INTRODUCTION

TOP ABSTRACT INTRODUCTION METHOD RESULTS DISCUSSION REFERENCES

PROXIMITY OF SITES of neurotransmitter release and neuronal electrical activity to the cerebrovasculature facilitates coupling of neuronal activity and cerebral blood flow (CBF). This communication depends on neuronal signals released synaptically [e.g., nitric oxide (14)] or directly linked to electric current flow through excitable membranes [e.g., K⁺ lost during action potential repolarization (23, 31)]. Because neuronal activity and synaptic transmission also cause immediate changes in tissue energy balance, coupling between CBF and neuronal activity may also depend on substances released in the abluminal space during periods of increased metabolic demand; the ATP-derived neurotransmitter and vasodilator adenosine has thus been referred to as a "metabolic" regulator of CBF (39). The resulting interactions of metabolic, synaptic, and ionic signals ensure local and dynamic regulation of CBF.

The anatomic relationship between cortical blood vessels and neurons suggests that small caliber arterioles are the target for neuronal-vascular interactions. Because of technical limitations, however, studies of cerebrovascular regulation have been limited to observations on large-diameter or superficial vessels (>100 µm (8, 30)]. Results obtained from these vessels suggested that K⁺-mediated dilations are mediated by voltage-dependent, Ba²⁺-sensitive inward rectifier channels [K_IR (30-32)]. Because Ba²⁺ also blocks metabolically regulated, ATP-sensitive channels expressed in vascular smooth muscle (VSM; see Ref. 31), we investigated the possible involvement of these ion channel mechanisms in K⁺-mediated dilations.

In addition to standard patch-clamp recording and cell isolation methodology, we took advantage of the technique developed originally by Dacey and Duling (4) to obtain resistance-size arterioles from rat neocortex; these vessels are responsible for the transduction of neuronal signals into vascular changes and thus regulate metabolic supply and oxygenation of deep cortical layers. We have previously shown that these vessels express ATP-sensitive K⁺ (K_ATP) channels and respond to drastic intracellular ATP concentration ([ATP]_i) changes (in vitro ischemia) by activation of glibenclamide-inhibitable dilations (17). It remains to be elucidated whether more physiological stimuli, such as modest elevation of parenchymal K⁺ comparable to those occurring during cortical activation (36), can cause similar coupling of metabolic changes to activation of K⁺ channels. Interestingly, McCarron and Halpern (26) reported two distinct mechanisms of cerebrovascular K⁺ dilations, one sensitive to ouabain, and presumably involving Na⁺-K⁺-ATPase activation by extracellular K⁺ concentration ([K⁺]_out), and one with a Ba²⁺-sensitive component mediated by ion channels. It was concluded that a synergistic mechanism involving activation of pump activity and ion channels underlies K⁺-induced dilations. However, a direct link between pump activity and K⁺ channels has only recently been demonstrated in cardiac and renal cells (20, 37).

The main goal of our study was to elucidate the mechanisms responsible for K⁺ dilations of resistance-size cerebral vessels. We tested the hypothesis that both K_IR- and K_ATP-dependent pathways are involved and that different VSM muscle ion channels mediate the dilations in response to prolonged or transient neuronal activity.

	METHOD

TOP ABSTRACT INTRODUCTION METHOD RESULTS DISCUSSION REFERENCES

Methodology for the in vitro isolation and cannulation of rat cerebral arterioles has been described in publications from this laboratory (17, 28, 29). Sprague-Dawley rats (300 g) were anesthetized with pentobarbital sodium (50 mg/kg ip) and decapitated. The brain was rapidly removed from the skull and immersed in cooled buffered saline solution (4°C) containing 1% dialyzed BSA and the following (in mM): 144 NaCl, 3.0 KCl, 2.5 CaCl₂, 1.5 MgSO₄, 5.0 glucose, 2.0 pyruvate, 0.02 EDTA, 2.0 MOPS, and 1.21 NaH₂PO₄. A section of cerebral cortex ~2 mm thick and containing the first portion of the middle cerebral artery was dissected from the brain. The pia mater and its attached penetrating intracerebral arterioles were separated from the parenchyma, and an unbranched distal segment of a vessel, ~0.5 mm in length, was severed from the pia. The vessel segment was then transferred to a temperature-controlled chamber (3.0-ml volume) mounted on the stage of an inverted microscope (Nikon). The isolated vessel was cannulated using a system of concentric glass pipettes (4) consisting of a perfusion pipette within a holding pipette. The pipettes were inserted into Plexiglas holders (White Instruments, Suitland, MD), mounted on micromanipulators, and attached to the microscope stage.

Before the development of physiological tone, "passive" vessel diameter was measured. The bath solution was then changed to one without albumin, and bath temperature was raised to 37°C. The rate of intraluminal perfusion (4 µl/min) was chosen to avoid flow-mediated effects on vessel diameter (29). The extraluminal bath medium was continuously circulated with a roller pump at 1 ml/min. After an equilibration period of ~30 min, viable arterioles develop vasomotion and contract spontaneously. The vessels must constrict to <70% of the passive diameter to be usable. All drugs, including K⁺ (as KCl, isomolar substitution for NaCl), were applied extraluminally. Glibenclamide was initially dissolved in DMSO and then was added to the perfusate.

VSM cells were isolated from the rat basilar artery as follows. Male Sprague-Dawley rats were killed, and the brain was removed. The basilar arteries were dissected in ice-cold, low-Ca²⁺ balanced salt solution (BSS). The BSS contained (in mM) 134 NaCl, 5.2 KCl, 1.2 MgSO₄, 0.05 CaCl₂, 10 HEPES, 11 glucose, 0.33 NaH₂PO₄, 4 dithiothreitol, and 0.06 papaverine hydrochloride, and 0.01% fatty acid-free BSA, pH 7.35. The enzymatic digestion protocol for the cerebral arteries was as follows. Vessels were transferred to a 15-ml centrifuge tube, and the following were added in 1 ml of BSS: collagenase (type II, 2 mg/ml, 159 U/mg; Worthington), elastase (porcine pancreas, 0.5 mg/ml, 4.8 U/mg; Worthington), and soybean trypsin inhibitor (type I-S, 1 mg/ml; Sigma). The vessel was then incubated at 36°C on an orbital shaker (45 rpm) for 30 min. The enzyme solution was replaced with a new enzyme solution (0.4 mg/ml BSA, 0.4 ml/ml trypsin inhibitor, and 0.4 mg/ml protease), and incubation continued for another 10 min. With the use of an Eppendorff pipette with the tip cut to 2 mm in diameter, the vessel was then transferred to ice-cold BSS (1 ml, in microfuge tube) and mechanically dissociated to release VSM cells. After a wash (200 g), the cell pellet was resuspended in BSS and Ca²⁺ raised slowly by dilution with BSS containing 1.6 mM Ca²⁺ to a final Ca²⁺ concentration of 0.2 mM. Cells are normally refrigerated until use.

Patch-clamp recordings were performed as described previously by us for endothelial cells (16, 18) or following procedures specific for VSM (3, 30, 31). VSM cells were bathed in artificial cerebrospinal fluid (aCSF) composed of (in mM) 120 NaCl, 3.1 KCl, 1 MgCl₂, 2 CaCl₂, 5 MOPS, 26 NaHCO₃, and 10 dextrose. Experiments were performed at room temperature (24-26°C). Experiments with high K⁺ were performed by adding potassium gluconate to aCSF; an isomolar concentration of NaCl was removed to maintain osmolarity. Patch-clamp recordings were obtained using an Axopatch 1C amplifier (Axon Instruments, Foster City, CA) in voltage- or current-clamp mode. Whole cell recordings were obtained with pipettes filled with (in mM) 140 potassium gluconate, 1 MgCl₂, 2 Na₂-ATP, 0.3 NaGTP, 10 HEPES, and 0.5 EGTA, final pH of 7.2 (with NaOH). Pipettes had a resistance of 5-10 M. Series resistance was monitored throughout the experiment and was usually around 15-30 M. Series resistance compensation was routinely performed up to 70-80% (lag time 10 µs). Recordings were digitized at 48 kHz, filtered at 2-10 kHz, displayed on an oscilloscope, recorded on tape, and acquired on a Pentium 266 computer by pCLAMP6 (Axon Instruments). VSM were selected for recording under visual control with a Nikon microscope equipped with Hoffman optics at ×400 magnification. Cell membrane potential was corrected for the tip potential determined upon withdrawal of the pipette from the cell.

	RESULTS

TOP ABSTRACT INTRODUCTION METHOD RESULTS DISCUSSION REFERENCES

The vessels used for our experiments had a mean diameter of 48 ± 2.1 µm (range 45-66 µm). Isolated penetrating pial vessels developed spontaneous tone when cannulated and perfused intraluminally (4, 17). As shown in Fig. 1A, vessel diameter was measured approximately halfway through the length of the cannulated vessel, and these determinations were then repeated at the desired intervals at the same cursor location. Similar to in vivo arterioles, these vessels readily respond to pH and various vasodilators. Vessels exposed abluminally to modest K⁺ increases dilated promptly and reversibly (Fig. 1). Cerebrovascular responses to elevated K⁺ were characterized by steep concentration dependency, and small K⁺ elevations (5-10 mM) caused dilation while, at K⁺ = 47 mM, a statistically significant constriction occurred. K⁺-induced dilations were persistent in nature and lasted for the entire duration of K⁺ application (Fig. 1C). These results are in agreement with previous findings obtained from larger cortical vessels (21).

View larger version (30K): [in this window] [in a new window]   Fig. 1.   Effects of increased abluminal K+ on the diameter of penetrating pial vessels. A: photomicrograph of an isolated vessel before and after exposure to elevated extracellular K+ concentration ([K+]out). B: biphasic changes observed after addition of 5, 10, and 47 mM to the bathing medium (external Na+ was reduced to maintain osmolarity constant). Vessel diameters measured at basal [K+]out (3 mM) were normalized and compared with steady-state responses measured after 5 min of incubation with different [K+]out. Six vessels from different animals were used for these experiments, and each vessel underwent applications of 5-47 mM [K+]out. Note that at 5-10 mM K+ caused dilation, whereas a pronounced vasoconstriction was observed at 47 mM (*P < 0.05 vs. 3 mM K+). C: K+ channel blocker Ba2+ (at 10 µM) prevents the development of dilations induced by addition of 5 mM K+. Note the rapid onset and plateau response () and the near complete abolishment of the response by the ion channel blocker (); *P < 0.05 and **P < 0.02. D: K+-induced dilations (5 mM K+ added; ) are reduced by Ba2+ in a dose-dependent fashion. K+-induced dilations were abolished by Ba2+ in the range 1-10 µM (*P < 0.02; **P < 0.05 compared with K+ dilations with no K+ channel blocker; n = 5). Response to Ba2+ applied to vessel exposed to basal concentrations of K+ is also shown (); at concentrations <50 µM, Ba2+ had little or no effect, whereas a vasoconstriction followed application of Ba2+ >50 µM.

Consistent with the persistent nature of K⁺ dilations, sudden diameter increases elicited by 5 mM K⁺ were followed by a plateau. Both instantaneous and sustained dilations to K⁺ were greatly attenuated by low concentrations of the K⁺ channel blocker Ba²⁺ (Fig. 1C); Ba²⁺ itself, at concentrations of <100 µM, had little or no effect (Fig. 1D). At concentrations >100 µM, Ba²⁺ caused constriction.

Similar to Ba²⁺, the K_ATP channel blocker glibenclamide also dramatically reduced dilations induced by small elevations of K⁺ (5 mM above baseline; Fig. 2A); at the concentration used (1 µM), the effects of glibenclamide are specific for blockade of K_ATP (31). In agreement with data by others (26), dilations induced by larger (10 mM) elevations of extracellular K⁺ were insensitive to glibenclamide (114.4 ± 2.7 vs. 110.04 ± 0.81 with 1 µM glibenclamide; n = 6), suggesting that K_ATP involvement is significant only for dilations induced by small increases in [K⁺]_out.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Resistance-size arterioles dilate in response to modest extracellular K+ by activation of ATP-sensitive K+ (KATP) channels. A: At concentrations highly specific for the sulfonylurea-receptor channel complex, glibenclamide (gli) dramatically reduced dilations to 5 mM K+. In contrast, the constriction observed at 47 mM K+ was totally unaffected. Glibenclamide alone had no effect. *Significant effect (P < 0.04; n = 8) of glibenclamide compared with dilations obtained after application of 5 mM K+ in the absence of the blocker. B: preferential effect of glibenclamide on steady-state vs. instantaneous dilations to K+ (5 mM) applied for the period indicated by arrow. *P < 0.05 vs. control. C: dose dependency of the K+ channel blocker Ba2+ on pinacidil-mediated dilations (10 µM). Note that the dilations induced by the KATP channel opener were almost completely abolished at Ba2+ concentrations >1 µM, analogous to K+-induced dilations (Fig. 1C). *P < 0.05 vs. Ba2+ alone.

The effects of glibenclamide on the time course of K⁺-induced dilation differed, however, from the effects of the mixed K_ATP-K_IR blocker Ba²⁺, because this specific K_ATP channel blocker completely prevented steady-state dilations to [K⁺]_out but had little effect on the early dilatatory response (compare Fig. 2B with Fig. 1C). These results suggested that two separate ion channel mechanisms are responsible for the vasodilatatory response to K⁺. Because of its sensitivity to Ba²⁺ and resistance to glibenclamide, the transient response appeared to be mediated almost exclusively by K_IR, whereas glibenclamide blockade of the delayed response was consistent with involvement of K_ATP.

We further tested whether the effects of Ba²⁺ could be consistent with blockade of K_ATP channels rather than an exclusive action on K_IR. To this end, vasodilations induced by the K_ATP agonist pinacidil (10 µM) were challenged with the same range of Ba²⁺ concentrations that abolished steady-state K⁺ dilations. The profound inhibition of K⁺ dilations by Ba²⁺ was paralleled by Ba²⁺ potency to block pinacidil-induced, K_ATP-mediated responses, demonstrating that Ba²⁺ actions on the plateau response to K⁺ were consistent with an effect on K_ATP channels.

Under physiological conditions, K_ATP channels are tonically inhibited by intracellular ATP (18, 24). However, even modest decreases in the ATP to ADP ratio cause opening of K_ATP (13). Tonic inhibition of K_ATP in resting vessels was confirmed in this and other studies, and glibenclamide had little or no effect on vessel diameter at physiological K⁺ concentrations (e.g., Fig. 2A; see also Refs. 6, 8, 17, 31). However, previous results have shown that reduction of ATP by metabolic poisoning with cyanide promptly results in robust dilations sensitive to 1 µM glibenclamide and thus mediated by K_ATP (17). Because high abluminal K⁺ and metabolic poisoning both cause comparable and glibenclamide-inhibitable opening of K_ATP, what could constitute the link between elevated K⁺ and reduction of intracellular ATP sufficient to open K_ATP?

In addition to effects on ion channel gating (21, 30), extracellular K⁺ acts as a powerful activator of Na⁺-K⁺-ATPase (7, 34). Because of the inherent dependency of the Na⁺-K⁺ pump on ATP hydrolysis, we speculated that extracellular K⁺-induced activation of the ATPase may have caused sufficient change in the ATP-to-ADP ratio to open K_ATP. A prediction of this hypothesis is that blockade of the pump concomitant to application of high K⁺ may result in preservation of intracellular ATP, thus preventing K_ATP-mediated vasodilations. This was directly tested by perfusion of the vessels with the pump inhibitor ouabain (0.2 mM) before and during application of K⁺ (Fig. 3A). Ouabain had per se little effect on vessel diameter but fully prevented the steady-state vasodilations induced by elevation of [K⁺]_out. To assess the specificity of ouabain's effects, and to rule out a direct effect on permeation through K_ATP, vessels were exposed to pinacidil, and pinacidil-induced dilations were challenged with the same concentration of the Na⁺-K⁺-ATPase blocker (Fig. 3B). Although ouabain fully prevented K⁺ dilations, responses elicited by the K_ATP opener pinacidil were unaffected.

View larger version (23K): [in this window] [in a new window]   Fig. 3.   K+ dilations are sensitive to preventive Na+-K+ blockade by ouabain. A: pretreatment of the vessels with ouabain (ouab; 200 µM) caused an almost complete abolishment of the dilatatory response to K+ (measured at steady state; the initial response was less sensitive). Note that ouabain per se exerted little effect on vessel diameter, ruling out a toxic effect of the drug. Furthermore, the effects of ouabain were fully reversible upon drug washout (for 15 min). * P < 0.05. B: effects of ouabain are not due to a direct, nonspecific effect on KATP conductance. Robust dilations to the KATP channel opener pinacidil (1 µM) were unaffected by the Na+-K+-ATPase inhibitor; in the same vessels, ouabain abolished the dilation induced by addition of K+ (2.5 mM above baseline). Results were pooled from 5 (A) or 4 (B) experiments. * P < 0.05; ** P < 0.02.

Further direct evidence linking the actions of extracellular K⁺ to opening of K_ATP was obtained by patch-clamp experiments. These experiments were performed on VSM cells isolated from the basilar artery or from penetrating pial vessels (n = 16 and 3, respectively; since the results obtained were identical, findings from these two populations are pooled together). These cells express K_ATP and could therefore be used to determine physiological coupling between K_ATP channels and ATP hydrolysis induced by activation of the Na⁺-K⁺-ATPase. As shown in Fig. 4A, VSM from the basilar artery responded to application of elevated K⁺ (from 3 to 8 mM) with membrane hyperpolarization from cell resting potential (43.5 ± 1.07 mV, n = 14). Furthermore, these membrane responses were attributable to opening of K_ATP since, at concentrations specific for blockade of these channels, glibenclamide greatly attenuated the membrane changes induced by elevated K⁺. Further increase in the concentration of glibenclamide (from 1 to 10 µM) did not significantly increase its blocking actions, suggesting that saturation of the effects was achieved at a concentration highly selective for K_ATP. For these experiments, the whole cell recording pipette contained 3 mM ATP, a concentration that causes complete inhibition of K_ATP. This was supported by the fact that glibenclamide per se did not cause any appreciable change in cell resting potential (n = 6).

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Patch-clamp recordings from VSM isolated from the basilar artery reveal KATP-mediated dilations to [K+]out. A: recordings performed with 2 mM ATP in the pipette, standard intracellular solution, revealed small hyperpolarizations induced by addition of K+ (5 mM above baseline, to a final concentration of 8 mM). These responses were blocked or reduced by glibenclamide (1 µM) applied for >5 min before subsequent challenge with high [K+]out. Glibenclamide per se had no effect on cell resting potential, in agreement with the hypothesis that KATP channels are tonically inhibited under these recording conditions. B: removal of ATP from the pipette solution caused disappearance of K+-mediated responses but revealed large depolarizations induced by glibenclamide. Because resting membrane potential in these cells was significantly more negative than in cells dialyzed with 2 mM ATP (C), positive current was injected to offset this change in passive properties. Cumulative results of these experiments are shown in C. ATP, intracellular ATP. *P < 0.04.

When similar patch-clamp experiments were performed with 0 mM ATP in the pipette, recordings from basilar artery VSM yielded resting potential values that were significantly more negative than those recorded with 3 mM [ATP]_i (64.5 mV, n = 4; Fig. 4C). This suggested that, under these recording conditions, increased conductance through K_ATP occurred. To test this hypothesis, cells were exposed to 1 µM glibenclamide. In contrast to what was observed in ATP-dialyzed cells, cells recorded with pipettes containing 0 mM ATP depolarized after exposure to the K_ATP blocker (+20 ± 5.2 mV on average, n = 4). If intracellular dialysis with 0 mM ATP promotes opening of K_ATP and if K⁺-mediated hyperpolarizations are due to opening of previously inhibited channels, then exposure of these cells to increased K⁺ should not cause a significant change in resting membrane potential. This was directly tested in three cells, and, as expected, addition of K⁺ to the recording medium did not cause any appreciable change of membrane potential. This was not solely due to the fact that these cells were more hyperpolarized than their ATP-containing correlates, since no response to elevated K⁺ could be recorded even at depolarized potentials (Fig. 4B).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHOD RESULTS DISCUSSION REFERENCES

New concepts have emerged to explain neuronal regulation of CBF, but the original theory of metabolic regulation (33) is still experimentally supported. CBF and neuronal activity can be coupled by metabolic by-products (i.e., adenosine) with vasoactive properties; alternatively, K⁺ dissipated by neuronal firing may act as a CBF regulator. On the basis of present knowledge, K⁺-mediated vasodilations occur exclusively by way of voltage-dependent K_IR channels and are thus largely independent from metabolic changes. We report that, in small, resistance-size cortical arterioles and in the basilar artery, K⁺ dilations are also mediated by a mechanism linking K_ATP channels to Na⁺-K⁺-pump activity. These results provide an additional mechanism for metabolic coupling of CBF to neuronal activity.

The coupling of brain cell function to the vascular system is the basis for a number of functional neuroimaging methods relevant for human studies. These methods map a specific localized brain activation through a vascular response, such as an increase in CBF or a change in blood oxygenation (38). Studies of the close interplay between neuronal activity and CBF have thus transcended basic science boundaries and have rapidly expanded into the field of clinical assessment of cerebral function. The exact mechanisms by which central nervous system neurons sense and regulate CBF have been elusive so far, but mechanisms involving K⁺ channels expressed in cerebral vasculature have recently received increasing attention (9, 22, 31). At least three temporally related events accompany neuronal activity: 1) rapid, transient changes in extracellular ion concentrations (e.g., K⁺); 2) release of potentially vasoactive neurotransmitters (e.g., adenosine or nitric oxide); and 3) a transient tissue hypoxia derived from increased metabolic demand. Although evidence linking adenosine release to metabolic deprivation induced by neuronal firing has been long available (39), K⁺-induced vasodilations have always been considered passive and dependent exclusively on K⁺ redistribution through voltage-dependent channels and subsequent changes in VSM resting potential.

The handling of extracellular K⁺ by glia and by cerebrovascular smooth muscle has been studied extensively (5, 15, 21, 25, 27, 30-32), but studies from relatively large cerebral vessels have failed to unmask a possible link between metabolic cytosolic changes in parenchymal vessels and K⁺ channel activity. This is somehow surprising since VSM cells are endowed with K⁺ channels regulated by subtle changes in cellular ATP content (1). K_ATP channels have been shown to mediate anoxic/ischemic vasodilations and may participate in cerebral autoregulation (12). The results presented herein expose an additional cerebrovascular regulatory mechanism propitiated by opening of K_ATP, i.e., coupling of brain activity to CBF by K⁺ lost during neuronal firing.

Abluminal application of K⁺ to isolated vessels may cause changes in resting potential in both endothelial and VSM cells. Furthermore, both K_ATP and K_IR are expressed in both cell types (18, 19, 21). It is thus possible that involvement of endothelial channels underlies some of the response to abluminal K⁺. Several considerations and experimental results rule against this hypothesis. 1) If K⁺ dilations are a mechanism designed to link CBF and net loss of K⁺ from parenchymal neurons, it is important that commonly occurring changes in blood K⁺ concentrations do not affect CBF, whereas even small changes in extravascular [K⁺]_out participate in the regulation of CBF. 2) Results from other laboratories (e.g., Ref. 21) clearly demonstrated that vessels with denuded endothelium still undergo K⁺ mediated dilations. 3) We have recently estimated the "tightness" of the transendothelial barrier of the penetrating pial vessels used for the experiments presented herein (10, 11). These experiments demonstrated the existence of a tight transendothelial barrier. Therefore, it is unlikely that spillover of K⁺ into the intraluminal compartment may underlie the observed responses, since the highly blood-brain barrier-impermeant ion, Ba²⁺, was capable of abolishing this response when applied abluminally.

It has to be noted that the sensitivity of K⁺-mediated dilations to ouabain and Ba²⁺ was first described by McCarron and Halpern (26). The novel finding in this study is twofold. 1) The blockade of dilations induced by K⁺ by the K⁺ channel blocker is not exclusively due to an effect on voltage-dependent channels (K_IR) but rather on a metabolically regulated type (K_ATP). 2) The effects of ouabain are not directly related to the inhibition of the Na⁺-K⁺-ATPase but rather to the subsequent hydrolysis of energy substrates.

Two questions remain. 1) How does extracellular K⁺ cause opening of K_ATP, the "metabolic sensors" of VSM? 2) Do voltage-dependent K_IR channels play any role in the regulation of cerebrovascular tone in resistance size vessels exposed to elevated [K⁺]_out? As shown in Fig. 2B, glibenclamide's effects on K⁺-induced dilations are more pronounced after steady-state K⁺ dilations; in contrast, Ba²⁺ abolished both the initial and plateau components. Thus it appears that Ba²⁺ acted by simultaneously blocking an early dilatatory response, insensitive to K_ATP blockers and in all likelihood mediated by K_IR, whereas the effects on the plateau could be mimicked by glibenclamide and hence appeared to be K_ATP mediated. This delayed sensitivity to K_ATP blockers suggested that the metabolic changes leading to opening of K_ATP channels required prolonged exposure to [K⁺]_out. Because the effects of Na⁺-K⁺-ATPase blockade by ouabain also overlapped with the effects of glibenclamide, we suggest that activation of the Na⁺-K⁺-ATPase acts as a link between changes in extracellular K⁺ and VSM ATP content (Fig. 5A).

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Events underlying K+-induced dilations in cerebral vessels and their relationship to neuronal activity. A: under physiological conditions, KATP channels are inhibited by intracellular ATP. This is possible because metabolic demand is matched and is due to the low level of Na+-K+-ATPase activity. Even small increases in extracellular K+ cause a significant activation of the Na+-K+-ATPase, leading to increased hydrolysis of intracellular ATP. Although the decrease of [ATP]i may not be sufficient to relieve the blockade of the channel, rising concentrations of intracellular ADP are sufficient to induce KATP channel opening, leading to hyperpolarization and vasodilation. B: synergistic contribution of metabolic (KATP) and purely voltage-dependent (KIR) K+ channels to the coupling of neuronal activity to cerebral blood flow. At low levels of above-baseline activity (and subsequent low increase in [K+]out), transient responses are mediated predominantly by KIR channels, whereas sustained dilations are maintained by KATP channels (our data, Fig. 2B). Hence, under conditions of phasic, brief activity, voltage-dependent mechanisms predominate, whereas tonic firing (as during sustained activity) leads to sufficient metabolic changes to allow for KATP-mediated dilations. Large changes in [K+]out lead to dilations (or constrictions) that are independent of KATP channels and thus appear to be entirely mediated by KIR channels or other mechanisms.

Recent insights into the intracellular mechanisms regulating K_ATP conductance have demonstrated that, in intact cells, even small decreases in intracellular ATP can lead to K⁺ flux through K_ATP (2, 35). Furthermore, in isolated membrane patches, activation of Na⁺-K⁺-ATPase causes sufficient ATP hydrolysis to cause activation of K_ATP, hence explaining the anomalous effect of the pump inhibitor ouabain on K_ATP-mediated currents (20). The latter effect could also be noticed in cell-free membrane patches, suggesting that nucleotide concentrations relevant to ion channel gating are closely associated with the plasma membrane, further strengthening the notion that localized changes in metabolic activity may be adequate signals for activation of a K_ATP conductance. Taken together, these results support a mechanism linking extracellular K⁺ increases, Na⁺-K⁺-ATPase, and K_ATP. Interestingly, K_ATP-mediated, glibenclamide-inhibitable dilations were observed in our study only after modest (5 mM above baseline) changes in [K⁺]_out; at [K⁺]_out >10 mM, Ba²⁺ sensitivity could be completely accounted for by a mechanism involving K_IR. Hence, a cooperation of two molecularly and biophysically distinct K⁺ channels seems to be involved.

In conclusion, our results demonstrate that, in addition to voltage-dependent K_IR channels, VSM respond to [K⁺]_out by activation of a K_ATP current. The implications of this finding warrant a reinterpretation of the mechanisms linking neuronal activation to the regulation of CBF.