Effects of early long-term treatment with antiepileptic drugs on development of seizures and depressive-like behavior in a rat genetic absence epilepsy model



This article is corrected by:

  1. Errata: Erratum Volume 53, Issue 5, 933, Article first published online: 25 April 2012

Address correspondence to Giovambattista De Sarro, Chair of Pharmacology, Department of Experimental and Clinical Medicine, School of Medicine, University of Catanzaro, Via T. Campanella, 115 88100 Catanzaro, Italy. E-mail: desarro@unicz.it


Purpose: Depression is most commonly associated with epilepsy. Recent reports have suggested a putative relationship between seizure development and onset of depressive behavior, whereas others proposed that two clinical entities might represent different neuropathologic aspects of the same neurologic disorder. The WAG/Rij rat absence epilepsy model has also been proposed as a suitable model to test antidepressant drugs. We previously reported on a long-term study of two antiepileptic drugs (AEDs) to assess their protective role in absence epileptogenesis. Here, we examined the effects of long-term treatment with several AEDs on absence seizure development and onset of depressive-like behavior in WAG/Rij rats at different ages, using a forced swimming test (FST).

Methods: Animals were divided into one untreated control group and four test groups, given ethosuximide, levetiracetam, zonisamide, or carbamazepine. Electroencephalography (EEG) readings were recorded at 6.5 months of age.

Key Findings: Ethosuximide-treated animals showed significant reductions in recorded spike-wave discharges (SWDs), and FST immobility time (IT) compared with untreated same age controls. However, zonisamide- and carbamazepine-treated animals had IT values similar to those of controls, but only zonisamide significantly decreased absence seizure development. Carbamazepine increased SWD incidence. Levetiracetam also protected against seizure development, while augmenting IT, suggesting a prodepressive effect.

Significance: Although treatment with ethosuximide, levetiracetam, or zonisamide reduced appearance of SWDs in WAG/Rij rats, this was not generally linked to a reduced onset of depressive characteristics, as assessed by FST. Therefore, expression of depressive-like behavior seems unrelated to seizure control in this model. Some possible alternative explanations for the observed data are discussed.

Depression is the most common comorbid psychiatric condition associated with epileptic syndromes, with a severe impact on quality of life (Edeh & Toone, 1987; Jacoby et al., 1996); there is also a significantly increased risk of suicide compared with healthy controls (Caplan et al., 2005; Verrotti et al., 2008; Bell & Sander, 2009). The prevalence of clinical depression in epilepsy is also higher than that for other chronic disorders such as asthma or diabetes (Ettinger et al., 2004). Between 20% and 55% of epilepsy patients are estimated to be affected, and the condition often remains underdiagnosed and undertreated (Kanner & Palac, 2000), mainly due to the misconception that antidepressant drugs are generally proconvulsant (see Jobe & Browning, 2005). Interestingly, longitudinal studies in epilepsy and depression have reported a bidirectional temporal association. Therefore, epilepsy is frequently followed by depression, whereas preexisting depression is approximately seven times more common among patients with new-onset epilepsy (usually preceding the seizure disorder), compared with age- and sex-matched controls. In addition, when analyses were restricted to cases with an unprovoked “localized onset” seizure, depression was 17 times more common among affected cases than among nonseizure controls (Forsgren & Nyström, 1990; Hesdorffer et al., 2000); the strong links between temporal or frontal lobe epilepsy and depression are particularly well recognized (Grabowska-Grzyb et al., 2006; García-Morales et al., 2008). Perhaps the more acute neurochemical/structural changes in the brain associated with a partial epileptic focus contribute to depression onset by disrupting neural circuits involved in emotional processing (Giovacchini et al., 2005; Chayasirisobhon, 2009).

Epilepsy is treated with a number of antiepileptic drugs (AEDs) according to etiology, age, and pattern of onset, some of which may themselves alter psychobiologic processes (Andersohn et al., 2010; Arana et al., 2010). Although effective seizure control improves the life quality of patients with epilepsy, the coexistence of depression seems a more important determinant of psychological well-being than seizure frequency/severity or AED-related side effects (Boylan et al., 2004); less is known about the role of AED treatments in comorbid depression onset and maintenance. In 2008, the U.S. Food and Drug Administration (FDA) conducted a study of patients taking AEDs and their increased risk of suicidal ideation and behavior, concluding that AED use was associated with a 1.8-fold higher likelihood of depression, compared to patients taking a placebo. However, this point has remained controversial and many neurologists have raised concerns over methodologic issues in the FDA’s evaluation (Bell et al., 2009; Hesdorffer & Kanner, 2009). Although AEDs might increase the risk of or induce depression, it remains difficult to distinguish between disease-induced depression and AED-induced depression in clinical practice, as also recently supported by other studies (Wen et al., 2010; Bagary, 2011). This confounding difficulty, however, has to be taken into account when studying epilepsy and depression in comorbidity, both in patients and animal models.

To study the pathogenic mechanisms of depression in epilepsy and to assist the development of adequate therapies, several animal models have been used. In particular, WAG/Rij rats, genetic absence epileptic rats from Strasbourg (GAERS), and Long-Evans rats with spontaneous spike-wave discharges (SWDs) represent well-validated animal models of absence-type epilepsy (Danober et al., 1998; Coenen & Van Luijtelaar, 2003; Shaw et al., 2009) with depression-like symptoms (Jones et al., 2008; Shaw et al., 2009; Sarkisova & van Luijtelaar, 2011). Other epilepsy/seizure models have been tested for depression-like behavior; however, results have been controversial and inconclusive (Mazarati et al., 2008; Müller et al., 2009). Furthermore, animal studies have never focused on a possible pharmacologic correlation between AED effects and the development of either seizures or depressive-like symptoms.

Absence-type seizures (van Luijtelaar & Sitnikova, 2006) are particularly prevalent in children, have strong genetic links, and are characterized by brief absences, where consciousness is compromised. Electroencephalography (EEG) recording in epileptic patients reveals bilateral, synchronous symmetrical SWDs (3–4 Hz), on a normal background (Commission, 1989). However, the precise origin of these oscillations is still unclear. Studies in WAG/Rij rats have revealed a cortical focus within the SmI (somatosensory cortex) responsible for SWD generation (Meeren et al., 2005). This model is now widely regarded as a valid animal model of human absence epilepsy (for a review see Coenen & Van Luijtelaar, 2003). We previously reported that early long-term treatment (from 1.5 to 5 months) of WAG/Rij rats with the antiabsence drug ethosuximide (ETH) or levetiracetam (LEV) suppressed absence seizure development (Russo et al., 2010). Recently, Sarkisova et al. (2010) reported that similar ETH treatment not only prevented SWDs, but also inhibited depressive-like behavior in these animals, concluding that seizures were responsible for depression onset. It remains unclear, however, whether this applies more generally to AEDs and depression, or merely a particular feature of the action of ETH. Here, we wished to test how some other AEDs [zonisamide (ZNS), carbamazepine (CBZ), and LEV] with proposed mechanisms of action different from those of ETH (Czapiński et al., 2005) would affect development of absence seizures and depression-like symptoms in WAG/Rij rats, with a view to a better understanding of depression management in epilepsy. LEV was chosen because of its proven efficacy in absence epilepsy (for a review see Hughes, 2009) and its known antiepileptogenic effects in this animal model (Russo et al., 2010). ZNS was chosen considering its efficacy in the treatment of absence epilepsy (Hughes, 2009), its neuroprotective/antiepileptogenic properties (Willmore, 2005), and a high associated risk of developing depression in patients (Mula & Sander, 2007). CBZ is known to aggravate absence seizures, both in human patients (Gansaeuer & Alsaadi, 2002) and experimentally (Liu et al., 2006); it was chosen as a negative control for absence epilepsy while having antiepileptogenic properties (Capella & Lemos, 2002), and also since its use has been associated with some beneficial effects in patients with major depressive disorder (Vigo & Baldessarini, 2009; for drug choice rationale see also Discussion).

Materials and Methods


Male WAG/Rij rats and age-matched Wistar rats were used. Rat progenitors were purchased from Charles River Laboratories s.r.l. (Calco, Lecco, Italy) at a body weight of approximately 60 g (4 weeks old). Following arrival, animals were housed three or four per cage and kept under controlled environmental conditions (60 ± 5% humidity; 22 ± 2°C; 12/12 h reversed light/dark cycle; lights on at 20.00 h). Female rats of all strains of 10 weeks of age were placed with same age group male rats for mating for 16 h in the ratio 1:1 and examined the next morning for the presence of a vaginal plug, a sign of successful copulation. Dams of all strains were housed two per cage, whereas, all offspring after weaning (P21) were housed three or four per cage. Animals were allowed free access to standard laboratory chow and water until the time of experiments. Procedures involving animals and their care were conducted in conformity with the international and national law and policies (European Communities Council Directive of 24 November 1986, 86/609EEC). All efforts were made to minimize animal suffering and to reduce the number of animal used.

Chronic drug administration protocol

The AEDs used in early long-term treatment were delivered via the drinking water in home cages. This type of approach was chosen considering our previous experience where long-term treatment was used (Russo et al., 2010), in order to avoid animal manipulation stress that might have influenced results for both the EEG recordings (Kovács et al., 2006) and forced swimming test (FST) (Bielajew et al., 2003; see below). No significant differences were observed in the quantity of drug solution or water drunk between any animal test group and control. Ethosuximide (ETH; Zarontin; Pfizer Italia Srl, Rome, Italy) 250 mg/5 ml syrup and levetiracetam (LEV; Keppra; UCB Pharma, Piazzetta, Torin, Italy) 100 mg/ml oral solution were used. ZNS and CBZ were purchased from Sigma-Aldrich Co. Ltd (Poole, United Kingdom). ETH or LEV was administered orally at a dose of approximately 80 mg/kg/day, by adding 4 ml of ETH syrup or 2 ml of LEV oral solution to 300 ml of drinking water, respectively. ZNS was orally administered at a dose of approximately 40 mg/kg/day by dissolving 40 mg in 120 ml of drinking water. CBZ solution for oral administration at a dose of approximately 20 mg/kg/day was prepared as detailed in Data S1. The CBZ-treated groups were administered with the following schedule: first week, 20 mg/kg; second week, 25 mg/kg; and for the remaining 15 weeks with a dose of 30 mg/kg (for rationale see Data S3). Doses were calculated on the basis of the knowledge that rats drink on average 10–12 ml/100 g/day (van Zutphen et al., 2001). Water bottles were wrapped in silver foil to exclude light, and solutions were freshly prepared and replaced twice a week. The doses used were chosen because of the lack of adverse events in previous reports in absence epilepsy animal models or preliminary experiments (Wallengren et al., 2005; Russo et al., 2010). Furthermore, these doses attained plasma concentrations similar to those seen clinically in patients (Neels et al., 2004; Russo et al., 2010).

Age-matched male rats of both strains (n = 12 per group; see Fig. S1) started treatment at 30 days of age and were kept on drugs for a further approximately 17 weeks up to the age of approximately 5 months; treatment was then stopped and animals normally housed. Control animals (n = 18) were kept under standard animal house conditions during the same time-window. During this period, animals were weighed weekly every Monday between 9:00 a.m. and 11:00 a.m. Every month, a blood sample of approximately 1 ml was obtained through the tail vein for later analysis of drug blood concentrations from three rats per group of CBZ-treated animals in a randomized order. Animals were only gently restrained during this process. Particular attention was given to the possible appearance of any obvious drug-induced side-effects such as marked ataxia or sedation; during the entire time-window of the experiments, no marked side effects were noted that might have influenced animal performance in the FST or EEG recordings.

Animal surgery and recording protocol in WAG/Rij rats

All WAG/Rij rats at the age of 6.5 months were chronically implanted, under chloral hydrate anesthesia (400 mg/kg, i.p.; Carlo Erba, Milan, Italy), using a Kopf stereotaxic instrument, with five cortical electrodes for EEG recordings (Russo et al., 2004). Stainless-steel screw electrodes were implanted on the dura mater over the cortex: two in the frontal region (AP = 2; L = ±2.5) and two in the parietal region (AP = −6; L = ±2.5) according to the atlas coordinates of Paxinos and Watson (1998). The ground electrode was placed over the cerebellum. All animals were allowed at least 1 week of recovery and handled twice a day. To habituate the animals to the recording conditions, the rats were connected to the recording cables for at least 3 days before the experiments. The animals were attached to a multichannel amplifier (Stellate Harmonie Electroencephalograph; Montreal, Quebec, Canada) by a flexible recording cable and an electric swivel, fixed above the cages, permitting free movements for the animals. All EEG signals were amplified and conditioned by analog filters (filtering: below 1 Hz and above 30 Hz at 6 dB/octave) and subjected to an analog-to-digital conversion with a sampling rate of 300 Hz. The quantification of absence seizures was based on the number and the duration of EEG SWDs (Russo et al., 2004).

Experimental protocol for chronically treated rats

Five separate groups of rats (n = 12 per group) were used to compare the effects of long-term drug treatment relative to control animals. At least 1 week after surgery, animals underwent three recording periods, starting at 9:00 a.m., for three consecutive days. Every recording session lasted 3 h without injection of any drug for every group.

Experimental protocol for acutely treated rats

Control animals, following the recordings of the chronic protocol (see above), were randomly subdivided into three groups of six rats and assigned to CBZ (20 and 40 mg/kg, i.p.), ZNS (40 and 80 mg/kg, i.p.), or CBZ focally injected in the perioral region of the somatosensory cortex (S1po; 20 μg/0.5 μl). In each group of systemically treated animals, every rat was then treated with the two doses of the drug under study or with vehicle, with an interval of at least 3 days between injections. Following this protocol, every animal was injected three times with the following order: vehicle, drug lower dose, and drug higher dose. The interval of 3 days between injections was sufficient for recovery, and this was confirmed by recordings with no injections on the second day of interval. Focally treated animals were additionally implanted with two guide cannulae into the S1po (AP = −2.1; L = ±5.5; H = 4 mm from bregma) according to the atlas by Paxinos and Watson (1998). Evaluation of the effect of focally injected CBZ was performed as described previously (Citraro et al., 2006). In this experimental group, animals were injected with vehicle (see Data S2) three days before and three days after CBZ, thereby obtaining two control recordings. EEG recordings for all treated groups, in the acute injection paradigm, lasted 4 h:1 h baseline without injection and 3 h after the injection of every drug or vehicle.

Determination of drug blood concentrations

CBZ serum levels were determined by means of an automated AxSYM immunochemistry analyzer (Abbott Laboratories, Abbott Park, IL, U.S.A.) routinely run in the laboratories of the Clinical Pharmacology Unit at the School of Medicine of the Catanzaro University (for results see Data S3).

Forced Swimming test

The FST (Porsolt et al., 1977) is currently the most reliable method for assessing behavioral “despair” and for screening antidepressant drug action in rodents by measuring the immobility time. The test can detect the activity of most established antidepressants with diverse mechanisms of action (Detke & Lucki, 1996; reviewed by Cryan et al., 2005); furthermore, WAG/Rij rats, which exhibit many depression-like behavioral symptoms under baseline conditions (in addition to absence epilepsy), were also recently validated as a useful test model of human low-grade depression (dysthymia), which can be reversed by chronic antidepressant treatment (Sarkisova & van Luijtelaar, 2011).

In the initial 15-min habituation session (excluded from the data analysis), animals were individually forced to swim in a clear plastic cylinder (47 cm in height; 38 cm in diameter) containing 38 cm of water (25 ± 1°C). At this water depth, animals could not touch the bottom of the cylinder and could not, therefore, modify the effects of the forced swim by developing behavioral adaptation (Detke & Lucki, 1996). After a period of vigorous swimming, all animals normally reduced their movements to only those necessary to maintain their head above the water level, with no other displacement. The 5-min test session began 24 h later. Test sessions were video-recorded by means of a digital camera (resolution: 640 × 480–30 fps; Medi@com Sport Cam Plus, Mediacom, Milan, Italy) fixed above the cylinders; the obtained files were later analyzed by three independent observers, two of them blinded to the treatment protocol. The total duration of immobilization, including passive swimming, was measured. The criterion for passive swimming was floating vertically in the water while making only those movements necessary to keep the head above the water. After the FST, animals were removed and dried with a towel before being placed in their home cages. Every experimental animal group was evaluated in the test according to the schedule reported in Fig. S1, starting at 9:00 a.m. and finishing before 11:00 a.m. in order to avoid possible circadian alteration of test results. Briefly, different groups of untreated WAG/Rij and Wistar rats were tested at the ages of 2, 4, and 6 months, whereas treated rats were tested at the age of 4 and 6 months only.

Statistical analysis

All statistical procedures were performed using SPSS 15.0.0 software (SPSS Inc., Chicago, IL, U.S.A.). EEG recordings were subdivided into 30-min epochs, and the duration and number of SWDs were treated separately for every epoch. Such values were averaged and data obtained were expressed as mean ± standard error of the mean (SEM) for every dose of compound. Long-term treated animals were compared by one-way analysis of variance (ANOVA), the treatment being the only variable, followed by a post hoc Bonferroni test. The significance of a drug’s acute effects in control animals was determined by repeated measures ANOVA followed by Tukey’s post hoc test, with the exception of focally injected CBZ, which was determined by Mann-Whitney U-test.

The mean immobility time (IT) of Wistar rats at 6 months of age was considered equal to 100% and every other time obtained in the test was reported with respect to this value and expressed as a percentage of it. Immobility times were compared by one-way ANOVA followed by Bonferroni’s post hoc test. The data are expressed as means ± SEM. All tests used were two-sided and p  0.05 was considered significant. Any behavioral responses, excluding immobility, were recorded, but not statistically analyzed for each animal.


Effects of acutely administered zonisamide or carbamazepine

Analysis of EEG recordings from control WAG/Rij rats at 6 months of age revealed that the mean number of SWDs (nSWDs) for a 30-min epoch was 11.27 ± 2.15 with a mean total duration (dSWDs) of 71.66 ± 8.95 s and a mean single duration (sSWD) of 5.76 ± 0.58 s (Table 1). Acute administration of ZNS (40 and 80 mg/kg) or CBZ (20 and 40 mg/kg; i.p.) produced opposite effects on SWD characteristics. Therefore, ZNS induced a dose-dependent and significant reduction of nSWDs and dSWDs (p < 0.05) rapidly after administration for up to 3 h after injection with both doses. In contrast, CBZ at both doses significantly and dose-dependently increased both nSWDs and dSWDs but not sSWDs (p < 0.05; Fig. 1); this agrees with previously reported proconvulsant effects of CBZ in GAERS (Wallengren et al., 2005). Interestingly, the bilateral focal injection of CBZ (20 μg/0.5 μl) directly into the S1po failed to affect any SWD parameters; a slight, transient, but nonsignificant (p = 0.56) reduction in nSWDs and dSWDs was noticed only during the first 30–60 min after injection (Fig. S2).

Table 1.   Effects of early treatment with AEDs on SWD parameters in WAG/ Rij rats
Animal group (n = 12)nSWDsdSWDs (s)sSWD (s)
  1. nSWDs indicates the mean number of SWDs for every 30-min epoch; dSWDs indicates the mean cumulative duration of SWDs for every 30-min epoch, expressed in seconds; sSWD indicates the mean duration of a single SWD expressed in seconds. Data are expressed as mean ± standard error of the mean (SEM) obtained by analyzing three different recordings from every single animal in three consecutive days, with a total recording duration of 3 h. Data marked with (*) are significantly different (p < 0.05) from respective control. n = 12 for every group.

  2. ETH, ethosuximide; LEV, levetiracetam; ZNS, zonisamide; CBZ, carbamazepine; s, seconds.

Control untreated group11.27 ± 2.1571.66 ± 8.955.76 ± 0.58
ETH-treated rats (approximately 80 mg/kg)7.81 ± 0.28*43.31 ± 3.52*4.12 ± 0.53*
LEV-treated rats (approximately 80 mg/kg)6.45 ± 0.32*33.81 ± 5.81*4.75 ± 0.85*
ZNS-treated rats (approximately 40 mg/kg)7.02 ± 0.73*38.88 ± 5.09*5.54 ± 0.49*
CBZ-treated rats (approximately 20 mg/kg)13.61 ± 0.6388.73 ± 10.606.76 ± 0.37
Figure 1.

Effects of acute systemic administration of zonisamide (ZNS) or carbamazepine (CBZ) in WAG/Rij rats (non-pretreated). Plots show time- and dose-dependent effects of ZNS and CBZ on the number (A) and duration (B) of epileptic SWDs. Data values are means ± SEM (n = 6). Every data point is significantly different (p < 0.05) from its own saline-injected control group, with the exception of 20 mg/kg CBZ from 90 min up to 180 min for both number and duration of SWDs.

Effects of early long-term drug-treatment on the development of SWDs

Early long-term treatment with either ETH or LEV (approximately 80 mg/kg/day) significantly reduced the development of absence seizures in adult WAG/Rij rats (p < 0.05: Table 1; Russo et al., 2010). ZNS chronically administered also significantly reduced absence seizure development, decreasing every SWD parameter with a similar efficacy to both ETH and LEV; chronic CBZ, however, was completely ineffective in preventing epileptogenesis and, therefore, appearance of absence seizures, since all measured SWD parameters were not significantly different from control (Table 1).

Effects of long-term AED treatment on immobility time in the forced swimming test

Video analysis of the FST of untreated animals at the age of 2, 4, and 6 months confirmed previous reports (Sarkisova et al., 2008, 2010) that WAG/Rij rats develop longer ITs with age than control Wistar rats, and, therefore, display depressive-like behavior over time. One-way ANOVA revealed that Wistar rats had similar ITs at every age studied, whereas, in untreated WAG/Rij rats there was an age-dependent increase, with the oldest animal group (6 months old) showing the longest IT relative to Wistar controls (Fig. 2).

Figure 2.

Effects of AED pretreatment on immobility times (ITs) in the forced swimming test. Bars indicate immobility times expressed as a percentage, considering the measured immobility time of Wistar rats at 6 months of age as control = 100%. Data values are means ± SEM (n = 6) *significantly different (p < 0.05) from age-matched Wistar rat controls; #significantly different (p < 0.05) from age-matched untreated WAG/Rij rats. WS, Wistar rats; WR, WAG/Rij rats; CTRL WS, control Wistar rats; CTRL WR, untreated WAG/Rij rats; Colored-bars show drug-pretreated WAG/Rij rats: ETH, ethosuximide; LEV, levetiracetam; CBZ, carbamazepine; ZNS, zonisamide.

Following long-term treatment with AEDs, Wistar rats did not show any significant change in ITs, either at 4 or 6 months of age (data not shown). In treated WAG/Rij rats, ETH was the only AED that significantly reduced IT at both ages relative to WAG/Rij controls; CBZ and ZNS where completely ineffective, whereas LEV induced a prolongation of IT that was already noticeable at 4 months (p = 0.12), and became statistically significant at 6 months following two-way ANOVA (Fig. 2). When rats displayed immobilization during the FST, facial twitching was not observed; rhythmic twitching of the vibrissae and facial muscles is generally associated with SWDs and, therefore, it is possible to exclude that the increased IT times in WAG/Rij rats was due to absence seizures. Furthermore, it has been demonstrated previously that acute ETH administration that completely suppresses SWDs was ineffective in the FST (Sarkisova et al., 2010).


Our aim was to test a possible correlation between the use of some established AEDs to control absence epileptogenesis and the development of depressive-like behavior in the WAG/Rij rats, two pathologic conditions that have recently been suggested to be directly related in this model (Sarkisova et al., 2010; Sarkisova & van Luijtelaar, 2011).

Effects of acutely administered zonisamide and carbamazepine

ZNS is a “new generation” broad-spectrum AED with multiple suggested mechanisms of action including modulation of voltage-sensitive Na+ and low-threshold T-type Ca2+ channels (Czapiński et al., 2005; Biton, 2007). It is “clinically effective against multiple seizure types” including absences (Hughes, 2009; Zaccara & Specchio, 2009). Experimentally, ZNS has never been tested in either GAERS or WAG/Rij rats; here, we demonstrated its dose-dependent efficacy in controlling absence seizures at doses free from side effects, very likely through blocking T-type Ca2+ currents in thalamocortical cells (Suzuki et al., 1992), in a fashion similar to that of ETH (Gören & Onat, 2007), although other contributing mechanisms cannot be excluded. In contrast, we also demonstrated for the first time in WAG/Rij rats that systemically administered CBZ dose-dependently increased the occurrence and total duration of SWDs. This result agrees with previous reports regarding the acute administration of CBZ in other absence models (McLean et al., 2004; Wallengren et al., 2005), most likely by potentiating γ-aminobutyric acid (GABA)A–evoked synaptic currents in the ventrobasal thalamus (Liu et al., 2006).

An epileptic focus within the perioral region of the primary somatosensory cortex (S1po), with upregulated levels of Na+ channel expression, has been identified in the brain of WAG/Rij rats, as being responsible for the generation of SWDs. In support of this, many drugs focally injected into this area can suppress absence seizures, including phenytoin, which is proconvulsant when administered systemically (Citraro et al., 2006; Gurbanova et al., 2006). On this basis, we tested the effects of focally injected CBZ into the S1po, but it failed to significantly modify the EEG pattern of WAG/Rij rats, although a slight, transient reduction in the number and duration of SWDs was noticeable in the first 30 min after injection. This lack of effect was surprising, considering both CBZ and phenytoin are known to block Na+ channels (Czapiński et al., 2005). Perhaps the high lipophilicity of CBZ facilitated its rapid diffusion to other brain areas with a proportional reduction of its local effective concentration. Kovács et al. (2006) reported that manipulation of WAG/Rij rats for drug injection can increase SWD parameters within the first 30 min after injection. Even if this stress-induced effect has never been systematically studied and quantified, it could have influenced the EEG recordings observed after focal CBZ administration. However, following the latter, an opposite decrease of SWD numbers and duration was noted; this contrasting result could indicate that the observed effect of CBZ is indeed due to a direct inhibition of Na+ channels (as for phenytoin) instead of stress manipulation.

Effects of early long-term AED treatment on the development of SWDs

Previous reports (Blumenfeld et al., 2008; Russo et al., 2010) demonstrated that an early long-term AED treatment, starting before seizure onset (approximately 2 months of age), can alter the development of absences in WAG/Rij rats. We found that both ETH and LEV treatments at approximately 80 mg/kg, starting at 6 weeks of age, reduced the number of SWDs at 6 months (Russo et al., 2010). Here, we showed that early long-term treatment with ZNS (but not CBZ), could also alter the development of absence pathology by significantly reducing seizure generation and synchronization (see Table 1).

Russo et al. (2010) concluded that the antiepileptogenic properties of a single AED might depend on the type of seizure/epilepsy/animal model considered, since ETH (which is not antiepileptogenic in any other model) possessed the same protective properties as LEV in the WAG/Rij rat model. This was confirmed by our present results, by the lack of efficacy of CBZ, which does show antiepileptogenic properties in some animal seizure models (Capella & Lemos, 2002) but not others (Schmutz et al., 1988; Pitkänen et al., 1996). ZNS, on the other hand, exhibited the same efficacy as ETH or LEV, and is known to be both antiepileptogenic and neuroprotective in other animal models (Hashimoto et al., 2003; Willmore, 2005). Taken together, these results suggest that the epileptogenic process in WAG/Rij rats does not depend on seizure control, since CBZ, which increased the severity of the seizure pathology during acute treatment, did not aggravate the disease outcome in the long term. Therefore, the antiepileptogenic mechanisms most likely reside in the intrinsic properties of the AEDs themselves, and remain to be clarified in detail.

AED treatment and the forced swimming test

WAG/Rij rats represent a validated model for mild human depression, and are reported to show depressive-like behavioral symptoms (despair) like a reduced sucrose intake (a measure of “hedonic” state), decreased exploratory activity in the open field arena, passive strategies under stress, cognitive defects, and an increased IT in the FST, but no difference in their anxiety profile compared to Wistar rats; this genetic model is, therefore, considered as particularly useful for studying absence epilepsy with depression comorbidity (Sarkisova & van Luijtelaar, 2011). We have confirmed that WAG/Rij rats at 6 months of age have longer IT in comparison to age-matched Wistar rats; moreover, we identified an age-dependent increase in IT for only WAG/Rij rats, which was not mirrored in control animals. In our tests, the AEDs used did not have any significant effect on Wistar rat behavior in the FST at any age considered; these animals cannot, therefore, be considered to show baseline depressive-like behavior, and the lack of effects of AEDs in this strain might be justified by the lack of a pathologic background. Furthermore, because AEDs might impair learning/memory processes and this is known to influence FST results, the lack of effect in Wistar rats suggests that, at the doses studied, the observed results in WAG/Rij rats are very likely solely due to their action on absence epileptogenesis or, as discussed below, their prodepressant side effects.

Sarkisova et al. (2010) recently reported that early long-term treatment with ETH, at a higher dose (300 mg/kg/day) than the one used by us (80 mg/kg/day), suppressed SWDs and increased IT values, concluding that SWD activity was necessary for the expression of depressive-like behavior. To examine the generality of this hypothesis, we tested whether some other AEDs, chronically administered in WAG/Rij rats, might also influence SWDs and IT values in parallel. In our previous paper (Russo et al., 2010), we showed that both ETH and LEV reduced absence epileptogenesis, reflected by a diminished number and duration of total absence seizures revealed in the EEG in comparison to control. This was confirmed in the present experiments; however, only ETH was able to significantly reduce IT at both age groups (cf. Sarkisova et al., 2010), whereas LEV induced a significant aggravation of depressive-like behavior. This latter effect has not, to our knowledge, been previously reported in other experimental animal models, and suggests that ETH may be more effective than LEV in comanaging depression in patients with absence epilepsy. Brandt et al. (2007) administered LEV for 4 weeks in rats following electrical induction of a self-sustained status epilepticus, but found no effect on IT in the FST. Interestingly, clinical use of LEV has only recently been associated with the development of depression in patients (Tamarelle et al., 2009; Vande Griend et al., 2009); however, this effect has only been reported in single studies of an open-label design and with small sample sizes or case reports (Miller et al., 2008). Because LEV did not have significant effects on IT in Wistar rats, it is likely that this effect is strain specific or depends on a pathologic background present only in WAG/Rij rats. The mechanism by which LEV exacerbates depressive symptoms in this model, therefore, remains to be determined.

ZNS was chosen to test because of its ability to increase brain dopamine and serotonin levels (Okada et al., 1992), while possessing clinical antiabsence properties (Hughes, 2009). Following early long-term treatment, it was as effective as ETH or LEV in preventing absence seizure onset; however, it did not significantly affect IT values in the FST. Its use in patients has been associated with a 7% probability to develop depressive mood disorders (Mula & Sander, 2007). We might, therefore, have expected an increase of depressive-like behavior, at least in some of the animals; the lack of observation of any effect on IT might be due to the small number of animals treated (n = 24) or to the dose used, since depression in human patients seems directly correlated to the use of high dosages (Mula & Sander, 2007).

CBZ was chosen because of its recognized lack of effect in absence epilepsy (Gansaeuer & Alsaadi, 2002; Liu et al., 2006), which was confirmed in our chronic treatment experiments by failing to modify absence seizure development. CBZ is also a mood stabilizer, which increases brain dopamine and serotonin levels in many areas (Baf et al., 1994; Ichikawa & Meltzer, 1999), and its use has been associated with some beneficial effects in patients with major depressive disorder (Vigo & Baldessarini, 2009). Recently, CBZ was tested in another animal model of epilepsy and depression comorbidity (Barbakadze et al., 2010). These authors found that CBZ, although suppressing kindled seizures, was unable to modify animal behavior, concluding that in their model, seizures and depression were unrelated. Redrobe and Bourin (1999) also observed that CBZ did not induce any significant antidepressant-like effects in the mouse FST. These animal models, together with the WAG/Rij rat model, therefore, seem unable to detect the antidepressant activity of CBZ; the reasons for this discrepancy are currently unclear.

Correlation between epilepsy, depression, and AED use

Because depression is the most common psychiatric comorbidity associated with epilepsy (Kanner, 2009), biologic linkages between the epilepsies and affective disorders have been the subject of increasing scientific attention (Kanner & Balabanov, 2002; Jobe, 2003; Caplan et al., 2005; Hajszan & MacLusky, 2006; Andersohn et al., 2010; Arana et al., 2010); however, any common mechanisms of pathogenesis remain unclear. In recent years, epilepsy and depression have been finally studied as two different hallmarks of a unique neurobiologic disorder, since there has been increasing evidence that serotoninergic (5-HT) neurotransmission modulates a wide variety of experimentally induced seizures (Bagdy et al., 2007). Elevation of extracellular 5-HT levels inhibits both focal and generalized seizures, possibly by hyperpolarization of glutamatergic neurons via 5-HT1A receptor activation and depolarization of GABAergic neurons by 5-HT2C receptor action. Very few studies have focused interest on pharmacologic treatments and their possible correlation with the development of depression in epileptic patients. To date, only a single study provides support for the notion that AEDs can impair performance in cognition, mood, and behavior, since duration of drug intake and the number of AEDs utilized are the main confounding variables (Shehata et al., 2009). Even if neurobiologic patterns leading to depression in epilepsy become more clearly elucidated, and a role for AEDs assumes greater prominence, few data support the hypothesis that different AEDs can either work in worsening depression or in protecting from suicide in epileptic patients. These analyses are controversial, even in experimental settings, since drug intake, pharmacokinetics, and receptor binding selectivity may change from rats to humans. The study of animal model results are, therefore, of extreme importance in determining the effects of drugs in comorbid conditions.

Our current working hypothesis is that in WAG/Rij rats, a cortical focus (already present before absence seizure onset), stimulates the thalamus and other brain areas mainly involved in absence seizure generation, inducing adaptive changes leading to the development of the disorder, as happens during experimental kindling (Morimoto et al., 2004) and that depressive-like behavior somehow originates from common pathogenic mechanisms, with seizures being necessary for the expression of negative behavioral symptoms (Sarkisova et al., 2010). Indeed, we found IT in WAG/Rij rats increased with age as did SWD parameters, suggesting that the two pathologies may follow the same developmental progression. In addition, in accordance with Sarkisova et al. (2010), we demonstrated that ETH (at 80 vs. 300 mg/kg/day) possesses antiabsence and antidepressant-like effects in the WAG/Rij rats, as it reduced the development of absence seizures and IT in the FST, with no effects in control Wistar rats. However, in direct contrast, ZNS decreased the development of absence epilepsy without affecting IT, whereas LEV induced an increase of IT, while also acting as an antiepileptogenic agent. Based on these results, it is clear that there was no obvious correlation between epilepsy and depression in this animal model, but possible single drug effects on behavior also need to be considered. Therefore, the observed dual effectiveness of ETH may be related to its low risk for inducing clinical depression (<1%; Mula & Sander, 2007). On the other hand, LEV and ZNS are associated with a higher depression risk; therefore, their ameliorating effects on absence epilepsy (which might also be reducing the development of behavioral alterations) might be compensated by their intrinsic behavioral effects seen only on a WAG/Rij background. The lack of effect of CBZ, which did not influence either absence epilepsy or depressive-like behavior in this or other animal models (Redrobe & Bourin, 1999; Barbakadze et al., 2010), would also fit in with this argument.


In conclusion, we confirmed that in WAG/Rij rats, an epileptogenic process underlies the development of absence epilepsy according to the concept of “seizures beget seizures,” which has been widely demonstrated in many convulsive experimental models (Morimoto et al., 2004; Ben-Ari & Holmes, 2006). Furthermore, we showed that early long-term treatments with ETH, LEV, or ZNS (but not CBZ) can ameliorate this epileptogenic process, but only ETH was clearly able to reduce codevelopment of depressive-like symptoms in the FST. We suggest that epilepsy and depression may well share a common pathogenic mechanism in this animal model, but ameliorating absence seizures with AEDs may not necessarily correlate with a reduced risk of developing depressive-like symptoms. Possible intrinsic behavioral effects of the AEDs themselves on a WAG/Rij background must also be considered.


This work was supported by the Italian Ministry of Education, University and Research (MIUR, Cofin 2007, Rome) and the National Research Council (CNR, Rome) is gratefully acknowledged. Dr. Donato Cosco is kindly acknowledged for the technical preparation of carbamazepine solution.


None of the authors has any conflict of interest to disclose. 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.