Lacosamide treatment following status epilepticus attenuates neuronal cell loss and alterations in hippocampal neurogenesis in a rat electrical status epilepticus model

Authors


Summary

Purpose

The antiepileptic drug, lacosamide, exerts its therapeutic activity by enhancing slow inactivation of voltage-gated sodium channels. Because putative preventive or disease-modifying effects of drugs may affect epileptogenesis, intrinsic severity, and comorbidities, it is of particular interest to assess the effect of lacosamide on the development of epilepsy and associated cellular alterations.

Methods

The effect of lacosamide was evaluated in an electrical rat status epilepticus (SE) model with a 24-day treatment phase following induction of SE. The impact of lacosamide on the development of spontaneous seizures based on continuous video–electroencephalography (EEG) monitoring, as well as the impact on neuronal cell loss and alterations in hippocampal neurogenesis, was assessed.

Key Findings

Neither low-dose nor high-dose lacosamide affected the development of spontaneous seizures. A dose-dependent neuroprotective effect of lacosamide with significant reduction of neuronal cell loss was observed in the hippocampal CA1 region, as well as in the piriform cortex. In addition, lacosamide attenuated the impact of SE on the rate of hippocampal cell neurogenesis. Moreover, lacosamide prevented a significant rise in the number of persistent basal dendrites.

Significance

Our data do not support an antiepileptogenic effect of lacosamide. However, because lacosamide reduced SE-associated cellular alterations, it would be of interest to determine whether these effects indicate a putative disease-modifying effect of lacosamide in future studies.

Lacosamide (LCM) is an antiepileptic drug (AED) widely used as adjunctive therapy for patients with refractory partial-onset seizures (Prunetti & Perucca, 2011; Stephen & Brodie, 2011). Electrophysiologic studies have indicated that LCM modulates voltage-gated sodium channels in a novel manner, selectively enhancing their slow inactivation with no effect on fast inactivation (Beyreuther et al., 2007; Errington et al., 2008). Based on this mechanism of action, LCM appears to regulate the long-term availability of sodium channels, thereby controlling pathophysiologic neuronal hyperexcitability (Beyreuther et al., 2007; Stephen et al., 2011). Contradictory findings exist regarding the putative interaction of LCM with collapsin response mediator protein 2 (CRMP-2). Although drug-binding assays clearly argue against specific binding of LCM to CRMP-2, another study showed that mutations in CRMP2 reduce the effects of LCM on neurite outgrowth (Wilson et al., 2012; Wolff et al., 2012).

AEDs are generally developed for symptomatic suppression of seizure activity. However, drugs with preventive and disease-modifying effects are of particular interest for several reasons; they may exert antiepileptogenic effects, reduce intrinsic disease severity, improve pharmacosensitivity, and exert beneficial effects on comorbidities such as cognitive dysfunction (Loscher & Brandt, 2010; Pitkanen, 2010; Pitkanen & Lukasiuk, 2011). By attenuating the cellular consequences of brain insults, the test compound may exert beneficial effects during epileptogenesis (Loscher & Brandt, 2010). In particular, neuronal cell loss and alterations in hippocampal neurogenesis post insult might be linked to epilepsy-associated impairment of learning and memory, and behavioral disturbances (Pekcec et al., 2008; Siebzehnrubl & Blumcke, 2008; Loscher & Brandt, 2010; Danzer, 2011; Pekcec et al., 2011).

Experimental data indicate that LCM might possess disease-modifying effects. In a rat-kindling paradigm, LCM delayed kindling progression (Brandt et al., 2006), whereas in an electrical rat status epilepticus (SE) model, LCM administered 10 or 40 min after stimulation of the perforant path reduced the development of spontaneous seizures (Wasterlain et al., 2011). However, early administration following onset of SE in this study might have resulted in an attenuation of the initial insult, thus confounding the interpretation of a true antiepileptogenic effect.

This study was designed to test the antiepileptogenic and disease-modifying effects of LCM by administering it at a later stage of SE, when the duration of continuous seizure activity is sufficient to trigger epileptogenesis. In addition, the effect of lacosamide on neurodegeneration in vulnerable hippocampal subregions, as well as on the number and fate of newborn granule cells, was evaluated.

Material and Methods

Female Sprague-Dawley rats with a body weight of 200–220 g (Harlan Winkelmann, An Venray, The Netherlands) were allowed to adapt to their new environment for at least 1 week before start of experiments. They were kept under controlled environmental conditions (20–22°C, 37–40% humidity, 12 h light/dark cycle) with free access to standard laboratory chow and tap water. All experiments were conducted in compliance with the European Communities Council Directive of 24 November 1986 (86/609/EEC) and following approval by the Government of Upper Bavaria. All efforts were made to minimize pain or discomfort of the animals and to reduce the number of animals.

Electrode implantation

A total of 125 animals were anesthetized with chloral hydrate (360 mg/kg, intraperitoneal [i.p.]) for chronic implantation of a Teflon-isolated bipolar stainless steel stimulation and recording electrode into the right basolateral amygdala (BLA) as described previously (Pekcec et al., 2007). The stereotactic coordinates in mm relative to Bregma according to the atlas of Paxinos and Watson (2005) were ap −2.2, lat −4.7, dv −8.4. Meloxicam (Metacam 5 mg/ml; Boehringer Ingelheim Vetmedica GmbH, Ingelheim/Rhein, Germany; 1 mg/kg) was administered subcutaneously (s.c.) 30 min prior and 24 h following surgery. In addition, bupivacaine 0.5% was applied s.c. for local anesthesia. One screw, placed above the left parietal cortex, served as the indifferent reference electrode. Additional skull screws and dental acrylic cement anchored the entire headset. After surgery, the animals were allowed to recover for at least 6 weeks. Antibiotics (0.1 ml per animal, Marbofloxacin, Marbocyl; Vétoquinol, Ravensburg, Germany) were administered s.c. twice daily (8.00 a.m. and 5.00 p.m.) from the day before surgery until day 7 postsurgery to prevent infection.

Animals were assigned randomly to undergo SE induction or not; within the two groups, animals were further randomized to receive low-dose LCM, high-dose LCM, or vehicle, resulting in six treatment groups in total.

Electrical induction of self-sustained status epilepticus

For induction of self-sustained SE, rats (n = 70) were electrically stimulated for 25 min via the BLA electrode using an Accupulser A310C stimulator connected to a Stimulus Isolator A365 (World Precision Instruments, Berlin, Germany). The stimulation parameters were as follows: duration 25 min, 100 msec trains of 1 msec alternating positive and negative square-wave pulses. Trains were applied at a frequency of 2 Hz, and the intratrain pulse frequency was 50 Hz. Peak pulse intensity was 700 μA. Following stimulation, development and duration of SE were monitored by 5-min electroencephalography (EEG) recording and by behavioral observation for 3.35 h. Seizure severity was classified according to Racine (1972) and was used as a basis for the classification of the SE type as described previously (Seeger et al., 2011). SE was limited to 4 h (including the 25 min of BLA stimulation) by i.p. injection of diazepam (Diazepam-ratiopharm; ratiopharm GmbH, Ulm, Germany) at a dose of 10 mg/kg body weight. For one animal (low-dose LCM treated SE) it was necessary to repeat the application of this dose until motor seizure activity was completely suppressed and the EEG was normalized. The time point was chosen based on a thorough evaluation of different experimental conditions in the model, which indicated that diazepam administered at this relatively late time point does not exert major effects on epileptogenesis (Brandt et al., 2003a). The rationale of our study design was to guarantee that there are no differences in the intensity of the initial insult between the groups. Rats exhibiting SE types II and III were distributed randomly to different treatment groups. Electrode-implanted rats without SE induction were not stimulated but were subjected to all handling and experimental procedures associated with the stimulation procedure, including the diazepam injection.

Treatment with lacosamide

LCM or vehicle was administered immediately after the diazepam injection in all animal groups (groups 1–6) and was continued for 23 days (Fig. 1). On the day of SE induction (day 0), all animals received twice daily oral administration of low-dose LCM (10 mg/kg), high-dose LCM (30 mg/kg), or vehicle (0.5% methylcellulose, 2 ml/kg) 5 min after diazepam administration and at 10:00 p.m. On days 1–4 post-SE, oral administration of LCM (low or high dose) or vehicle was repeated three times daily (8:00 a.m., 4:00 p.m., and 12:00 p.m.). In the morning of day 5 post-SE, animals received one oral administration of LCM or vehicle at 8:00 a.m. Subsequently, osmotic minipumps (Alzet osmotic pump, model 2ML2; DURECT Corporation, Cupertino, CA, U.S.A.) were subcutaneously implanted into the neck of the animals to allow continuous administration of LCM (0.73 mg/kg/h), LCM (2.1 mg/kg/h) or vehicle (N-methyl-2-pyrrolidone with phosphate buffer pH 7.4) for 2 weeks. For implantation of the osmotic minipumps, animals (n = 43) received a chloral hydrate anesthesia (360 mg/kg, i.p.). Because of problems with apnea during implantation of the minipumps, the type of anesthesia was changed to isoflurane anesthesia in the remaining 48 rats. All animals received bupivacaine 0.5% (1 ml/animal), s.c., for local anesthesia. The minipump had a reservoir volume of 2 ml and delivered solutions continuously for 14 days with a pumping rate of 4.68 ml/h. After 14 days of treatment by minipumps (days 5–19 post SE) the pumps were removed and treatment was continued for 5 days (days 19–23) with oral LCM (low or high dose) or vehicle three times daily (8:00 a.m., 4:00 p.m., and 12:00 p.m.). For plasma LCM analysis, blood was sampled by tail vein puncture (V. caudalis mediana, anesthesia with chloral hydrate or isoflurane) or in case of problems by retroorbital puncture (anesthesia with chloral hydrate or isoflurane) and local anesthesia with tetracaine hydrochloride (Opthocain N; Dr. Winzer Pharma GmbH, Berlin, Germany). Sampling took place at three different time points: on day 5 associated with the minipump implantation procedure, on day 19 associated with the removal of the minipumps, and on day 22 during the final oral administration phase. Plasma samples were sent to UCB, Braine L'Alleud, Belgium, for analysis of LCM concentrations. Mean plasma concentrations amounted to 3.58 ± 0.13 μg/ml (low-dose LCM, n = 17) and 10.55 ± 0.68 μg/ml (high-dose LCM, n = 19) (mean of three repeated analyses/animal used to calculate group means ± standard error of the mean [SEM]), which are within the therapeutic range of LCM (Biton et al., 2008; Cawello et al., 2010).

Figure 1.

Schematic illustration of the experimental protocol.

BrdU labeling

As a thymidine analog, 5-bromodeoxyuridine (BrdU; Sigma-Aldrich, Taufkirchen, Germany) is incorporated into the DNA during the S phase of the cell cycle and is available for approximately 30 min (Packard et al., 1973; del Rio & Soriano, 1989). Therefore, each injection of BrdU labels those proliferating cells that are in the DNA-synthetic phase of the cell cycle. Rats received twice daily i.p. injections of 50 mg/kg BrdU (in 3 ml/kg saline) on days 4–9 following SE induction (interval 8 h; 12 injections in total).

Monitoring of spontaneous seizures

Nine weeks following SE induction, the development of spontaneous seizures was assessed by continuous video-EEG monitoring over a period of 3 weeks (24 h/day for 7 days a week) using a combined full digital video- and EEG-detection system. To allow optimal video recordings, animals were housed in modified glass cages (40 × 40 × 40 cm) under controlled conditions (see above). For continuous day and night video monitoring, several infrared light sensitive cameras and a multiple-channel PCI analog-digital converter (ABUS Security-Tech, Affing, Germany) were used. The sampling rate was 25 pictures/s. Data analysis was performed using the Digi-Protect Searcher 6.275 beta software (ABUS Security-Tech). EEGs were simultaneously recorded using 1-channel bioamplifiers (BioAmps) and analog-digital converters (PowerLab/800s) (all from AD Instruments, Hastings, East Sussex, United Kingdom). Data acquisition and analysis was performed using the Chart5 for windows software (AD Instruments). Seizure frequency was evaluated in animals randomly chosen from each treatment group with SE: group 4, vehicle (n = 13); group 5, low-dose LCM (n = 12); group 6, high-dose LCM (n = 11). Due to limited capacity, animals were monitored in two phases (first phase 16 rats, second phase 20 rats). Seven animals with SE induction (group 4, vehicle-treated [n = 4]; group 5, low-dose LCM [n = 2]; group 6, high-dose LCM [n = 1]) lost their electrode during video-EEG monitoring and were excluded from further analyses (e.g., seizure frequency, seizure duration, immunohistochemistry).

Rats without SE induction (groups 1–3) were also transiently placed in the monitoring cages and handled in a comparable way (group 1, vehicle-treated [n = 9]; group 2, low-dose LCM [n = 10]; group 3, high-dose LCM [n = 9]). One electrode implanted rat treated with low-dose LCM lost its electrode and was therefore excluded from the analysis. For detection of generalized spontaneous seizures, EEG recordings were visually analyzed for characteristic seizure-like activity. For evaluation of the severity of an EEG-detected seizure, the corresponding video recording was viewed.

Tissue preparation and immunostaining

Following video-EEG monitoring, animals were deeply anesthetized with pentobarbital (500 mg/kg) and perfused transcardially with saline followed by 4% paraformaldehyde in 0.1 m phosphate-buffered saline (pH 7.4). The brains were removed and transferred into 30% sucrose and stored at 4°C until six series of coronal sections (40 μm) were cut using a freezing microtome. Sections were processed for immunohistochemical detection of BrdU/NeuN and doublecortin as described previously (Seeger et al., 2011).

Histologic and immunohistologic evaluation

Thionin-stained sections of the CA1, CA3, and piriform cortex subregions of the hippocampal formation were inspected visually for damage. Approximately 14 sections of the hippocampus (located between −2.4 and −5.8 mm posterior to bregma at 240 μm intervals) were analyzed from each rat, and both the ipsilateral (right) and contralateral hemispheres (with respect to the BLA electrode) were inspected. Severity of neuronal damage was assessed semiquantitatively using a grading system as described previously (Brandt et al., 2003b; Seeger et al., 2011): score 0, no obvious damage; score 1, slight lesions involving <20% of neurons; score 2, lesions involving 20–50% of neurons; score 3, lesions involving >50% of neurons. It is noteworthy that neuronal loss must exceed 15–20% before it is reliably detected by visual inspection (Fujikawa, 2005). In each rat and for each area the highest score (ipsi- and contralateral hemispheres considered separately) was given as a value for neuronal damage. The number of thionin-positive cells in the dentate hilus, the number of doublecortin-labeled cells in the dentate gyrus, as well as the number of persistent basal dendrites extending into the hilus from the granule cell layer were quantified by unbiased stereologic analysis using the computer-assisted imaging system StereoInvestigator 6.0 (Microbrightfield Europe, Magdeburg, Germany) as described previously (Unkruer et al., 2009). An experimenter not aware of the treatment conditions performed the counting of cells using the optical fractionator method. The number of doublecortin-labeled cells was quantified by stereologic analysis in six sections per animal. The values were summed up for each set of sections. About 14 sections per animal were used to assess the severity of neuronal damage in thionin-stained sections.

Cell counts of fluorescent signals from double-staining for BrdU and NeuN were performed at 40× magnification in an area encompassing the entire dentate granule cell layer (superior and inferior blades) and extending approximately two cell bodies deep into the hilus. The total number of BrdU and BrdU/NeuN double-labeled cells per animal was summed up for each set of sections (six sections per animal with 240 μm interval). Double-labeling was verified by careful analysis of the confocal z-series of multiple cells per animal. Double-labeled cells of the dentate hilus of the hippocampal formation were counted separately, but using the same method and the same sections. The hilus was defined as the inner border of the granule cell layer and two straight lines connecting the tips of the granule cell layer and the proximal end of the CA3 region.

To ensure that the antibody actually penetrated the entire section, the free floating method was used, and distribution was verified by microscope z-scanning.

Statistics

Depending on whether data were normally distributed or not, either parametric or nonparametric tests were used for statistical evaluation. One-way analysis of variance (ANOVA) was used for calculation of significant differences in spontaneous seizures (seizure frequency and seizure duration) in the three different treatment groups of rats with SE (vehicle, low-dose LCM, high-dose LCM), followed by testing for individual differences by the Student′s t-test. Data from neurodegeneration scores were analyzed using two-way ANOVA for nonparametric data, followed by testing for differences between subgroups by the Mann–Whitney U-test. Statistical analyses of group differences in cell counts of thionin-, DCX- and BrdU-positive cells and BrdU/NeuN double-labeled cells were performed using two-way ANOVA followed by testing for individual differences in subgroups by the Student′s t-test. All tests were used two-sided; a p < 0.05 was considered as significant.

One animal in the low-dose LCM treatment group with SE induction was excluded from all analyses based on the 3-sigma-rule of outliers (spontaneous seizures in video-EEG monitoring: n = 285). Two further animals without SE induction were excluded due to inconsistent findings regarding LCM plasma concentrations (group 1: vehicle-treated [n = 1], group 2: low-dose LCM [n = 1]).

Results

Development of spontaneous seizures

Considering handling-associated seizures, the mean latency to first chronic seizures reached 29.14 + 4.75 days in the control group. No significant differences were evident between the groups. In the 3-week monitoring phase, spontaneous seizures were detected in 8 of 9 vehicle-treated rats, 8 of 10 treated with low-dose LCM, and 8 of 10 treated with high-dose LCM. Most detected seizures were generalized. The three treatment groups did not differ in the number of seizures observed during the 3-week monitoring phase (vehicle 5.00 ± 1.44; low-dose LCM 7.13 ± 2.27; high-dose LCM with SE: 6.38 ± 2.65). As mentioned, one animal in the low-dose LCM group was an outlier due to high seizure density (285 seizures) during the monitoring phase.

The mean duration of spontaneous seizures was not affected by low- or high-dose LCM treatment (vehicle 54.81 ± 2.92 s; low-dose LCM 61.04 ± 6.28 s; high-dose LCM 58.31 ± 2.98 s).

In addition, the number of handling-associated seizures was analyzed during the time frame between SE and the monitoring phase. The number of handling-associated seizures was very low during the treatment phase, that is, weeks 1–3 following SE without any difference between vehicle and LCM-treated groups (vehicle 0.56 ± 0.24, low-dose LCM 0.78 ± 0.43, high-dose LCM 0.70 ± 0.34). Moreover, no differences among groups were identified in the number of handling-associated seizures counted between the treatment phase and start of monitoring, that is, weeks 4–8 following SE (vehicle 1.56 ± 0.77, low-dose LCM 1.78 ± 0.60, high-dose LCM 2.30 ± 1.01).

Impact of lacosamide on hippocampal neurodegeneration and neurogenesis

Neurodegeneration scores in the CA1 and CA3 (Fig. 2A–C) regions of the hippocampus, as well as in the piriform cortex (Fig. 2D–F), confirmed the typical SE-associated neuronal cell loss in vehicle-treated rats (with induction of SE) (Fig. 2G–I). Administration of a low-dose of LCM during the weeks following SE did not affect neuronal cell loss in a significant manner. In contrast administration of high-dose LCM exerted neuroprotective effects. In the CA1 region (Fig. 2G) and piriform cortex (Fig. 2I), high-dose LCM prevented neuronal cell loss in a significant manner.

Figure 2.

Representative thionin-stained sections of the hippocampal formation (AC) and the piriform cortex (DF) of a vehicle-treated rat without SE induction (A, D), a vehicle-treated rat with SE induction (B, E), and a high-dose LCM-treated rat with SE induction (C, F) (scale bar = 200 μm). The graphs (GI) illustrate the neurodegeneration score in thionin-stained sections assessed using a grading system of vehicle-treated (n = 8), low-dose LCM-treated (n = 8), and high-dose LCM-treated (n = 9) electrode-implanted rats without SE induction (non-SE) as well as vehicle-treated (n = 9), low-dose LCM-treated (n = 9), and high-dose LCM-treated (n = 10) rats with SE induction. Data are shown separately for the CA1 (G) and CA3 subregions (H) of the hippocampus and the piriform cortex (I). Data are shown as means ± SEM. Significant differences from post hoc analysis between non-SE and SE animals are indicated by asterisks (compared with respective groups without SE induction) or # (within the non-SE or SE group) (two-way ANOVA, post hoc Mann-Whitney U-test, p < 0.05).

Neuronal cell density was analyzed in the hilus of the hippocampus. However, unbiased stereologic cell counts did not reveal SE-associated neuronal loss in this hippocampal subregion (vehicle non-SE 3,145 ± 146; low-dose LCM non-SE 3,471 ± 124; high-dose LCM non-SE 3,441 ± 104 and vehicle with SE 2,996 ± 222; low-dose LCM with SE 3,031 ± 319; high-dose LCM with SE 3,250 ± 186).

Impact of lacosamide on neuronal progenitor cells and hippocampal neurogenesis

We analyzed the number of BrdU- as well as BrdU/NeuN labeled cells in the granule cell layer to obtain data that reflect cell proliferation during the time of administration (BrdU injected twice daily on days 4–9) as well as neuronal differentiation and survival of newborn neurons.

The number of BrdU-labeled cells detected in the granule cell layer (Fig. 3A–C) of vehicle-treated rats with SE significantly exceeded that in vehicle-treated rats without SE by 653% (Fig. 3E). Analyses of BrdU/NeuN-labeled cells (Fig 3D) in rats treated with vehicle revealed the typical SE-associated increase in hippocampal neurogenesis (Fig. 3F). Low-dose and high-dose LCM attenuated both the SE-associated increase in newborn granule cells in a significant manner.

Figure 3.

(AC) show representative BrdU/NeuN-immunostained sections of the hippocampal granule cell layer of a vehicle-treated rat without SE induction (A), a vehicle-treated rat with SE induction (B) and a high-dose LCM-treated rat with induction of SE (C) (BrdU in red, NeuN in green; scale bar = 100 μm). (D) shows double-labeling verification by carefully performed confocal z-series (scale bar = 5 μm). The graphs illustrate the number of BrdU-positive cells in the hippocampal granule cell layer (E) and the number of BrdU/NeuN double-labeled cells in the hippocampal granule cell layer (F) in vehicle-treated (n = 8), low-dose LCM-treated (n = 8), and high-dose LCM-treated (n = 9) electrode-implanted rats without SE induction (non-SE) as well as in vehicle-treated (n = 9), low-dose LCM-treated (n = 9) and high-dose LCM-treated (n = 10) rats with induction of SE. Data are shown as means ± SEM. Significant differences from post hoc analysis between non-SE and SE animals are indicated by asterisks (compared to respective groups without SE induction) or # (within the non-SE or SE group), p < 0.05).

SE can trigger an aberrant migration of newborn neurons into the hilus. A respective rise in the number of new hilar neurons was demonstrated as a consequence of SE. Moreover, the number of BrdU-labeled cells was increased by SE in the hippocampal hilus. Neither low- nor high-dose LCM exerted significant effects on the number of hilar BrdU- and BrdU/NeuN-labeled cells (number of hilar BrdU-labeled cells: vehicle non-SE 2.13 ± 1.17; low-dose LCM non-SE 2.63 ± 0.78; high-dose LCM non-SE 2.56 ± 0.71 and vehicle with SE 17.33 ± 3.92; low-dose LCM with SE 13.11 ± 3.45; high-dose LCM with SE 12.30 ± 3.53; number of hilar BrdU/NeuN-labeled cells: vehicle non-SE 1.13 ± 0.61; low-dose LCM non-SE 1.00 ± 0.50; high-dose LCM non-SE 1.11 ± 0.26 and vehicle with SE 10.33 ± 1.85; low-dose LCM with SE 9.00 ± 2.53; high-dose LCM with SE 7.60 ± 1.63).

We additionally assessed the number of doublecortin-labeled cells, which reflect cell proliferation and neuronal differentiation in the chronic phase and allow conclusions about long-term effects of LCM on cells that have probably been born during the last 2 weeks before taking the brain samples (animals were sacrificed in week 12 following SE, i.e., 9 weeks following end of treatment). The number of doublecortin-labeled cells can additionally be affected by disease-associated alterations in the speed of development of newborn cells.

Twelve weeks following SE induction, a high number of doublecortin-labeled cells were detected in the subgranular zone and granule cell layer (Fig. 4A–C) of vehicle and low-dose LCM-treated rats significantly exceeding that in respective control groups (Fig. 4E). As another long-term consequence of SE, the number of persistent basal dendrites extending into the hilus (Fig. 4D) was increased significantly by 597% and 635% in animals receiving vehicle or low-dose LCM administration (Fig. 4F). Please note that we analyzed basal dendrites based on doublecortin labeling, because this marker perfectly shows the neurites, but one needs to consider that doublecortin is transiently expressed during development of new neurons. As we recently reported, genetic fate mapping might offer a novel perfect method to assess the long-term outcome in more detail in the future (Jafari et al., 2012).

Figure 4.

Representative doublecortin-immunostained sections of the hippocampal formation of a vehicle-treated rat without SE induction (A), a vehicle-treated rat with SE induction (B), and a high-dose LCM-treated rat with SE induction (C) (scale bar 100 μm). (D) illustrates a high magnification image of doublecortin-labeled cells with persistent basal dendrites in the granule cell layer (GCL) of a vehicle-treated rat with SE induction. The cell bodies are positioned at the border of the GCL and the subgranular zone (SGZ) and have long basal dendrites (black arrows) extending deep into the hilus (H). The apical dendrites extend into the GCL (scale bar = 20 μm). (E) illustrates the number of doublecortin-expressing cells in the rat dentate gyrus of vehicle-treated (n = 8), low-dose LCM-treated (n = 8), and high-dose LCM-treated (n = 9) rats without induction of SE (non-SE) as well as of vehicle-treated (n = 9), low-dose LCM-treated (n = 9), and high-dose LCM-treated (n = 10) rats with induction of SE. (F) represents the proportion of doublecortin-positive cells with persistent basal dendrites arising from the dentate gyrus and extending into the hilus of vehicle-treated (n = 8), low-dose LCM-treated (n = 8), and high-dose LCM-treated (n = 9) non-SE rats as well as of vehicle-treated (n = 9), low-dose LCM-treated (n = 9), and high-dose LCM-treated (n = 10) rats with induction of SE. Data are shown as means ± SEM. Significant differences from post hoc analysis between non-SE and SE animals are indicated by asterisks (p < 0.05).

High-dose LCM treatment prevented the SE-induced long-term alterations in the number of doublecortin-expressing neuronal progenitor cells, early postmitotic neurons, and persistent basal dendrites (Fig. 4E,F).

Discussion

Whereas neither low-dose nor high-dose LCM affected the development of spontaneous seizures, LCM treatment initiated following SE exerted pronounced effects on the cellular consequences. A considerable part of brain damage develops following initial insults, with delayed types of cell death contributing to neuronal cell loss (Fujikawa, 2005; Loscher & Brandt, 2010). Therefore, there is a window of opportunity for the application of neuroprotective strategies in patients with brain insults. Moreover, repeated single seizures as well as seizure clusters can also contribute to progressive neuronal cell loss and sclerosis, thereby aggravating the brain pathology and its consequences in patients with chronic epilepsy (Fujikawa, 2005). Therefore, it is of particular interest to determine whether novel antiepileptic drugs possess relevant properties that can protect neurons from seizure-associated damage.

A dose-dependent neuroprotective effect of LCM was demonstrated in the hippocampal CA1 region and the piriform cortex. Several experimental studies have already argued against a close functional link between neuronal cell loss and epileptogenesis, demonstrating that several compounds protect neurons without prevention of seizures in rodent models (Brandt et al., 2003a, 2006; Andre et al., 2007; Nehlig, 2007). In line with these studies, neuroprotection by LCM was not associated with a relevant impact on epileptogenesis. Regarding the mechanism of LCM contributing to its neuroprotective effects, the enhancement of slow inactivation of sodium channels might be considered as a contributing factor. Targeting voltage-gated sodium channels is considered as a general concept for neuroprotection in different central nervous system (CNS) diseases (Gribkoff & Winquist, 2005). In line with this concept, a neuroprotective effect has been demonstrated for different AEDs that promote fast inactivation of voltage-gated sodium channels (Loscher & Brandt, 2010). The present data indicate that prevention of seizure-induced neurotoxic effects can also be achieved using compounds affecting slow inactivation of these cation channels. Although an interaction of LCM with CRMP2 might also contribute, as noted in the introduction, drug-binding assays did not reveal specific binding of LCM to CRMP2 (Wilson et al., 2012; Wolff et al., 2012).

Regarding the differences between LCM effects in the CA1 and CA3 region, we can only speculate about the mechanisms. Recent studies evaluating the gene activation in CA1 and CA3 neurons have demonstrated that CA1 and CA3 neurons show distinct gene expression patterns (Yoshioka et al., 2012). Moreover, images showing immunoreactivity of voltage-gated sodium channel subtypes in rat hippocampal neuronal cell cultures derived from control and epileptic rats indicate that this might also apply to voltage-gated sodium channel subtypes (Guo et al., 2013). However, definite conclusions require further quantitative investigations. Functional differences associated with differences in the expression pattern might have contributed to the differential effects of LCM on CA1 and CA3 neuronal damage and loss. Further studies would be necessary to definitely elucidate the exact mechanisms.

Of interest, we did not observe any SE-associated neuronal loss in the hilus of the hippocampus. This finding is consistent with recent data from our group and other groups, indicating a surprising resistance of the hilus to seizure-induced neuronal cell loss (Loscher & Brandt, 2010; Seeger et al., 2011). It has been hypothesized that the reduced sensitivity of hilar neurons observed in recent studies compared with earlier studies is related to a genetic drift in the rat strain used, also considering that animal facilities and breeding sites have been reorganized at Harlan Winkelmann (Loscher & Brandt, 2010).

In addition to its neuroprotective effects, LCM significantly attenuated the impact of SE on hippocampal neurogenesis. Several studies have demonstrated that the induction of prolonged or repeated seizure activity in naive animals results in a pronounced rise in the formation and integration of new granule cells into the hippocampal dentate gyrus (Parent & Murphy, 2008; Scharfman & McCloskey, 2009). Both, low-dose and high-dose LCM significantly attenuated the SE-associated increase in hippocampal neurogenesis. This impact of LCM treatment on the number of BrdU/NeuN-double-labeled cells can be due to an effect on the birth of new cells, and their survival and neuronal differentiation. Analyses of the number of neuronal progenitor cells at the end of the experiment, that is, 12 weeks following SE and 9 weeks following end of treatment, revealed long-lasting effects of high-dose LCM administration. Whereas SE resulted in a number of doublecortin-labeled progenitor cells exceeding that in respective control animals, the number of neuronal progenitor cells was in the control range in animals treated with high-dose LCM. The findings might be of interest, as we recently demonstrated that partial normalization of neurogenesis in a post-SE model can be associated with relevant attenuation of long-term cognitive deficits (Pekcec et al., 2008). Therefore, future studies might be of interest to evaluate the functional consequences of LCM treatment on cognitive development following brain insults.

The increase in the number of doublecortin-positive cells in the chronic phase is in contrast to findings from several other studies in post-SE models. The number of neuronal progenitor cells was often described to be unchanged or decreased in the chronic phase (Hattiangady & Shetty, 2008, 2010). However, the discrepancy might reflect differences between the models used. In particular, the severity of epilepsy, seizure frequency, and the degree of neuroinflammation could have a major impact.

Seizure activity in rodents can enhance persistent extension of granule cell basal dendrites into the hilus instead of the molecular layer (Buckmaster & Dudek, 1999; Shapiro et al., 2008). These basal dendrites are considered as a potential route for recurrent excitation among granule cells (Shapiro et al., 2008). Therefore, it is of interest that high-dose LCM efficaciously prevented a significant rise in the number of persistent basal dendrites in animals with SE. It needs to be emphasized that this effect was observed 9 weeks following the end of LCM treatment.

Based on electrophysiologic recordings and morphologic analyses of newborn granule cells in chronic epilepsy models, it has been suggested that the impact of brain insults on neurogenesis as well as on the fate of the newborn cells might have functional implications on epileptogenesis (Parent & Murphy, 2008; Scharfman & McCloskey, 2009; Danzer, 2011; Kokaia, 2011). However, the lack of any impact of LCM on seizure development on the one hand, and the efficacious modulation of neurogenesis on the other, argues against a central role of brain insult–induced alterations of neurogenesis for epileptogenesis. The findings are in line with those of recent experimental studies, in which different strategies efficaciously modulated neurogenesis but failed to affect epileptogenesis or progression of seizures in a kindling paradigm (Pekcec et al., 2007, 2008, 2011). Regarding conclusions from the present study, it needs to be taken into consideration that the aberrant migration of newborn granule cells into the hilus was not affected by LCM in a significant manner. Recordings from hilar granule cells in chronic epilepsy models have indicated that these cells might also contribute to hyperexcitability (Scharfman et al., 2000).

Neither the number of spontaneous recurrent seizures nor their severity or duration was affected by LCM treatment. Therefore, the data do not support an antiepileptogenic effect of LCM. In contrast, subchronic treatment with LCM during kindling acquisition exhibited a dose-dependent effect with significant retardation of kindling at 10 and 30 mg/kg daily doses (Brandt et al., 2006). Based on these data, an antiepileptogenic or disease-modifying potential was suggested for LCM; however, it was also pointed out that this observation needed to be further evaluated in other models of acquired epilepsy. The discrepancy between the findings might reflect differences in the predictive validity of the kindling and the post-SE model. In general, testing for an impact on the progression of kindled seizures more often reveals positive results compared with testing for prevention of spontaneous seizures in a post-SE model (Loscher & Brandt, 2010). In a previous study using a post-SE model with prolonged perforant path stimulation, a single LCM injection resulted in a dose-dependent reduction in the number of spontaneous seizures (Wasterlain et al., 2011). However, as mentioned above, LCM was administered 10 or 40 min following onset of SE, and the SE was not limited in duration in the control group. Therefore, it is questionable whether the effects observed were based on insult modulation or a true antiepileptogenic effect. On the other hand, one also has to consider the duration of treatment, the relatively low baseline seizure frequency, and the duration of seizure monitoring in the present study. In this context, we cannot exclude the possibility that prolonged treatment and/or monitoring phases might have revealed evidence for attenuating effects on epileptogenesis. However, it is important to note that additional analyses of handling-associated seizures in the time frame between treatment and start of monitoring also argues against a relevant antiepileptogenic effect.

In conclusion, although our data do not support an antiepileptogenic effect of LCM, its use was associated with a reduction in SE-related cellular alterations. Whether these effects indicate a putative disease-modifying effect of lacosamide remains to be determined. Moreover, it would also be of interest to narrow down the relevant time window by evaluating the consequences of shortened treatment periods.

Acknowledgments

The authors thank Marion Fisch, Angela Vicidomini, and Andrea Wehmeyer for their technical assistance, and Dr. Pierre Boulanger and the UCB Bioanalytical Team for analysis of LCM plasma concentrations. The authors would also like to thank Azita Tofighy, Ph.D. from UCB Pharma for editorial assistance.

Disclosure

Heidrun Potschka and her group have received fees and funding from different pharmaceutical companies including UCB Pharma for talks, consulting, and scientific collaborations. Alain Matagne is full employee of UCB Pharma. 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.

Ancillary