Abstract
Absence seizures are characterized by impairment of consciousness associated with bilaterally synchronous spike-and-wave discharges (SWDs) in the electroencephalogram (EEG), which reflect paroxysmal oscillations in thalamocortical networks. Although recent studies suggest that the subthalamic nucleus (STN) provides an endogenous control system that influences the occurrence of absence seizures, the mechanisms of propagation of cortical epileptic discharges in the STN have never been explored. The present study provides the first description of the electrophysiological activity in the cortico-subthalamo-pallidal network during absence seizures in the genetic absence epilepsy rats from Strasbourg, a well established model of absence epilepsy. In corticosubthalamic neurons, the SWDs were associated with repetitive suprathreshold depolarizations correlated with EEG spikes. These cortical paroxysms were reflected in the STN by synchronized, rhythmic, high-frequency bursts of action potentials. Intracellular recordings revealed that the intraburst pattern in STN neurons was sculpted by an early depolarizing synaptic potential, followed by a short hyperpolarization and a rebound of excitation. The rhythmic hyperpolarizations in STN neurons during SWDs likely originate from a subpopulation of pallidal neurons exhibiting rhythmic bursting temporally correlated with the EEG spikes. The repetitive discharges in STN neurons accompanying absence seizures might convey powerful excitation to basal ganglia output nuclei and, consequently, may participate in the control of thalamocortical SWDs.
Introduction
Absence epilepsy is a generalized, nonconvulsive epilepsy of multifactorial genetic origin (Panayiotopoulos, 1997; Crunelli and Leresche, 2002). Absence seizures consist of a brief, sudden impairment of consciousness concomitant with bilateral synchronized spike-and-wave discharges (SWDs) in the electroencephalogram (EEG) over wide cortical areas (Panayiotopoulos, 1997). Electrophysiological recordings in patients (Williams, 1953) and animal models of absence epilepsy (for review, see Danober et al., 1998; Crunelli and Leresche, 2002; Timofeev and Steriade, 2004) show that SWDs result from abnormal synchronized oscillations in thalamocortical networks. A number of studies performed in genetic absence epilepsy rats from Strasbourg (GAERS), a well established genetic model of absence epilepsy (Marescaux et al., 1992; Danober et al., 1998), suggest that the basal ganglia provides an endogenous control system for absence seizures via a modulation of the glutamatergic subthalamonigral pathway (for review, see Danober et al., 1998; Deransart et al., 1998, 2002). Specifically, alteration of the excitatory synaptic influence of the subthalamic nucleus (STN) on the substantia nigra pars reticulata (SNr), either by bilateral injection of a GABA agonist in the STN or a blockade of nigral glutamatergic NMDA receptors, has been shown to suppress absence seizures (Deransart et al., 1996). Furthermore, either bilateral excitotoxic lesions or highfrequency stimulation of the STN dramatically affects the occurrence of cortical paroxysms in GAERS (Vercueil et al., 1998). Consistent with the antiepileptic effect of an interruption of the excitatory subthalamonigral transmission, intranigral injection of GABAA receptor antagonists increases the occurrence of SWDs in GAERS (Danober et al., 1998; Deransart et al., 1998). Together, these findings suggest that the occurrence of absence seizures is modulated by changes in the activity of subthalamonigral neurons.
STN neurons integrate powerful monosynaptic glutamatergic inputs originating from corticosubthalamic (CSth) neurons and relay these excitatory signals to the SNr and the globus pallidus (GP) (Deniau et al., 1978; Van Der Kooy and Hattori, 1980; Kitai and Deniau, 1981; Afsharpour, 1985; Rouzaire-Dubois and Scarnati, 1987; Maurice et al., 1998; Magill et al., 2000). GABAergic neurons of the GP shape the activity of STN neurons by a disinhibitory mechanism involving corticostriatal and striatopallidal pathways (Maurice et al., 1998; Bevan et al., 2002b) and via reciprocal connections with the STN (Magill et al., 2000; Bevan et al., 2002b). Recent studies indicate that rhythmic activity in the cerebral cortex can promote oscillations in STN and GP, the phase relationships of which are controlled by the reciprocal excitatory and inhibitory connections between these two nuclei (Plenz and Kitai, 1999; Magill et al., 2000, 2001; Bevan et al., 2002b).
The effect of cortical SWDs on the activity of STN neurons remains unknown, and the electrical events occurring in corticosubthalamo-pallidal networks during absence seizures have never been studied. Because these questions are crucial to our understanding of how the basal ganglia influence absence seizures, we examined in vivo in GAERS the intracellular and/or extracellular activity of CSth neurons, STN neurons, and GP neurons simultaneously with the spontaneous SWDs occurring in the EEG of the functionally related cortical region.
Materials and Methods
All experiments were performed in accordance with local Ethical Committee and European Union guidelines (directive 86/609/European Economic Community), and every precaution was taken to minimize stress and the number of animals used in each series of experiments.
Animal preparation. Experiments were performed in vivo on 24 adult (3-12 months of age) rats from the GAERS strain. Animals were anesthetized initially with sodium pentobarbital (40 mg/kg, i.p.; Sanofi, Libourne, France) and ketamine (100 mg/kg, i.m.; Imalgène, Rhone Mérieux, France). A cannula was inserted into the trachea, and the animal was placed in a stereotaxic frame. Wounds and pressure points were (every 2 h) infiltrated repeatedly with lignocaine (2%). Once the surgical procedures had been completed (see below), ear bars were removed, and the head was held via a metallic rod cemented to the skull. Rats were maintained subsequently in a narcotized and sedated state by injections of fentanyl (4 μg/kg, i.p.; Janssen-Cilag, Issy-Les-Moulineaux, France), repeated every 20-30 min (Simons and Carvell, 1989; Pinault et al., 1998; Charpier et al., 1999; Slaght et al., 2002a, 2004; Bruno et al., 2003). To obtain long-lasting stable intracellular recordings, rats were immobilized with gallamine triethiodide (40 mg, i.m.; every 2 h; Specia, Paris, France) and artificially ventilated. The degree of anesthesia was assessed by continuously monitoring the EEG and heart rate, and additional doses of fentanyl were administered at the slightest change toward a waking pattern (i.e., an increase in the frequency and reduction in amplitude of EEG waves and/or an increase in the heart rate). Body temperature was maintained (36.5-37.5°C) with a homeothermic blanket. At the end of the experiments, animals received an overdose of sodium pentobarbital (200 mg/kg, i.p.).
Electrophysiological recordings. EEG recordings were obtained with a low-impedance (≈60 KΩ) silver electrode placed on the dura above the orofacial motor cortex (12.5 mm anterior to the interaural line; 3.5-4 mm lateral to the midline) (Hall and Lindholm, 1974; Neafsey et al., 1986; Paxinos and Watson, 1986), and the reference electrode was placed in the muscle on the opposite side of the head.
Intracellular recordings were performed using glass micropipettes filled with 2 m potassium acetate (40-70 MΩ). Measurements of apparent membrane input resistance and time constant were based on the linear electrical cable theory applied to an idealized isopotential neuron (Rall, 1969). The voltagecurrent (V-I) relationship was measured from variations of the membrane potential in response to intracellular injections of hyperpolarizing current pulses (-0.2 to -1.2 nA; 100 ms duration; every 1.55 s; n ≥ 6) applied through the recording electrode. Apparent input resistance was assessed by measurement of the slope of the linear portion of the V-I curve. The membrane time constant, assessed from -0.4 nA current pulses, was the time taken for the membrane potential to reach 63% of its final value. The values of membrane potential were corrected according to the tip potential recorded extracellularly immediately after termination of the intracellular recording.
For single-unit extracellular recordings and juxtacellular labeling (see below), glass electrodes were filled with 0.5 m NaCl and 1.7% neurobiotin (15-20 MΩ; Vector Laboratories, Burlingame, CA).
Cortical cells, located in the orofacial motor cortex, were recorded within 300 μm of the EEG electrode at the following coordinates: 12.5 mm anterior to the interaural line, 3-3.8 mm lateral to the midline, and 1491-2025 μm below the cortical surface. They were electrophysiologically identified as CSth neurons by their antidromic activation by electrical stimulation within the ipsilateral STN (5.1-5.4 mm anterior to the interaural line, 2.5-2.7 mm lateral to the midline, and 7.3-7.6 mm ventral to the brain surface). STN stimuli used to test antidromic activation (200 μs duration; 2-27 V) were applied with a bipolar concentric electrode (NE-100; Rhodes Medical Instruments, Woodland Hills, CA). To ensure correct positioning of the stimulating electrode in the STN, multiunit responses to whiskers stimulation were recorded through the stimulating electrode (Maurice et al., 2003). The criteria used to identify antidromic action potentials were: (1) a constant latency of antidromic responses despite imposed changes of membrane potential, (2) collision of antidromic spikes with spontaneously occurring orthodromic action potentials, and (3) the all-or-none property of the evoked spikes when the stimulation was just below threshold for antidromic activation.
Intracellular, single-unit and multiunit extracellular recordings in the STN were obtained from the subthalamic projection field of the orofacial motor cortex (Kolomiets et al., 2001). The corresponding stereotaxic coordinates were 4.9-5.4 mm anterior to the interaural line, 2.4-2.8 mm lateral to the midline, and 7042-7962 μm ventral to the brain surface. Extracellular recordings from GP neurons were obtained at the following coordinates: 7.6-8.1 mm anterior to the interaural line, 3-3.7 mm lateral to the midline, and 5025-6724 μm ventral to the brain surface.
In all experiments, the intracellular or single-unit and multiunit extracellular recordings were simultaneously performed with the corresponding ipsilateral cortical EEG.
Morphological identification. Extracellularly recorded neurons were labeled by juxtacellular injection of neurobiotin (Pinault, 1996; Mailly et al., 2003). Briefly, positive current pulses (1-8 nA; 200 ms) were applied at a frequency of 2.5 Hz through the bridge circuit of the amplifier. The current was slowly increased while the electrode was advanced toward the neuron in 1 μm steps (LSS-1000 Inchworm Motor Positioning System; Burleigh Instruments, Fishers, NY) until the cell discharge was driven by the injected current (see Fig. 3A2, inset). Current pulses were applied for a 10-15 min period to obtain a reliable labeling of neuronal processes. For intracellular recordings, depolarizing current pulses (0.2-1 nA; 100-200 ms) were applied at a frequency of 2.5 Hz at the end of the recording period. The histochemical methods used to reveal the morphology of neurobiotin-filled neurons were described in detail previously (Slaght et al., 2002a). The position of labeled neurons within the structures was confirmed using the atlas of Paxinos and Watson (1986).
Three-dimensional neuronal reconstruction. Soma, dendrites, and axons of some STN neurons were precisely drawn under 25-63× oil immersion objectives and plotted in three dimensions using the video computer Neurolucida system (MicroBrightField, Colchester, VT). Three-dimensional (3-D) models of neurons were visualized using Lightwave software (Newtek, San Antonio, TX) as described previously by Mailly et al. (2001). 3-D reconstructed models of neurons could be rotated around the x-, y-, and z-axes, and the sources of light and camera were adjusted to enhance the 3-D appearance of reconstructed neurons in two-dimensional images (see Fig. 3A2).
Data acquisition and analysis. Intracellular records were obtained under current-clamp conditions using the active bridge mode of an Axoclamp-2B amplifier (Axon Instruments, Union City, CA). Data were stored on-line on a DRA 800 digital tape recorder (Biologic, Claix, France) and then digitized with a sampling rate of 20 kHz (intracellular signals), 10 kHz (extracellular signals), or 300 Hz (EEG) for off-line analysis. To perform spectral analysis of EEG potentials, fast Fourier transforms were applied using Spike 2 (CED Software; Cambridge Electronic Design, Cambridge, UK). Cross-correlograms of the firing between two simultaneously recorded units were obtained by first encoding the position of the peak of the action potentials into separate channels using the memory buffer function of Spike 2; the event correlation function of Spike 2 was then used to produce the cross-correlogram. The amplitude of action potentials was calculated as the potential difference between their voltage threshold, measured as the membrane potential at which the dV/dt exceeded 10 V·s-1 (Mahon et al., 2003) and the peak of the waveform. Numerical values are given as means ± SEM unless stated otherwise. Statistical significance was assessed by performing a one-way ANOVA. In some measurements, the normality of distributions was tested using the Kolmogorov-Smirnov test, and a Gaussian-Laplace fit was performed. Statistical analysis and curve fitting were performed with Origin 7.0 (Microcal Software, Northampton, MA).
Results
Properties of SWDs
The SWDs (n = 1303) recorded in the present study had a mean duration of 5.7 ± 0.2 s (from 0.8 to 129.5 s) and recurred once every 22.6 ± 1.2 s. The intra-SWD frequency, revealed by spectral analysis of the EEG signal, ranged from 7 to 9.7 Hz (7.9 ± 0.1 Hz) (Fig. 1D). These temporal properties of SWDs, as well as the shape of individual spike-wave complexes, are similar to those described previously under analogous experimental conditions (Pinault et al., 1998; Charpier et al., 1999; Slaght et al., 2002a,b, 2004; Pinault, 2003) and in freely moving GAERS (Marescaux et al., 1992; Deransart et al., 2003).
Intracellular activity of CSth neurons
We recorded from CSth neurons (n = 7 from six GAERS) located in the orofacial motor cortex. The depth of intracellular recordings, between 1491 and 2025 μm from the cortical surface, suggested that CSth neurons were located in the deep part of layer V (Hall and Lindholm, 1974), a laminar localization that is consistent with previous findings obtained from anterograde and retrograde labeling of rat CSth neurons (Orieux et al., 2002). CSth neurons were identified electrophysiologically by antidromic activation after electrical stimulation of the ipsilateral STN (Fig. 1A). Antidromic latencies ranged between 1.1 and 3 ms (1.6 ± 0.2 ms; n = 7 cells) (Fig. 1A), consistent with the fast (2-4 ms) orthodromic activation of STN neurons in response to motor cortex stimulation (Kitai and Deniau, 1981; Kolomiets et al., 2001). The passive membrane properties of CSth neurons, measured during interictal periods, included a membrane potential of -61.3 ± 0.7 (from -64 to -59.2 mV; n = 7 cells) (Figs. 1B1,C, 2A1,A2) and a membrane time constant of 13 ± 0.8 ms (range, 9.2-15.8 ms; n = 7 cells) (Fig. 1B1). The apparent input resistance, measured from the slope of the linear V-I curve (Fig. 1B2), was 18.5 ± 1.2 MΩ (from 15 to 24.1 MΩ; n = 7 cells). In five CSth neurons, the firing pattern evoked by intracellular injection of suprathreshold current pulses was characteristic of “regular-spiking” neocortical neurons (Fig. 1B1) (Connors and Gutnick, 1990; Steriade, 2004). The two remaining cells exhibited the firing mode of the “intrinsic-bursting” neocortical neurons. The firing frequency-current (F-I) relationship, which followed a sigmoidal function (Fig. 1B2), indicated that CSth neurons reached a maximal firing rate of 78.3 ± 20.4 Hz (SD, from 50 to 100 Hz; n = 6 cells) in response to current pulses of +0.8 nA. In most CSth neurons (six of seven cells), large-amplitude hyperpolarizing current pulses induced a depolarizing “sag” of membrane potential (Fig. 1B1, arrowhead), likely caused by a hyperpolarization-activated inward cationic current (Ih). The negative current pulse was followed immediately by a postinhibitory rebound of depolarization (Fig. 1B1, arrow), possibly caused by the slow kinetics of Ih and/or by a low voltage-activated calcium potential.
Five recordings from CSth neurons (of seven) permitted observations on spontaneous transitions between interictal and ictal activities. During SWDs, CSth neurons exhibited suprathreshold rhythmic depolarizations (Figs. 1C, 2B1), which were temporally correlated with spike-wave complexes (Fig. 2B1,C). These repetitive membrane depolarizations were superimposed on a tonic hyperpolarization of 11.3 ± 0.1 mV in amplitude (range, 8.7-15.5 mV; n = 103 from five cells) that lasted for the entire SWD (Fig. 1C, dashed line). Spontaneous membrane potential fluctuations in CSth neurons followed a unimodal distribution during interictal periods (Fig. 2A2) and changed to a bimodal distribution, with sustained depolarized and hyperpolarized potentials, during ictal epochs (Fig. 2B2). Between SWDs, the mean value of membrane potential was -60 ± 4.55 (SD, from -64 to -59.2 mV; n = 5 cells) (Fig. 2A2). During cortical paroxysms, the average potential value reached during the rhythmic depolarizations and hyperpolarizations was -55.7 ± 1.4 mV (SD, from -58 to -54.4 mV; n = 5 cells) and -67.9 ± 2.6 mV (SD, from -71 to -64.8 mV; n = 5 cells), respectively (Fig. 2B2). The low occurrence of intermediate potentials demonstrates the rapid transitions made between the two preferred levels of membrane polarization (Fig. 2B1,B2). The appearance of ictal activity was also associated with a marked change in the firing pattern of CSth neurons, from an irregular activity to a rhythmic bursting mode (Figs. 1C, 2, compare A1, B1). Moreover, the mean spontaneous firing rate in CSth neurons increased by 92% during the SWDs compared with the interictal periods (interictal, 5 ± 1.2 Hz vs SWDs, 9.2 ± 0.3 Hz; n = 5 cells; p < 0.005), corresponding to a probability of firing in association with the EEG spike of 0.76 (from 0.66 to 0.91; n = 4311 EEG spikes from 227 SWDs; n = 5 cells).
To assess the temporal relationship between the firing of CSth neurons and the surface cortical paroxysms, we measured the timing of individual action potentials in CSth neurons relative to the peak negativity of the corresponding spike component in the EEG (Fig. 2C1). The latencies of the first action potential were distributed bimodally with two preferential values close to -40 and -12 ms (n = 3171 action potentials from five cells) (Fig. 2C2). A similar analysis using all action potentials (from one to six action potentials per depolarization cycle) also led to a bimodal distribution with two principal values at -38 and -10 ms (n = 4482 action potentials from five cells) (Fig. 2C2).
Extracellular recordings of STN neurons
Changes in the electrical activity of STN neurons during SWDs were first examined by extracellular recordings (n = 30 cells from six GAERS animals). Recorded cells were morphologically identified by juxtacellular injection of neurobiotin (see Materials and Methods) (Fig. 3A2). Labeled cells were located in the subthalamic projection field of the orofacial motor cortex (Kolomiets et al., 2001) and exhibited morphological characteristics of STN neurons, with a soma of diameter ∼20 μm, four to five primary dendrites, and axonal projections to the GP and the SNr (Beurrier et al., 1999; Bevan and Wilson, 1999) (Fig. 3A2) (see also Fig. 6A).
Extracellularly recorded STN neurons showed various patterns of background firing (Fig. 3A1,B), characterized by either low-frequency single action potentials (Fig. 3A1) or a sustained irregular firing (Fig. 3B). Pooling all cells, the mean firing rate during interictal periods was 12.6 ± 1.6 Hz (from 0.3 to 34.4 Hz; n = 30 cells) (Fig. 3C). During SWDs, the firing frequency of STN neurons reached a mean value of 21.8 ± 0.6 Hz (range, 3-48 Hz; from 292 SWDs; n = 30 cells) (Fig. 3C), corresponding to a probability of discharge in association with individual EEG spikes of 0.88 ± 0.02 (from 0.68 to 1; n = 30 cells). We further characterized the SWDs-induced changes in the STN firing rate by measuring, for each cell, the increase in firing frequency as a function of its interictal value. A clear inverse relationship was found with the percentage increase decaying exponentially as the function of the mean interictal firing frequency (Fig. 3D). The paroxysmal activity resulted in a profound change in the firing pattern of STN neurons characterized by the occurrence of high-frequency bursts of action potentials, synchronous with the spike component of the EEG (Figs. 3A1,B, 4A-D, 5A,B3). STN bursts, which included 2-14 action potentials (4.11 ± 0.02 action potentials; n = 10,510 bursts from 441 SWDs; n = 30 cells), often consisted of one or two action potentials followed by a short electrical silence and a high-frequency cluster of spikes (Fig. 4C,D). The burst duration was bimodally distributed around 10 and 25 ms (Fig. 4E), with a mean value of 16.8 ± 0.1 ms (range, 2.1-75.6 ms; n = 10,510 bursts from 441 SWDs; n = 30 cells) (Fig. 4E). The firing frequency during STN bursts was 206.8 ± 0.9 Hz (range, 58-670 Hz; n = 10,510 bursts from 441 SWDs; n = 30 cells).
As observed in CSth neurons, the latency of the first action potential in the STN burst was distributed bimodally (Fig. 4F) with respect to the peak negativity of the corresponding EEG spike, with two principal values at -18.7 and -1.9 ms (n = 15643 action potentials from 441 SWDs; n = 30 cells). A similar analysis of all the action potentials in a burst also led to a bimodal distribution with two peak values of -18.5 and +2.5 ms (n = 49,259 action potentials from 441 SWDs; n = 30 cells) (Fig. 4G).
Multiunit recordings (n = 5) revealed similar firing properties, both during and outside periods of SWDs (Fig. 5A). These recordings also highlighted how the unrelated firing patterns of two neighboring STN cells in the absence of seizures (Fig. 5B1,B3) became highly correlated with each other and tightly time-locked to the EEG spike during SWDs (Fig. 5B2,B4). These recordings also showed that the first action potential in a burst from one STN neuron could either precede or follow the first action potential in the concomitant burst of the neighboring cell, with a mean delay between these two action potentials of 2.3 ± 0.2 (from 0.3 to 37.9 ms; n = 2234 action potentials from 86 SWDs; five double unit recordings from two GAERS) (Fig. 5C).
Intracellular recordings of STN neurons: electrical membrane properties
Intracellular recordings were made from STN neurons (n = 9 cells from nine GAERS) to explore the membrane and synaptic mechanisms responsible for the generation of STN bursts during SWDs. The passive membrane properties of STN neurons included an interictal membrane potential of -59.3 ± 0.4 mV (n = 9), an apparent input resistance of 23.8 ± 3.1 MΩ (n = 6), and a membrane time constant of 5.31 ± 0.97 ms (n = 6) (Fig. 6B). Action potentials of intracellularly recorded STN neurons had an amplitude and duration of 47.5 ± 1.9 mV and 0.75 ± 0.07 ms (n = 9 cells), respectively, with a voltage threshold of -50.8 ± 0.5 mV (n = 9 cells). The V-I relationship, obtained from negative current pulses of amplitude -0.2 to -0.8 nA, was linear, indicating the absence of membrane rectification in the hyperpolarizing direction (data not shown). The location and the morphological features of these neurons (Fig. 6A) were similar to those described after juxtacellular dye injection (see Fig. 3A2), including a soma of 25 μm in diameter, four primary dendrites, and a bifurcated axon projecting to both SNr and GP, indicating the identity of extracellularly and intracellularly recorded STN neurons.
Although their spontaneous interictal firing was slow (<35 Hz), STN neurons were capable of firing at rates as high as several hundred Hertz during SWD-associated spontaneous bursting (see above) or in response to depolarizing current pulses (Fig. 6B,D). The current-driven discharge of STN neurons was characterized by a slight spike frequency adaptation and showed a very wide dynamic range, with peak firing rates as high as 500 Hz (Fig. 6C,D). Despite the passage of large depolarizing currents, the membrane potential between action potentials traversed approximately the same voltage range seen during spontaneous firing. Interestingly, high-frequency bursts similar to those discharged during SWDs (Figs. 4A-D) occurred as rebound responses after hyperpolarizing current pulses (Fig. 6B, gray trace).
These membrane properties of the GAERS STN neurons closely resemble those previously described in vitro from nonepileptic rats (Bevan and Wilson, 1999; Beurrier et al., 1999; Bevan et al., 2000; Hallworth et al., 2003). Thus, it is very unlikely that the epileptiform discharges of STN neurons during absence seizures were caused by a specific alteration in their intrinsic excitability (see Discussion).
Intracellular recordings of STN neurons: cellular events associated with SWDs
In five intracellularly recorded STN neurons, spontaneous transitions between interictal and ictal periods were characterized by a switch from single-spike activity to burst firing mode similar to that observed in extracellular records. The rhythmic bursting of STN neurons during SWDs was accompanied by a triphasic sequence of intracellular events (Fig. 7B): (1) an early short-duration depolarizing potential that could generate an action potential, (2) a transient, large membrane hyperpolarization that lasted 16.6 ± 0.1 ms (from 4.2 to 30.5 ms; n = 1801 from five cells), and (3) a postinhibitory rebound of depolarization leading to a burst with a relatively fixed latencies for the first three action potentials. The rhythmic hyperpolarizations were increased in amplitude under DC depolarization (Fig. 7C1) and reversed in polarity for membrane potentials less than -73 mV (Fig. 7C3). These observations strongly suggest that the rhythmic hyperpolarizations in STN neurons during SWDs represent Cl--dependant, GABAA-mediated, hyperpolarizing postsynaptic potentials (IPSPs) (see Discussion). These presumed IPSPs started 10.36 ± 0.13 ms (range, -40 to + 6.17 ms; n = 1801 IPSPs from five cells) before the spike component of SWDs and reached their peak potential at -2.73 ± 0.13 ms (range, -30 to +14 ms; n = 1801 IPSPs from five cells) (Figs. 7B,9B).
Extracellular recordings of GP neurons
We made extracellular recordings of GP neurons (n = 13 from five GAERS) to examine whether these neurons were involved in the rhythmic bursting of the STN during SWDs. We found two distinct GP cell populations, in term of firing pattern, without any evidence for a differential topographical distribution within the nucleus. The smaller group of GP neurons [called “early discharging” GP (ED-GP) neurons; n = 3 of 13 cells] was characterized by a slow irregular background firing that shifted, during SWDs, to rhythmic single spikes or short bursts (two to four action potentials) correlated with the EEG spike (Fig. 8A1,A2). This rhythmic firing was maintained throughout the cortical seizure. The mean latency of all action potentials to the peak negativity of the EEG spike was -18.67 ± 0.14 ms (n = 2409 action potentials from three cells) (Fig. 8A2). Simultaneous recordings (n = 2 from two GAERS) of ED-GP neurons with STN neurons showed that their random firing patterns in the absence of SWDs (Fig. 8C1,C2) became highly correlated with each other (Fig. 8D2) and tightly time-locked to the spike-wave complexes (Fig. 8D1) during seizures.
In contrast, the main class of GP neurons, called “pausing” GP (P-GP) neurons (n = 10 of 13 cells), displayed a tonic interictal activity. During SWDs, discharges were transformed into a burst firing pattern with interburst periods that coincided with cortical EEG spikes (Fig. 8B1,B2). These periods of electrical silence in P-GP neurons occurred 14.76 ± 0.21 ms before the EEG spike (n = 1260 “pauses” from 10 cells) (Fig. 8B2) and lasted for 32.1 ± 11.5 ms (SD, range, 15.7-49.2 ms; n = 1260 pauses from 10 cells). The recovery after the electrical silence was systematically associated with a transient increase in the firing rate (Figs. 8B2, inset, 9A, bottom trace).
Discussion
The present study provides the first description of electrical activities in the cortico-subthalamo-pallidal networks during absence seizures. Our principal findings are: (1) during SWDs, CSth neurons exhibit rhythmic suprathreshold depolarizations superimposed on a sustained hyperpolarization lasting for the entire cortical paroxysm, (2) STN neurons display synchronized repetitive bursts of action potentials in-phase with EEG spikes, (3) intracellular analysis indicates that rhythmic firing in STN neurons results from an early excitation, followed by an hyperpolarization and then a late excitation, and (4) during SWDs, GP neurons display either a rhythmic bursting synchronous with the cortical paroxysm or a sustained firing momentarily interrupted by EEG spikes. These results indicate that the propagation of cortical paroxysms through the cortico-subthalamo-pallidal pathway generates a bursting pattern in the STN that might produce a powerful phasic synaptic excitation of the basal ganglia output nuclei.
Rhythmic bursting of CSth neurons during SWDs
To our knowledge, the present experiments provide the first description of the intracellular activity of identified CSth neurons. The occurrence of SWDs in the EEG was concomitant with a sudden, pronounced modification in the activity of CSth neurons. The interictal, small-amplitude, irregular synaptic activity in CSth neurons, generating an erratic firing pattern, changed, during an SWD, into a step-like behavior with suprathreshold depolarizations in-phase with the EEG spikes. As previously observed in other GAERS cortical neurons (Charpier et al., 1999; Slaght et al., 2002a,b), the SWD-associated rhythmic depolarizations in CSth neurons were concomitant with a tonic membrane hyperpolarization, likely attributable to a synaptic disfacilitation responsible for an increase in membrane resistance (Contreras et al., 1996; Charpier et al., 1999; Slaght et al., 2002b).
Interestingly, the synaptic activity as well as the firing rate of CSth neurons during SWDs differ significantly from those of GAERS corticostriatal neurons (Slaght et al., 2002b, 2004), which exhibit oscillating membrane depolarizations that remain subthreshold or elicit single action potentials during seizures. The increased excitability of CSth neurons could result from interactions between synaptic inputs and specific voltage-gated intrinsic channels. Consistent with this hypothesis, we observed in CSth neurons, in response to hyperpolarizing current injection, a depolarizing sag of membrane potential followed by a postanodal rebound of excitation. These observations provide indirect evidence that CSth neurons may express Ih and a low-voltage activated calcium current (Fig. 1B1), which act together to promote oscillations and repetitive firing in cortical neurons (for review, see Reyes, 2001; Migliore and Shepherd, 2002).
Origin of the rhythmic excitation and inhibition in STN neurons during SWDs
The GAERS STN neurons, intracellularly recorded in the course of this study, displayed membrane properties including membrane potential, postanodal rebound of excitation, action potential properties, and V-I and F-I relationships similar to those previously described in vitro from normal rats (Beurrier et al., 1999, 2000; Bevan and Wilson, 1999; Bevan et al., 2000, 2002a). These important observations indicate that the increased firing rate in STN neurons accompanying the SWDs does not result from an altered intrinsic excitability.
The rhythmic bursts in STN neurons, observed in extracellular recordings during SWDs, were characterized frequently by an early action potential, usually preceding the corresponding EEG spike, followed by a short electrical silence (<25 ms) and then a high-frequency cluster of action potentials. This pattern of electrical events in the STN is very similar to that observed after electrical stimulation of the motor cortex (Kitai and Deniau, 1981; Kita, 1994; Kolomiets et al., 2001). Intracellular recordings let us characterize membrane and synaptic events underlying this triphasic response. The initial discharge of STN neurons during cortical paroxysms was elicited by a short depolarizing potential likely induced by the early discharge of CSth neurons (Fig. 9), which provide powerful excitatory synaptic inputs to STN neurons (Kitai and Deniau, 1981; Kita, 1994; Magill et al., 2000; Kolomiets et al., 2001). The brief silence within the STN bursts associated with the spike-wave complexes resulted from a transient (∼17 ms) hyperpolarization that reversed in polarity at membrane potentials less than -73 mV (Fig. 7C3). This value is very close to the chloride equilibrium potential of GABAA current measured in STN neurons from normal rats (Bevan et al., 2000, 2002a). Thus, rhythmic hyperpolarizations observed during SWDs are probably GABAA-mediated IPSPs that may be induced by a subpopulation of GABAergic pallidosubthalamic neurons (∼25% of the GP neurons described in this study), which generate bursts of action potentials just before (∼16 ms) the peak of the STN inhibitory potentials (Fig. 9, ED-GP neurons). The burst activity in ED-GP neurons may well be initiated, at least in part, by a short-latency excitation attributable to the early firing of excitatory subthalamopallidal neurons (Kita and Kitai, 1991; Magill et al., 2000; Nambu et al., 2000) (Fig. 9). The postinhibitory rebound of firing in STN neurons may result from synergistic interactions between active membrane properties and excitatory synaptic inputs. Indeed, STN neurons possess low-threshold calcium channels responsible, at least in part, for the generation of burst activity after membrane hyperpolarization (Beurrier et al., 1999, 2000; Bevan and Wilson, 1999; Bevan et al., 2000; Song et al., 2000). Such intrinsic rebound properties in STN neurons, which were also found in the present study on removal of large-amplitude hyperpolarizing current pulses (Fig. 6B), might contribute to the postinhibitory (late) bursting in STN neurons (Fig. 9). Moreover, in accordance with previous in vitro studies (Bevan et al., 2002b), we found that the repetitive action potentials that followed large-amplitude IPSPs impinging on STN neurons were almost exactly in-phase (Fig. 7B), an intrinsic process that would tend to promote coherent bursting in the STN during absence seizures (Figs. 5B2, 4). The delayed firing of CSth neurons would provide an additional excitatory synaptic mechanism promoting the late STN bursting during SWDs (Fig. 9).
The functional impact of P-GP neurons on STN activity remains unclear. However, we can propose that the transient interruption of P-GP neuron firing during EEG spikes, potentially caused by intra-GP inhibitory inputs (Kita and Kitai, 1994) from ED-GP neurons (Fig. 9), provides a permissive mechanism to reinforce the relative weight of the corticosubthalamic excitatory inputs. Moreover, the sustained desynchronized firing in P-GP neurons between EEG spikes could contribute to the interburst silence in the STN by increasing the STN neurons membrane conductance.
Pathophysiology
The main finding of the present study is the synchronized rhythmic bursting of STN neurons during absence seizures. This novel result, together with the lack of action potential firing in GAERS striatal projection neurons during ictal activity (Slaght et al., 2004), strongly suggests a change in the balance between excitation and inhibition in the basal ganglia output nuclei that leads to a sudden reinforcement of the synaptic excitation originating from the STN. This hypothesis is supported by recent findings from freely moving GAERS showing that SNr neurons exhibit a synchronized bursting concomitant with SWDs (Deransart and Depaulis, 2002; Deransart et al., 2003).
A potential implication of the STN in the control of absence seizures has recently emerged from pharmacological studies and deep brain stimulation in GAERS (Deransart et al., 1998; Vercueil et al., 1998), showing that neurotoxic lesions or high-frequency electrical stimulation of STN negatively modulate cortical paroxysms. These data are consistent with the antiepileptic effects of intranigral injections of GABAA agonists or NMDA antagonists in the GAERS SNr (Depaulis et al., 1988, 1989; Deransart et al., 1996, 1998, 2001) and the aggravation of absence seizures induced by application of GABAA antagonist in the SNr (Deransart et al., 1998). Changes in the output of STN neurons associated with cortical SWDs could, therefore, provide a subcortical mechanism participating in the control of the abnormal oscillations in the thalamocortical loop. First, the synchronized bursting of subthalamonigral neurons during SWDs, associated with the electrical silence in GABAergic striatonigral cells (Slaght et al., 2004), might prolong seizures via a repetitive excitation of SNr neurons (Deransart and Depaulis, 2002; Deransart et al., 2003). Such a positive modulation of SWDs could result from the repetitive inhibition of thalamocortical neurons by SNr (Di Chiara et al., 1979; Deniau and Chevalier, 1985). Indeed, if the nigral-induced thalamic rhythmicity is in-phase with SWDs, it could, by a resonance phenomenon, maintain the paroxysmal thalamocortical oscillations. Second, the recovery of the irregular, desynchronized, firing in STN neurons at the end of the SWD, together with the rebound of excitation in striatal projection neurons (Slaght et al., 2004), could initiate the postictal decrease in the activity of nigrothalamic neurons (Deransart et al., 2003) and so participate in the termination of the seizure.
Footnotes
This work was supported by the Ministère Français de la Recherche. We are grateful to Drs. S. Mahon and R. Miles for thoughtful discussion and critical reading of this manuscript. We also thank Dr. P. Mailly for neuronal 3-D reconstruction and A. Menetrey for assistance with the histological processing.
Correspondence should be addressed to J. T. Paz, Institut National de la Santé et de la Recherche Médicale U114, Chaire de Neuropharmacologie, Collège de France, 11 Place Marcelin Berthelot, 75231 Paris Cedex 05, France. E-mail: jeanne.paz{at}college-de-france.fr.
Copyright © 2005 Society for Neuroscience 0270-6474/05/252092-10$15.00/0