Address correspondence and reprint requests to Dr. H.R. Cock at Clinical Neurosciences (Epilepsy), Department of Cardiological Sciences, St. George's Hospital Medical School, London SW17 0RE, U.K. E-mail: firstname.lastname@example.org
Summary: Purpose: To characterize in detail a model of focal neocortical epilepsy.
Methods: Chronic focal epilepsy was induced by injecting 25–50 ng of tetanus toxin or vehicle alone (controls) into the motor neocortex of rats. EEG activity was recorded from electrodes implanted at the injection site, along with facial muscle electromyographic (EMG) activity and behavioral monitoring intermittently for up to 5 months in some animals. Drug responsiveness was assessed by using the antiepileptic drugs (AEDs) diazepam (DZP) and phenytoin (PHT) delivered systemically, while 1,2,3,4-tetrahydro-6-nitro-2,3-dioxo-benzo[f]quinoxaline-7-sulfonamide (NBQX), a competitive antagonist at AMPA receptors, was administered directly to the brain to investigate the potential benefits of focal drug delivery.
Results: Tetanus toxin induced mild behavioral seizures that persisted indefinitely in all animals. EEG spiking activity, occurring up to 80% of the time, correlated with clinical seizures consisting of interrupted behavioral activity, rhythmic bilateral facial twitching, and periods of abrupt motor arrest. Seizures were refractory to systemic administration of DZP and PHT. However, focal delivery of NBQX to the seizure site reversibly reduced EEG and behavioral seizure activity without detectable side effects.
Conclusions: This study provides a long-term detailed characterisation of the tetanus toxin model. Spontaneous, almost continuous, well-tolerated seizures occur and persist, resembling those seen in neocortical epilepsy, including cortical myoclonus and epilepsia partialis continua. The seizures appear to be similarly resistant to conventional AEDs. The consistency, frequency, and clinical similarity of the seizures to refractory epilepsy in humans make this an ideal model for investigation of both mechanisms of seizure activity and new therapeutic approaches.
Focal neocortical epilepsy is a frequent cause of refractory epilepsy and has a considerable associated morbidity (1). In its most severe form, almost continuous seizure activity can occur. Continuous focal motor twitching was described in 1895 by the Russian neurologist Kojewnikoff (2), who termed this condition epilepsia partialis continua (EPC). This condition represents the status epilepticus equivalent of simple partial motor seizures. EPC can be defined as regular or irregular clonic muscular twitching affecting a limited part of the body, occurring for a minimum of 1 h, and recurring at intervals of ≤10 s (3). Movement frequency varies from 0.5 to 10 Hz and is often of brief duration. EPC is of cortical origin and can result from structural abnormalities or from more generalised encephalitides. The development of treatment strategies for focal neocortical epilepsy has, however, been hampered by the lack of good animal models of this condition. Indeed, the need for good animal models of pharmacoresistant epilepsy has been highlighted as a current research priority by leading authorities (4).
In 1990, focal cortical application of tetanus toxin to cats was reported to result in a syndrome similar to EPC in humans (5). Tetanus toxin is known to induce acute and spontaneous seizures, seemingly through inhibiting the release of neurotransmitters, predominantly from inhibitory interneurons (6,7). An extensive range of behavioral and electrophysiologic studies investigated the seizure activity induced by tetanus toxin injected into the hippocampus (8–11), which has been shown to induce long-term changes in patterns of synaptic excitation while causing no or minimal neuronal damage. Intrahippocampal tetanus toxin can thus induce a chronic epileptic syndrome, in which spontaneous seizures can occur long after the toxin has been cleared from the system (11). Although some studies have shown that seizures can reoccur up to months later after intrahippocampal toxin injection (9,12,13), the seizures are usually self-limiting, lasting only between 1 and 3 weeks (14–16), thus precluding long-term treatment studies.
A few studies have applied tetanus toxin neocortically where, similarly, it has been shown to decrease functional inhibition (17) and produce a chronic epileptic focus (17–22), but which seems to persist for longer than seen in intrahippocampal treatments. Most of these studies have concentrated on investigating mechanisms behind seizure activity in vitro in brain slices or included only limited behavioral monitoring. Where clinical seizures were reported, they were found to range in severity and frequency between animals and with time. In addition, the seizures were often very severe, with secondary generalization in the acute phase and a significant mortality rate associated with this (5,19,22). Detailed long-term electromyography (EMG) and behavioral monitoring in parallel with EEG recordings in tetanus toxin–induced neocortical epilepsy have not previously been reported. In this study, we set out to develop a consistent model of refractory neocortical epilepsy by using the application of tetanus toxin. Our preliminary investigations demonstrated that tetanus toxin injection into the neocortex produced chronic but mild motor seizures that were extremely consistent between animals. We then characterized this model in detail by using EEG, EMG, and behavioral recordings for 5≥ months in 32 animals. Finally, we asked whether this model shows the same antiepileptic drug (AED) resistance typical of refractory neocortical epilepsy, including EPC, in humans, and whether these seizures respond to focal drug application of the a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)/kainate receptor antagonist 1,2,3,4-tetrahydro-6-nitro-2,3-dioxo-benzo[f]quinoxaline-7-sulfonamide (NBQX), a compound that inhibits excitatory synaptic transmission.
MATERIALS AND METHODS
All animal procedures followed the UK Animal (Scientific Procedures) Act, 1986 and were subject to local ethical review.
Tetanus toxin was provided by Quadratech UK Ltd. Epsom, Surrey, UK and used at a working concentration of 25–50 ng/0.5 μl, pH 7.5. Drugs and reagents were from Sigma UK, Gllingham, Dorset, UK, unless otherwise stated. Electrodes were from Plastics One, Seven Oaks, Kent, UK.
Thirty-two male Sprague–Dawley rats of initial weights between 240 and 320 g (from Tuck & Son, Ltd.) were used. Rats were kept in a 12-h light/dark cycle with free access to food and water. An additional nine rats were injected with a vehicle solution of the inactivated form of the toxin, tetanus toxoid (Quadratech UK Ltd. Epsom).
The rats were kept under 1.5–2.5% halothane inhalation anaesthesia. Tetanus toxin or the vehicle control was stereotaxically injected from a 2-l Hamilton syringe mounted on a microinjector unit into the right motor cortex (1.1 mm ventral from bregma and 2.5 mm lateral from bregma (23)). A volume of 0.5 μl was injected over a period of 5 min through an internal cannula, which was left in place for a further 10–15 min to allow binding of the toxin and to minimize reflux. A guide cannula and EEG electrode were permanently implanted at the injection site and fixed in place with dental acrylate. A reference electrode was placed beneath the skin behind the neck, and an additional electrode was placed subcutaneously over the right facial muscles in some animals for EMG recordings. Immediately after surgery, animals were given 0.05 mg/kg subcutaneous (s.c.) buprenorphine analgesia and 5 ml s.c. saline. No further administrations of analgesia were required, and the animals all recovered completely within 24 h.
Two days to a week after tetanus toxin injection, EEG and EMG activity was recorded continually for 4 h either daily or weekly. Individual rats were recorded at the same time of day for each session to minimize diurnal variation. Behavioral responses were monitored simultaneously and correlated to the EEG traces. Recording sessions were repeated for 5≤ months. Data were acquired and sampled by using CED (Cambridge Electronic Design UK, Ltd. Cambridge, UK) Spike 2 software. Seizure activity was defined as rhythmic EEG spiking associated with stereotyped behavioral change, supported by EMG recordings. All EEG/EMG recordings were visually reviewed and correlated to behavioral monitoring by an investigator blinded to the animal status. Seizure onsets and offsets were manually defined, and a tailor-made analysis script (devised in collaboration with David Crick, CED) was then used to calculate the percentage of seizure time and the frequency spectrum of the EEG activity that occurred during the measurement period and to undertake frequency spectrum analyses. Statistical analysis was performed by using a nonparametric Mann–Whitney U test in SPSS for Windows and one-way analysis of variance (ANOVA) with repeated measures. Results were accepted as significant if p ≤ 0.05.
Pharmacologic studies were carried out 2–13 weeks after the initial surgery. Phenytoin (PHT) and diazepam (DZP) studies were conducted ≥4 weeks after surgery; NBQX studies were carried out 2–4 weeks after surgery. In a preliminary study of the responsiveness of the model to conventional AEDs, a group of five rats (one vehicle control and four epileptic rats) were given single high-dose intraperitoneal injections of PHT (50 mg/kg) and DZP (5 mg/kg and 20 mg/kg) on separate occasions. In addition, to investigate whether agents can affect seizure activity when delivered directly to the epileptic focus, a single dose of NBQX (100 nmol), a competitive antagonist at AMPA receptors, was perfused over a 5-min period via the implanted cannula above the seizure focus. A focal cortical injection of saline was used as a control for this procedure. Each rat received each of these treatments in a randomized order at intervals of ≥2 weeks. The EEG and EMG activity was recorded for ≥2 hours before all injections, such that each animal also served as its own control; injections were then performed without sedation and followed by 4–5 h of further recording. Statistical comparisons of total seizure discharge duration and amplitude were undertaken by using Student's t test.
A single application of tetanus toxin induced frequent, mild facial seizures in all animals, which persisted indefinitely. From a total of 46 rats that underwent the surgical procedure, five died during or immediately after the operation (mortality, 10.8%), nine animals received the control injection of tetanus toxoid and exhibited no clinical seizures, and the remaining 32 animals injected with the tetanus toxin all subsequently developed behavioral seizures, as described later, within a few days. Animals recovered from the surgery within a few hours, and aside from the seizures, the rats behaved normally and were in good health. No mortality occurred because of the seizures themselves, which were mild and well tolerated. Additional mortality in animals kept for >3 months was 10.9% and occurred because of infection around the fixed head unit or detachment of the head unit. No recordings were made from rats where infection was present.
In four of the nine control animals, rare interictal spiking was seen (Figs. 1A and 2C), but never with any associated behavioral change, or with an evolving pattern suggestive of subclinical seizures. In control animals where interictal spiking was seen, it was present on average during ≤12% of recording time (contrasting to the 40–60% spiking seen in tetanus toxin animals), and with a different frequency pattern from that seen in tetanus toxin animals (Fig. 3), being predominantly of low frequency (<2Hz) and with very few fast (>10 Hz) components.
In tetanus toxin animals, epileptic activity occurred throughout the period of each recording session (4 h) in clearly identifiable episodes (see Fig. 2A–C). Discrete seizures (Fig. 2Cii) started with an initial abrupt change in background EEG, followed by a burst of fast activity evolving into a rhythmic spike discharge, typically between 0.7 and 1.9 mV in amplitude, associated with interruption in activity, and either behavioral and EMG arrest, or more commonly with synchronous EMG discharges (Fig. 2B) and visible bilateral facial myoclonic jerking. Occasionally facial jerking was not discernable on inspection but was evident subclinically on EMG recordings. All animals displayed all three behavioral manifestations (behavioral and EMG arrest; arrest with visible facial twitching; arrest with EMG spiking) at times throughout the course of experiments, although data on exact proportions have not been collected. Seizure episodes typically lasted 1–15 min, during which seizure activity on the EEG would be present semicontinuously, with breaks of ≤1–2 s. Fourier analysis showed that in contrast to the pattern seen in control animals (Fig. 3i and ii), three distinct EEG frequency bands were associated with seizure activity: slow spiking of 0.5–2 Hz, intermediate spiking from 10 to 15 Hz, and faster spiking >20 Hz that tended to be of shorter duration and occurred less frequently (Fig. 3iii), although often preceding discrete seizures. Normal behaviour between these episodes, including eating, sleeping, grooming, drinking, and exploratory behavior, was correlated with “quiet” EEG periods (Fig. 1i) with amplitudes ≤0.3mV and with absence of epileptic spiking. Interictal and ictal periods in tetanus toxin animals were easily identifiable behaviorally and by the abrupt cessation of epileptic discharge activity (Fig. 2A) on the EEG and EMG recordings (Fig. 2Bii).
In tetanus toxin animals over the first 4 weeks, no effect of time on seizure activity was noted (p = 0.7; Fig. 4i), and total activity between animals also was very consistent (Fig. 4ii), the overall mean seizure time being 62.6% (95% confidence limits, 61.1–64.2%). After the first month, a gradual decrease in the amount of seizure activity was seen in some animals over time (Fig. 4iii), although seizures were still occupying ≤40% of the time 5 months after initial tetanus toxin application (Fig. 4iii).
Effects of anticonvulsants on tetanus toxin–induced seizure activity
Because focal neocortical epilepsy and EPC in humans are largely resistant to AEDs, we asked whether our model exhibited a similar resistance. We used high doses of PHT and DZP, two drugs that are commonly used to treat status epilepticus. The systemic administration of the anticonvulsant PHT (50 mg/kg) had no effect on the seizure frequency, with no significant differences in the percentage seizure time (43.6 ± 7.26% preinjection, 45.3 ± 6.14% postinjection; n = 4; Fig. 5i) or in EEG discharge amplitude during seizures (1.58 ± 0.13 mV preinjection, 1.57 ± 0.14 mV postinjection; n = 4; Fig. 5ii). In addition, no change was noted in the behavioural seizures observed. Similarly, systemic administration of DZP did not decrease the percentage amount of clinical and electrographic seizure frequency (35.8 ± 3.7% preinjection, 37.7 ± 3.5% postinjection; n = 4; Fig. 5), even at the highest dose (20 mg/kg). A decrease in the amplitude of seizure-related EEG discharges was observed at high doses of DZP, although significance was not reached (1.83 ± 0.14 mV preinjection, 1.57 ± 0.15 mV 30 min postinjection; n = 4; Fig. 5ii).
We next asked whether inhibiting excitatory synaptic transmission with the AMPA receptor antagonist, NBQX, would stop the seizure activity. We used a dose of NBQX (100 nmol in 4 μl) that had previously been shown to inhibit hippocampal activity damage by glutamate-receptor agonists (24). First we confirmed that an injection of vehicle (saline) into the neocortical seizure focus had no effect on the EEG pattern (Fig. 4); both seizure frequency and EEG discharge amplitude remained stable (39.9 ± 2.0% and 1.89 ± 0.21 mV preinjection, 40.9 ± 3.05%, and 1.84 ± 0.15 mV 30 min postinjection; n = 4). In contrast, administration of NBQX to the seizure focus reduced the frequency of behavioural and seizure-related EEG discharge activity (55.7 ± 7.8% preinjection, 3.1 ± 0.87% postinjection; n = 4; p ≥ 0.05; Fig. 5). Where seizures occurred after NBQX, the amplitude of seizure-related EEG discharge activity was also reduced (1.47 ± 0.15 mV preinjection, 0.40 ± 0.10 mV postinjection; n = 4, p ≤ 0.05), and a substantial reduction occurred in high-frequency components in all animals tested (a typical example is shown in Fig. 6). Continued recording demonstrated that the suppression of seizure activity decreased with time after injection, so that after 2 h postinjection, the seizure frequency had almost returned to pretreatment values (55.7 ± 7.8% and 1.47 ± 0.15 mV preinjection, 50.0 ± 10.5% and 1.01 ± 0.07 mV measured 2 h postinjection; n = 3; Fig. 4ii). Note that the baseline percentage seizure time varies with the different drug applications; this is because each was carried out at a different time point from the initial tetanus toxin application, and the interval between each time point was ≥2 weeks (Fig. 1iii). As the percentage seizure time remains high, however, each animal can be used as its own control with seizure activity directly compared before and after a drug application.
We have further characterized a model of focal neocortical epilepsy in rats that produces almost continuous long-term recurrent motor seizures that are spontaneous and resistant to systemic treatment with DZP and PHT, two AEDs commonly used to treat status epilepticus. In contrast, the seizures were controlled after focal injection of the glutamate-receptor antagonist, NBQX.
Although AEDs are effective in the majority of epilepsy patients, systemic drug administration is often limited by intolerable side effects due to widespread CNS and extra-CNS actions. Furthermore, up to a third of patients, many of whom have focal neocortical epilepsy (1,25), are refractory to existing therapies and continue to have intractable seizures. Refractory epilepsy is associated with a significant morbidity, mortality, and widespread psychosocial and economic consequences. It is recognized that preclinical models of refractory epilepsy are essential in addition to existing processes of therapy discovery and development to enable more predictive evaluation of clinical success (4). In particular, the need for better in vivo animal models has been highlighted, with the following characteristics: that are representative of the human condition (i.e., those that exhibit long-term recurrent spontaneous seizures and so are an indication of underlying epileptogenic brain pathology rather than short-term induced seizures in normal brain); that are refractory, not responding or poorly responding to currently available AEDs; that have a high frequency of seizure activity to provide a robust and reliable model for multiple compound screening; and that can also be used for basic epilepsy research to increase understanding of the underlying mechanisms epileptic brain areas. This study demonstrates that the tetanus toxin applied to the neocortex results in a model of human epilepsy that fulfils all these criteria and is thus ideal to investigate the underlying mechanisms of seizures and evaluate different treatment approaches.
It was previously demonstrated that tetanus toxin induces spontaneous seizures that can continue long after the toxin has been cleared from the system, but a long-term detailed behavioral and EMG study with parallel EEG recordings has not been previously reported. We have demonstrated that the application of tetanus toxin into the motor cortex results in a consistent chronic epileptic focus with the electroclinical features of EPC, including discrete focal motor seizures, persisting for ≤5 months after application of tetanus toxin to the motor cortex and that seizures remain frequent over this time. The seizures resemble those seen focal neocortical epilepsy and EPC: they are neocortical in origin, spontaneous, frequent as in EPC, and continue for long periods including through sleep (3); the clinical manifestation of the seizures also are similar, consisting of rhythmic repetitive jerking of a localised area, with intervals between individual seizures in a session mostly lasting less than a few seconds; seizures can occur in isolation as well as in clusters (26); seizure activity is mostly made up of frequencies <15 Hz, but some high-frequency components are seen. The pattern of high-frequency components evolving into rhythmic spiking is also similar to that seen during depth recordings of focal clonic seizures in human epilepsy (27).
That previous studies have reported variable seizure severity and mortality may be explained by methodologic differences including the age/weight and strain of rat, tetanus toxin [dose, suppliers, inclusion of bovine serum albumin (BSA) in the vehicle], and depth of injection into the cortex. However, as described here, we found the model highly reproducible and hence ideal to test not only short-term treatments but also for long-term studies. Importantly, the seizures are mild, causing little or no distress to the animals, which are healthy with normal behaviour between seizures. Behavioral, EMG, and EEG recordings in freely moving animals are straightforward, as are applications of agents both directly to the seizure site or systemically, allowing comprehensive evaluation of seizure activity and treatment effects.
The interictal EEG discharges seen in control animals may have occurred as a result of direct cortical trauma from the surgical implantation of the cannula/electrode unit; the epileptogenicity of injecting any substance into the cortex; or the incomplete inactivation of the tetanus toxoid, although we consider this unlikely because of manufacturing testing and guarantees. It is possible that the chronic epileptic focus results from the combination of local mechanical damage and superimposed tetanus toxin, and that tetanus toxin alone would be insufficient. However, as injection of toxin inevitably involves mechanical disruption, it is not possible to test this hypothesis. It is clear, however, that the mechanical disruption alone was not sufficient to establish a true epileptic focus, as the discharges clearly distinguishable from the seizures seen in tetanus toxin animals, and seizures themselves were never seen in control animals.
We did not carry out EEG recordings from sites distant to the focus of tetanus toxin injection, as this has been undertaken by previous groups, who have demonstrated rapid spread of seizure activity from the injection focus in motor cortex ipsilateral neocortical structures and ipsilateral cingulate gyrus (5), and to contralateral structures (19). Brain-slice studies as early as 16 h after toxin injection also consistently demonstrated independent interictal changes in the contralateral hemisphere (19,20), but a contralateral seizure focus has not been described. The mechanisms of independent contralateral changes are unclear, although thought to reflect transynaptic movement of toxin and axonal transport, rather than simple diffusion (19,28). The observation of bilateral facial jerking in our animals is thus consistent with the known spread of seizure activity from a unilateral focal area, although additional foci cannot be absolutely excluded, given the limited recording points.
The seizures in this model also appear to be largely resistant to DZP and PHT, mimicking focal neocortical epilepsy and EPC, which are notoriously pharmacologically resistant. The extracellular fluid concentrations in brain after a single doses, as used in this study, would be expected to reach therapeutic levels (29,30), and marked sedation of the animals was observed, but without any significant effect on seizure activity. A more comprehensive study of drug responsiveness is under way to extend the preliminary data presented here. The consistency of this model and frequency of spontaneous seizures allows easy comparison of EEG, behavioral, and EMG parameters immediately before and after application of agents. This contrasts with other models with spontaneous neocortical (31–33) or hippocampal seizures (34) in which seizures may occur relatively inconsistently or infrequently, are often subclinical or spontaneously improve, making it difficult to assess the effects of short-term drug administration, as the pharmacologic effect may have worn off before meaningful recordings can be made. Injecting control substances such as saline does not seem to alter the seizure activity, as little or no effect is observed on seizure frequency and EEG discharge amplitude, whereas focal injections of NBQX, a competitive antagonist at AMPA/kainate receptors, had a profound reversible effect on seizure activity, providing proof-in-principle that focal treatment can significantly reduce both EEG and behavioral seizure activity where systemic agents have failed. The general behavior of the rats did not seem to be adversely affected by the application of NBQX, in contrast to the high incidence of side effects, including sedation and motor impairment, that did occur with systemic PHT, and would be anticipated with systemic administration of glutamate antagonists at a high enough dose to provide antiepileptic effects (35).
Therefore this model permits the testing of putative AEDs in vivo that are too toxic to give systemically or that do not cross the blood–brain barrier. In addition, the use of specific agonists and antagonists will permit a greater understanding of the mechanisms underlying seizure generation. Furthermore, the consistency and stability of the model make it ideal to test other focal treatment options such as neuronal grafting or long-term focal drug delivery that are currently being developed in the hope of providing new therapeutic options for patients with refractory neocortical epilepsy (36). Other treatment approaches such as focal cooling devices, recently shown in anaesthetised animals to reduce significantly immediately induced cortical seizures (37) might also be assessed in this model.
We thus established and characterized in detail a model of refractory focal cortical epilepsy, including EPC, that, as in the human situation, is resistant to systemic AEDs. In contrast, the excellent response to focal injection of NBQX, despite refractoriness to systemic conventional agents, provides a promising approach to treatment of this devastating condition. More selective and enduring focal treatments are presently under investigation.
Acknowledgment: This work was supported by a grant from The Brain Research Trust. Karen Nilsen is currently supported by the Epilepsy Research Foundation. Hannah Cock was funded by the Wellcome Trust. Analysis scripts were designed in collaboration with David Crick from Cambridge Electronic Designs Ltd. Experimental assistance was provided by Hannah Sleven.