Inhibition of I h in Striatal Cholinergic Interneurons Early after Transient Forebrain Ischemia

Abstract

Striatal cholinergic interneurons are relatively resistant to ischemic insults. These neurons express hyperpolarization-activated cation current (I_h) that profoundly regulates neuronal excitability. Changes in neuronal excitability early after ischemia may be crucial for determining neuronal injury. Here we report that I_h in cholinergic interneurons was decreased 3 h after transient forebrain ischemia, which was accompanied by a negative shift of the voltage dependence of activation. The inhibition of I_h might be due to the tonic activation of adenosine A1 receptors, as blockade of A1 receptors significantly increased I_h in postischemic neurons, but had no effect on control neurons. Consistent with the inhibition of I_h, postischemic neurons showed a reduction in both spontaneous firing and hyperpolarization-induced rebound depolarization. These findings indicate that I_h may play excitatory roles in striatal cholinergic interneurons. Postischemic inhibition of I_h might be a novel mechanism by which adenosine confers neuronal resistance to cerebral ischemia.

Introduction

Transient cerebral ischemia induces selective neuronal death in certain brain regions; however, its mechanisms are still unclear. In the neostriatum, forebrain ischemia of 20 to 30mins causes irreversible injury of most medium spiny neurons within 24 h, but cholinergic interneurons remain intact (Chesselet et al, 1990; Francis and Pulsinelli, 1982; Pulsinelli et al, 1982). Previous studies have shown that, in addition to excitatory synaptic inputs (Pang et al, 2002; Zhang et al, 2006b), the intrinsic membrane properties are differentially altered between ischemia-sensitive and resistant striatal neurons in response to ischemic insults (Calabresi et al, 1997; Pisani et al, 1999). For instance, during in vitro hypoxia/hypoglycemia, medium spiny neurons are depolarized presumably due to the activation of voltage-dependent Na⁺ channels and Ca²⁺ entry, whereas cholinergic interneurons are hyperpolarized by the activation of ATP- and Ca²⁺-dependent potassium channels (Centonze et al, 2001). Furthermore, the delayed rectifier potassium currents in cholinergic interneurons are enhanced after transient forebrain ischemia and lead to a shortening of spike width (Deng et al, 2005). These results suggest that the intrinsic neuronal excitability may contribute to the different vulnerabilities to cerebral ischemia among striatal neurons.

Striatal cholinergic interneurons express hyperpolarization-activated cation current (I_h) (Jiang and North, 1991; Kawaguchi, 1993). Unlike medium spiny neurons, most of the cholinergic interneurons are spontaneously active during this period I_h provides subthreshold depolarizing currents to drive firings (Bennett et al, 2000; Bennett and Wilson, 1999). Additionally, they exhibit I_h-dependent rebound depolarizations (RD) upon termination of hyperpolarizations (Bennett et al, 2000). Thus, regulation of I_h may cause profound changes in the membrane excitability of cholinergic interneurons. Several lines of evidence have shown that I_h in neurons is very sensitive to ischemic insults (Erdemli and Crunelli, 1998; Gao et al, 2006; Mironov et al, 2000; Ray et al, 2003; Zhang et al, 2006a). However, the role of I_h in ischemia-induced neuronal damage remains to be elucidated. With a focus on the early phase after transient forebrain ischemia in vivo, this study was therefore designed to detect the postischemic changes of I_h in cholinergic interneurons and its possible roles in regulating neuronal excitability.

Materials and methods

Male Wistar rats (weighing 100 to 200 g, Charles River Laboratories, Wilmington, MA, USA) were used in this study. Experimental protocols were approved by the Institutional Animal Care and Use Committee of Indiana University School of Medicine in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Transient Forebrain Ischemia Model

Transient forebrain ischemia was induced using the four-vessel occlusion method (Pulsinelli and Brierley, 1979) with modifications (Ren et al, 1997). Briefly, the rats were fasted overnight to provide uniform blood glucose levels and anesthetized with a mixture of 1% to 2% halothane, 33% O₂ and 66% N₂ through a nasal mask. The common carotid arteries were isolated after which an occluding device was placed loosely around each common carotid artery for subsequent occlusion of these vessels. The animal was then placed on a stereotaxic frame and both vertebral arteries were electrocauterized. A tiny temperature probe was inserted beneath the skull in the extradural space, and the brain temperature was maintained at 37°C with a heating lamp connected to a temperature control system (BAT-10; Physitemp, Clifton, NJ, USA). A skull window was opened at 9.5 mm anterior to the interaural line, 3.0 mm from the midline. Glass microelectrodes (5 to 8 μm in diameter of tip, filled with 2mol/L NaCl) were used to record ischemic depolarization, which is a reliable indication of complete ischemia (Ren et al, 1997). The microelectrode was advanced 3.0 mm below dura into the neostriatum. The recordings were performed with a neuroprobe amplifier (Model 1600; A-M Systems, Carlsborg, WA, USA). Transient forebrain ischemia was produced by occluding both common carotid arteries to induce ischemic depolarization for 20 to 22 mins. The duration of ischemic depolarization was determined by measuring the period from the beginning of the extracellular direct current potential reaching −20 mV to the point where the potential started to repolarize after recirculation. The rats were returned to the cages after recovering from ischemia and allowed free access to water and food.

Brain Slice Preparation

Brain slices were prepared from animals of control, 3 and 24 h after ischemia using procedures similar to those previously described (Deng et al, 2005). Briefly, the animals were anesthetized with ketamine-HCl (1 mg/kg, intraperitoneally) and perfused transcardially with an ice-cold (4°C) sucrose solution containing (in mmol/L) 230 sucrose, 26 NaHCO₃, 2.5 KCl, 1.25 NaH₂PO₄, 0.5 CaCl₂, 10 MgSO₄, and 10 glucose, pH 7.4, 290 to 305 mosM/L, equilibrated with 95% O₂ and 5% CO₂. The brains were quickly removed and transverse striatal slices (300 μm) cut using a vibratome (VT1000S; Leica, Nussloch, Germany) in the sucrose solution. The slices were maintained in an artificial cerebrospinal fluid containing (in mmol/L) 130 NaCl, 3 KCl, 2 CaCl₂, 2 MgCl₂, 1.25 NaH₂PO₄, 26 NaHCO₃, and 10 glucose, pH 7.4, 295 to 305mosM/L. The artificial cerebrospinal fluid was continuously equilibrated with 95% O₂ and 5% CO₂, and slices were incubated for >1h before recording.

Electrophysiological Recording

Recording electrodes were prepared from borosilicate glass (World Precision Instruments, Sarasota, FL, USA) using a horizontal electrode puller (P-97; Sutter Instruments, Novato, CA, USA). Electrodes had resistances of 2 to 4MΩ when filled with an intracellular solution containing (in mM) 120 KMeSO₄, 12 KCl, 1 MgCl₂, 1 EGTA (ethylene glycol bis (2-aminoethylether)-N,N,N′N′,-tetraacetic acid), 0.2 CaCl₂, 10 HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid), 2 Mg-ATP, and 0.4 Na-GTP, pH 7.4, 295 to 300 mosM/L. Oxygenated artificial cerebrospinal fluid was used as bath solution, and the flow rate was adjusted to 2 to 3mL/min. Neurons in brain slices were visualized with an infrared differential interference contrast microscope (BX50WI; Olympus Optical, Tokyo, Japan) and a charge-coupled device camera. All recordings were performed at 32°C ± 1°C with an Axopatch 200B amplifier (Molecular Devices, Foster City, CA, USA). After tight-seal (> 1 GΩ) formation, the electrode capacitance was compensated. For cell-attached recordings, a voltage-clamp mode was used, and data acquisition was terminated if the seal resistance decreased to < 1 GΩ. During whole-cell recordings, series resistance (8 to 15 MΩ) was monitored periodically, and cells with >15% change were excluded from the analysis. For current-clamp recordings, the fast I-clamp mode was used. Signals were filtered at 2 kHz and digitized at a sampling rate of 5 kHz using a data-acquisition program (Axograph 4.6; Molecular Devices).

Drug Application

All drugs were purchased from Sigma (St Louis, MO, USA) unless otherwise noted and were bath applied. Drugs were dissolved as concentrated stocks in either water or dimethylsulfoxide and stored at −20°C. Working solutions were prepared immediately before use. When dimethylsulfoxide was used to prepare drug solution, equivalent amounts were added to artificial cerebrospinal fluid as controls. To characterize I_h in cholinergic interneurons, 4-ethylphenylamino-1,2-dimethyl-6-methylamino-pyrimidinium chloride (ZD7288, 30 μmol/L; Tocris Bioscience, Ellisville, MO, USA) and CsCl (2 mmol/L) were used as blockers of I_h channels. For recording I_h, CdCl₂ (300 μmol/L), 4-aminopyridine (2 mmol/L), and tetraethylammonium (5 mmol/L) were added to bath solution to block voltage-dependent Ca²⁺ and K⁺ channels. Tetrodotoxin (TTX, 0.5 to 1 μmol/L) was also used to block Na⁺ channels. In some experiments, BaCl₂ (200 μmol/L) was used to eliminate hyperpolarization-activated inward rectifier K⁺ currents. To examine the modulation of I_h, membrane-permeable 8-bromo-cyclc AMP (8-Br-cAMP, 100 μmol/L) and Rp-cAMP (50 μmol/L) were used as analog and inhibitor of cAMP signaling pathway, respectively. In addition, 8-cyclopentyl-1,3-dipropylxanthine (DPCPX, 200 nmol/L) was used as a selective antagonist of adenosine A1 receptor. To reduce intracellular Ca²⁺ concentration, 1,2-bis(2-aminophenoxy) ethane-N,N,N′N′-tetraacetic acid tetrakis(acetoxymethyl ester) (BAPTA-AM, 50 μmol/L) was applied in bath solution. Synaptic blockers were not used in this study because the spontaneous activity of cholinergic interneurons is independent of synaptic input (Bennett and Wilson, 1999).

Data Analysis

Recording data were analyzed with Axograph 4.6. To evoke I_h, a series of hyperpolarizing voltage commands (from −60 to −150 in 10 mV steps, 2 sees) were applied from a holding potential of −50 mV. The amplitude of I_h was calculated as the current difference between the instantaneous current (measured just after the decay of capacitive transient) and the steady-state current (mean current of 50 ms before the termination of each voltage step). Voltage-dependent activation of I_h was determined by normalizing the peak amplitudes of tail currents at −50 mV after various hyperpolarizing voltage commands and then by fitting the current-voltage relationship with a Boltzmann function. The time constants of activation were estimated by fitting I_h traces evoked by various hyperpolarizing steps with a single exponential function. To examine the spontaneous activity, the firing rate was obtained from a 2-min sample of a cell-attached recording. The coefficient of variation (CV), a measure of irregularity in interspike intervals, was calculated for cells with a firing rate > 1 Hz. With current-clamp recordings, the amplitude of RD was measured as the voltage difference between the holding potential (approximately −55 mV) and the RD peak. The voltage response to a depolarizing current pulse was measured as the mean voltage of 50 ms (from 100 to 150 ms of current pulse), which was comparable to that of RD reaching the peak amplitude.

The data are presented as mean ± s.e.m. Statistical difference was detected using paired or unpaired Student's t-test (StatView 5.0; Abacus Concepts, Berkeley, CA, USA). Changes were considered significant when P <0.05.

Results

In striatal slices of both control and ischemia, only neurons with large somata (> 20 μm in diameter) and thick primary dendrites were selected for recordings (Figure 1A). These neurons exhibited electrophysiological characteristics that are typical of cholinergic interneurons (Jiang and North, 1991; Kawaguchi, 1993). During current-clamp recordings, although injection of depolarizing current evoked repetitive spiking followed by large-amplitude and long-duration after hyperpolarization, negative current induced an initial hyperpolarization followed by a depolarizing sag in the membrane potential (Figure 1B). These morphologic and electrophysiological features show that the recorded cells are cholinergic interneurons.

Figure 1

Postischemic decrease of I_h in cholinergic interneurons. (A, B) Identification of cholinergic interneuron based on morphological and electrophysiological features. (A) Under infrared differential interference contrast microscope, this neuron had large soma and thick primary dendrites. (B) With whole-cell current-clamp recording, the cholinergic interneuron exhibited large-amplitude and long-duration afterhyperpolarization after each evoked spike and a depolarizing voltage sag during a hyperpolarizing pulse. The holding potential was −59 mV. (C) Representative traces of I_h evoked in neurons of control and 3 h after ischemia. From a holding potential of −50 mV, the voltage- and time-dependent inward currents were evoked by a series of hyperpolarizing voltage commands (from −60 to −140 in 20 mV step, 2 secs). (D) Plots of I_h amplitude as a function of voltage for control and postischemic neurons. The amplitudes of I_h were significantly decreased 3 h after ischemia when evoked with voltages more negative than −60 mV (P < 0.01 versus control).

Postischemic Decrease of Ih in Cholinergic Interneurons

In both control and postischemic neurons, hyperpolarizing voltage steps more negative than −50 mV evoked voltage-dependent, slowly activating inward currents (Figure 1C). Identification of these currents as ih was supported by their high sensitivity to I_b channel blockers such as ZD7288 and Cs⁺, but not to Ba²⁺. For example, application of ZD7288 at a concentration of 30 μmol/L blocked >90% of the currents (n = 5).

To test whether I_h was altered after ischemia, we first examined the amplitude of I_h in cholinergic interneurons. Notably, the I_h amplitude was decreased in neurons 3 h after ischemia and then recovered to control levels by 24 h (Figures 1C and 1D). When evoked with a command voltage of −90mV, the amplitudes were 372.6 ± 15.2 pA in control neurons (n = 45), 220.7 ± 14.8 pA in neurons 3 h after ischemia (n = 38; P <0.01 compared to control), and 343.5 ± 19.2 pA in neurons 24 h after ischemia (n = 13; P > 0.05 compared to control). It is possible that changes in rapidly activating inward rectifier K⁺ currents might affect the measure of I_h amplitude. We therefore performed another set of experiments, in which Ba²⁺ (200 μmol/L) was applied to reduce the contamination of instantaneous currents. Consistent with the above data, a significant reduction of I_h amplitude was detected in neurons 3 h after ischemia (by 37.7% when evoked at −90mV, n = 11 in each group; P <0.01). These results show an inhibition of I_h early after ischemia. Thus, the following experiments were focused on neurons of control and 3 h after ischemia.

We then examined the voltage dependence and kinetics of I_h activation. Analysis of the tail currents revealed a negative shift of the half-activation voltage (control: −87.3 ± 2.1 mV, n = 17; ischemia: −94.0 ± 1.7 mV, n = 18; P <0.01; Figure 2A). In both control and postischemic neurons, the activation kinetics of I_h was voltage dependent, as the time constant of activation decreased sharply with more hyperpolarization. However, when evoked with the same voltage, the activation of I_h in postischemic neurons was much slower than that of control. At a voltage step of −90 mV, the activating time constants were 391.2 ± 21.3 ms in control (n = 27) and 645.6 ± 20.8 ms in postischemic neurons (n = 23; P <0.01; Figures 2B and 2C). The above data show coincident changes in the amplitude and the voltage-dependent activation of I_h after ischemia, and suggest that the negative shift of voltage dependence might contribute to the postischemic decrease of I_h amplitude.

Figure 2

Ischemia-induced changes in voltage dependence and kinetics of I_h activation. (A) Plots of normalized tail current showing the voltage-dependent activation of I_h. The half-activation voltage in postischemic neurons was more negative than that of control neurons (approximately −7mV, thin lines, P <0.01). (B, C) Deceleration of I_h activation after ischemia. (B) Typical fittings of I_h traces with a single exponential function. Both traces were evoked at a voltage step of −90 mV. (C) Histogram showing that the time constants of I_h activation (at voltages of −90 and −120 mV) were increased in postischemic neurons. ⁺P< 0.01.

Differential Modulation of Ih by Intracellular cAMP and Ca2+

Intracellular cAMP modulates the voltage-dependent activation of I_h (Frere et al, 2004). It is reasonable to speculate that cAMP might be involved in the postischemic decrease of I_h in cholinergic interneurons. To test this hypothesis, we examined the modulation of cAMP on I_h amplitude in control and postischemic neurons. Application of 8-Br-cAMP (100 μmol/L), an analog of cAMP, led to a dramatic increase in the I_h amplitude in both control (by 44.0% ± 5.6%, n = 6; P <0.01) and postischemic neurons (by 21.1% ± 3.4%, n = 5; P <0.01; Figures 3A and 3D), with less efficacy after ischemia (P<0.05). However, application of Rp-cAMP (50 μmol/L), an inhibitor of cAMP pathway, caused a decrease in I_h amplitude in control neurons (by 31.7% ± 4.4%, n = 5; P <0.01), but not in postischemic neurons (by 1.0% ± 5.2%, n = 5; P > 0.05; Figures 3B and 3D). These results indicate that the cAMP modulation of I_h was tonically suppressed in cholinergic interneurons after ischemia, which might underlie the postischemic decrease of I_h.

Figure 3

Regulation of I_h by intracellular cAMP and Ca²⁺. (A to C) Representative traces of I_h recorded before and during application of 8-Br-cAMP (A), Rp-cAMP (B), and BAPTA-AM (C). All the traces were evoked at a voltage step of −90 mV. (D) Group data showing the differential effects of 8-Br-cAMP (100 μmol/L), Rp-cAMP (50 μmol/L), and BAPTA-AM (50 μmol/L) on the amplitude of I_h (evoked with a −90 mV voltage command) in control and postischemic neurons, respectively. The amplitudes of I_h in the presence of drugs were normalized to those before drug application (dash line). *P < 0.01; *P < 0.05.

We also tested whether Ca²⁺ was involved in the postischemic changes of I_h, considering that intracellular Ca²⁺ regulates I_h in other neurons (Luthi and McCormick, 1998). As shown in Figures 3C and 3D, membrane-permeable Ca²⁺ chelator BAPTA-AM (50 μmol/L) significantly decreased I_h in postischemic neurons (by 17.9% ± 6.3%, n = 5; P <0.05), but not in control neurons (by 0.2% ± 3.3%, n = 5; P > 0.05). These data suggest that, rather than reducing, intracellular Ca²⁺ upregulates I_h in cholinergic interneurons after ischemia.

Adenosine Modulation of Ih in Cholinergic Interneurons

Striatal cholinergic interneurons express adenosine A1 receptors (Song et al, 2000). Activation of A1 receptors has been shown to reduce I_h in mesopontine cholinergic neurons and geniculocortical neurons, presumably by inhibiting adenylyl cyclase activity and thus decreasing the intracellular cAMP levels (Pape, 1992; Rainnie et al, 1994). It is possible that the same pathway might be responsible for the ischemia-induced decrease in I_h. As expected, the application of selective A1 receptor antagonist DPCPX (200nmol/L) led to a significant increase in I_h in postischemic neurons (by 23.3% ± 5.9%, n = 7; P <0.01), but not in control neurons (by 3.0% ± 6.1%, n = 7; P > 0.05; Figure 4). In addition, application of DPCPX caused a positive shift of the half-activation voltage of I_h in postischemic neurons (5.1 ± 1.9mV, n = 5; P <0.05). Furthermore, when DPCPX (200nmol/L) was applied in the presence of 8-Br-cAMP (100mmol/L), no obvious increase in I_h was detected in postischemic neurons (by 1.2% ± 2.2%, n = 5; P > 0.05; Figure 4B). This is in agreement with the fact that the actions of A1 receptors were mediated by cAMP pathway. Thus, our findings strongly support the conclusion that through a cAMP pathway, tonic activation of adenosine A1 receptors contributes, at least in part, to the postischemic decrease of I_h in cholinergic interneurons.

Figure 4

Involvement of adenosine A1 receptor in I_h inhibition after ischemia. (A) Representative traces of I_h recorded before and during application of selective A1 receptor antagonist DPCPX (200 nmol/L). Traces were evoked at a voltage step of −90 mV. (B) Group data showing that application of DPCPX significantly increased I_h in postischemic neurons but had no effect in control neurons. Moreover, the effects of DPCPX in postischemic neurons could be completely occluded in the presence of 8-Br-cAMP (100 μmol/L). The amplitudes of I_h in the presence of DPCPX were normalized to those before DPCPX application (dash line). *P<0.01.

Postischemic Changes of Neuronal Excitability in Cholinergic Interneurons

Activity of I_h is critical to the intrinsic membrane properties (Robinson and Siegelbaum, 2003). Previous studies have shown that the spontaneous firing in cholinergic interneurons is largely controlled by I_h. Specifically, a partial blockade of I_h causes a decrease in both firing rate and regularity (Bennett et al, 2000; Deng et al, 2007). Consistent with the inhibition of I_h, we found a significantly decreased firing rate in postischemic neurons (control: 3.08 ± 0.33 Hz, n = 19; ischemia: 1.20 ± 0.38 Hz, n = 5; P <0.01), accompanied by an increase in the CV of interspike interval (control: 0.29 ± 0.02, n = 19; ischemia: 0.53 ± 0.14, n = 5; P <0.01; Figure 5). Changes of TTX-sensitive Na⁺ currents may also modulate spontaneous firing. However, these currents have been shown to be unchanged in cholinergic interneurons in response to ischemic insults (Pisani et al, 1999). Therefore, inhibition of I_h might be the major mechanism underlying the postischemic suppression of spontaneous activities, although other membrane conductances (e.g., K⁺ currents) cannot be completely excluded.

Figure 5

Suppression of spontaneous activity after ischemia. (A) Representative cell-attached recordings of spontaneous firing in control and postischemic neurons. The spikes were fewer and less regular after ischemia. (B) Group data showing the decreased firing rate (upper panel) and the increased CV of interspike interval (lower panel) in postischemic neurons. *P<0.01.

In cholinergic interneurons, hyperpolarizing currents can induce rebound excitation (i.e., RD) with increasing firings (Deng et al, 2007; Reynolds et al, 2004). The RD is primarily produced by I_h because these neurons do not express T-type Ca²⁺ currents (Bennett et al, 2000). To test whether the rebound excitation was altered after ischemia, the RD amplitude was measured in the presence of TTX (0.5 μmol/L) to eliminate the contribution of Na⁺ currents and to suppress spontaneous spikes. With the continuous application of TTX, the membrane potentials were −61.4 ± 2.4 mV (n = 11) and −63.7 ± 3.8mV (n = 7) in control and postischemic neurons (P > 0.05), respectively. The neurons were then held at −55 mV to ensure that most I_h channels were deactivated before the onset of hyperpolarization. Under these conditions, a prominent RD could be evoked by the offset of a hyperpolarizing current pulse in both control and postischemic neurons (Figure 6A). Moreover, the RD could be completely abolished with the I_h channel blocker ZD7288 (30 μmol/L; n = 6 in each group). Because membrane resistance at depolarizing voltages (more than −55 mV) may affect the amplitude of RD, the voltage responses to depolarizing current pulses (30 to 120 pA) were compared between control and postischemic neurons, and no obvious change was detected after ischemia. For example, when elicited with a current of 90 pA, the voltages were 11.6 ± 1.8 mV (n = 5) in control and 11.4 ± 1.5 mV in postischemic neurons (n = 6; P>0.05; Figures 6A and 6B). We then examined the amplitude of RD evoked by hyperpolarizing pulses (from −30 to −300 pA) with different durations (120, 600, and 1,000ms). In both control and postischemic neurons, the RD amplitude was positively related to the amplitude and duration of hyperpolarizing pulses, increased with increasing amplitude or duration of current pulse (Figures 6C and 6D). Most importantly, the RD amplitude was dramatically decreased after ischemia (by 45 to 60%, n = 6 in each group; P < 0.01; Figures 6C and 6D). Together with previous studies (Bennett et al, 2000; Reynolds et al, 2004), our data indicate that postischemic inhibition of I_b is capable of reducing rebound excitation in cholinergic interneurons.

Figure 6

Postischemic reduction of RD in cholinergic interneurons. (A) Typical voltage responses to depolarizing and hyperpolarizing current pulses. Neurons were held at −55 mV in the presence of TTX (0.5 μmol/L). In the postischemic neuron, both voltage sag during hyperpolarization and RD after hyperpolarization were dramatically reduced, whereas the voltage response to a depolarizing current was similar to that of control. (B) Group data showing that the voltage responses to depolarizing current pulses were not changed after ischemia (P > 0.05 versus control). (C, D) Reduced RD in postischemic neurons. (C) Plots of RD amplitudes as a function of the hyperpolarizing currents (600 ms). In both control and postischemic neurons, the RD amplitude was positively dependent on the amplitude of hyperpolarizing current. A significant decrease in the amplitude of RD, produced by each pulse, was detected in postischemic neurons (P < 0.01 versus control). (D) Histogram showing that the RD amplitude was also dependent on the duration of hyperpolarizing current (–180 pA). The RD amplitudes were significantly decreased after ischemia, across all durations tested (P < 0.01 versus control).

Discussion

The major findings of this study are that I_h in striatal cholinergic interneurons is decreased early after transient forebrain ischemia, which results, at least in part, from the tonic activation of adenosine A1 receptors via a cAMP-dependent pathway. Postischemic inhibition of I_h is well correlated with the decreased neuronal excitability, suggesting that adenosine-mediated I_h inhibition may be one of the mechanisms underlying neuronal resistance to cerebral ischemia.

Mechanisms of Postischemic Decrease of Ih in Cholinergic Interneurons

It has long been recognized that striatal cholinergic interneurons express I_h, which is essential for neuronal excitability (Bennett et al, 2000; Jiang and North, 1991; Kawaguchi, 1993). Several studies have shown that the activity of I_h was altered in other neurons in response to hypoxia. For instance, I_h in thalamocortical neurons was increased during hypoxia in vitro, which might be due to a positive shift of the voltage dependence and an acceleration of activation (Erdemli and Crunelli, 1998). On the contrary, acute hypoxia caused a disappearance of I_h in inspiratory brainstem neurons and dorsal root ganglion neurons (Gao et al, 2006; Mironov et al, 2000). In addition, a partial inhibition ofI_h, without significant shift of voltage-dependent activation, was observed in hippocampal CA1 pyramidal neurons after neonatal hypoxia in vivo (Zhang et al, 2006a). Different from the above observations, this study revealed that I_h in striatal cholinergic interneurons was significantly decreased early (3h) after transient forebrain ischemia and suggested that the postischemic inhibition of I_h might be caused by a negative shift of the voltage dependence of I_h activation. It is therefore likely that the responses of I_h to ischemia/hypoxia vary dramatically among particular neurons and may depend on specific insults.

Although little is known about the mechanisms underlying the changes of I_h in response to ischemic insults, altered intracellular cAMP levels might be involved in the postischemic inhibition of I_h in striatal cholinergic interneurons. It has been well established that intracellular cAMP regulates I_h by shifting its voltage-dependent activation (Frere et al, 2004; Robinson and Siegelbaum, 2003). In particular, decreases in the intracellular cAMP levels cause negative shifts of the voltage-dependent activation and thus reduce I_h at a given voltage. Morphologic studies have shown that striatal cholinergic interneurons express cAMP-sensitive I_h channel (HCN) subunits, including HCN2–4 (Notomi and Shigemoto, 2004; Santoro et al, 2000). Indeed, I_h in these neurons is enhanced in the presence of membrane-permeable cAMP analog (8-Br-cAMP), whereas application of the cAMP pathway inhibitor (Rp-cAMP) decreases the current amplitude (Deng et al, 2007). Furthermore, as detected in this study, the modulation of I_h through cAMP pathway was remarkably suppressed in postischemic neurons. These observations strongly support the idea that a decrease of intracellular cAMP levels is involved in the ischemia-induced I_h inhibition in cholinergic interneurons. On the basis of our previous results, it is further suggested that the postischemic inhibition of I_h might be independent of protein kinase A, as its selective inhibitors failed to affect the cAMP-mediated modulation of I_h (Deng et al, 2007). If a reduction of intracellular cAMP levels was the only mechanism mediating the postischemic inhibition of I_h, the application of 8-Br-cAMP would produce a greater effect on I_h in postischemic neurons. However, our data revealed that 8-Br-cAMP was less effective in postischemic neurons, suggesting that other pathways (cAMP-independent) might also contribute to the postischemic inhibition of I_h. One of the possible candidates is phosphatidylinositol-4,5-bisphosphate, because this phosphoinositide modulates I_h channels independent of cAMP, and a decrease of phosphoinositides may inhibit I_h in neurons (Zolles et al, 2006). It is also possible that ischemic insults may cause a loss of functional I_h channels, leading to a decrease in I_h amplitude.

Adenosine is a potential modulator of striatal cholinergic interneurons (Song et al, 2000). Extracellular adenosine concentration is dramatically increased during and after ischemia in vitro (Pearson et al, 2006). In addition, it appears that adenosine A1 receptors in the neostriatum are tonically active early (4 to 6h) after transient forebrain ischemia (Pang et al, 2002). One of the signaling pathways of A1 receptors is inhibition of adenylyl cyclase, leading to a decrease in cAMP (Ralevic and Burnstock, 1998). This pathway has been believed to be a mechanism by which adenosine inhibits I_h (Pape, 1992). In this study, we found that the blockade of A1 receptors with a selective antagonist (DPCPX) produced an increase of I_h in postischemic neurons but had no effect in control neurons. Moreover, the effects of DPCPX could be completely occluded by a cAMP analog 8-Br-cAMP. These findings indicate that the tonic activation of A1 receptors might contribute to the postischemic inhibition of I_h by decreasing intracellular cAMP. Interestingly, our findings also suggest that I_h in postischemic neurons was substantially upregulated by intracellular Ca²⁺. Given the decreased neuronal excitability of striatal cholinergic interneurons after ischemia, the Ca²⁺-mediated modulation of I_h might be related to an intracellular Ca²⁺ mobilization rather than an extracellular Ca²⁺ influx. If this is true, A1 receptors might also play a role in this process because their activation is able to stimulate the release of Ca²⁺ from intracellular stores, such as endoplasmic reticulum, by activating phospholipase C (Ralevic and Burnstock, 1998). Overall, it seems that activation of A1 receptors may exhibit opposite effects on I_h through distinct signaling pathways, with an overwhelming inhibition in striatal cholinergic interneurons after ischemia.

However, it remains uncertain as to whether the postischemic inhibition of I_h is solely mediated by adenosine. Neurotransmitters and neuromodulators that regulate cAMP levels may produce a reduction of I_h in postischemic neurons. Activation of dopamine D2 receptors has been shown to inhibit I_h in cholinergic interneurons (Deng et al, 2007); however, it is unlikely that dopamine may contribute to the postischemic inhibition of I_h, as the elevated dopamine release is returned to control levels within 30 mins after ischemia (Globus et al, 1988). Also unlikely is serotonin, considering its lack of effect on I_h in these neurons (Blomeley and Bracci, 2005). Nevertheless, our findings support the fact that adenosine, at least partially, mediates the postischemic inhibition of I_h in striatal cholinergic interneurons.

Functional Implications of Ih in Ischemia-Induced Neuronal Damage

Accumulating evidence indicates that I_h is involved in neurologic disorders such as seizure and stroke. In addition to regulating intrinsic membrane properties, including the resting membrane potential and the input resistance, I_h modulates neuronal responses to synaptic inputs (Robinson and Siegelbaum, 2003). The functional significance of I_h in neurologic disorders may depend on the subcellular compartmentalization and property of I_h channels, the nature and location of synaptic inputs, and the electrophysiological activity of particular neurons (Santoro and Baram, 2003). For instance, I_h may exhibit opposite roles in epileptogenesis. An enhanced I_h in CA1 pyramidal neurons (most likely in soma) can convert the potentiated synaptic inhibition to hyperexcitability (i.e., an increase of rebound firing) in a frequency-dependent manner, which promotes epileptogenic process (Chen et al, 2001). In contrast, an increase of dendritic (but not somatic) I_h attenuates the summation of excitatory synaptic inputs from distal dendrites, leading to a decrease of neuronal excitability and hence to the prevention of epileptogenesis (Poolos et al, 2002).

Studies on the involvement of I_h in cerebral ischemia are less extensive. This study has shown that the spontaneous activity and hyperpolarization-induced RD in striatal cholinergic interneurons were dramatically suppressed after ischemia. As both spontaneous firing and RD are largely controlled by I_h in these neurons, our findings indicate that postischemic inhibition of I_h is sufficient to decrease neuronal excitability and that I_h inhibition may reduce hyperexcitability induced by inhibitory synaptic inputs in postischemic neurons. In cholinergic interneurons, I_h may have little impact on excitatory synaptic inputs because I_h channels are mainly located in the soma, as shown by the immunostaining against HCN3 (Notomi and Shigemoto, 2004). Additionally, both membrane potentials and voltage responses to depolarizing currents were not changed significantly after ischemia. The simplest explanation is that cholinergic interneurons have less negative membrane potentials (more than −64 mV in the presence of TTX), and thus only a small proportion of I_h is constitutively active. The roles of I_h in cholinergic interneurons may differ from those in other neurons. Both ischemia-sensitive and -resistant neurons express I_h. The functional significance of I_h in neuronal damage after cerebral ischemia may be related to the nature of its changes (i.e., excitatory or inhibitory). It is reasonable to speculate that changes of I_h with inhibitory effects on neuronal excitability might be neuroprotective.

One of the most intriguing findings in this study is that a tonic activation of A1 receptors produced an inhibition of I_h after ischemia, leading to a suppression of neuronal excitability. Endogenous adenosine is believed to be neuroprotective against cerebral ischemia (Fredholm et al, 2005). Activation of A1 receptors has been shown to cause the presynaptic inhibition of excitatory neurotransmission (Pang et al, 2002; Pearson et al, 2006). However, intrinsic membrane excitability is also critical to neuronal injury after ischemic insults. Therefore, the postischemic inhibition of I_h may represent a novel mechanism by which adenosine protects neurons from cerebral ischemia.

Disclosure/Conflict of Interest

The authors declare no competing financial interests.