Function of inhibitory micronetworks is spared by Na+ channel-acting anticonvulsant drugs

Abstract

The mechanisms of action of many CNS drugs have been studied extensively on the level of their target proteins, but the effects of these compounds on the level of complex CNS networks that are composed of different types of excitatory and inhibitory neurons are not well understood. Many currently used anticonvulsant drugs are known to exert potent use-dependent blocking effects on voltage-gated Na(+) channels, which are thought to underlie the inhibition of pathological high-frequency firing. However, some GABAergic inhibitory neurons are capable of firing at very high rates, suggesting that these anticonvulsants should cause impaired GABAergic inhibition. We have, therefore, studied the effects of anticonvulsant drugs acting via use-dependent block of voltage-gated Na(+) channels on GABAergic inhibitory micronetworks in the rodent hippocampus. We find that firing of pyramidal neurons is reliably inhibited in a use-dependent manner by the prototypical Na(+) channel blocker carbamazepine. In contrast, a combination of intrinsic and synaptic properties renders synaptically driven firing of interneurons essentially insensitive to this anticonvulsant. In addition, a combination of voltage imaging and electrophysiological experiments reveal that GABAergic feedforward and feedback inhibition is unaffected by carbamazepine and additional commonly used Na(+) channel-acting anticonvulsants, both in control and epileptic animals. Moreover, inhibition in control and epileptic rats recruited by in vivo activity patterns was similarly unaffected. These results suggest that sparing of inhibition is an important principle underlying the powerful reduction of CNS excitability exerted by anticonvulsant drugs.

Keywords: anticonvulsants; carbamazepine; epilepsy; inhibition; interneurons.

Figure 1.

Carbamazepine effects on firing properties of principal neurons. A, Morphological reconstruction of a representative pyramidal neuron. SLM, Stratum lacunosum moleculare; SO, Stratum oriens; SP, Stratum pyramidale; SR, Stratum radiatum; Alv, alveus. B, Effects of 30 μm CBZ on intrinsic firing induced by current injection (lowermost traces; 225, 400, and 700 pA, 1000 ms). C, Corresponding input–output relationship of the average firing rate during the 1 s current injection versus the magnitude of the current injection. D, Effects of CBZ on the maximal discharge frequency measured during the 1 s current injection (**p < 0.01). E, CBZ effects on the maximal firing rate measured during the first 50, 100, 200, 500, and 1000 ms of current duration (see insets; *p < 0.05, **p < 0.01, Wilcoxon signed rank test). F, Percent reduction in firing frequency by CBZ within the different time intervals shown in E (*p < 0.05, **p < 0.01, Wilcoxon signed rank test). G, Top, Recording configuration to examine CBZ effects on synaptically induced firing. Schaffer collaterals were stimulated with a bipolar steel electrode placed in the radiatum while firing was monitored with cell-attached recordings. To prevent recurrent excitation, a cut was made between CA1 and CA3. Bottom, Raster plot of action potential firing during a 50 Hz stimulus (Stim) train of 500 ms duration. All three conditions (control, application, and washout) were conducted in the presence of the GABA_A blocker gabazine (10 μm). H, I, CBZ effects on the average firing frequency during consecutive 100 ms intervals (H) and the corresponding magnitude of reduction in synaptically driven firing frequency in percent (I; **p < 0.01, Wilcoxon signed rank test for both I and H). J–L, Effects of CBZ on Schaffer collateral EPSCs (25 stimuli at 50 Hz) recorded in the presence of 10 μm gabazine. Stimulation was performed as in G, and EPSCs were recorded in the whole-cell configuration. J, Representative recordings under control conditions (top trace), in 30 μm CBZ (middle trace), and after washout (bottom trace). Traces show averages from 10 consecutive sweeps, and stimulus artifacts were truncated. K, Quantification of peak EPSC amplitude showed no effect of CBZ on excitatory Schaffer collateral input. L, Peak EPSCs after CBZ application normalized to the mean of control and washout. Gray traces correspond to individual cells; the black trace shows average ± SEM.

Figure 2.

Carbamazepine effects on intrinsic firing properties of different types of GABAergic interneurons. A, Effects of CBZ on a representative BC, a bistratified cell innervating the proximal dendrites of pyramidal cells (PD), and an OLM interneuron (top, reconstructions of representative neurons; middle, current-clamp recordings during 1000 ms current injections; bottom, corresponding input–output relationship of the average firing rate during the 1 s current injection vs the magnitude of the current injection). SLM, Stratum lacunosum moleculare; SO, Stratum oriens; SP, Stratum pyramidale; SR, Stratum radiatum; Alv, alveus. B, CBZ effect on maximal firing rates of BCs (left), proximal dendritic targeting cells (PD; middle), and distal dendritic interneurons (OLM; right; *p < 0.05) during 1 s current injections. C, CBZ effects for different current injection durations measured from the onset of the current injection. D, Percent reduction in the maximal firing rate for the different time intervals from C.

Figure 3.

Comparison of the CBZ effects on intrinsic firing of different types of GABAergic interneurons. At the longest current injection durations (500 and 1000 ms), PD interneurons are affected significantly more compared with other types of neurons. *p < 0.05, ANOVA with post hoc Tukey's test.

Figure 4.

Carbamazepine effects on synaptically induced firing of interneurons. A, Left, Recording configuration to examine the impact of CBZ on synaptically induced firing of interneurons. Right, Cell-attached recordings of a pyramidal cell (PC), a basket cell (BC), a proximal dendritic targeting cell (PD), and an OLM interneuron. B, The latency of antidromic spikes in PCs is significantly shorter than those observed in interneurons as estimated with an ANOVA and a subsequent post hoc Tukey's test. *p < 0.05; **p < 0.01; ***p < 0.001. C, Example recordings of EPSPs from different interneuron types during alveus stimulation. EPSP amplitudes of a basket cell (top trace) and a proximal dendritic cell (middle trace) decrease during a 50 Hz stimulus train, whereas an OLM interneuron (bottom trace) shows facilitation. D, EPSP amplitudes from interneurons in the BC and PD groups that show depression were normalized to the first EPSP; data from OLM interneurons that show facilitation were normalized to the mean amplitude of the last three EPSPs. The results clearly show the difference in EPSP dynamics onto OLM interneurons versus other types of interneurons. E, Synaptically induced firing of interneurons is not affected by CBZ during a 200 ms stimulus train at 50 Hz (top: raster plots of action potential firing, 50 Hz stimulus train; bottom: average firing probability during trains of 10 stimuli at 50 Hz). F, Synaptically induced firing of interneurons by stimulation trains composed of five stimuli at 100 Hz. Insets depict the total number of action potentials induced during stimulus trains. G, Raster plots of synaptically induced firing of OLM cells with longer stimulation trains of 500 ms at 50 Hz under control conditions, during application of CBZ, and after washout. H, Left, Average firing frequency binned over 100 ms periods of stimulation. The inset shows effects on the number of action potentials over the whole stimulus train. Right, Percent reduction of firing for the different time intervals. These experiments reveal a minor effect on synaptically driven OLM neuron firing (*p < 0.05). Stim, Time points of alveus stimulation.

Figure 5.

Firing of CA1 neurons during alveus stimulation is undisturbed by CBZ. A, B, Representative cell-attached recording (A) and corresponding raster plots (B) of antidromically elicited action potentials in a CA1 pyramidal cell (25 stimuli at 50 Hz). Stimulation (Stim) was performed as in Figure 4A, in the presence of 10 μm gabazine. C, Average firing frequency during direct stimulation of pyramidal cell axons remains unaffected by CBZ.

Figure 6.

Carbamazepine effects on GABAergic feedback inhibition. A, Recording configuration used to elicit feedback inhibition. B, C, Representative recordings of feedback PSCs with sequential application of 30 μm CBZ, washout of CBZ, and application of 10 μm gabazine (leftmost traces). Feedback IPSCs were isolated by subtracting traces in the presence of gabazine from all other traces (rightmost traces; B, 50 Hz stimulation; C, 100 Hz stimulation). D, Quantification of the CBZ effect on peak IPSCs during 50 and 100 Hz stimulation. Top, Averaged IPSC amplitudes under control conditions, in the presence of CBZ and after washout. Bottom, Peak IPSCs after CBZ application normalized to the mean of control and washout. Gray traces correspond to individual cells; the black trace shows average ± SEM.

Figure 7.

Carbamazepine effects on GABAergic feedforward inhibition. A, Recording configuration to elicit feedforward inhibition. B, C, Representative recordings of feedforward IPSCs under control conditions, after application and washout of CBZ, and after the subsequent application of 10 μm gabazine during 50 Hz (B) and 100 Hz (C) stimulus trains. D, Top, Quantification of the peak IPSCs. Bottom, Peak IPSCs after CBZ application normalized to the mean of control and washout. Gray traces correspond to individual cells; the black trace shows average ± SEM.

Figure 8.

Voltage-dependent dye imaging. A, Stimulation paradigm for voltage imaging (top trace). The raw trace of the fluorescent signal (black) shows the first and second inhibitory events with the interspersed high-frequency stimulation train. Double-exponential fits were used for bleaching correction; these are superimposed in orange on the raw fluorescent traces. Right, Magnification of processed signal (% ΔF/F, with bleaching correction). Arrows indicate peak inhibition (positive peak) and peak excitation (negative peak). B, C, Effects of CBZ on feedback (B) and feedforward (C) inhibition measured using voltage imaging. Pseudo color images indicate inhibitory voltage signals across the CA1 subfield (color scale is from −1 to 0.04% ΔF/F). The traces correspond to the voltage signals (calibration: 0.01% ΔF/F, 50 ms) obtained in stratum pyramidale (sp; red trace), stratum oriens/radiatum (so/sr; blue trace), and stratum lacunosum moleculare (slm; green trace). D, Averaged peak inhibitory voltage signals for feedback inhibition (n = 10) and feedforward inhibition (n = 9) within the different laminae. Top, first inhibitory event; bottom, second inhibitory event. E, Averaged data for the first excitatory voltage signal and the second excitatory voltage signal evoked by alveus stimulation (feedback, left, n = 10) and CA3 stimulation (feedforward, right, n = 9).

Figure 9.

Effects of additional anticonvulsants on pyramidal cell firing and GABAergic inhibition. A, Representative recordings before, during application, and after washout of either LTG (25 μm; left) or PHT (50 μm; right). B, LTG (left; n = 7) and PHT (right; n = 6) reduce the maximal average firing rate of pyramidal cells by about 25.2 and 48.1%, respectively. *p < 0.05. C, D, Effects of LTG (C) and PHT (D) on feedback inhibition measured as in Figure 6, A–C. E, F, Quantification of the peak IPSCs for 50 Hz (E) and 100 Hz (F) stimulation under control conditions, in the presence of CBZ and after washout. Bottom, Peak IPSCs after CBZ application normalized to the mean of control and washout. Gray traces correspond to individual cells; the black trace shows average ± SEM.

Figure 10.

Effects of CBZ on pyramidal neurons in the chronically epileptic hippocampus. A, Effects of CBZ on the maximal discharge frequency measured during 1 s current injections in pilocarpine-treated rats (**p < 0.01). B, CBZ effects on the maximal firing rate measured during the first 50, 100, 200, 500, and 1000 ms of current duration (see Fig. 1; **p < 0.01, Wilcoxon signed rank test). C, Percent reduction in firing frequency by CBZ within the different time intervals shown in B (*p < 0.05, **p < 0.01, Wilcoxon signed rank test). For comparison, data from control animals (see Fig. 1) are indicated in gray.

Figure 11.

Effects of CBZ on feedback inhibition in the chronically epileptic hippocampus. A, Recording configuration used to elicit feedback inhibition. B, C, Representative recordings of feedback PSCs with sequential application of 30 μm CBZ, washout of CBZ, and application of 10 μm gabazine (leftmost traces). Feedback IPSCs were isolated by subtracting traces recorded in the presence of gabazine (rightmost traces; B, 50 Hz stimulation; C, 100 Hz stimulation). D, Quantification of the CBZ effect on peak IPSCs during 50 and 100 Hz stimulation. Bottom, Peak IPSCs after CBZ application normalized to the mean of control and washout. Gray traces correspond to individual cells; the black trace shows average ± SEM.

Figure 12.

Effects of CBZ on feedback inhibition recruited by activity patterns generated from in vivo juxtacellular recordings. A, In vivo firing behavior of CA1 pyramidal cells in a representative control (left) and an epileptic (right) animal during HFOs in control (ripples) and kainate-treated (fast ripples) rats. Top, Averaged local field potential (LFP) recordings of HFOs during which single-cell spike patterns were obtained. HFOs were aligned to the HFO peaks before averaging (see Materials and Methods). Middle, Raster plot of action potential firing of a representative pyramidal cell ±1 s relative to the HFO peak. To generate stimulus patterns for in vitro experiments, only sweeps with more than two action potentials were selected (red dots). Sweeps in which an HFO event occurred within 1 s after a previous HFO event were also not included into the stimulus pattern (corresponding action potentials are marked in gray). Bottom, Histogram of all spikes relative to the HFO peak. Bin size, 10 ms. Red lines indicate distribution of action potentials selected for in vitro stimulus patterns. B, In vitro recordings of feedback IPSCs obtained with the selected stimulus patterns based on in vivo juxtacellular recordings (A, red dots). The traces shown are derived from averaging of individual current traces obtained by stimulation with the different firing patterns shown in A. Examples are shown for a cell from a sham-injected (top traces) and a cell from a chronically epileptic pilocarpine-treated (bottom traces) animal. Left, Stimulation was performed with the control stimulus patterns. Right, A recording obtained with the epileptic stimulus patterns. For all panels, averaged traces in control conditions (black trace), in the presence of CBZ (red trace), and after washout (gray trace) are shown. Stimulus artifacts were cut out. Stimulation was performed as in Figure 6. C, Quantification of the charge transfer during control (left) and epileptic (right) stimulus patterns. The inhibitory response during the HFO (B, dashed black box) and during the baseline surrounding the HFO was analyzed separately. D, Ratio of the charge transfer after application of CBZ relative to control conditions (mean of ACSF and washout). *p < 0.05, significant change relative to mean of ACSF and washout, Wilcoxon signed rank test.