Address correspondence to Frances E. Jensen, M.D., Director of Epilepsy Research, Neurology, Enders 348, Children’s Hospital Boston, 300 Longwood Ave, Boston, MA 02115, U.S.A. E-mail: email@example.com
Purpose: To test the efficacy of the novel candidate anticonvulsant talampanel (GYKI 53773) in a rodent model of hypoxic neonatal seizures. Talampanel is a noncompetitive antagonist of the alpha-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid subtype of the glutamate receptor (AMPAR). We have previously shown that AMPARs play a critical role in the generation of acute seizures and later-life seizure susceptibility in this model of neonatal seizures.
Methods: Seizures were induced in postnatal day (P) 10 Long-Evans rat pups by a 15 min exposure to global hypoxia. Acute seizure activity at P10 and subsequent susceptibility to seizure-induced neuronal injury with a “second-hit” kainate-induced seizure at P30–31 were compared between animals receiving talampanel (1, 5, 7.5, or 10 mg/kg) intraperitoneally (i.p.) versus saline vehicle treatment.
Results: Talampanel treatment suppressed seizures in a dose-dependent manner, with maximal effect at 7.5 and 10 mg/kg. In addition, talampanel treatment 30 min before hypoxia prevented later-life increases in seizure-induced neuronal injury as assessed by in situ DNA nick end-labeling (ISEL).
Discussion: We have previously demonstrated efficacy of other AMPAR antagonists such as NBQX and topiramate in this model. The present finding shows that the novel agent talampanel, under evaluation as an antiepileptic drug in children and adults, may have clinical potential in the treatment of neonatal seizures, particularly those occurring in the context of hypoxic encephalopathy.
To date, there have been no new drugs that have been proven effective in clinical trials for suppression of neonatal seizures. We have previously developed a model of hypoxia-induced seizures in the immature rat, in which hypoxia induces early refractory seizures as well as long-term effects, including spontaneous seizures and increased susceptibility to seizure-induced neuronal injury in later life (Jensen et al., 1991). In this model, activation of the alpha-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid subtype of the glutamate receptor (AMPAR) is a critical factor in the epileptogenic effect of hypoxia (Sanchez et al., 2001, 2005). AMPAR antagonists such as NBQX and topiramate effectively suppress the acute and long-term epileptogenic effects of hypoxia in the perinatal rodent (Jensen et al., 1995; Koh & Jensen, 2001; Koh et al., 2004). We have shown that AMPARs are overexpressed on cortical neurons in the neonatal rat and term human, when the susceptibility to hypoxic seizures is highest (Talos et al., 2006a, 2006b). In addition, the AMPAR antagonists topiramate and NBQX do not cause an increase in constitutive apoptotic neuronal death in the immature brain, unlike some conventional AEDs (phenobarbital, phenytoin, diazepam, and valproate), suggesting a better safety profile for use in the neonatal population (Silverstein & Jensen, 2007). Current obstacles to clinical trial development of AMPAR antagonists in the neonatal population include the fact that NBQX is not being developed for clinical use and topiramate is not available in parenteral form. Experimental and human tissue studies suggest that the AMPAR plays a critical role in the generation of neonatal seizures (Talos et al., 2006a, 2006b; Silverstein & Jensen, 2007). Recently, more specific AMPAR antagonists have been investigated for clinical use, and include the novel noncompetitive AMPAR antagonist talampanel (GYKI-53773), which has been shown to be anticonvulsant in experimental models of seizures in the adult brain (Belayev et al., 2001; Jakus et al., 2004). Clinical trials in adults show anticonvulsant efficacy as monotherapy in refractory epilepsy (Langan et al., 2003) and as an add-on for therapy for partial complex seizures (Chappell et al., 2002; Howes & Bell, 2007). Given that talampanel has demonstrated efficacy in adult epilepsy trials, we aimed to examine its potential efficacy for neonatal seizures. As a first step to gather preclinical evidence and to further define a specific role for AMPAR in epileptogenesis in the developing brain, we examined the efficacy of talampanel in a rodent model of neonatal hypoxia-induced seizures.
Materials and Methods
Seizures were induced in Long-Evans male rats at postnatal (P) day 10. Each litter was divided into three groups: control, hypoxic seizures with vehicle treatment, and hypoxic seizures with talampanel treatment. Animals were maintained in a 12-h light/dark cycle. Pups were returned to their dam following induction of hypoxic seizures or sham, and adult female rats were housed with their pups in cages until pups were weaned at P21. All procedures were undertaken in accordance with the guidelines of the animal care and use committee at Children’s Hospital, Boston.
Long-Evans male rats at postnatal P10 were exposed to global hypoxia for 15 min in an airtight chamber. As per our previously published methods, the oxygen concentration was gradually decreased to 7% for 8 min, 5% for 6 min, and 4% for the final minute (Jensen et al., 1991; Koh et al., 2004). During hypoxia, seizures were videotaped to record seizure activity. The tapes were scored by an investigator, who was blinded to treatment condition for the number of tonic–clonic head and limb movements and cumulative seizure duration during hypoxia. The talampanel/hypoxia group received talampanel by intraperitoneal (i.p.) injection at a dose of 1, 5, 7.5, or 10 mg/kg. Talampanel or vehicle [dimethyl sulfoxide (DMSO) in 0.5% methyl cellulose] was administered 30 min before hypoxia. Litter mate control pups were kept at room air, and the body temperature of all pups were maintained at 32°–34° on a warming blanket. The number of seizure episodes during hypoxia was compared by a one-way analysis of variance (ANOVA).
Following induction of hypoxic seizures, animals were permitted to survive until P30–31, when they underwent a “second-hit” challenge of kainate-induced status epilepticus (10 mg/kg i.p.). Seizure activity and severity was recorded over a 3-h period, and latency to seizure onset and maximal severity was determined per our previously described modified Racine scale from 0 to V as follows: 0, no response; I, wet-dog shake and/or behavioral arrest; II, wet-dog shake, staring, pawing, clonic jerks; III, bilateral forelimb clonus with rearing falling; IV continuous grade III seizures for >30 min; and V, death (Koh et al., 2004). P30–31 rats with prior hypoxic seizures at P10 were compared to their age-matched normoxic littermates. On the basis of our prior studies, we have found that this dose of kainate results in 100% of animals reaching grade IV on this modified Racine scale.
DNA fragmentation by in situ end-labeling histochemistry
We compared neuronal death in hippocampus (areas CA1 and CA3) and amygdala in rats permitted to survive 72-h postkainate-induced seizures. As per our previous studies, neuronal injury was assessed by in situ DNA nick end-labeling (ISEL) staining (Koh & Jensen, 2001; Koh et al., 2004). At 72 h after kainate-induced seizures, rats were killed and then perfused transcardially with 4% paraformaldehyde. Normoxic littermates were also sacrificed at the same age. Each brain was removed, postfixed, cryoprotected in 20% sucrose overnight, and mounted onto a freezing microtome. Three 50-μm–thick coronal sections/per rat at stereotactically identical locations between rats were processed for in situ nick translation using a modification of the protocol developed by Wijsman et al. (Wijsman et al., 1993; Koh et al., 2004). To quantify DNA fragmentation, counts of positively stained cells were compared at a final magnification of ×100. Positively stained cells were counted using a grid reticule over areas CA1, CA3, and the amygdala and the values were compared by a one-way ANOVA. Similar to Koh et al. (2004), counts were made by averaging three fields (4.16 mm2 per region × 3 = 12.5 mm2 total): both left and right hippocampi in CA1, CA3 (−3.0 to −3.3 mm from Bregma and 0–3 mm lateral to midline), and basolateral/basomedial amygdaloid nucleus (−3.0 to −3.3 mm from Bregma and 3–6 mm lateral to midline) on three coronal sections per animal (Paxinos & Watson, 1986). A total of 126 sections from 42 animals were analyzed (three sections per rat and two hippocampi per section).
Assessment of cell death after talampanel administration to normal P10 rats
Because a number of AEDs have been shown to acutely increase constitutive apoptosis in the developing rodent brain following single dosing, we examined whether the optimal dose of talampanel would cause cell death when administered at P10. Normal P10 Long-Evans rats were given one dose of talampanel (i.p.) at 7.5 mg/kg and were sacrificed 24 h after the injection, as described previously (Bittigau et al., 2002). Briefly, each brain was dissected, postfixed, cryoprotected in 20% sucrose overnight, mounted onto a freezing microtome, and processed for in situ end-labeling as described previously. Counts (cells/12.5 mm2) were made by counting total number of ISEL-positive cells in coronal sections at −2.0 to −2.6 mm from Bregma (Sherwood & Timiras, 1970). Total section counts of talampanel versus vehicle injection were compared with Student’s t-test.
Talampanel suppresses hypoxia-induced seizures
Acute hypoxia-induced seizures were suppressed by talampanel in a dose-related manner within the range of administration from 1–10 mg/kg. Similar to our previously published reports, vehicle-treated rats responded initially to hypoxia with myoclonic jerks, followed by the onset of tonic–clonic head and trunk movement (Jensen et al., 1991; Koh & Jensen, 2001; Koh et al., 2004). To determine treatment efficacy, we compared the number or episodes of tonic–clonic seizures between groups. Compared to vehicle-treated animals, the anticonvulsant activity of talampanel was maximal at 7.5 and 10 mg/kg, where seizures were blocked 74.6% at 10 mg/kg (25.4 ± 7.3, n = 17; p < 0.001) and 86.7% at 7.5 mg/kg (13.4 ± 3.2, n = 17; p < 0.001) (Fig. 1). The effect on time spent in tonic–clonic seizure activity was less at the lower doses of 1 mg/kg (52.6 ± 11.3, n = 7; p = 0.056) and 5 mg/kg (44.28 ± 10.4, n = 17; p = 0.002). There was no difference between groups in the number of myoclonic jerks exhibited during hypoxia. Using the data from all the doses, linear regression analysis of the percentage inhibition of tonic–clonic seizure activity yielded a median effective dose (ED50) of 0.57 mg/kg (SigmaPlot 9.0, SYSTAT Software, Richmond, CA, U.S.A.) (Fig. S1).
Talampanel attenuates later-life seizure-induced neuronal injury following hypoxia-induced seizures
We compared differences in status-induced neuronal injury in rats at P30–31 with prior hypoxic seizures at P10 treated with vehicle or the optimal talampanel dose of 7.5 mg/kg, and naive P30–31 litter mates with no prior hypoxic seizures. Both groups of rats were treated with kainate (10 mg/kg, i.p.) at P30–31, and all animals reached grade IV status epilepticus. Mean latency to onset of seizures was 23.1 ± 2.2 min (±SEM) across all groups, without differences between treatment groups. However, there were differences in the extent of injury between the naive controls, vehicle-, and talampanel-treated hypoxic seizure rats when examined at 72 h after the “second-hit” kainate seizure. The total number of ISEL-positive cells in sections from stereotactically identical regions was compared between naive controls, vehicle-treated, and talampanel- (7.5 mg/kg) treated rats. Naive control litter mate rats showed cell death in hippocampal CA1 and CA3 regions (Figs 2A–2D), as well as throughout basal amygdala (basolateral and basomedial amygdaloid nucleus) (Figs 2E–F). Rats pretreated with vehicle prior to hypoxia at P10 showed significantly more (4-fold) combined neuronal injury in amygdala, CA1, and CA3 following kainate-induced seizure at P30–31 (105.8 ± 30.4, n = 14) compared to naïve controls (20.1 ± 2.2 n = 13; p < 0.001) (Fig. 3). Talampanel treatment (7.5 mg/kg) prior to hypoxia at P10 resulted in significant attenuation of cell death in these regions (41.3 ± 16.8, n = 15; p < 0.001) compared to vehicle treatment, representing about a 60% decrease. Notably, the talampanel-treated animals had injury counts comparable to those of naive controls (Fig. 3).
Assessment of cell death after talampanel administration to normal P10 rats
To determine whether the optimal dose of talampanel caused acute cell death at P10, normal P10 male Long-Evans rats were given one dose of talampanel (i.p.) or vehicle at 7.5 mg/kg and perfused 24 h later at P11. Consistent with prior descriptions of constitutive apoptosis during development (Ikonomidou et al., 1999; Bittigau et al., 2002), there was some scattered baseline cell death in the vehicle-treated P11 rats upon inspection of three stereotactically chosen areas. Over the total coronal section in vehicle-treated pups, 24.9 ± 2.8 (±SEM) ISEL-positive cells were counted per 12.5 mm2 (n = 10 pups) (Fig. 4). No significant difference was observed in the ISEL-positive cell counts in the rodent brain sections following a single dose of talampanel (7.5 mg/kg) [28.8 ± 5.0 (±SEM), n = 10] (Fig. 4). These counts are similar to those reported in our other studies in control groups from animals of this age (Koh & Jensen, 2001).
We and others have previously shown that AMPARs play a critical role in the induction and potential epileptogenic effects of neonatal seizures. Using our model of hypoxia-induced neonatal seizures in the P10 rat, we show here that the noncompetitive AMPAR antagonist talampanel (GYKI-53773) effectively suppresses seizures in a dose-dependent manner. In addition, early talampanel treatment also prevents the enhanced susceptibility to later-life seizure-induced neuronal injury that has been shown to occur following early life seizures. We also show that therapeutic doses of talampanel do not result in any increase in constitutive apoptosis at P10, consistent with other observations regarding the safety of AMPAR antagonists in the developing brain. Taken together, these findings suggest that the specific AMPAR antagonist talampanel may have potential for treatment of neonatal seizures.
Despite preclinical data suggesting efficacy of AMPAR antagonists in a variety of models of epilepsy and excitotoxicity in the developing brain, no AMPAR antagonist is available for clinical study in human infants. Both NBQX and topiramate are not available for parenteral administration, which would be a necessity in this population of infants in intensive care with variable gastrointestinal absorption. Talampanel is a selective antagonist for AMPAR, acting in a noncompetitive manner at an allosteric site referred to as the GYKI receptor (Solyom & Tarnawa, 2002). Doses of talampanel in adult trials are in the range of 1–4 mg/kg day, and prior adult mouse models have shown anticonvulsant efficacy at 8–24 mg/kg and neuroprotection against hypoxic/ischemic injury at 12 mg/kg. In this study we report acute seizure suppression in the immature rat brain that is dose dependent, maximal at 7.5 mg/kg, with an ED50 of 0.57 mg/kg. These data are in a similar range to that of the 8.0 mg/kg, i.p., dose of talampanel reported to be neuroprotective in a P7 rat model of excitotoxic injury (Vilagi et al., 2002).
In addition to acute anticonvulsant efficacy, talampanel treatment prior to hypoxic seizures also attenuated the long-term increases in susceptibility to seizure-induced neuronal injury seen in this model. Previous data reveals that adult rats with prior hypoxic seizures show a 2- to 4-fold increase in neuronal injury compared to naive rats. Both NBQX and topiramate administered to rat pups in a pre- or post-treatment paradigm, result in an approximate 60% decrease in the amount of neuronal injury seen after kainate seizures in adulthood in this group of sensitized rats (Koh & Jensen, 2001; Koh et al., 2004). NBQX is an AMPAR and kainate receptor antagonist, with preferential action at AMPARs, and topiramate acts as an AMPAR antagonist but has other potential mechanisms of action, such as GABA agonism, and Ca2+ and Na+ channel block (Shank et al., 2000). Both NBQX and topiramate have been shown to be effective at blocking long-term effects of early life seizures (Jensen et al., 1995; Koh & Jensen, 2001; Cha et al., 2002; Koh et al., 2004). Talampanel is a selective AMPAR antagonist, and early prehypoxia treatment results in a quantitatively similar protective profile (approximately 60% reduction in status-induced cell loss) to these other agents, supporting a specific role for AMPARs in sensitizing neuronal populations to later seizure-induced injury. Given the similarities in the pretreatment results, these data suggest that postseizure treatment trials with talampanel are warranted in this model. Talampanel may not only be effective as an anticonvulsant but, like topiramate, might exhibit antiepileptogenic efficacy in the developing brain. Indeed, we have recently reported that talampanel treatment following hypoxia-induced seizures reverses seizure-induced increases in network and neuronal excitability in hippocampal CA1 neurons in P10 rats (Rakhade et al., 2008).
The pattern of injury following “second hit” seizures has previously been reported to include hippocampal regions (Koh et al., 1999; Koh & Jensen, 2001; Cha et al., 2002; Stafstrom & Sasaki-Adams, 2003; Koh et al., 2004). This susceptibility to injury has been suggested to reflect seizure-induced changes that result in long-term hyperexcitabilty of hippocampal networks. Electrophysiologic recordings of hippocampal neurons reveal both network hyperexcitability and a decrease in GABAergic inhibition on pyramidal neurons in area CA1 (Swann et al., 1989; Brooks-Kayal et al., 1998; Sanchez et al., 2001; Khazipov et al., 2004; Sanchez et al., 2005; Ben Ari et al., 2007; Rakhade et al., 2008). Here we show that talampanel suppression of acute seizure activity results in protection from these long-term increases seen in the fifth postnatal week of life. Another novel finding is that early life hypoxia-induced seizures alter long-term susceptibility to seizure-induced neuronal injury in the amygdala, and that talampanel also is protective against this long-term change. The observation that there is increased sensitivity to later seizure-induced injury in the amygdala is potentially relevant to clinical data suggesting that early life seizures and asphyxia may predispose to amygdala dysfunction. Early life injury to limbic structures such as the amygdala are thought to increase the risk of mental disorders such as autism and schizophrenia (Daenen et al., 2002; Diergaarde et al., 2004; Shaw et al., 2004).
Recent studies have cautioned against the use of anticonvulsant drugs and certain glutamate receptor antagonists in early postnatal life. NMDARs are essential for normal synaptogenesis and plasticity, and exposure to NMDAR antagonists during development results in increased constitutive apoptosis as well as later-life deficits in neurobehavior and learning (Tandon et al., 1996; Bittigau et al., 2002). A number of the conventional anticonvulsants have also been shown to affect later-life learning (Wirrell, 2005; Kim et al., 2007) and phenobarbital, carbamazepine, valproate, and diazepam have all been reported to increase neuronal apoptosis, even after one administration in early postnatal development in the rodent (Bittigau et al., 2002). Notably, the AMPAR antagonists NBQX and topiramate do not show this increased apoptosis. In addition, AMPAR antagonists appear to have a favorable safety profile in the developing brain, without inducing apoptosis (Bittigau et al., 2002) or affecting later-life learning and behavior (Zhao et al., 2005). Likewise, we show here that talampanel treatment at P10 with the effective dose does not affect this normal developmental phenomenon of apoptosis. Future studies could address additional potential effects on behavior and learning in this model.
In summary, this study supports further investigation of talampanel in the prevention and possibly the long-term effects of neonatal seizures. Here we show that talampanel treatment appears to have similar efficacy to that previously demonstrated for NBQX and topiramate. In addition, talampanel does not alter the normal pattern of apoptosis present in early development. Another finding of note in this study is that early life hypoxia-induced seizures alter susceptibility of amygdala to later-life seizure-induced injury, and this susceptibility to neuronal injury is attenuated by talampanel. A parenteral formulation of talampanel would have excellent potential for translation to the clinic for the indication of neonatal seizures. Given the present data regarding efficacy and safety, additional preclinical trials are justified to evaluate the more clinically relevant postseizure treatment efficacy, as well as additional safety trials to assess more subtle adverse effects on neuromotor development.
This work was supported by NIH/NINDS RO1 NS31718 (FEJ) and a grant from the Epilepsy Therapy Development Project/Epilepsy Foundation (FEJ and MF), in addition to core support from the Mental Retardation Research Center Grant (NIH NICHHD P30 HD18655). We also acknowledge the generous gift of talampanel compound and product information from Teva Neuroscience.
We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.
Disclosures: Dr. Jensen received a grant from the Epilepsy Therapy Development Project and the Epilepsy Foundation for this work. Dr. Fetell is an employee of Teva Neuroscience, and provided the compound to Dr. Jensen for study. P. Aujla has no conflicts to disclose.