Fast activity at seizure onset is mediated by inhibitory circuits in the entorhinal cortex in vitro.код для вставкиСкачать
Fast Activity at Seizure Onset Is Mediated by Inhibitory Circuits in the Entorhinal Cortex In Vitro Vadym Gnatkovsky, MD, PhD, Laura Librizzi, PhD, Federica Trombin, PhD, and Marco de Curtis, MD Objective: Network mechanisms responsible for focal seizure initiation are still largely unknown. One of the prevalent seizure patterns observed during diagnostic intracranial recordings performed in patients with mesial temporal lobe epilepsy is characterized by fast activity at 20 to 30Hz. We reproduced 20 to 30Hz oscillations at seizure onset in the temporal lobe of the in vitro isolated guinea pig brain to study cellular and network mechanisms involved in its generation. Methods: Seizure-like activity was induced in the isolated brain by 3-minute arterial perfusion of 50M bicuculline. Intracellular, extracellular, and ion-selective electrophysiological recordings were performed simultaneously in the entorhinal cortex (EC) during interictal-ictal transition. Results: Principal neurons in deep and superficial layers of the EC did not generate action potentials during fast activity at ictal onset, whereas sustained firing was observed in putative interneurons. Within 5 to 10 seconds from seizure initiation, principal neurons generated a prominent firing that correlated with the appearance of extracellular hypersynchronous bursting discharges. In superficial neurons, fast activity correlated with rhythmic IPSPs that progressively decreased in amplitude during the development of a slow depolarization associated with an increase in extracellular potassium. Interpretation: We conclude that in an acute model of temporal lobe ictogenesis, sustained inhibition without firing of EC principal neurons correlates with the onset of a focal seizure. The progression of the ictal discharge is contributed by a potassiumdependent change in reversal potential of inhibitory postsynaptic potentials. These findings demonstrate a prominent role of inhibitory networks during the transition to seizure in the EC. Ann Neurol 2008;64:674 – 686 The cure of epilepsies depends on how effectively available treatments control seizure onset and propagation. In 30% of focal epilepsies, seizures are resistant to antiepileptic drugs. New strategies to treat pharmacoresistant epilepsies will benefit from the identification of the mechanisms involved in the transition from the interictal to the ictal state, which, despite many studies, remain elusive. The most common surface electroencephalographic correlate of seizure onset in human focal epilepsies originating from the temporal lobe is the occurrence of small-amplitude fast activity (electroencephalographic flattening) in the temporal region.1,2 Such activity may evolve into large-amplitude, rhythmic discharges that secondarily diffuse to adjacent cortical areas. Invasive presurgical studies with intracranial depth electrodes have been utilized to circumscribe the epileptogenic zone in patients suffering from mesial temporal lobe epilepsy resistant to pharmacological treatment.3,4 Such invasive diagnostic procedures dem- onstrated that the ictal discharge associated with a temporal lobe seizure most often initiates with a sequence of fast activity at 20 to 30Hz in the hippocampus and in the parahippocampal region.1,5–12 Even though alternative ictal onset patterns have been described,8,13,14 it has been proposed that fast 20 to 30Hz activity has the greatest localizing value for the identification of the epileptogenic region.6,11,15 Network and cellular mechanisms associated with the generation of fast, small-amplitude cortical activity in the beta-gamma range that initiates a seizure in the mesial temporal lobe are not yet clearly identified. Seizure-like fast ictal discharges in mesial temporal lobe structures can be experimentally reproduced in animal models of seizures.16 Ictogenesis can be reliably induced in the temporal lobe of the in vitro isolated guinea pig brain preparation17,18 by acute, partial, and transient disinhibition with the GABAA receptor antagonist bicuculline.19,20 This procedure was utilized to From the Unit of Experimental Epileptology and Neurophysiology, Fondazione Istituto Neurologico Carlo Besta, Milan, Italy. Potential conflict of interest: Nothing to report. Received Nov 21, 2007, and in revised form Jul 3. Accepted for publication Jul 8, 2008. Published in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.21519 674 Address correspondance to Dr de Curtis, Unit of Experimental Epileptology and Neurophysiology, Fondazione Istituto Neurologico Carlo Besta, via Celoria 11, 20133 Milano, Italy. E-mail: email@example.com © 2008 American Neurological Association Published by Wiley-Liss, Inc., through Wiley Subscription Services study network mechanisms that regulate the generation of 20 to 30Hz activity (also designated as beta-gamma activity) at seizure onset in the entorhinal cortex (EC) of the in vitro isolated guinea pig brain. We focused on the EC because this temporal lobe region is primarily involved in seizure generation in humans,9,10 and in experimental models of temporal lobe epilepsy and seizures.19,21–24 In mammals, the EC is reciprocally connected with the hippocampus proper.25–27 The network interactions within the EC and between this region and the hippocampus have been analyzed extensively.28 –32 Neurons in superficial EC layers II and III project to the dentate gyrus and to the CA1 field, and deep EC neurons receive a feedback projection from CA1/subiculum. A recently demonstrated direct feed-forward projection from these areas to superficial EC interneurons33,34 may play a crucial role in controlling EC excitability by preventing reentrant activation of the EChippocampal-EC circuit.34 Studies on temporal lobe ictogenesis and epileptogenesis identified enhanced network interactions and altered intrinsic neuronal properties in superficial and deep EC neurons.16,24,35,36 This study demonstrates for the first time that seizure onset in the EC is associated with a complete interruption of neuronal firing in principal neurons and is supported by the activation of an inhibitory network. Materials and Methods Guinea pig brains were isolated in vitro according to the previously described procedure.17,18 After administration of barbiturate anesthesia (80mg/kg sodium thiopental intraperitoneally), intracardiac perfusion with cold (15°C) saline solution (see later) was performed for 3 minutes to reduce brain temperature during dissection. The entire brain was isolated and transferred to a perfusion chamber. A poliethylene cannula was inserted in the basilar artery to restore brain perfusion with a solution composed of 126mM NaCl, 3mM KCl, 1.2mM KH2PO4, 1.3mM MgSO4, 2.4mM CaCl2, 26mM NaHCO3, 15mM glucose, and 3% dextran molecular weight 70,000, oxygenated with a 95% O2/5% CO2 gas mixture (pH 7.3). Experiments were performed at 32°C. The experimental protocol was reviewed and approved by the Committee on Animal Care and Use and by the Ethics Committee of the Istituto Nazionale Neurologico, in accordance with national and international guidelines on care and use of laboratory animals. Extracellular recordings were performed with glass pipettes filled with 0.9M NaCl (2–5M⍀ resistance). Intracellular recordings were performed with sharp electrodes filled with 3M potassium acetate and 2% biocytine (60 –120M⍀ input resistance). Electrophysiological signals were amplified via a multichannel differential amplifier (Biomedical Engineering, Thornwood, NY) and an intracellular amplifier (Neurodata, New York, NY). Data were acquired and analyzed utilizing software (ELPHO) that Dr Vadym Gnatkovsky developed in our laboratory. Joint time-frequency analysis was applied to show how the frequency content of the seizures evolves over Fig 1. (A, left) Schematic drawing of the positions of the recording electrodes in PC and in medial and lateral EC, and of the stimulating LOT electrode. (A, right) Pairing test in the PC: (a) PC response to a single stimulus, characterized by a monosynaptic (asterisk) and a disynaptic potential (solid circle); (b) PC response to a paired test with a 25-millisecond interstimulus interval; (b-a) subtracted traces enhance the inhibition of the disynaptic response. (B) Efficacy of inhibition measurement in sampled traced recorded in PC (left) and lateral EC (right). (C) Time course of the average inhibition efficacy values in the PC (left) and in the lateral EC (right) during the arterial perfusion of bicuculline (gray shaded area). Pairing tests were performed every 10 seconds. Seizure onset is marked by the arrows. Time values were normalized between experiments with reference to the time of seizure onset recorded in the medial EC. time (see Figs 2A, 2B, 3A). Cross-correlation coefficient was calculated in sequences of 100-millisecond time windows and represented as correlogram over time (see Fig 3B). Ictal discharges in the EC were induced by brief (3minute) arterial perfusions with 50M of the GABAA receptor antagonist bicuculline methiodide19,20,37,38 (SigmaAldrich, St. Louis, MO). Pairing of the responses evoked by stimulation of the lateral olfactory tract (LOT) in the piriform and entorhinal cortices was performed to evaluate the reduction of inhibition efficacy during perfusion with bicuculline. As discussed elsewhere,39 recurrent GABAergic inhibition is responsible for the depression of the disynaptic po- Gnatkovsky et al: Inhibition at Seizure Onset 675 In brief, ion-selective electrodes (tip diameter, 3–5m) were filled with the potassium ionophore I cocktail A (Fluka 60031, Hamburg, Germany). Absolute [K⫹]o values were calculated by solving the equation y ⫽ a ⫹ b logx, where x is the [K⫹]o, y is the measured voltage reading induced by the changes in [K⫹ ]o, and a ⫹ b is the slope coefficient derived from the calibration curve performed for each K⫹sensitive electrode (calibration solutions with K⫹ concentrations of 1, 2.5, 6, 12.5, and 48mM). Only electrodes with a response of 30 to 40mV for 10mM K⫹ were utilized. Ionselective signals were amplified with a high-input impedance head-stage amplifier (Biomedical Engineering, Thornwood, NY), and field potential values were subtracted. Principal neurons in layers II to III (n ⫽ 24), layers V to VI (n ⫽ 6), and putative interneurons located in the superficial 500m (n ⫽ 10) were identified from their response to the hippocampal input driven by stimulation of the olfactory area.41 Thirteen of 30 principal cells were further identified morphologically as stellate or pyramidal cells. At the end of the electrophysiological experiment, brains were fixed in 4% paraformaldheide, and the standard horseradish peroxidase protocol (ABC kit; Vector Laboratories, Burlingame, CA) was utilized to show neurons injected with biocytine. Sections were counterstained with thionine to identify cortical layers. Fig 2. Extracellular recording of a typical ictal discharge in the medial (EC). Ictal onset was characterized by fast activity (in the expansion below) consistently followed by irregular spiking (small stars) that progressively evolved into a regular bursting pattern. (inset) Postictal depression is shown. Average frequency content from 5-second recordings just before seizure and after the termination of rhythmic bursting (n ⫽ 7 experiments) show a marked reduction of the activity in all frequency ranges after the seizure (black shading) in comparison with preictal activity (gray shading). The fast activity at seizure onset is shown in the expanded trace. The relative spectrogram highlights the power of fast activity at 20-30 Hz (and subharmonics). (B) Averaged spectrogram of seizure onset (n ⫽ 7). tential in the conditioned response evoked within 30 milliseconds after the conditioning stimulus. The ratio between the amplitude of the disynaptic components of the potentials evoked by a conditioning and a conditioned LOT stimuli separated by 25 milliseconds was utilized to monitor changes in inhibition efficacy (see Results). Statistical T test was used to evaluate changes of inhibition efficacy. Recording of extracellular potassium concentration ([K⫹]o) in the EC was conducted as described previously.40 676 Annals of Neurology Vol 64 No 6 December 2008 Results Seizure-like Activity Correlates with Transient and Partial Entorhinal Cortex Disinhibition Seizure-like, ictal activity was reliably induced in the medial EC of the in vitro isolated guinea pig brain by a brief arterial perfusion with bicuculline methiodide (50M). Three-minute perfusions of bicuculline induced transient and partial disinhibition of the isolated brain, as demonstrated by paired-pulse test42 (Fig 1). The depression of the conditioned responses evoked by paired stimulation of the LOT were analyzed in the piriform cortex (PC) and in the lateral EC, a region that receives a direct olfactory input19 bordering the seizure onset area in the medial EC (see scheme in Fig 1A). In control solutions, pairing with a 25-millisecond interstimulus interval determined a 56.3 ⫾ 0.8 and 67.1 ⫾ 1.9% (mean ⫾ standard error of the mean; n ⫽ 10) reduction of the conditioned disynaptic responses recorded in both PC and lateral EC (see Figs 1A, B, filled circles). This parameter of inhibition efficacy was measured as the difference between the amplitude of the disynaptic response to a single stimulus (see Fig 1A, part a) and the amplitude of the subtracted disynaptic paired response (see Figs 1A, part b-a, and 1B; n ⫽ 12). Responses in which control/ conditioned monosynaptic potentials (see Figs 1A, B, asterisks) did not match in amplitude were excluded from the analysis. The pairing test was repeated every 10 seconds, and the time course of the changes in inhibition efficacy were evaluated during bicuculline arterial perfusion (see Fig 1C). Sample point values were syn- Fig 3. (A) Simultaneous extracellular and intracellular recordings from a principal neuron of the superficial layers of the medial EC (top trace) during the transition into seizure-like discharge. Recording segments outlined by boxes a, b, and c are expanded in (B). Resting membrane potential (rmp) was ⫺64mV. (B) At the onset of the seizure, the membrane potential of the neuron was depolarized by 5mV, to enhance the abrupt hyperpolarization that correlates with seizure onset (arrowhead). Cross-correlation coefficient for intracellular and extracellular recording was calculated in sequences of 100-millisecond time windows and represented as correlogram over time. Phases of the seizure identified as irregular firing (b) and regular bursting (c) are illustrated, respectively. (C) Power content of intracellular (solid line) and extracellular (dotted line) signals during the seizure shown in (A). High-pass filter was set at 10Hz. Bottom graph shows the average power content of intracellular (solid line) and extracellular (dotted line) fast activities recorded at seizure onset in 17 superficial layer neurons. chronized between experiments (n ⫽ 12) with reference to the seizure onset in the medial EC. At the time of seizure initiation (see Fig 1C, arrows), measurements of inhibition efficacy were reduced by 3.9 and 33.5% in the PC and lateral EC, respectively ( p ⬍ 0.02). We conclude from the pairing test that the efficacy of inhibition is only partially reduced in the isolated brain preparation when EC seizures start. An abrupt and more robust reduction of inhibition efficacy was observed during the bursting phase of the ictal discharge and after the sei- zure. In the lateral EC, paired-pulse inhibition reverted into excitation 30 seconds after seizure onset, and inhibition efficacy could not be measured after this time. The reduction of inhibition efficacy could still be evaluated in the PC, a brain region usually not involved in seizure discharge.20 Only experiments in which ictal events initiated with fast activity at 20 to 30Hz (27/32 tests) were selected for this study (Figs 2A, B). Two to 5 seconds after fast ictal onset, bursts of high-amplitude poten- Gnatkovsky et al: Inhibition at Seizure Onset 677 Fig 4. Reversal of inhibitory postsynaptic potentials (IPSPs) evoked by LOT stimulation (a; arrowhead) and fast activity (b) in superficial EC neurons. Simultaneous extracellular potentials are shown for each recording (extra). (A) The membrane of the neuron was hyperpolarized from resting potential (⫺63mV; dotted line) by injection of a steady negative current. The reversal potentials of the LOT-evoked IPSP (a) and the intracellular correlate of the fast IPSP activity (b) are illustrated; the arrow points at the intracellular correlate of the spike that initiates the seizure. (left) Two different traces recorded at ⫺47 and ⫺80mV are illustrated to show the reversal potential of the LOT-evoked responses. In a different neuron in (B), a positive current was injected to depolarize the membrane potential during the LOT-evoked response (a) and during the ictal-onset IPSPs (b; rmp ⫺61mV). Arrowhead in (a) mark the LOT stimulation. Arrow points to the onset of the seizure-like discharge. (C) Correlation between membrane potential depolarization observed during seizure onset and the amplitude changes of IPSPs in five experiments, identified by the different symbols. Linear regression lines were calculated and plotted for each experiment. The absolute changes in membrane potential depolarization (MP) were utilized to plot IPSP changes, regardless of the initial resting potential. (D) Average of extrapolated linear regression plots represent changes in IPSPs during membrane potential depolarization (n ⫽ 5, mean ⫾ SD). tials appeared and progressively increased with time in amplitude, regularity, and duration19,20 (see Fig 2A; see also Figs 3–5 and 8). Seizure-like events (mean duration ⫾ standard deviation, 8.4 ⫾ 1.4 minutes; n ⫽ 22) were typically followed by postictal depression, characterized by a decrease in the global activity content measured by signal frequency analysis (see Fig 2A, top). As demonstrated previously,19 interictal spikes could be observed ahead of an ictal discharge (see Figs 2A, 4, and 7). In this study, we focus on the characterization of specific networks that generated fast 20 to 30Hz activity by recording different types of neurons in the medial EC. 678 Annals of Neurology Vol 64 No 6 December 2008 Cellular Correlates of Fast Activity at Ictal Onset: Superfial Entorhinal Cortex Neurons Stellate and pyramidal neurons in the superficial layers II and III of the medial EC were characterized by electrophysiological and morphological features, as reported previously.34,41 In correlation with the appearance of fast activity at seizure onset, superficial layers principal cells did not generate action potential firing. When the neuron was depolarized above firing threshold before seizure onset, as shown in Figures 3A and B, action potential generation ceased and inhibitory rhythmic potentials appeared during the extracellular fast oscillation (see Fig 3B, thin arrows). The specto- gram in the lower part of Figure 3A shows the frequency content of the extracellular signal during the seizure. Small-amplitude intracellular potentials were correlated with extracellular fast activity, as demonstrated by the correlogram in the bottom of Figure 3B (part a). Mean correlation value calculated between pairs of intracellular and extracellular recordings during fast activity was 0.63 ⫾ 0.11 (n ⫽ 17). The bottom part of Figure 3C also shows the average frequency power of fast activity in superficial neurons (see Fig 3C, continuous line) and the corresponding extracellular signals (see Fig 3C, dotted line) (n ⫽ 17). During fast activity, membrane potential showed an abrupt hyperpolarization (see Fig 3B, part a, arrowhead) followed by a slow depolarization (see later). In coincidence with the fragmentation of fast activity, erratic firing resumed in superficial neurons (see Fig 3B, part b, thick arrow) in parallel with the emergence of irregular bursting in the extracellular trace (see Figs 3A, 3B, part b). Within a few seconds, bursts of action potentials superimposed on paroxysmal depolarizing shift appeared in coincidence with the occurrence of regular bursts in the extracellular trace (see Fig 3A). These bursts and their extracellular correlates became progressively more robust and less frequent (see Fig 3A and 3B, part c), and were followed by postictal depression. To better analyze the intracellular correlates of fast activity at seizure onset, we induced seizures when the membrane potential of neurons was hyperpolarized/depolarized by injection of steady current via the intracellular recording electrode. When the membrane potential of the recorded neuron was hyperpolarized, no spontaneous firing was observed (see Fig 4A, part b). In this condition, the fast rhythms observed at seizure onset correlated with a depolarizaion of the membrane potential and often triggered a field event (see Figs 4A, part b, and 4B, part b, arrows) similar to that observed during a preictal spike (see Figs 4A, part a, and 7). The amplitude of the intracellular fast activity progressively decreased, whereas membrane potential slowly depolarized (see Fig 4A, part b). In other experiments in which membrane potential was depolarized to values more positive than ⫺55mV, fast potentials at seizure onset were hyperpolarizing (see Fig 4B, part b); rebound spikes could be generated at the break of each fast hyperpolarizing potential. The polarity of the fast activity deflection matched the polarity of the inhibitory postsynaptic potential (IPSP) evoked by LOT stimulation34 recorded just before seizure onset (n ⫽ 20; see Figs 4A, part a, and 4B, part a); it was depolarizing for membrane potentials more negative than ⫺65mV (n ⫽ 5) and hyperpolarizing when the membrane potential was depolarized to values positive to ⫺60mV (n ⫽ 15). In all experiments, the ictal discharge was preceeded by interictal spikes (see Fig 4A, arrow). Preictal spikes correlated with a membrane po- tential deflection that showed the same reversal of LOT-evoked IPSPs (see Figs 4A, part b, and 7). Based on this evidence, we conclude that the intracellular correlates of both preictal spikes and repetitive smallamplitude potentials coupled with the extracellular fast activity are IPSPs. These data also confirmed that synaptic inhibition is preserved at the onset of seizures induced by brief bicuculline perfusion (see Discussion). A gradual reduction of the fast IPSPs was consistently observed in parallel with the spontaneous membrane depolarization that occurred at the onset of the seizure. The changes of fast IPSP amplitude during membrane depolarization is shown in Figure 4C. Each experiment (n ⫽ 5) is identified by a different symbol. The progressive reduction of fast IPSPs with membrane depolarization is further illustrated in the averaged plot in Figure 4D. Cellular Correlates of Fast Activity at Ictal Onset: Deep Entorhinal Cortex Neurons In principle, ictal fast activity could be generated by neurons located in deep layers of the medial EC. Intracellular recordings from layer V and VI principal cells identified from previously described electrophysiological features34,41 demonstrated that these neurons do not generate neuronal firing at seizure onset. During extracellular fast activity, they showed an abrupt steplike membrane hyperpolarization (see Fig 5A), followed by a slow depolarization that lasted 5 to 10 seconds. To demonstrate the inhibitory correlates, we depolarized the membrane potential of the deep neurons by a steady positive current injected via the intracellular pipette, as shown in the representative neuron illustrated in Figure 5. Unlike superficial neurons, no fast IPSPs were observed in deep principal cells (n ⫽ 6). Neuronal firing reappeared in deep neurons during the extracellular burst activity that developed later on during the seizure (see Fig 5B, C; n ⫽ 6). We conclude that, as for superficial principal neurons, firing activity was dampened in deep-layer EC principal cells, and membrane potential was transiently hyperpolarized in temporal correlation with the appearance of fast activity at seizure onset. Cellular Correlates of Fast Activity at Ictal Onset: Putative Entorhinal Cortex Interneurons Finally, we recorded from 10 cells in superficial layers (superficial 600m of the medial EC) identified as putative interneurons from the previoulsy described electrophysiological criteria.41 They generated fast-burst firing in response to LOT stimulation and generated nonadapting firing at more than 100Hz during spontaneous or stimulus-induced depolarizing events. Three of these putative interneurons were recorded for 10 minutes before bicuculline application and during the initiation of a seizure. As illustrated in Figure 6, puta- Gnatkovsky et al: Inhibition at Seizure Onset 679 Fig 5. Intracellular correlate of seizure activity in a principal neuron of the deep layers of the medial entorhinal cortex (EC). Simultaneous extracellular recording (bottom trace, extra) and intracellular recording from a principal neuron of the deep layers of the medial EC (top trace) during the transition into seizure-like discharge. Parts a, b, and c, outlined by the squares, are expanded in the bottom part of the figure. Arrow marks the abrupt hyperpolarization at the onset of seizure, characterized by fast activity in the extracellular recording. The resting membrane potential (rmp) of the neuron (⫺67mV) is outlined by the dotted line. tive interneurons generated a barrage of high-frequency action potentials at more than 150Hz either at the onset (see Fig 6A) or few milliseconds ahead of the seizure (see Fig 6B). Continuous firing gradually evolved into phasic bursting that gradually became time locked with the extracellular discharge pattern (see Fig 6A, part b). Preictal Spikes in Different Types of Entorhinal Cortex Neurons Overall, the intracellular findings demonstrate that putative interneurons generate firing during fast activity at seizure onset, whereas principal neurons in all EC layers are inhibited and silent. To further analyze the role of inhibitory EC circuits during the transition to the ictal state, we evaluated the intracellular correlates of the preictal spikes that occur 30 seconds ahead of the initiation of the seizure-like discharge. Figure 7 shows representative examples of intracellular recordings from principal neurons in superficial and deep layers, and from putative interneurons during the preictal spikes (right traces) and during responses evoked by 680 Annals of Neurology Vol 64 No 6 December 2008 LOT stimulation in the preictal state34 (left traces). Superficial principal cells (n ⫽ 17) generated a direct IPSP in response to LOT stimulation (see Fig 7, top panel) and a negative potential with time course and membrane reversal similar to the evoked IPSP in correspondence to a preictal spike. Also, deep principal neurons (n ⫽ 6) generated a pronounced hyperpolarizing potential both during the preictal spike and in response to LOT stimulation (see Fig 7, middle panel). Unlike principal neurons, the correlate of preictal spikes in the three putative interneurons was a marked bursting discharge (see Fig 7, bottom right traces). A similar bursting response was also observed in response to LOT stimulation (see Fig 7, bottom left traces). These findings confirm that, in the medial EC, inhibition is preserved in the preictal state, just ahead of seizure discharge. Extracellular Potassium Changes during Fast Activity at Ictal Onset Finally, we investigated the mechanisms that promote the gradual switch from fast activity to burst discharges Fig 6. Firing features of putative interneurons during seizure-like activity induced by bicuculline. (A) Simultaneous extracellular (bottom traces, extra) and intracellular recordings from a putative interneuron of the entorhinal cortex (EC) superficial layers (top trace). Expansions of parts a and b are shown in the bottom part of the panel. (B) The firing of another putative interneuron is illustrated. As shown in the expanded trace (a) on the right, the cell firing started before the onset of the fast oscillatory activity that initiates the seizure. rmp ⫽ resting membrane potential. during the ictal event. Extracellular field recordings and intracellular recordings from superficial principal neurons were performed in close proximity to a two-barrel electrode positioned in layers II to III that recorded extracellular potassium changes, [K⫹]o, during a seizure, close to the fast activity generators (Fig 8). [K⫹]o gradually increased with the appearance of fast 20 to 30Hz activity, whereas the superficial cell was still silent. The [K⫹]o increase was closely paralleled by membrane potential depolarization of the neuron and correlated with the slow downward shift in the extracellular trace. When both [K⫹]o and slow membrane depolarization reached a plateau value, irregular firing initiated both in the intracellular and extracellular recordings (see Fig 8A, top panel). The time course of the membrane potential depolarization, the [K⫹]o increase, and the gradual decrease in amplitude of the fast IPSPs are shown in the expanded traces in the bottom panel of Figure 8A. The IPSPs associated with the fast activity progressively decreased in amplitude in parallel with membrane depolarization and [K⫹]o increase (see Fig 8, traces a, b, and c in the right panel). Correlation data obtained in three experiments performed with the same protocol are illustrated in Figure 8B. Finally, the plot of the IPSP amplitude as a function of the changes in [K⫹]o is illustrated for the three neurons in Figure 8C. Discussion We demonstrate that transient and partial disinhibition of an in vitro isolated guinea pig brain induces in the medial EC seizure-like discharges that initiate with 20to 30Hz oscillations. Such fast activity correlated with the following conditions: (1) no firing in principal neurons of superficial and deep layers, (2) sustained firing in putative interneurons, (3) fast IPSPs at 20 to 30Hz in superficial principal neurons, and (4) a slow [K⫹]o increase associated with a progressive decrease of fast IPSPs in superficial neurons. A few seconds after the onset of fast activity, the ictal discharge procedes with the appearance of erratic bursting that becomes more regular and robust with time, and gradually ceases within 10 minutes. As expected after a seizure, postictal depression was consistently demonstrated by signal fre- Gnatkovsky et al: Inhibition at Seizure Onset 681 Fig 7. Intracellular-extracellular correlates of preictal spikes (right panels) compared with the potentials evoked by stimulation of the lateral olfactory tract (LOT; arrowheads in left panels) recorded in superficial (top traces) and deep (middle traces) principal cells, and in putative interneurons (bottom traces) of the medial entorhinal cortex (EC). Resting membrane potentials (rmp) were ⫺57mV for the superficial neuron, ⫺58mV for the deep neuron, and ⫺56mV for the putative interneuron. In the superficial cell, the LOT-evoked intracellular traces were recorded at resting membrane potential (top trace) and during steady membrane hyperpolarization to ⫺82mV. The inset on the right in the middle panel shows a longer extract of the deep neuron response during the preictal spike. quency analysis. The mechanisms of ictal onset generation is the main focus of this study. Simulation studies based on intracranial recordings from human hippocampus and EC proposed that epileptic fast activity at 20 to 30Hz can be explained by a transient and partial GABAergic impairment.15,43 This hypothesis was recently confirmed by reproducing in a computer model of the EC the seizure patterns observed in the isolated guinea pig brain with the experimental protocol utilized in this study.44 We demonstrate now that, in this model, 3-minute arterial perfusion of bicuculline induced a moderate reduction of GABAergic transmission that results in a paradoxical transitory reinforcement of inhibitory networks. By evaluating paired-pulse depression of disynaptic activity, we observed a reduction of 3.9 and 33.5% in the efficacy of inhibition at the time of seizure onset in the 682 Annals of Neurology Vol 64 No 6 December 2008 EC and in the PC, respectively, where depression of disynaptic inhibitory recurrent circuits could be analyzed accurately. Such a measure could not be performed in the medial EC because this region does not receive a direct input from LOT.45 Local stimulation into the medial EC to study inhibition efficacy was not performed to avoid stimulus-induced changes in local excitability. Intracellular recordings demonstrated that inhibition prevails when seizures initiate and in the preictal state. In principal neurons of the EC, seizures start out with an abrupt membrane potential hyperpolarization that coincides with robust firing in putative interneurons. Moreover, we observed that interictal spikes recorded during the 30 seconds that precede a seizure (preictal spikes) correlate with an inhibitory potential in superficial neurons and a burst discharge in putative interneurons. Because these responses were similar to those responsible for the generation of GABAA receptor–dependent IPSPs evoked in normal excitability conditions by LOT stimulation,34 we propose that the intracellular correlates of the preictal spikes and the abrupt hyperpolarization at seizure onset are mediated by GABAergic inhibition. We conclude that the transition from interictal to ictal in EC is promoted by synchronous inhibitory events. Avoli and coworkers46 – 48 demonstrated that, in hippocampal and EC slices in vitro, the initiation of an ictal discharge induced by either 4-aminopyridine or low-magnesium solution correlates with the generation of a large-amplitude, depolarizing, GABAA receptor–mediated potential that lasts approximately 1 second.49 The inhibitory responses observed in our experiments retain the classic features of IPSPs (short duration and hyperpolarizing in nature), suggesting that the mechanisms of ictal generation in the 4-aminopyridine and low-Mg2⫹ slice models may be different. Moreover, unlike EC slice studies, principal EC neurons in our experiments do not generate firing in the early phases of the ictal discharge and, therefore, cannot be responsible for seizure onset. The sustained firing observed in the imminence of a seizure in the limited number of putative interneurons recorded in our close-to-in-vivo condition suggest that a paradoxical reinforcement of inhibition associated with seizure initiation may be enforced by a prevalent reciprocal release of inhibition between inhibitory neurons induced by bicuculline. Long sequences (approximately 5–10 seconds) of fast activity at 20 to 30Hz were consistently observed in superficial EC neurons (but not in deep neurons) during the initial phase of the ictal discharge. These oscillations showed a membrane potential reversal similar to the LOT-evoked IPSP (see also Gnatkovsky and Curtis’s article34) and should, therefore, be considered fast inhibitory synaptic potentials. What is the source of such activity? We recently demonstrated that the hip- Fig 8. Changes in the extracellular potassium ([K⫹]o) during seizure onset in the medial entorhinal cortex (EC). (A) Simultaneous recordings of a principal neuron in superficial layers (top trace), extracellular potential (middle trace), and changes in [K⫹]o (bottom trace) during an ictal event induced by 3-minute perfusion of bicuculline. The tract outlined by the box is expanded in the lower left part in (A). Resting membrane potential (rmp) was ⫺59mV. Expanded sweeps of fast activity sequences marked by a, b, and c are illustrated in the right. Values of potassium concentration are reported for each trace. (B) Average values (⫾ standard error) of the changes in inhibitory postsynaptic potential (IPSP) amplitude (solid squares; left scale) and potassium concentration (filled circle; right scale) during seizure onsets recorded in three experiments. The time after the onset of the fast activity is reported in the abscissae. (C) The dependence of the IPSP reduction on the changes in [K⫹]o are illustrated with different symbols for the three experiments. pocampal output generates direct IPSPs into the superfical neurons of the medial EC, via a feedforward inhibition mediated by EC interneurons located in superficial layers.34 According to this observation, we propose that the 20- to 30Hz oscillations in superficial neurons are generated by feedforward inhibition derived from fast activity that originates in the hippocampus, possibly in CA1 and subiculum. A more detailed simultaneous intracellular analysis of hippocampal and superficial medial EC neurons will allow to test this hypothesis. If this will prove to be the case, seizure activity in the medial EC may be considered to be secondary to hippocampal activation. Yet, the mechanisms by which seizures are generated within the medial EC rely on the reinforcement of activity in local inhibitory circuits. Our findings suggest the possibility that both EC and hippocampus may be necessary to produce seizures in the temporal lobe, possibly through different local network mechanisms. Fast activities at 20 to 80Hz recorded in vivo50,51 or in vitro either by tetanic stimulation52,53 or by pharmacological manipulation in the hippocampus54 –56 and in the EC57,58 are sustained by reciprocal interactions between inhibitory and excitatory networks, with a prevalent role played by the synchronous activation of networks of interconnected interneurons.54 Fast oscillations in the beta-gamma band were also observed during seizure-like activities induced in the hippocampus in vitro by either tetanic stimulation59,60 or by double-pulse stimulation in EC slices of chronic epileptic rats.61 It has been proposed that these oscillations may arise from excitatory GABAergic depolarizing potentials59 mediated by hypersynchronization of GABAergic interneuronal networks.60,62 Brief runs of ultrafast activity (200 – 600Hz), denominated fast ripples, were observed in the hippocampus and in the parahippocampal cortex of patients affected by temporal lobe epilepsy either in coincidence with an interictal Gnatkovsky et al: Inhibition at Seizure Onset 683 spike or in isolation63,64 but were never observed during a seizure. Fast ripples can be reproduced in chronic animal models of temporal lobe epilepsy65 and were proposed to be generated by the synchronous activation of clusters of highly interconnected neurons capable of overcoming interneuron feedback inhibition. No direct relation between fast ripples and the 20- to 30Hz epileptiform activity has been reported. Therefore, these two patterns of fast activity associated with temporal lobe epileptic conditions are probably mediated by different mechanisms. If interneuron activation is the prevalent event at seizure onset, what are the mechanisms that promote the transition from seizure onset into the massive and highly synchronous bursting typically observed during the advanced phase of an ictal event? Gradual inactivation of IPSPs mediated by extracellular ion changes may be the main factor.66 In our experimental conditions, interneurons may contribute to ictal transition by generating pronounced bursting just ahead of and at the onset of seizures. This interneuronal hyperactivity occurs in the absence of principal neuron activation and is probably responsible for the large extracellular potassium changes observed in our experiments. At the onset of a seizure-like event, the increase in [K⫹]o correlated with membrane potential depolarization and a gradual decrease in fast IPSP amplitude in layer II to III principal neurons. IPSP amplitude should increase when the membrane depolarizes above reversal potential for GABAA-mediated chloride current, because of the increase of ionic drive. Interestingly, we observed that fast IPSPs decreased in amplitude during the membrane depolarization associated with seizure onset and potassium increase. This suggests the possibility that the reversal potential of the current associated with the fast IPSPs may gradually change because of the potassium changes. A depolarizing shift in the reversal potential of the GABAA receptor–mediated chloride current is expected, indeed, when [K⫹]o increases.67 When GABAA reversal depolarizes above resting membrane potential, the inhibitory efficacy diminishes, chloride conductance becomes inward, and GABAA receptor activation may turn into excitatory.43 This sequence of events could restore neuronal firing in principal neurons and may favor neuronal hypersynchronization, therefore promoting synchronous bursting discharges and the progression of seizure activity. High correlation between IPSPs and field potential oscillations at seizures onset indicates strong synchronization of interneuronal activity. Decrease in interneuronal synchrony with the progression of the seizure in addition to changes in excitability related to potassium increase could also contribute to the observed IPSP amplitude reduction. 684 Annals of Neurology Vol 64 No 6 December 2008 This study was supported by the Italian Health Ministry (Ricerca Corrente e Ricerca Finalizzata RF 64) and by the Mariani Foundation (R06-50) (V. G., L. L., F. T., M. D.). References 1. Lieb JP, Walsh GO, Babb TL. A comparison of EEG seizure patterns recorded with surface and depth electrodes in patients with temporal lobe epilepsy. Epilepsia 1976;17:137–160. 2. Quesney LF, Gloor P. Localization of epileptic foci. Electroencephalogr Clin Neurophysiol Suppl 1985;37:165–200. 3. Engel JJ, Rausch R, Lieb JP, et al. Correlation of criteria used for localizing epileptic foci in patients considered for surgical therapy of epilepsy. Ann Neurol 1981;9:215–224. 4. Bancaud J, Angelergues R, Bernouilli C, et al. Functional stereotaxic exploration (stereo-electroencephalography) in epilepsies. Rev Neurol (Paris) 1969;120:448. 5. Fisher RS, Webber WR, Lesser RP, et al. High-frequency EEG activity at the start of seizures. J Clin Neurophysiol 1992;9: 441– 448. 6. Gotman J, Levtova V, Olivier A. Frequency of the electroencephalographic discharge in seizures of focal and widespread onset in intracerebral recordings. Epilepsia 1995;36:697–703. 7. Cendes F, Dubeau F, Andermann F, et al. Significance of mesial temporal atrophy in relation to intracranial ictal and interictal stereo EEG abnormalities. Brain 1996;119:1317–1326. 8. Pacia SV, Ebersole JS. Intracranial EEG in temporal lobe epilepsy. J Clin Neurophysiol 1999;16:399 – 407. 9. Spencer SS, Spencer DD. Entorhinal-hippocampal interactions in medial temporal lobe epilepsy. Epilepsia 1994;35:721–727. 10. Bartolomei F, Khalil M, Wendling F, et al. Entorhinal cortex involvement in human mesial temporal lobe epilepsy: an electrophysiologic and volumetric study. Epilepsia 2005;46: 677– 687. 11. Bartolomei F, Wendling F, Vignal JP, et al. Seizures of temporal lobe epilepsy: identification of subtypes by coherence analysis using stereo-electro-encephalography. Clin Neurophysiol 1999;110:1741–1754. 12. Kahane P, Chabardes S, Minotti L, et al. The role of the temporal pole in the genesis of temporal lobe seizures. Epileptic Disord 2002;4(suppl 1):S51–S58. 13. Mintzer S, Cendes F, Soss J, et al. Unilateral hippocampal sclerosis with contralateral temporal scalp ictal onset. Epilepsia 2004;45:792– 802. 14. Spencer SS, Kim J, Spencer DD. Ictal spikes: a marker of specific hippocampal cell loss. Electroencephalogr Clin Neurophysiol 1992;83:104 –111. 15. Wendling F, Bartolomei F, Bellanger JJ, Chauvel P. Epileptic fast activity can be explained by a model of impaired GABAergic dendritic inhibition. Eur J Neurosci 2002;15:1499 –1508. 16. Avoli M, D’Antuono M, Louvel J, et al. Network pharmacological mechanisms leading to epileptiform synchronization in the limbic system in vitro. Progr Neurobiol 2002;68:167–207. 17. Muhlethaler M, de Curtis M, Walton K, Llinas R. The isolated and perfused brain of the guinea-pig in vitro. Eur J Neurosci 1993;5:915–926. 18. de Curtis M, Biella G, Buccellati C, Folco G. Simultaneous investigation of the neuronal and vascular compartments in the guinea pig brain isolated in vitro. Brain Res Protoc 1998;3: 221–228. 19. Uva L, Librizzi L, Wendling F, de Curtis M. Propagation dynamics of epileptiform activity acutely induced by bicuculline in the hippocampal-parahippocampal region of the isolated Guinea pig brain. Epilepsia 2005;46:1914 –1925. 20. Librizzi L, de Curtis M. Epileptiform ictal discharges are prevented by periodic interictal spiking in the olfactory cortex. Ann Neurol 2003;53:382–389. 21. Avoli M, D’Antuono M, Louvel J, et al. Network and pharmacological mechanisms leading to epileptiform synchronization in the limbic system in vitro. Prog Neurobiol 2002;68: 167–207. 22. Bragin A, Csicsvari J, Penttonen M, Buzsaki G. Epileptic afterdischarge in the hippocampal-entorhinal system: current source density and unit studies. Neuroscience 1997;76:1187–1203. 23. Pare D, de Curtis M, Llinas R. Role of the hippocampalentorhinal loop in temporal lobe epilepsy: extra- and intracellular study in the isolated guinea pig brain in vitro. J Neurosci 1992;12:1867–1881. 24. Wozny C, Gabriel S, Jandova K, et al. Entorhinal cortex entrains epileptiform activity in CA1 in pilocarpine-treated rats. Neurobiol Dis 2005;19:451– 460. 25. Witter MP, Groenewegen HJ, Lopes da Silva FH, Lohman AH. Functional organization of the extrinsic and intrinsic circuitry of the parahippocampal region. Prog Neurobiol 1989;33: 161–253. 26. Lopes da Silva FH, Witter MP, Boeijinga PH, Lohman AH. Anatomic organization and physiology of the limbic cortex. Physiol Rev 1990;70:453–511. 27. Witter MP, Wouterlood F. The parahippocampal region. Organization and role in cognitive functions. Oxford: Oxford University Press, 2002. 28. Bartesaghi R, Gessi T, Sperti L. Electrophysiological analysis of the hippocampal projections to the entorhinal area. Neuroscience 1989;30:51– 62. 29. Deadwyler SA, West JR, Cotman CW, Lynch G. Physiological studies of the reciprocal connections between the hippocampus and entorhinal cortex. Exp Neurol 1975;49:35–57. 30. Jones RSG, Heinemann V. Synaptic and intrinsic responses of medial entorhinal cortical cells in normal and magnesium-free medium in vitro. J Neurophysiol 1988;59:1476 –1496. 31. Finch DM, Tan AM, Isokawa-Akesson M. Feedforward inhibition of the rat entorhinal cortex and subicular complex. J Neurosci 1988;8:2213–2226. 32. Kloosterman F, Van Haeften T, Witter MP, Lopes Da Silva FH. Electrophysiological characterization of interlaminar entorhinal connections: an essential link for re-entrance in the hippocampal-entorhinal system. Eur J Neurosci 2003;18: 3037–3052. 33. van Haeften T, Baks-te-Bulte L, Goede PH, et al. Morphological and numerical analysis of synaptic interactions between neurons in deep and superficial layers of the entorhinal cortex of the rat. Hippocampus 2003;13:943–952. 34. Gnatkovsky V, de Curtis M. Hippocampus-mediated activation of superficial and deep layer neurons in the medial entorhinal cortex of the isolated guinea pig brain. J Neurosci 2006;26: 873– 881. 35. Kobayashi M, Wen X, Buckmaster PS. Reduced inhibition and increased output of layer II neurons in the medial entorhinal cortex in a model of temporal lobe epilepsy. J Neurosci 2003; 23:8471– 8479. 36. De Guzman P, Inaba Y, Baldelli E, et al. Network hyperexcitability within the deep layers of the pilocarpine-treated rat entorhinal cortex J Physiol 2008;586:1867–1883. 37. Forti M, Biella G, Caccia S, de Curtis M. Persistent excitability changes in the piriform cortex of the isolated guinea-pig brain after transient exposure to bicuculline. Eur J Neurosci 1997;9: 435– 451. 38. de Curtis M, Manfridi A, Biella G. Activity-dependent pH shifts and periodic recurrence of spontaneous interictal spikes in a model of focal epileptogenesis. J Neurosci 1998;18: 7543–7551. 39. Biella G, Panzica F, de Curtis M. Interactions between associative synaptic potentials in the piriform cortex of the in vitro isolated guinea pig brain. Eur J Neurosci 1996;8:1350 –1357. 40. Librizzi L, Janigro D, De Biasi S, de Curtis M. Blood-brain barrier preservation in the in vitro isolated guinea pig brain preparation. J Neurosci Res 2001;66:289 –297. 41. Gnatkovsky V, Wendling F, de Curtis M. Cellular correlates of spontaneous periodic events in the medial entorhinal cortex of the in vitro isolated guinea pig brain. Eur J Neurosci 2007;26: 302–311. 42. Dunwiddie TV, Worth TS, Olsen RW. Facilitation of recurrent inhibition in rat hippocampus by barbiturate and related nonbarbiturate depressant drugs. J Pharmacol Exp Ther 1986; 238:564 –575. 43. Wendling F, Hernandez A, Bellanger JJ, et al. Interictal to ictal transition in human temporal lobe epilepsy: insights from a computational model of intracerebral EEG. J Clin Neurophysiol 2005;22:343–356. 44. Labyt E, Uva L, de Curtis M, Wendling F. Realistic modeling of entorhinal cortex field potentials and interpretation of epileptic activity in the guinea pig isolated brain preparation. J Neurophysiol 2006;96:363–377. 45. Biella G, de Curtis M. Olfactory inputs activate the medial entorhinal cortex via the hippocampus. J Neurophysiol 2000;83: 1924 –1931. 46. Perrault P, Avoli M. 4-Aminopyridine-induced activity in hilar neurons in the guinea pig hippocampal slice. J Neurosci 1992; 12:104 –115. 47. Avoli M, Barbarosie M, Lücke A, et al. Synchronous GABAmediated potentials and epileptiform discharges in the rat limbic system in vitro. J Neurosci 1996;16:3912–3924. 48. Lopantsev V, Avoli M. Laminar organization of epileptiform discharges in the rat entorhinal cortex in vitro. J Physiol 1998; 509:785–796. 49. Lopantsev V, Avoli M. Participation of GABAA-mediated inhibition in ictallike discharges in the rat entorhinal cortex. J Neurophysiol 1998;79:352–360. 50. Chrobak JJ, Buzsáki G. Gamma oscillations in the entorhinal cortex of the freely behaving rat. J Neurosci 1998;18:388 –398. 51. Csicsvari J, Jamieson B, Wise KD, Buzsaki G. Mechanisms of gamma oscillations in the hippocampus of the behaving rat. Neuron 2003;37:311–322. 52. Traub RD, Whittington MA, Buhl EH, et al. On the mechanism of the gamma 3 beta frequency shift in neuronal oscillations induced in rat hippocampal slices by tetanic stimulation. J Neurosci 1999;19:1088 –1105. 53. Bracci E, Vreugdenhil M, Hack SP, Jefferys JGR. On the synchronizing mechanisms of tetanically induced hippocampal oscillations. J Neurosci 1999;19:8104 – 8113. 54. Traub RD, Jefferys JGR, Whittington MA. Fast oscillations in cortical circuits. Cambridge, MA: MIT Press, 1999. 55. Fisahn A, Pike FG, Buhl EH, Paulsen O. Cholinergic induction of network oscillations at 40 Hz in the hippocampus in vitro. Nature 1998;394:186 –189. 56. Williams JH, Kauer JA. Properties of carcachol-induced oscillatory activity in the rat hippocampus. J Neurophysiol 1997;78: 2631–2640. 57. van der Linden S, de Curtis M, Panzica F. Carbachol induces fast oscillations in the medial but not in the later entorhinal cortex of the isolated guinea pig brain. J Neurophysiol 1999; 82:2441–2450. 58. Dickson CT, Biella G, de Curtis M. Evidence for spatial modules mediated by temporal synchronisation of carbachol induced gamma rhythm in medial entorhinal cortex. J Neurosci 2000;20:7846 –7854. 59. Köhling R, Vreugdenhl M, Bracci E, Jefferys JGR. Ictal epileptiform activity is facilitated by hippocampal GABAA receptormediated oscillations. J Neurosci 2000;20:6820 – 6829. Gnatkovsky et al: Inhibition at Seizure Onset 685 60. Velazquez JLP, Carlen PL. Synchronization of GABAergic interneuronal networks during seizure-like actvity in the rat horizontal hippocampal slice. Eur J Neurosci 1999;11:4110 – 4118. 61. Tolner EA, Kloosterman F, Kalitzin SN, et al. Physiological changes in chronic epileptic rats are prominent in superficial layers of the medial entorhinal area. Epilepsia 2005;46(suppl 5):72– 81. 62. Fujiwara-Tsukamoto Y, Isomura Y, Kaneda K, Takada M. Synaptic interactions between pyramidal cells and interneurone subtypes during seizure-like activity in the rat hippocampus. J Physiol 2004;557:961–979. 63. Bragin A, Wilson CL, Staba RJ, et al. Interictal high-frequency oscillations (80-500 Hz) in the human epileptic brain: entorhinal cortex. Ann Neurol 2002;52:407– 415. 686 Annals of Neurology Vol 64 No 6 December 2008 64. Staba RJ, Wilson CL, Bragin A, et al. Quantitative analysis of high-frequency oscillations (80-500 Hz) recorded in human epileptic hippocampus and entorhinal cortex. J Neurophysiol 2002;88:1743–1752. 65. Bragin A, Engel J, Wison CL. Hippocampal and entorhinal cortex high-frequency oscillations in human epileptic brain and kainic acid treated rats. Epilepsia 1999;40:127–137. 66. Farrant M, Kaila K. The cellular, molecular and ionic basis of GABA(A) receptor signalling. Prog Brain Res 2007;160: 59 – 87. 67. Thompson SM, Gahwiler BH. Activity-dependent disinhibition. II. Effects of extracellular potassium, furosemide, and membrane potential on ECl⫺ in hippocampal CA3 neurons. J Neurophysiol 1989;61:512–523.