Activation of a G Protein–coupled Inwardly Rectifying K⁺Current and Suppression of I  ^hContribute to Dexmedetomidine-induced Inhibition of Rat Hypothalamic Paraventricular Nucleus Neurons

The study was approved by the Animal Investigation Committee of Miyazaki University (Miyazaki, Japan) and was conducted according to the animal use guidelines of the American Physiologic Society (Bethesda, Maryland).

Male Wistar rats (postnatal days 14–21) were anesthetized with isoflurane and then killed by decapitation. The brain was rapidly removed and placed in cooled (2°–4°C) standard artificial cerebrospinal fluid (ACSF) containing 140 mm NaCl, 3 mm KCl, 1.3 mm MgSO⁴, 1.4 mm NaH²PO⁴, 2.4 mm CaCl², 5 mm HEPES, 3.25 mm NaOH, and 11 mm d-glucose, which had been oxygenated with 100% O². The osmolarity was maintained at 290–300 mOsm with a pH of 7.4. This solution was also used as the extracellular solution in most experiments. In some cells, where we tested the effects of dexmedetomidine on I  ^h, we used a modified ACSF that contained tetraethylammonium (20 mm; blocking the delayed rectifier K⁺current), BaCl²(1 mm; blocking the inwardly rectifying K⁺current, I  ^Kir), 4-aminopyridine (4-AP; 2 mm; blocking the transient outward K⁺current), tetrodotoxin (1 μm; blocking the Na⁺current), and nominally zero Ca²⁺(0 mm CaCl²; preventing Ca²⁺influx). Coronal slices 270 μm in thickness, which included the PVN, were prepared using a vibrating brain slicer (DSK-2000; Dosaka, Kyoto, Japan) and equilibrated for 1–2 h at room temperature in a tissue storage chamber containing ACSF saturated with 100% O².

The electrodes were made with a puller (PB-7; Narishige, Tokyo, Japan) from thick-wall borosilicate glass (GD-1.5; Narishige). Their resistance was 3–5 MΩ in the bath, with access resistances in the range of 8–12 MΩ. The electrode solution contained 130 mm K gluconate, 1 mm NaCl, 10 mm EGTA, 1 mm MgCl², 2 mm Na²ATP, 1 mm CaCl², 10 mm HEPES, and 0.5 mm Na³GTP, pH adjusted to 7.3 with KOH. In some cells, where we tested the effects of cyclic adenosine monophosphate (cAMP), perforated patch recordings were made using amphotericin B. In this case, the patch pipettes were filled with a solution containing 150 mm KCl and 10 mm HEPES, adjusted to pH 7.3 with KOH. An amphotericin B solubilized preparation (Sigma, St. Louis, MO) was added into the pipette solution just before use. The final concentration of amphotericin B in the pipette solution was 300 μg/ml.17The series resistance was compensated by 80%. Slices were transferred into the recording chamber and continuously perfused (1.5–2.0 ml/min) with ACSF at room temperature (24°–26°C). Whole cell recordings were made from microscopically identified cells. In each slice, only one neuron was recorded. The transmembrane voltage and current were recorded using a patch clamp amplifier (Axopatch 200B; Axon Instruments, Foster City, CA) and stored on a hard disk via  an analog–digital converter for later analysis. The signals were also displayed on a thermal rectigraph (model 8M14; San-Ei, Tokyo, Japan). The stored data were analyzed using AxoGraph software (version 8.0; Axon Instruments). The membrane potential and current were low-pass filtered at 2 and 10 kHz, respectively. The liquid-junction potential between the pipette and ACSF was corrected at the beginning of the experiment. After stable recording conditions were obtained, a PVN neuron was identified electrophysiologically as a transient outward rectification-expressing (magnocellular; type 1) neuron or a nonoutward rectification-expressing (parvocellular; type 2) neuron according to previously established criteria.18 

The effects of dexmedetomidine on the membrane potentials were examined at the resting membrane potential. The membrane currents were recorded at a holding potential of −60 mV. Drugs were applied after recording at least 180 s of control events. In evaluating the dose-dependent responses of membrane events, the same neurons were tested from one to three times in another concentration. In these cases, the concentration of the applied drugs was selected at random, and subsequent drug applications were made after the effects of the first drug application had fully reversed.

Tetrodotoxin, tetraethylammonium, 4-AP, CsCl, yohimbine, prazosin, 8Br-cAMP, amphotericin B, forskolin, and SQ-22536 were purchased from Sigma Chemical (St. Louis, MO), and ZD7288 was obtained from Tocris Cookson (Ballwin, MO). Dexmedetomidine was provided by Abbott Japan (Osaka, Japan). All other reagents were from Nacalai Tesque (Osaka, Japan). ZD7288 was prepared as a 50 mm stock solution (in H²O) and stored at −20°C until use. Forskolin was initially solubilized in dimethylsulfoxide and then diluted in ACSF. All other drugs were dissolved in ACSF.

Measurements were acquired using the pCLAMP system (Axon Instruments) and stored on a computer hard disk for off-line analysis. Off-line analysis was performed using Axograph software (Axon Instruments). The amplitude of I  ^hwas obtained by subtraction of the instantaneous currents (I  ^INS) at the end of the capacitative current transient (5–10 ms after stimulus onset) from the steady state currents at the end of the 1-s voltage step (I  ^SS). I  ^INSand I  ^SSwere plotted against the membrane potential. To examine the effect of dexmedetomidine on the voltage-dependent activation of I  ^h, an activation curve was constructed by a tail current analysis, as described by McCormick and Pape.19The tail current amplitudes were normalized (I /I  ^max) and plotted as a function of the membrane potential during the initial hyperpolarizing voltage steps. The resulting data were fitted by the Boltzmann equation: I /I  ^max={1 + exp[(V^m− V^1/2)/k]}⁻¹, where I  ^maxis the maximal tail current, V^mis the membrane potential during the initial voltage steps, V^1/2is the voltage at half-maximal conductance, and k is the slope factor. Tail currents were normalized, plotted as a function of the preceding hyperpolarization step voltage, and fitted with Boltzmann curves for derivation of V^1/2and k by using a least squares analysis program (Origin; Microcal Software, Northampton, MA).

Data were analyzed statistically using one-way analysis of variance followed by Scheffé F test, Student t  test, or linear regression. All values described in this study are expressed as mean ± SD. P < 0.05 was considered statistically significant.

A representative sample of 34 PVN magnocellular neurons showed a resting membrane potential of −59.2 ± 10.5 mV (range, −56 to −64 mV). The mean input resistance was 932 ± 306 MΩ. Overshooting action potentials were observed in all neurons recorded in current clamp mode.

The effect of dexmedetomidine was reversible on washing the tissue with a drug-free solution for periods up to 1 h. There was no evidence of desensitization with repeated (up to 3 h) or continuous (up to 1 h) applications of dexmedetomidine. A bath application of dexmedetomidine (1 μm) hyperpolarized PVN magnocellular neurons (12.2 ± 1.1 mV, n = 8) was accompanied by decreasing input resistance (24 ± 11.3%). Of all PVN magnocellular neurons tested, only the hyperpolarization of the membrane potentials induced by dexmedetomidine (10 nm–6 μm) was dose related (fig. 1). Figures 1A and Bshow the effect of two incremental concentrations of dexmedetomidine on a typical neuron in the PVN magnocellular neuron. The most sensitive of these effects, which showed a remarkable change even at low dexmedetomidine concentrations, was the decrease in the firing rate. Dexmedetomidine at 10 and 30 nm decreased the neuronal firing rate by 38% and 72%, respectively (fig. 1A); however, little hyperpolarization and only a small reduction of the input resistance of the neurons were seen. Higher concentrations of dexmedetomidine (300 nm–6 μm) resulted not only in a greater inhibition of firing but also in the hyperpolarization of the membrane and a decreased input resistance (figs. 1B and C).

A bath application of dexmedetomidine (3 μm) produced an outward current (154 ± 66 pA, n = 6) that peaked within 2 min (fig. 2A). To examine the reversal potential for dexmedetomidine-induced hyperpolarization, we calculated from the intersection of the control current–voltage curve and the curve obtained in the presence of dexmedetomidine (1 μm). Figure 2Bshows the currents of a PVN neuron in the voltage clamp mode elicited by 1-s ramp pulses between −50 and −130 mV from a holding potential of −60 mV before (control) and after application of dexmedetomidine (1 μm). The current–voltage relation of the PVN magnocellular neurons revealed inward rectification. The reversal potential for dexmedetomidine-induced hyperpolarization in each of eight neurons was deduced from the intersection of the two current–voltage curves. The dexmedetomidine-induced current reversed at −94.1 ± 7.1 mV (fig. 2C), which was close to the K⁺equilibrium potential calculated by the Nernst equation (E^K=−95.2 mV). By the exchange of the external K⁺concentration from 3.0 to 6.0 and 12.0 mm, the reversal potential of the current was shifted to −79.3 ± 11.6 mV and −62.1 ± 16.1 mV, respectively (fig. 2C). The predicted reversal potentials for K⁺conductance are −77.7 and −60.2 mV, respectively, under these conditions. These results suggest that dexmedetomidine elicited an inwardly rectifying K⁺(IRK) current.

The blocking effects of Ba²⁺(0.3 mm) on dexmedetomidine-induced hyperpolarization were tested on eight PVN magnocellular neurons (fig. 3). After a bath application with BaCl²(0.3 mm) for 5 min, dexmedetomidine (100 nm) was applied. Ba²⁺completely blocked dexmedetomidine-induced hyperpolarization in six cells, but a partial inhibition of the baseline firing rate was still observed in two cells (fig. 3B). Ba²⁺is relatively specific for blocking IRK channels, including IRK and G protein–coupled inwardly rectifying K⁺(GIRK) channels. In neurons, the peptide toxin tertiapin should specifically block GIRK channels, with little effect on IRK channels at concentrations up to 2 μm.20To further test whether the IRK channels were the GIRK channel, we applied tertiapin 5 min before the dexmedetomidine application. In the tertiapin (100 nm)–treated neurons, tertiapin eliminated the dexmedetomidine (100 nm)–induced hyperpolarization (n = 8; fig. 3).

Next, we studied whether hyperpolarization induced by dexmedetomidine was mediated by the α²-adrenergic receptor. The α²-adrenergic antagonist yohimbine (100 nm) partially blocked the dexmedetomidine (100 nm)–induced hyperpolarization (42.4 ± 29.6%, n = 6; fig. 4). The application of yohimbine (100 or 1,000 nm) produced a parallel, dose-related shift to the right of the dexmedetomidine dose–response curve (fig. 4B).

The neuron having I  ^hchannels in the PVN exhibited a lack of transient outward rectification in response to a series of depolarizing current pulses delivered at a hyperpolarizing membrane potential (fig. 5A). These neurons were parvocellular neurons in the PVN.18More than 80% of parvocellular neurons (91 of 112) displayed a time-dependent inward rectification “sag” during the hyperpolarizing pulses (fig. 5B), which was blocked by 50 μm ZD7288. These characteristics are consistent with I  ^hconductance.21 

Figure 6shows typical recordings of I  ^hin a PVN parvocellular neuron during activation voltage protocols in a modified ACSF that contained tetraethylammonium, BaCl², 4-AP, and tetrodotoxin and nominally zero Ca²⁺. Hyperpolarizing step pulses of 1-s duration between −50 and −120 mV from a holding potential of −50 mV in 10-mV increments were used as the activation voltage protocols. A dexmedetomidine (10 μm) application reversibly reduced the I  ^hamplitude within 5 min. In addition, the tail currents indicated by asterisks in figure 6A, which appeared after stepping back to the holding potential of −50 mV, were reduced in parallel with the reduction of I  ^h. These effects were stable at approximately 3 min after the application of dexmedetomidine. The effect of dexmedetomidine was not recoverable even after a wash with normal ACSF for half an hour. I  ^INS, as shown in figure 6A(○= control; •= dexmedetomidine), and I  ^SS, as shown in figure 6A(□= control; ▪= dexmedetomidine), were plotted against the membrane potential (fig. 6B). The difference between I  ^INSand I  ^SSin the control (▵) and during the application of dexmedetomidine (10 μm; ▴) was plotted against the membrane potential (fig. 6C). The difference between I  ^INSand I  ^SSrepresents I  ^h. The mean amplitude of I  ^hat the −120 mV command voltage was 341 ± 54 pA (n = 9). As shown in figure 6C, the activation of I  ^hwas voltage dependent. The application of dexmedetomidine (0.1–20 μm) suppressed I  ^hin a dose-dependent manner (fig. 6D).

To examine the effect of dexmedetomidine on the voltage-dependent activation of I  ^h, an activation curve was constructed by a tail current analysis. A sample activation curve fitted with the Boltzmann equation is shown in figure 6C(inset). I  ^hwas half-activated at −94 mV in the control, and dexmedetomidine caused a small change in the V^1/2(to −102 mV). The mean values were as follows: V^1/2=−96 ± 4.5 mV, k = 9.2 ± 1.5 in the control and V^1/2=−104.6 ± 3.6 mV, k = 8.9 ± 3.3 during the application of dexmedetomidine (10 μm, n = 9). These data revealed that dexmedetomidine produced a significant shift in the V^1/2to a more hyperpolarized potential (P < 0.01). The slope values were not altered by dexmedetomidine (P > 0.05).

The reversal potential of the dexmedetomidine-sensitive component of I  ^hwas determined by tail-current analysis, as described previously.22As illustrated in figure 7A, after a step up to −120 mV (250 ms duration) from a holding potential of −60 mV, the membrane voltage was stepped in the range of −70 to −20 mV (50-ms duration, 10-mV increments). The plot of the instantaneous current at each test potential yielded the fully activated current–voltage relation, which was linear (fig. 7B). The reversal potential was −34.8 ± 10.5 mV (n = 6) in standard external solution ([K⁺]^o= 3 mm, [Na⁺]^o= 144.6 mm). It has been reported that in other central neurons, I  ^his carried by Na⁺and K⁺.16To investigate the ionic basis of I  ^h, we tested the effects of changing [K⁺]^oand [Na⁺]^oon I  ^h. We measured the mean reversal potential of the dexmedetomidine-sensitive current in two other extracellular solutions, one with a low Na⁺concentration ([K⁺]^o= 3 mm, [Na⁺]^o= 80 mm) and the other with a high K⁺concentration ([K⁺]^o= 20 mm, [Na⁺]^o= 144.6 mm), using the same method. The reversal potential was shifted toward a more negative potential (−49 ± 2.4 mV; □; n = 6) in the low-Na⁺solution, whereas it was shifted toward a more positive potential (−29 ± 2.2 mV; ▵; n = 6) in the high-K⁺solution. These results suggest that the channel modulated by dexmedetomidine is permeable to both Na⁺and K⁺. I  ^his a slowly activating inward current evoked by hyperpolarization, and the current–voltage curve shows rectification in the inward direction; the reversal potential of I  ^his close to −35 mV in media of physiologic composition, and I  ^his mixed Na⁺and K⁺current; furthermore, the relation between the chord conductance of I  ^hand voltage can be fitted by a Boltzmann equation. Therefore, we conclude that the dexmedetomidine-sensitive current corresponds to I  ^h, i.e. , dexmedetomidine inhibits I  ^hin rat PVN parvocellular neurons.

The pharmacologic properties and characterization of the adrenoceptors involved in the inhibition of I  ^hwere explored in PVN parvocellular neurons. We recorded the drug-induced change in the amplitude of I  ^hevoked by 1-s hyperpolarizing voltage steps from −60 mV to −120 mV. The application of 1 mm Cs⁺decreased the peak amplitude of I  ^hby 86 ± 26.4% (n = 6), and the effects of dexmedetomidine (20 μm) on I  ^hwere occluded completely by Cs⁺-induced blockage of I  ^h(fig. 8A).

Prazosin (20 μm), an α¹-adrenergic receptor antagonist, had little effect on the I  ^hamplitude (6.0 ± 10.5%, n = 6) in the same neurons in which dexmedetomidine inhibited I  ^h(fig. 8B).

Figure 8Cshows the antagonism of the dexmedetomidine-induced inhibition of the I  ^hby yohimbine. When standard ACSF contained 2 mm yohimbine, the application of 20 μm dexmedetomidine resulted in a small reduction of the I  ^hamplitude (13 ± 7.5%, n = 6), whereas, once the yohimbine had been washed off the neurons, dexmedetomidine produced a significantly larger reduction of the I  ^hamplitude (67 ± 17.6%, n = 6). These results suggest that the dexmedetomidine-induced inhibition of I  ^his mediated by α²adrenoceptors.

Finally, we examined the regulation of I  ^hsensitive to dexmedetomidine by cAMP. To avoid washing out the cytoplasmic biochemicals required for second messenger–mediated responses, perforated patch recordings were performed using amphotericin B. As shown in figure 9, bath application of 8Br-cAMP (1 mm) increased the amplitude of I  ^hby an average of 7.3 ± 4.5% (P < 0.05, n = 8). The value of V^1/2shifted by an average of 2.6 ± 1.1 mV (P < 0.05, n = 8) in the positive direction (fig. 9B). The application of forskolin (50–100 μm) caused a slight increase in the amplitude of I  ^h(data not shown; n = 8). This increase in the amplitude of I  ^hwas always less than the increase induced by 1 mm 8Br-cAMP. To test whether physiologic changes in the intracellular cAMP concentration were sufficient to cause maximal activation of I  ^hchannels, we bath-applied the adenylate cyclase inhibitor SQ22536 (50 μm). SQ22536 caused no significant changes in the amplitude or activation curves of I  ^h(n = 8). The amplitude of I  ^hwas changed to 104.7 ± 12.2% of the control level (P > 0.05, n = 8). The value of V^1/2was changed from −95.8 ± 2.3 to −95.2 ± 2.5 mV (P > 0.05, n = 8). From these results, we concluded that I  ^hsensitive to dexmedetomidine in hypothalamic PVN parvocellular neurons is only weakly modulated by the intracellular cAMP system.

Under whole cell voltage clamp conditions in a neonatal rat hypothalamic slice preparation, we demonstrated that dexmedetomidine, a highly selective α²-adrenoceptor agonist, activated an inwardly rectifying K⁺current leading to membrane hyperpolarization in PVN magnocellular neurons. A second component of the dexmedetomidine current was due to inhibition of I  ^hcarried by a mixture of Na⁺and K⁺in PVN parvocellular neurons. This effect was blocked by external Cs⁺and ZD7288, but not by Ba²⁺, and was primarily due to a decrease in the maximal I  ^h, although it was also accompanied by a slight hyperpolarizing shift in the half-activation potential. This I  ^hwas weakly activated by intracellular cAMP.

We showed that dexmedetomidine activated a GIRK in PVN magnocellular neurons. This facilitates significant hyperpolarization of the membrane potential and a decrease in input resistance. The functional significance of the modulation of K⁺currents in the PVN magnocellular neurons has not been established; however, intravenous administration of α²-adrenoceptor agonists produces a decrease in plasma9and cerebrospinal fluid23concentration of AVP. Intravenous administration of clonidine induces water diuresis and decreases plasma AVP levels, probably because of central inhibition of AVP release.23Many of the magnocellular neurons in the PVN project to the posterior pituitary and release AVP into the peripheral circulation.10It has been suggested that the PVN contains a large number of α²adrenoceptors12and may be involved in the regulation of the autonomic nervous and endocrine systems.11Therefore, it is possible that the diuretic responses induced by α²-adrenoceptor agonists are, at least in part, mediated by the inhibitory actions on PVN magnocellular neurons. Our results are consistent with another report regarding the activating effect on IRK channels of dexmedetomidine in the locus coeruleus.24The effects of dexmedetomidine on these IRK channels were inhibited by Ba²⁺. The peptide toxin tertiapin also blocked the effect of dexmedetomidine. These results suggest that dexmedetomidine activated GIRK channels, which is supported by the fact that the hypnotic–anesthetic action of dexmedetomidine can be attenuated by pretreatment with either an intracerebroventricular injection of pertussis toxin, a specific inactivator of inhibitory G proteins, or an intraperitoneal injection of 4-AP, a K⁺channel blocker.25The effect of dexmedetomidine seems to be mediated via α²adrenoceptors, because yohimbine, antagonists at α²adrenoceptors, produced a parallel, dose-related shift to the right of the dexmedetomidine dose–response curve. These results suggest that dexmedetomidine activated GIRK channels mediated by α²adrenoceptors and contributed to both the reduction in spike frequency and the hyperpolarization of the membrane potential of PVN magnocellular neurons. The best known function of GIRK channels is to provide a mechanism for the slow inhibitory modulation of the overall excitability of the cell. It has been suggested that the noradrenergic input to AVP neurons may regulate cardiovascular function, either by AVP effects on CNS autonomic centers26,27or by direct effects of circulating AVP on blood pressure.28The cardiovascular depression induced by dexmedetomidine may be, at least in part, mediated by inhibition of PVN magnocellular neurons. Diuretic and natriuretic action of clonidine is due to both central inhibition of AVP release in combination with a second action of the drug that leads to inhibition of tubular reabsorption of sodium and water either in the thin loop of Henle or in the proximal tubule.9Intravenous administration of dexmedetomidine during the perioperative period may prevent oliguria mediated by inhibition of PVN magnocellular neurons.

This is the first study showing the effects of dexmedetomidine on I  ^h. The autonomic beating of the heart and a considerable number of rhythmic activities in the brain are controlled by the I  ^h.16,20,29Originally described in cardiac atrial cells, I  ^his a noninactivating, inwardly rectifying current carried by both Na⁺and K⁺ions.16Four different isoforms (HCN1–4) of the I  ^hchannels have been cloned,30,31and we previously reported that the messenger RNA of three isoforms (HCN1–3) is expressed in rat PVN parvocellular neurons.13Dexmedetomidine dose-dependently elicited a decrease in I  ^hthat was due primarily to inhibition of the maximal current amplitude in the PVN parvocellular neurons tested. Dexmedetomidine also produced a statistically significant hyperpolarizing shift in the voltage dependence of I  ^hactivation that was similar to that observed in rat hypoglossal motoneurons induced by clonidine.32The reversal potential (approximately −35 mV) of I  ^hin PVN parvocellular neurons was consistent with the mixed Na⁺and K⁺permeability and close to the reversal potential values reported for I  ^hin other central neurons.15,33The preapplication of Cs⁺blocked the inhibition of I  ^hinduced by dexmedetomidine. This is consistent with the report that the extracellular application of Cs⁺in the millimolar concentration range blocks I  ^hin various central neurons in the brain.16The attenuation of the dexmedetomidine inhibition by yohimbine supports the existence of an α²mechanism. This is consistent with other central neurons15,32,33relevant to clonidine.

The molecular identification of genes encoding I  ^hchannels34has enabled studies of the electrophysiologic properties of specific channel types that mediate I  ^h. The regulation of these HCN channels may occur via  cAMP, cyclic guanosine monophosphate, and/or Ca²⁺.16,29,35,36Among the HCN channel family (HCN1–4), HCN1 and HCN2 are most prominently expressed in the CNS, with the HCN1 subunit more selectively localized than HCN2.31The activation of HCN2 is highly accelerated by increases of cytoplasmic cAMP. By contrast, HCN1 is only weakly modulated by cAMP.31,34The half activation concentration of cAMP for HCN2 channels is 0.5 μm.30In the current study, we showed that a relatively high concentration of 8Br-cAMP (1 mm) caused a slight increase in the amplitude of I  ^hwith a positive shift of the voltage dependence. Forskolin caused smaller effects on I  ^hthan those induced by a high concentration of 8Br-cAMP. These results suggested that the sensitivity of the I  ^hchannel subtype in PVN parvocellular neurons to intracellular cAMP could be much less than that of HCN2 channels. We also found that there were no significant changes in either the amplitude or the activation kinetics of I  ^hduring the application of adenylate cyclase inhibitor SQ22536,37suggesting that the likelihood of the tonic up-regulation of I  ^hchannels by the intracellular cAMP system is even smaller. The intravenous anesthetic propofol selectively inhibits HCN channels containing HCN1 subunits in cortical pyramidal neurons.38The inhalational anesthetic halothane modulates neuronal I  ^h, and the corresponding HCN channel subunits are HCN1 and HCN2.39The precise contribution of HCN channel inhibition to dexmedetomidine action in PVN parvocellular neurons will require modeling in networks or experiments in animals with targeted disruption of HCN subunits.

Among the various autonomic responses, increases or decreases in blood pressure, heart rate, and sympathetic nerve activity are elicited by various stimulations in the rat PVN,40indicating a clear relation between outputs from the PVN and autonomic activity. Dexmedetomidine5and clonidine41induce cardiovascular depression and sympatholytic action mediated by a central mechanism. Dexmedetomidine may affect the autonomic nervous system activity mediated by the inhibition of I  ^hin PVN parvocellular neurons. Because I  ^hin PVN neurons can be modulated weakly by the cAMP system, any chemical substances or neurotransmitters that alter cAMP levels will have a small effect on the resting potential, spontaneous activity, and rebound potential latency. This feature of I  ^hmay stabilize spontaneous discharge rates in PVN parvocellular neurons. In contrast to these compounds, dexmedetomidine and other anesthetics such as halothane,39enflurane,42and propofol38can directly reduce the I  ^hconductance. These drugs should induce a hyperpolarization of the membrane potential, a decrease in the slope steepness of slow depolarization, and a longer latency of rebound action potentials. As a result, the spontaneous discharge rate of neurons should decrease when neurons are exposed to these drugs. Perineural administration of the I  ^hchannel blocker ZD7288 significantly reduced mechanical allodynia induced by partial sciatic nerve injury and hind paw incision.43It has been suggested that the medial parvocellular subdivision of the PVN is also involved in the arousal mechanisms.44These results suggest that antinociceptive, hypnotic, and sedative effects induced by anesthetics may be, at least in part, mediated by activation of GIRK and/or blockage of the I  ^hchannel.

In conclusion, the current study demonstrated that dexmedetomidine binds to α²adrenoceptors on the cell membrane of PVN neurons, which leads to activation of the GIRK channels and inhibition of I  ^hchannels, resulting in hyperpolarization of the membrane.