Gabapentin Increases the Hyperpolarization-activated Cation Current Ih in Rat CA1 Pyramidal Cells

Authors


Address correspondence and reprint requests to Dr. R. Surges at Department of Neurology, Section of Clinical Neuropharmacology, Neurozentrum, Breisacher Str. 64, 79106 Freiburg i.Br., Germany. E-mail: surges@nz.ukl.uni-freiburg.de

Abstract

Summary:  Purpose: Gabapentin (GBP) is a commonly used drug in the treatment of partial seizures, but its mode of action is still unclear. The genesis of seizures in temporal lobe epilepsy is thought to be crucially influenced by intrinsic membrane properties. Because the Ih substantially contributes to the intrinsic membrane properties of neurons, the effects of GBP on the Ih were investigated in CA1 pyramidal cells of rat hippocampus.

Methods: CA1 pyramidal cells in hippocampal slices were examined by using the whole-cell patch-clamp technique.

Results: GBP increased the Ih amplitude in a concentration-dependent manner mainly by increasing the conductance, without significant changes in the activation properties or in the time course of Ih. The effects ranged from ∼20% at 50 μM, ∼25% at 75 μM, to ∼35% at 100 μM GBP (at –110 mV). In the presence of intracellular cyclic adenosine monophosphate (cAMP), the effects of GBP on Ih were similar to those obtained in the absence of cAMP.

Conclusions: These results suggest that GBP increases the Ih through a cAMP-independent mechanism. Because the applied GBP concentrations were in a clinically relevant range, the observed effect may contribute to the anticonvulsant action of GBP in partial seizures and may represent a new concept of how this anticonvulsant drug works.

Gabapentin (GBP, Neurontin) has been used for several years to treat neurologic diseases. It has been demonstrated to be useful in the treatment of focal epilepsies, essential tremor, chronic neuropathic pain syndromes, and migraine (1–4). Although its clinical relevance in treating partial seizures has been proven, the molecular mechanisms underlying its efficacy are still unclear. GBP exerts effects on the extrinsic properties of neurons (i.e., on different neurotransmitter systems at the presynaptic and the postsynaptic site), as well as on the intrinsic properties (e.g., on voltage-dependent ionic channels) (for review, see (5,6). GBP binds to the α2δ subunit of voltage-gated calcium channels and may thereby induce an inhibition of calcium currents (7–12, but see 13). Recently GBP also was suggested to act as a KATP channel agonist (14).

Epilepsy is one of the most common neurologic diseases. Among the different types of epilepsy, temporal lobe epilepsy (TLE) is the most frequent. There is increasing evidence that the genesis of epileptiform activity within the hippocampus is due to alterations in extrinsic as well as intrinsic properties of neurons (15–17). The hyperpolarization-activated cation current (Ih) contributes substantially to the intrinsic membrane properties of hippocampal neurons (18,19). The Ih is a slowly developing inward current, carried by K+ and Na+, which shows no inactivation and is modulated by internal cyclic adenosine-3′-5′-monophospate (cAMP; 20,21) as well as by external K+(22). Because the Ih has been found in a variety of neuronal tissues, it is thought to be an ubiquitous component of the nervous system (21,23). Several recent findings pointed out that Ih is essentially involved in pathophysiologic processes such as epilepsy (24–26). Therefore we tested the effect of the GBP on the biophysical properties of Ih in CA1 pyramidal cells with the help of the whole-cell patch-clamp technique in hippocampal slices. We demonstrate here that GBP increases Ih in a concentration-dependent manner. At clinically relevant concentrations, GBP mainly increases the conductance of Ih without significant changes in the activation properties or in the time course of Ih. This effect may contribute to the anticonvulsant action of GBP in partial seizures and may represent a new concept of how this anticonvulsant drug (ACD), and perhaps also others, work.

MATERIALS AND METHODS

Slice preparation

Hippocampal slices were prepared from 14- to 26-day-old Wistar rats of either sex. Rats were decapitated in agreement with national and institutional guidelines, and the brain was rapidly removed, hemisected, and submerged in ice-cold physiologic Ringer's solution (Biometra, Göttingen, Germany), which was continuously bubbled with 95% O2 and 5% CO2 and which contained (in mM): NaCl, 125; NaHCO3, 25; KCl, 2.5; NaH2PO4, 1.25; MgCl2, 1; glucose, 25; and CaCl2, 2. Then the hemispheres were blocked in the horizontal plane and glued to the stage of a vibratome (Leica VT 1000S, Wetzlar, Germany). Horizontal sections through the hippocampus were cut at a thickness of 350 or 400 μm, and slices were then transferred to a holding chamber where they were submerged in Ringer's solution. This solution was continuously bubbled with 95% O2 and 5% CO2 and was maintained at 35°C for the first 30 min and then at room temperature (20–22°C). Slices were kept in the holding chamber for ≥1 h before recordings were performed.

Electrophysiologic recordings

Slices were transferred, one at a time, into the recording chamber, fixed with a nylon-grid, and were continuously perfused with oxygenated Ringer's solution at room temperature. Neurons were then visualized by infrared differential interference contrast (IR-DIC) videomicroscopy with a Newvicon camera (C2400; Hamamatsu, Hamamatsu City, Japan) and an infrared filter (RG9; Schott, Mainz, Germany) to an upright microscope (Axioskop 2 FS; Zeiss, Oberkochen, Germany) equipped with a ×40 water-immersion objective. Recordings were made in the whole-cell configuration of the patch-clamp technique by using an Axopatch 200B amplifier and pClamp 8 (Axon Instruments, Foster City, CA, U.S.A.). Signals were filtered online at 10 kHz and digitized at 50 kHz during current-clamp experiments, and at 1 kHz and 5 kHz, respectively, during voltage-clamp experiments (Digidata 1200; Axon Instruments). Pipettes were pulled from nonfilamented borosilicate glass capillaries (Hilgenberg, Malsfeld, Germany; outer diameter, 1.5 mm; inner diameter, 1.0 mm) on a P-97 Flaming/Brown horizontal puller (Sutter Instruments, Novato, CA, U.S.A.) and had resistances between 3.5 and 5.5 MΩ when filled with pipette solution. Standard pipette solution contained (in mM): K-gluconate, 130; KCl, 20; MgCl2, 1, CaCl2, 1; HEPES, 10; EGTA, 1; Mg-ATP, 2, Na-GTP, 0.4; pH at 7.25 adjusted with KOH. The offset potential between the patch-pipette and the reference electrode was zeroed before the tight-seal (>2 GΩ) was established. The membrane and series resistance were determined by using the “membrane test” protocol provided by the software pClamp 8.0. Series resistance was between 8 and 15 MΩ and was monitored during the experiments (compensation ≤60%). Leakage currents were reduced by addition of external Ba2+. Furthermore, a short pulse from –40 mV to –45 mV for 200 ms was applied after each protocol, and the leakage current amplitudes were substracted offline from the current peaks.

Solutions and drugs

All drugs were purchased, if not otherwise indicated, from Sigma (Taufkirchen, Germany). The standard bath solution consisted of (in mM): NaCl, 120; NaHCO3, 25; KCl, 2.5; MgCl2, 1; glucose, 25; CaCl2, 2. To isolate Ih during voltage-clamp experiments, the following substances were added (in mM): tetrodotoxin (TTX; Tocris, Biotrend Chemikalien, Köln, Germany), 0.01; NiCl2, 1.0; tetraethylammoniumchlorid (TEA), 10; 4-aminopyridine (4-AP), 2; BaCl2, 0.5. GBP (Tocris) was dissolved in H2O in a stock solution of 10 mM. ZD7288 (Tocris) was dissolved in H2O in a stock solution of 25 mM.

Data acquisition and analysis

Data were acquired with the help of pCLAMP 8.0 and analyzed offline by using the software-packages Clampfit 8.0 (Axon Instruments) and GraphPadPRISM 2.0 (GraphPad Software Inc., San Diego, CA, U.S.A.). To determine the time constants of activation and deactivation, the following double exponential term was used with the Chebyshev fitting routine of Clampfit 8.0:

image

with A being the current amplitude, and τ the time constant, and C a constant. The fittings were made from a time point at which there was no interference with capacitive transients. Boltzmann fits were made by using a least-squares nonlinear curve-fitting routine in GraphPadPRISM to determine the activation curve with the following equation:

image

with V50 being the midpoint activation voltage and s being the slope. Linear regression analysis also was performed with GraphPadPRISM. Statistical significance (p < 0.05) was assessed by using the two-tailed paired or unpaired Student's t test. All data were expressed as mean ± SEM.

RESULTS

Basic properties

At the beginning of each experiment, a family of current pulses was applied to the cells in the current-clamp mode from +50 pA to –150 pA in 50-pA steps (Fig. 1). On tonic depolarizing currents, most cells presented a train of action potentials with increasing intervals between the spikes. This adaptation of action-potential frequency is typical for CA1 pyramidal neurons. On tonic hyperpolarizing current injections, the cells showed a phenomenon called “voltage sag,” which consists of a slow depolarization of the cell membrane due to the activation of Ih. Only cells that displayed both characteristics, frequency adaptation and voltage sag, were included in this study. The CA1 pyramidal cells had a resting potential of –60.5 ± 0.65 mV and a membrane resistance of 232.9 ± 17.64 MΩ (n = 29). In the voltage-clamp mode, currents were evoked by hyperpolarizing voltage steps of a duration of 2 s from a holding potential of –50 mV up to –120 mV in 10-mV increments (Fig. 2A, upper panel). The amplitude of Ih was determined by substracting the instantaneous current at the beginning of the voltage step from the steady-state current at the end (Fig. 2A, arrows in the upper traces). The current–voltage relation revealed that with increasing amplitude of the voltage step, the Ih amplitude increases (Fig. 2B, control), from –11.0 ± 4.3 pA at –60 mV to -243.4 ± 82.1 pA at –120 mV (n = 5). To characterize further the Ih, ZD7288, a specific known blocker of Ih(27), was applied with the bath solution at a concentration of 50 μM (Fig. 2A, lower traces). In the presence of ZD7288, the amplitude of Ih was significantly reduced at potentials more negative than –70 mV (e.g., by 70.6 ± 2.2% at –110 mV) (Figs. 2B and 3C; paired t test, p < 0.003, n = 5).

Figure 1.

Electrophysiologic characteristics of CA1 pyramidal cells. Tonic depolarizing currents of 50 pA evoked a train of action potentials with increasing spike intervals (upper trace). During injection of hyperpolarizing currents from –50 to –150 pA, the so-called voltage sag, which is due to the activation of Ih, can be observed (lower traces).

Figure 2.

Characterization of Ih. A: Current traces were evoked by hyperpolarizing voltage steps of a duration of 2 s from a holding potential of –50 mV up to –120 mV in 10-mV increments (upper panel). The amplitude of Ih was determined by substracting the instantaneous current at the beginning of the voltage step from the steady-state current at the end (arrows in the control traces). Application of ZD7288 (50 μM) reduced Ih (lower current traces). B: The current–voltage relation under control conditions shows that the Ih amplitude increases with increasing amplitude of the command potential. Application of ZD7288 significantly reduced the Ih (paired t test, p < 0.05; n = 5).

Figure 3.

Gabapentin (GBP) increases Ih in a concentration-dependent manner mainly by increasing G(Ih) without significant modification of the activation properties. A: Original current traces before (left traces) and during application of GBP at a concentration of 100 μM (right traces). B: The current–voltage relation before and during GBP influence (100 μM) shows that the Ih amplitude was significantly increased by GBP (paired t test, p < 0.05; n = 5). Inset: The data points in the voltage range between –90 and –120 mV were subjected to a linear regression analysis whose slope represents G(Ih). G(Ih) was significantly increased by 45.9 ± 11.9% by GBP application (paired t test, p < 0.029; n = 5). C: GBP increased the Ih amplitude in a concentration-dependent manner at concentrations of 50, 75, and 100 μM (white bars, n in parenthesis). In the presence of intracellular cyclic adenosine monophosphate (cAMP), GBP (100 μM, dotted bar) increased the Ih to the same amount as in the absence of cAMP. ZD7288 reduced the Ih amplitude to ∼30% (black bar). All values at a potential of –110 mV. D: Analysis of the tail currents (see inset) revealed a nonsignificant shift of the activation curve during GBP influence (n = 5). Symbols are the same as in B.

Effect of GBP on the Ih amplitude and the conductance

GBP was applied via the bath perfusion system. Continuous application of GBP at a concentration of 100 μM led to a substantial increase of the Ih amplitude (Fig. 3A, left and right traces). The effect of GBP on the Ih amplitude began within 5–8 min and reached its maximum after ∼30 min. On washout, the Ih only partially recovered from its GBP-induced increase (not shown). The current–voltage relation shows that the Ih amplitude was significantly increased by application of GBP between –90 and –120 mV (Fig. 3 B). At a potential of –110 mV, the Ih amplitude was increased from –312.0 ± 19.0 pA by 34.0 ± 3.4% to –415.4 ± 18.4 pA (paired t test, p < 0.0001, n = 7). To test whether this effect was concentration dependent, GBP also was administered in a concentration of 50 and 75 μM (current traces not shown). At –110 mV, Ih was increased by 19.1 ± 7.7% at 50 μM (n = 5) and by 25.3 ± 2.1% at 75 μM GBP (n = 3) (Fig. 3C). To determine the conductance of Ih[G(Ih)], the data points within the linear portion of the current–voltage relation (between –90 and –120 mV) were subjected to a linear regression analyis: Its slope represents G(Ih) (not shown). Under control conditions, G(Ih) was 5.45 ± 0.60 nS and was significantly increased by 45.9 ± 11.9% to 7.78 ± 0.79 nS at 100 μM GBP (paired t test, p < 0.029, n = 5; Fig. 3B, inset).

Activation properties

To show the voltage dependency of Ih activation, the tail currents on relaxation of the different voltage steps to the holding potential were analyzed (Fig. 2A, marked with a box in the upper traces, and Fig. 3D, inset). Therefore the tail-current amplitudes were normalized to the maximal amplitude evoked by a preceeding voltage step to –120 mV and plotted against the corresponding preceeding command potential. Then the resulting data points were fitted with a Boltzmann equation (Fig. 3D). This analysis revealed the potential at which the Ih is activated to its half-maximal value (V50), and the steepness of the activation curve expressed by its slope s. V50 was –80.72 ± 0.73 mV under control conditions and not significantly changed by application of 100 μM GBP (V50 = –84.5 ± 0.64 mV; paired t test, p = 0.1750; n = 5). The slope s was –10.7 ± 0.66 mV−1 under control conditions and –10.0 ± 0.56 mV−1 under GBP influence (paired t test, p > 0.05; n = 5).

Effects of GBP in the presence of intracellular cAMP

Further to investigate the mechanism by which GBP exerts the increase of Ih, cAMP (10 μM) was added to the pipette solution. The Ih was measured under control conditions and during GBP application (100 μM, current traces not shown). In the presence of cAMP, the Ih amplitude also was increased by 100 μM GBP, from –301.3 ± 49.8 pA by 30.1 ± 7.9% to –396 ± 77.8 pA (at –110 mV; n = 4; Fig. 3C). The comparison between the effect of GBP on the Ih (at –110 mV) in the presence and absence of intracellular cAMP did not reveal a significant difference (Fig. 3C).

Time course of Ih

To describe the time course of Ih activation, the current traces were best fitted with the sum of two exponentials. Figure 4A exemplifies current traces at –80 and –120 mV (dotted black line) under control conditions as well as during GBP application (100 μM) together with the superimposed corresponding fit curves (white line). Both the short time constant τshort and the long time constant τlong became shorter with increasing amplitude of the command potential (Fig. 4B). τshort was 53.0 ± 2.4 ms under control (at –120 mV) and 56.5 ± 4.9 ms during GBP application (paired t test, NS; n = 5). τlong was 517.3 ± 35.1 ms under control conditions and 537.8 ± 44.4 ms under GBP influence (paired t test, NS; n = 5). The time constants of deactivation were determined by fitting the tail-current trace evoked on the return from –120 mV to the holding potential. Under control conditions, τshort was 133.5 ± 4.6 ms and 122.3 ± 10.5 ms under GBP influence (paired t test, NS; n = 5). τlong was 650 ± 153.5 ms under control conditions and 606 ± 75.2 ms under GBP influence (paired t test, NS; n = 5;Fig. 4C).

Figure 4.

The time course of Ih activation and deactivation is not significantly changed by gabapentin (GBP). A: Original current traces at –80 and –120 mV (dotted black line) under control conditions and during GBP-application (100 μM) are shown together with the superimposed double exponential fit curves (white line). B: Both τshort and τlong became shorter with increasing amplitude of the command potential. No significant differences between control (filled circles) and test conditions (GBP 100 μM open circles, n = 5 for each data point). C: The time constants of deactivation were determined by fitting the tail-current trace evoked on the return from –120 mV to the holding potential. No significant differences between the two groups (GBP, 100 μM; n = 5 for each bar). See text for details.

DISCUSSION

We demonstrated for the first time that GBP increases the Ih mainly by increasing G(Ih) without significant changes in the activation properties or in the time course of Ih. The observed effect occured in a clinically relevant concentration range, because the serum level of GBP was determined to be between 12 and 88 μM(28).

Effect of GBP on the biophysical properties of Ih

The Ih amplitude was increased by GBP in a concentration-dependent manner. Because our experiments were performed with a slow application of the substances via the bath-solution system, we cannot elucidate the real time course of the onset of the GBP effect. In the present experimental setup, the increase of the Ih amplitude began within 5 to 8 min and reached its maximum after ∼30 min. Additionally, we only observed a partial recovery of the GBP effect in our experiments that might also be related to the slow perfusion system together with an intracellular site of action of GBP, which has been assumed to be due to its delayed onset of action (5).

The Ih amplitude can be increased in two ways, by (a) increasing G(Ih), and/or (b) by shifting the activation curve in the positive direction. Linear regression analysis of the linear portion of the current–voltage relation revealed a significant increase of G(Ih) on GBP application, whereas there was no significant effect on the activation curve. Because intracellular cAMP is known to increase Ih by shifting the activation curve in the positive direction (20), one would expect a shift in the positive direction if cAMP were mediating the GBP effect. Furthermore, the effect of GBP on the Ih amplitude in the absence and presence of intracellular cAMP was similar. Therefore we conclude that a second-messenger pathway via cAMP is not involved in mediating the GBP effect and that it is more likely that Ih channels possess a binding site for GBP. Interestingly, Poolos et al. (29) showed that lamotrigine (LTG), another novel ACD, increased the Ih of CA1 dendrites by shifting the activation curve in the positive direction without modifications of G(Ih). Because GBP and LTG are different molecules, they probably possess different binding sites and may act via different mechanisms. However, taken together, these data suggest that Ih channels represent a novel target for ACD therapy.

Functional implications

The Ih has a considerable impact on the resting properties of hippocampal neurons by slightly depolarizing the cell membrane and by substantially decreasing the input resistance (18,19). Thus an increased Ih leads to a decreased input resistance. Because a decrease in the input resistance makes cells less sensitive to synaptic input, it is tempting to speculate that the GBP-induced increase in Ih protects the CA1 pyramidal cells against excessive synaptic or intrinsic activity and may stabilize thereby the neuronal network within the hippocampus. The hypothesis that the Ih contributes to neuronal stabilization is also supported by two recent findings. First, in CA1 pyramidal cells, Ih channels are predominantly located on dendrites, where they have been shown to dampen dendritic excitability by decreasing the apparent input resistance (30,31). Second, Brauer et al. (24) reported that after a kainate-induced lesion within the entorhinal cortex (i.e., under pathophysiologic conditions), the Ih amplitude was decreased in hippocampal hilar and mossy cells. In contrast to the idea that Ih has a stabilizing function, Chen et al. (25) worked out another hypothesis of the role of pathophysiologically altered Ih in their animal model of febrile seizures. It has been reported that γ-aminobutyric acid subtype A (GABAA)-receptor sensitivity and/or GABAergic neurotransmission are increased in some types of epilepsy (32–36). In the adult nervous system, GABAergic neurotransmission is considered to inhibit neuronal excitability by hyperpolarizing the cell membrane. Because the depolarizing Ih is activated by hyperpolarization, Chen et al. (25) suggested that the pathophysiologically increased Ih accounts for the reversion of the increased GABAergic synaptic input into neuronal excitation. Further investigations are required to learn under which conditions an increase in Ih (e.g., induced by GBP) contributes to stabilization or destabilization of neurons and neuronal networks. At least, that GBP is effective in treating partial seizures may be in favor of a stabilization of hippocampal networks. Conversely, the fact that GBP may precipitate or aggravate absence seizures that are known to be promoted by hyperpolarization of thalamic neurons suggests a destabilizing action of GBP in this special condition of absence seizures (37).

CONCLUSIONS

We showed that GBP augments the Ih mainly by increasing the conductance via a cAMP-independent way. Because the applied GBP concentrations were in a clinically relevant range, our data suggest that the GBP-induced increase in Ih contributes to its anticonvulsant effects. Thus the activation of Ih channels may represent a new concept of how GBP and perhaps other novel ACDs work.

Acknowledgment: This work was supported by the Deutsche Forschungsgemeinschaft (SFB 505/ TP C8).

Ancillary