kcnma1^+/+, upper trace) or a mouse lacking the α subunit of the BK_Ca channel (Kcnma1^−/−, lower trace) during the perfusion of acidic aCSF under hypercapnic or normocapnic conditions, or in presence of the BK_Ca channel agonist NS11021, in the absence or presence of the BK_Ca channel blocker paxilline. (B) Summary data showing vasodilator responses related to initial level of myogenic tone (0%) and fully relaxed diameter (100%; 0 Ca²⁺, papaverine 100 μmol/L), with the number of experiments indicated in parentheses. *P < 0.001

Acidosis-induced dilation is not affected by K_ATP channel inhibition in mouse PAs

K_ATP channels are directly activated by protons in VSMCs ¹⁸. However, in normocapnic conditions, the selective K_ATP channel inhibitor glibenclamide (1 μmol/L) failed to inhibit pH 7.0-induced dilation (n=6, Figure 6A). Higher glibenclamide concentrations (3 μmol/L or 5 μmol/L) also did not alter the vasodilation (n=4 and n=3 respectively, data not shown). Similarly, hypercapnia dilated PAs by 69.4 ± 2.9% and by 68.2 ± 2.0% in the absence and presence of glibenclamide (1 μmol/L) respectively (n=6). To test for the presence of functional K_ATP channels, cumulative additions of the K_ATP channel opener cromakalim were applied to PAs and to third order mesenteric arteries from the same animal (Figure 6B, 6C, and 6D). Cromakalim (3 μmol/L) relaxed mesenteric arteries up to 65.1 ± 9.8% and this action was fully reversed by glibenclamide (1 μmol/L) (Figure 6C and 6D). In contrast, cromakalim had no effect on PA diameter, suggesting these arterioles lack functional K_ATP channels.

Figure 6.

Acidic pH-induced dilation of PAs is not affected by glibenclamide. (A) Typical recording of the internal diameter of a pressurized (40 mm Hg) parenchymal arteriole (PA) during the perfusion of acidic aCSF under normocapnic conditions with and without the K_ATP blocker glibenclamide (1 μmol/L). (B) Absence of effect of K_ATP channel agonist cromakalim on a parenchymal arteriole. (C) Vasodilation of a pressurized (60 mm Hg) third order mesenteric artery (Mesenteric A) induced by cumulative addition of Cromakalim (10⁻⁸ to 3.10⁻⁸ mol/L). Glibenclamide (1 μmol/L) completely blocked the cromakalim-induced dilation of mesenteric artery. (D) Cromakalim concentration response relationships in parenchymal arterioles (n = 5) and mesenteric arteries (n = 5). Dilation is related to initial level of myogenic tone (0%) and fully relaxed diameter (100%).

Inhibition of endothelial nitric oxide synthase (eNOS) does not significantly affect acidosis induced-dilation

The eNOS inhibitor L-NAME (100 μmol/L) induced a modest constriction of PAs under normal conditions of pH, temperature, and pressure, which is not surprising given the prominent tonic influence of NO on PA tone ^{7, 21}. Nevertheless, the dilations induced by normocapnic or hypercapnic acidosis were not different in the absence and presence of L-NAME (Figure 7).

Figure 7.

NOS-inhibition does not alter dilations induced by acidic pH_o in normocapnic and hypercapnic conditions. (A) Typical recording of the internal diameter of a pressurized parenchymal arteriole (40 mm Hg) during the perfusion of acidic aCSF induced by reduced bicarbonate concentration (normocapnia) with and without the eNOS inhibitor L-NAME (100 μmol/L). (B) Summary data (n = 4). (C) Typical recording of the internal diameter of a pressurized parenchymal arteriole (40 mm Hg) during the perfusion of acidic aCSF induced by increased pCO₂ (hypercapnia) with and without the eNOS inhibitor L-NAME (100 μmol/L). (D) Summary data (n = 6). Dilation is expressed as percent of maximum dilation (0 Ca²⁺, papaverine 100 μmol/L).

Decreasing extracellular pH inhibits calcium channel activation by Bay K 8644

In the presence of paxilline 1 μM, we measured the vasoconstriction induced by increasing concentrations of the L-type channel agonist Bay K 8644, at pH 7.4 and pH 7.0. Decreasing extracellular pH significantly increased the Bay K 8644 EC₅₀ (p = 0.0005) from 16.3 nmol/L (95% CI 9.1 nmol/L to 29.1 nmol/L) at pH 7.4, to 65.4 nmol/L (95% CI 47.3 nmol/L to 90.4 nmol/L) at pH 7 (Online Figure I). This result suggests an inhibition of the gating propriety of the channel by protons.

Discussion

Acidosis activates Ca²⁺ sparks and BK_Ca channels

Acidification of brain tissue can occur with reduced respiratory rate, during CO₂ inhalation, or during cerebral lactic acidosis related to ischemia ²⁶ and hypoxia ²⁷. Under those conditions, a decrease in extracellular pH of 0.4 units occurs rapidly, e.g. within seconds during global ischemia. This tissue acidification leads to cerebral vasodilation. We found that acidification from pH 7.4 to 7.0 rapidly dilated parenchymal arterioles (Figure 4 and 5) and that this dilator response is correlated with a dramatic increase in Ca²⁺ spark activity and BK_Ca channel function (figure 3). Each Ca²⁺ spark delivers micromolar Ca²⁺ to the adjacent BK_Ca channels to increase its open probability ~10⁵- to 10⁶-fold ²⁸, causing up to a 20 mV hyperpolarization ²⁹. Our results are consistent with the idea that acidification increases Ca²⁺ spark activity, which in turn activates BK_Ca channels to hyperpolarize and dilate PAs. The present report is the first investigation of the effects of acidosis on Ca²⁺ release from the SR in vascular smooth muscle and it implies a prominent role for Ca²⁺ sparks as a mediator of cerebral vasodilation associated with acidosis.

Reshaping of RyRs-dependent Ca²⁺ release by protons

At physiological pH, Ca²⁺ spark activity in PA myocytes is low, and Ca²⁺ wave activity predominates. The slight vasoconstriction observed in the presence of 1 μmol/L paxilline or 100 nmol/L IBTX in the present study suggests that RyR-mediated Ca²⁺ release in the form of waves does not cause sufficient BK_Ca channel activation to modulate membrane potential, and hence diameter. The simplest interpretation of this result is that the necessary micromolar Ca²⁺ concentrations are not reached at the BK_Ca channel, likely because the waves disperse the released Ca²⁺ ions whereas the sparks concentrate Ca²⁺ in spatially limited areas. Acidosis causes a fundamental shift of this Ca²⁺ signaling dynamic from waves to sparks. While we have not directly assessed the mechanism by which this reshaping of the Ca²⁺ signal events occurs, we propose the following as a possible explanation. The kinetic properties of Ca²⁺ waves and their inhibition by both ryanodine and tetracaine strongly suggest that Ca²⁺ waves travel through the VSMCs by a RyR-dependent Ca²⁺-induced Ca²⁺ release mechanism (CICR) ³⁰. By this mechanism, a local increase in Ca²⁺ activates neighboring RyRs because the resting open probability (P₀) of the RyRs is high ³¹. However, if the resting P₀ is low, RyR excitability is not sufficient to induce the regenerating process described above and Ca²⁺ release remains limited to activation of single RyR (Ca²⁺ quark) or clusters of RyRs (Ca²⁺ sparks) ³¹. From these observations and because PAs show more Ca²⁺ waves than the pial arteries, we propose that the resting P₀ of RyRs in PAs is higher than in large diameter pial arteries, possibly due to altered activity of one or more RyR modulators including [Ca²⁺]_i, the FK506-binding protein, cyclic ADP-ribose, the RyR phosphorylation state ³² or perhaps intracellular pH. Recent studies have suggested that spontaneous Ca²⁺ sparks in smooth muscle are likely encoded by the RyR2 isoform ^{33, 34}, which we have found to be expressed in PA myocytes, along with RyR1 and RyR3 (data not shown). At physiological pH, the lack of Ca²⁺ sparks, and thereby diminished contribution of BK_Ca channels to control of membrane potential could explain the enhanced myogenic tone observed in parenchymal arterioles. Physiologically, enhanced myogenic tone may allow a higher “vasodilator reserve” during functional hyperemia but also a greater tendency towards vasospasm associated with pathologies.

The P₀ of RyRs is modulated by intracellular pH. For instance in cardiac sarcoplasmic reticulum vesicles ³⁵ and in planar lipid bilayers ^{36, 37} an increase in proton concentration decreases P₀. Further, acidic pH decreases the efficiency of CICR in skeletal muscle ³⁸ and reduces Ca²⁺ spark frequency in rat ventricular myocytes ³⁹. Conversely alkalinization increases the Ca²⁺ sensitivity of RyRs ⁴⁰. We previously demonstrated that increases in pH to 7.6 and beyond changed the primary Ca²⁺ signaling modality from Ca²⁺ sparks to ryanodine-sensitive Ca²⁺ waves in large diameter pial artery smooth muscle ⁴¹. However, although our previous study demonstrated the apparent pH sensitivity of RyRs in vascular smooth muscle cells, the functional link between altered Ca²⁺ signaling dynamics and vascular tone appeared minimal since the constriction induced by alkalosis was mediated mostly (82%) by direct enhancement of voltage-dependent Ca²⁺ channel activity. The mechanism described in the present study relies on the same regulation of RyR P₀ by protons. However, here, we provide genetic and pharmacological evidence that the reversible activation of Ca²⁺ sparks and BK_Ca channels is a major mechanism by which acidosis, over a narrow and physiologically relevant pH range, induces cerebral vasodilation.

Role of potassium channels in acidic-pH induced dilation

Our results indicate that substantial portion (60%) of acid pH-induced dilation depends on RyR-mediated Ca²⁺ spark activation of BK_Ca channels. A number of studies have also suggested involvement of BK_Ca channels in acidic pH-induced vasodilation ⁴². On the contrary, in rat penetrating arterioles, vasodilation to acid was not reduced by the BK_Ca blocker TEA ¹². However, species and methodological differences between this previous study and ours, particularly in the perfusion solution composition and the drugs used, make direct comparisons difficult.

Many studies support a role for K_ATP channels in the response to acidosis in mesenteric and cerebral arteries, including brain penetrating arterioles ^{12, 18, 42}. In light of this previous work, the lack of effect of glibenclamide on both normocapnic and hypercapnic acidosis-induced vasodilation shown here was not expected. However the absence of dilation in response to the K_ATP agonist cromakalim, observed in the present study, strongly suggests a lack of functional K_ATP channels in mouse PAs. Interestingly, rat arterioles, lacking perivascular nerves, also are _43, ⁴⁴ (for review see ⁴⁵). Another study resistant to the activation of K_ATP channels by cromakalim found that glibenclamide attenuates the response of rabbit pial arterioles in vivo to low level of hypercapnia but it is ineffective with more intense hypercapnia ⁴⁶. This is consistent with the suggestion that K_ATP channels are activated only by mild hypercapnia ⁴⁷. However, other studies have found that vasodilation induced by direct, strong acidification is inhibited by glibenclamide ₁₂⁴². It appears that additional studies will be needed to clarify the specific role of K_ATP channels in the vasodilator response to reduced pH.

The RyR- and BK_Ca-independent component of dilation induced by acidosis

Multiple in vivo studies report a role for NOS in the increase of cerebral blood flow elicited by hypercapnia ^{48, 46}. Interestingly, using an in vivo endothelial injury model, Wang and coworkers reported, that the origin of NO in hypercapnic cerebrovascular dilation was nonendothelial, and likely due to neuronal NOS activation ⁴⁹. Their conclusion is supported by the absence of effect of a NOS inhibitor on the CO₂-induced increased blood flow in mice lacking neuronal NOS ⁵⁰. In vitro, relaxation of rat mesenteric arteries induced by hypercapnia is not altered by endothelial removal ¹⁵. A more recent study using isolated pial arteries implies that the regulation of the arteriolar diameter by pH involves NO released from perivascular nerves and not from endothelial cells ¹⁹. Since we used isolated brain parenchymal arterioles, which lack extrinsic innervation, our observations support the concept that the NOS-dependent component of increased cerebral blood flow elicited by hypercapnia in vivo, results from neuronal NOS activation. In situ, this neuronal source of NO is likely derived from neurons that encompass the “neurovascular unit” along with arterioles and astrocytes.

In the present study, vasodilation induced by normocapnic or hypercapnic acidosis was not significantly different when eNOS or K_ATP channels were inhibited. However, a residual dilation, about 40% of the initial response, remained when RyR or BK_Ca channel blockers were applied during acidic pH induced-dilation. Vascular L-type Ca²⁺ channels are inhibited by protons both internally and externally ^{17, 51} and this mechanism has been proposed as a contributor to the decrease in [Ca²⁺]_i and tone observed when pH is reduced ¹⁶. Therefore the BK_Ca- and RyR-independent dilation observed at pH 7.0 (figure 4 and 5), is likely induced by the inhibition of L-type channels by protons. This hypothesis is supported by the observed decrease in the vasoconstriction induced by L-type channel agonist Bay K 8644 during acidosis.

Summary

Inhibition of RyRs or BK_Ca channels has little effect on the diameter of pressurized parenchymal arterioles at physiological pH. Normocapnic or hypercapnic acidosis converts smooth muscle Ca²⁺ waves to sparks, and causes profound vasodilation, which is significantly reduced by inhibitors of RyRs or BK_Ca channels (Figure 8). Our results support the novel concept that pH modulates local Ca²⁺ signaling (“Ca²⁺ sparks”) to regulate BK_Ca channel activity, and thereby dramatically influence arteriolar tone.

Figure 8.

Proposed mechanism for RyRs- and BK_Ca-dependent, acidosis-induced dilation of brain parenchymal arterioles. By decreasing the P₀ of RyRs in VSMCs, protons reshape intracellular Ca²⁺ signaling from waves to sparks which activate BK_Ca channels and then cause dilation.

Novelty and Significance.

What Is Known?

• Cerebral blood flow (CBF) is highly regulated by pH, with acidification causing profound vasodilation.

• In arteries on the surface of the brain (pial arteries), activation of large conductance potassium calcium-activated channels (BK_Ca) channels by local Ca²⁺ signals (Ca²⁺ sparks) through ryanodine receptors opposes vasoconstriction.

• Unlike pial arteries, inhibition of BK_Ca channels has little effect on the diameter of (parenchymal) arterioles within the brain, even though their smooth muscle cells have a significant BK_Ca channel density.

What New Information Does This Article Contribute?

• Ca²⁺ waves are the predominant spontaneous Ca²⁺ signals in parenchymal arteriolar smooth muscles cells.

• Acidosis reshapes spontaneous Ca²⁺ waves into Ca²⁺ sparks which activate BK_Ca channels.

• This mechanism accounts for 60% of acidosis-induced dilation of parenchymal arterioles.

Summary.

Regulation of CBF by changes in pH has been recognized as a critical homeostatic mechanism for more than a century, yet little is known about this effect in parenchymal arterioles. Here, we provide genetic and pharmacological evidences that the reversible activation of Ca²⁺ sparks and BK_Ca channels is a prominent mechanism by which acidosis, over a narrow and physiologically relevant pH range, induces cerebral vasodilation. Our study supports the novel concept that protons can induce the conversion of smooth muscle Ca²⁺ waves to Ca²⁺ sparks, which triggers activation of BK_Ca channels, causing dilation. When pH is shifted from 7.4 to 7.0, this mechanism accounts for about 60% of a substantial and maintained vasodilator response. A change of 0.4 units in pH can occur within seconds in brain tissue during reduced respiratory rate, CO₂ inhalation, or cerebral lactic acidosis related to ischemia and hypoxia. The mechanism presented here likely is relevant to these pathological situations. Moreover, the parenchymal arterioles, along with neurons and astrocytes, constitute a functional neurovascular unit that coordinates blood flow and neuronal activity. The present study leads us to propose that protons, through modulation of Ca²⁺ signaling, may mediate or modulate neurovascular coupling under normal and pathophysiological conditions.

Non-standard abbreviations

PAs

parenchymal arterioles

BK_Ca

large conductance potassium channel

VSMC

vascular smooth muscle cell

RyR

ryanodine receptor

CBF

cerebral blood flow

VDCC

voltage-dependent Ca²⁺ channel

SR

sarcoplasmic reticulum

aCSF

artificial cerebrospinal fluid

IBTX

iberiotoxin

CICR

Ca²⁺-induced Ca²⁺ release

P_o

open probability