The influence of strain and housing on two types of spike-wave discharges in rats

Authors


U. Schridde, NICI, Department of Biological Psychology, University of Nijmegen, PO BOX 9104, 6500 HE Nijmegen, the Netherlands. E-mail: u.schridde@nici.kun.nl

Abstract

WAG/Rij rats, a genetic model of absence epilepsy, show two types of spike-wave discharges (Type 1 and Type 2) in their EEG activity. The large interindividual variation in the expression of the phenotypes (number and mean duration of spike-wave discharges) suggests that as well as genetic, environmental factors also play a role. The aim of our study was to establish effects of strain and housing on the incidence and expression of both types of paroxysms. Therefore, WAG/Rij and ACI rats were housed from weaning in either an enriched or impoverished environment for 60 days. At three months of age the EEG of the rats was recorded for four hours to examine the effects of strain and housing on the incidence and expression of the two types of paroxysms. Generally, enriched housing led to worsening of Type 1 and Type 2 spike-wave discharges (SWD). However, the number of affected rats and the expression (number and mean duration) of Type 1 and Type 2 spike-wave discharges were differently influenced by strain and housing. This suggests that Type 1 and Type 2 spike-wave discharges are independent phenomena and that number and mean duration of these paroxysms are controlled by different mechanisms. Finally, the worsening of absence seizures after enrichment is different from what has been found for convulsive seizures.

Many different laboratory rat strains develop absence seizures, especially when they are getting older (Inoue et al. 1990; Willoughby & Mackenzie 1992). The seizures are characterized by a sudden paroxysmal loss of consciousness (Drinkenburg et al. 2003), behavioral immobility with twitches of the vibrissae with concomitant bilateral generalized synchronous spike-wave discharges (SWD) around 8 Hz in the EEG, that last about five seconds (Van Luijtelaar & Coenen 1986). Hence, they closely resemble the SWDs characterizing childhood absence epilepsy; although in humans the frequency of the SWD is lower (∼3Hz) and the onset earlier in life.

Since the discovery that rats of the WAG/Rij strain have a substantial amount of SWD (Van Luijtelaar & Coenen 1986), many studies were devoted to establishing the validity of the model, including studies on the underlying causes of the discharges (Van Luijtelaar & Coenen 1997). Nowadays the WAG/Rij rat is regarded as a well-established rat model for childhood absence epilepsy (Coenen et al. 1992; Crunelli & Leresche 2002). Nevertheless, the definite cause of the SWD is still a matter of debate. Although mainly thalamic structures were formerly thought to be the primary source of the rhythmic oscillations (McCormick & Bal 1997; Steriade et al. 1993), recent results also point to a leading role of the neocortex in the initiation of SWD (Meeren et al. 2002).

Cross-breeding WAG/Rij rats with ACI rats, a strain commonly used as nonepileptic control because most members of this strain are nearly free of absence seizures (Inoue et al. 1990) or show substantial less SWD (De Bruin et al. 2000), revealed an autosomal dominant monogenetic inheritance for the presence of SWD (Peeters et al. 1990, 1992). However, complex inheritance patterns were suggested for the number and duration of the discharges, implying that SWD are influenced by more than one gene. Comparable results were also found for other rat strains (Danober et al. 1998; Jandóet al. 1995; Vadász et al. 1995).

In addition to the generalized SWD (Type 1 SWD), WAG/Rij rats show another type of SWD (Type 2 SWD) not accompanied by concomitant behavioral signs. The Type 2 SWD are characterized by a lower amplitude of opposite polarity compared to the Type 1 SWD. Also, they have a shorter duration (∼1 second) and their frequency is slightly lower (∼6 Hz) (Van Luijtelaar & Coenen 1986). Furthermore, they seem to be a more localised phenomenon than the Type 1 SWD and are, again compared to type 1 SWD, oppositely sensitive for dopaminergic agents such as haloperidol and apomorphine (Midzianovskaia et al. 2001). How far these two types of paroxysms are related, or share a common underlying oscillatory mechanism is, however, still an open question.

WAG/Rij rats have a quite large interindividual variation in their expression (number and mean duration) of Type 1 SWD. Because WAG/Rij rats are fully inbred for more than 130 generations, the observed phenotypic variations between rats must be due to environmental factors. The finding that the first onset of Type 1 SWD is age-related and varies between rats (Coenen & Van Luijtelaar 1987) further suggests an environmental contribution. In animal models such as EL mice it was shown that environmental factors could alter seizure susceptibility (Todorova et al. 1999). Furthermore, Vadász and colleagues proposed that in addition to genetic, developmental-environmental factors are also involved in the generation of SWD (Vadász et al. 1995).

A commonly used method to study environmental influences is the manipulation of housing conditions (Würbel 2002). It has been shown that rats and mice housed in enriched or impoverished environments differ in a variety of brain and behavior related variables (Rampon et al. 2000; Renner & Rosenzweig 1987; Van Praag et al. 2000), including the dopaminergic and GABAergic system (Cordoba et al. 1984; Jones et al. 1992). These amines are known to play an important role in the pathogenesis of SWD (Danober et al. 1998). Furthermore, it was recently shown that housing in an enriched environment could prevent convulsive seizures in rats (Auvergne et al. 2002; Young et al. 1999).

The aim of the present study was to investigate whether strain (WAG/Rij vs. ACI) and housing (enriched vs. impoverished) would have an influence on the number of rats that develop Type 1 or Type 2 SWD as well as on the phenotypic expression at three months of age, when SWD showed up in the EEG for the first time. This would allow one to gain more insight into possible shared or non-shared mechanisms in the pathology of the two types of discharges.

Materials and Methods

Animals

A total of 39 WAG/Rij and 40 ACI rats were used. The rats were bred and born in our laboratory. Experimental handlings were performed with the aim of minimizing discomfort. They were approved by the local animal ethical committee of Nijmegen University and performed according to local guidelines and the European Communities Council Directive of 24 November 1986 (86/609/EEC).

At weaning, male rats were selected and divided into an enriched environment (EC) housed (nWAG/Rij (EC) = 20; nACI (EC) = 20) and an impoverished environment (IC) housed group (nWAG/Rij (IC) = 19; nACI (IC) = 20) using a split-litter design.

From weaning onward, all rats were maintained on a reversed 12 h light-dark regime with white lights on at 21.00. All rats of a given strain were housed in the same room. The air temperature was kept at 20 °C. Once a week all rats were handled briefly. Every two weeks the bedding material of the cages (EC and IC) was changed. Fresh tap water and standard rat chow (Hope Farms, Woerden, the Netherlands) were always available ad lib.

The rats remained in the different housing conditions for a period of 60 days. At three month of age, all animals underwent surgery and EEG recordings were made. During this time the rats were single housed.

Housing

Enriched environment

Rats were housed in groups of 8–10 in a large cage (75 × 150 × 80 cm). Only animals from the same strain were housed together. Each EC cage had metal sidewalls and a wire mesh roof. The front consisted of two Plexiglas doors. On one side, each cage was equipped with two food hoppers and a bottle of tap water. The floor was divided into two compartments (75 × 75 cm): one side contained standard bedding material, the other one consisted of a wire mesh grid. Another wire mesh grid (75 × 75 cm) hung approximately 40 cm from the ceiling at one side of each cage making up a ‘second floor’. This floor could be reached via a wooden plank and a branch spanned across the other compartment of the cage. Furthermore, an EC cage always contained two smaller metal ‘sleeping boxes’ (12 × 32 × 12 cm). The side of the hanging compartment and the grid floor, as well as the places of the metal boxes, were changed every two weeks according to a predefined schedule. The plank and branch were rearranged every week. A variety of 10 different objects (tubes with various openings, ceramic pots, wooden/plastic and metal toys, running wheels for exercise, toy cars, etc.) from a pool of 50 objects were always present, scattered throughout the cage according to a predefined arrangement. Those objects were changed and/or rearranged every other day.

Impoverished environment

In the IC condition all rats were singly housed in standard Makrolon type III cages (21 × 37 × 15 cm).

Surgery

Rats were anaesthetized with isoflurane, mounted in a stereotaxic apparatus, and received 0.1 ml atropine intramuscular (i.m.) as a parasympatholyticum to avoid excessive salivation. Before incisions were made, lidocaine was applied as a local anaesthetic. Maintenance anaesthesia was provided via permanent isoflurane inhalation applied through a nose mask. The depth of the anaesthesia was periodically assessed via the limb withdrawal reflex. The rat's body temperature was monitored via a sensing probe and maintained at 37 °C using a heating pad. Rats were equipped with unilateral epidural electrodes (Plastic One MS-332/2-A, Roanoke, VA, USA) over the frontal (A-P: 2.0; M-L: 3.5) and parietal (A-P: −6.0; M-L: 4.0) cortex. A ground electrode was placed over the cerebellum. All coordinates were determined with the aid of the stereotaxic atlas of Paxinos and Watson (1998) (Bregma zero-zero, skull surface flat). The electrode socket was fixed to the skull using two stainless steel jewellery screws and dental acrylic. After the operation the rats were given a recovery period of 14 days.

EEG/SWD recordings

Apparatus

The EEG was recorded and stored on disk for off-line analysis by use of a microcomputer running a windaq system (version 1.48, DATAQ Instruments, Akron, OH, USA). The signals were amplified and filtered (1–100 Hz) and than digitised at a sample rate of 512 Hz. During recordings, animals were exposed to background white noise (53.2 dB).

Procedure

Before the recordings all rats were briefly handled daily during a period of three days. One day before the EEG recordings the rats were placed in transparent Plexiglas recording cages (25 × 30 × 35 cm), equipped with food and tap water ad lib, and connected to EEG leads. The EEG leads were connected to the EEG amplifier via a swivel to guarantee free movements. The rational for this procedure was to ensure that the rats were completely adapted to the recording situation before the start of the actual recordings, and hence prevent any putative confounding behavioral effects on the EEG recordings. This was also validated by analysing the behavior scored during EEG recordings, which did not differ between EC and IC housed WAG/Rij rats (data not shown). The EEG was recorded during the dark phase for four hours between 11.00 and 15.00. During this period the number of paroxysms is fairly high in rats (Van Luijtelaar & Coenen 1988).

Dependent variables

The numbers of animals that have developed Type 1 SWD or Type 2 SWD per strain and housing condition were counted and percentages were calculated. For all rats the Type 1 SWD and the Type 2 SWD were automatically detected using swauto (Westerhuis et al. 1996), and subsequently checked by one of the authors using criteria described elsewhere (Van Luijtelaar & Coenen 1986). The number and mean duration for both types of SWD during the four hour EEG recordings were determined. For rats without Type 1 or Type 2 SWD the respective number of paroxysms was included in the analysis as zero. For the analysis of the mean duration of Type 1 and Type 2 SWD, only those rats that had Type 1 or Type 2 SWD, respectively, were included.

Statistical analyses

The software package spss 9.0 (SPSS Inc., Chicago, IL, USA) was used. The data were analysed using non-parametric and parametric tests, a P-value < 0.05 was regarded as significant.

The percentages of rats that showed Type 1 SWD or Type 2 SWD were analysed by Pearson Chi-square with either strain (WAG/Rij vs. ACI) or housing (EC vs. IC) as between factor. Also Pearson Chi-square with housing (EC vs. IC) as between factor was applied per strain separately. The variables ‘number of Type 1 SWD’, ‘number of Type 2 SWD’ and ‘mean duration of Type 2 SWD’ were analysed using a two-way analysis of variance (anova) with factors strain (WAG/Rij vs. ACI) and housing (EC vs. IC). Due to the absence of any ACI rat showing Type 1 SWD (see Results), the variable ‘mean duration of Type 1 SWD’ was analysed for WAG/Rij rats only using Student's t-test for independent samples. The grouping factor was housing (EC vs. IC).

Results

EEG/SWD recordings

General

Figure 1 shows that none of the ACI, but 84.6% of the WAG/Rij rats showed Type 1 SWD ( χ21 = 47.22, P < 0.001). The percentages of WAG/Rij rats (69.2%) or ACI rats (80.6%) that showed Type 2 SWD were not statistically different. The distribution of rats showing either Type 1 SWD or Type 2 SWD was not different between housing conditions. Figure 2 gives an example of a typical Type 1 SWD from a rat of the WAG/Rij strain and typical examples of Type 2 SWD for both strains. A Type 1 SWD was characterized by an ongoing pattern of asymmetric sharp spikes followed by a slow wave. The spikes had a positive deflection and an amplitude several times, but at least twice, the background EEG activity. The frequency of a Type 1 SWD was about 7–8 Hz. Type 2 SWD also had an asymmetric spike-wave pattern albeit the amplitude always had an opposite deflection and was generally smaller than for Type 1 SWD. The frequency of Type 2 SWD was around 6–7 Hz.

Figure 1.

Percentages of rats showing (a) Type 1 SWD or (b) Type 2 SWD per strain (WAG/Rij vs. ACI) and housing condition (EC vs. IC).*P < 0.001 indicates level of significance for difference between strains. 12 out of 14 (85.7%) of the EC housed WAG/Rij rats and 10 out of 12 (83.3%) of the IC housed WAG/Rij rats showed Type 1 SWD, but none of the ACI rats (0 out of 18 for both housing conditions) showed Type 1 SWD. 9 out of 14 (64.3%) of the EC housed WAG/Rij rats and 9 out of 12 (75%) of the IC housed WAG/Rij rats showed Type 2 SWD. In the ACI rats 14 out of 18 (77.8%) of the EC housed and 15 out of 18 (83.3%) of the IC housed rats showed Type 2 SWD.

Figure 2.

Examples of a typical Type 1 SWD (WAG/Rij) and Type 2 SWD (WAG/Rij & ACI). Type 1 SWD had an ongoing pattern of asymmetric sharp spikes followed by a slow wave. The spikes had a positive deflection and amplitude of at least twice the background EEG activity. The frequency of a Type 1 SWD was about 7–8 Hz. The Type 2 SWD had an asymmetric spike-wave pattern albeit the amplitude had an opposite deflection and was generally smaller than for the Type 1 SWD. The frequency of the Type 2 SWD was around 6–7 Hz. Scale bars indicate time 1 second (abscissa) and amplitude 1 mV (ordinate).

Type 1 SWD

An overview of the results for the variables ‘number of Type 1 SWD’ and ‘mean duration of Type 1 SWD’ is depicted in Fig. 3. The anova for the ‘number of Type 1 SWD’ revealed a main effect for strain (F1,58 = 21.35, P < 0.001). The WAG/Rij rats had many Type 1 SWD, the ACI rats none and this was independent of housing. The variable ‘mean duration of Type 1 SWD’ was analysed for WAG/Rij rats only. Student's t-test revealed an effect for housing (t16 = 2.56, P < 0.05): WAG/Rij rats that were EC housed had longer discharges than their IC housed counterparts.

Figure 3.

Mean (± SE) for (a) number of Type 1 SWD per strain (WAG/Rij vs. ACI) and housing condition (EC vs. IC); and (b) mean duration of Type 1 SWD in seconds per housing condition (EC vs. IC) for WAG/Rij rats only. WAG/Rij rats had a significant higher number of Type 1 SWD compared to ACI rats. EC housed WAG/Rij rats had longer Type 1 SWD than IC housed WAG/Rij rats. **P < 0.001; *P < 0.05 indicates level of significance for differences between strains or housing condition, respectively.

Type 2 SWD

An overview of the results for the variables ‘number of Type 2 SWD’ and ‘mean duration of Type 2 SWD’ is depicted in Fig. 4. The anova for the ‘number of Type 2 SWD’ revealed a main effect for housing (F1,58 = 7.42, P < 0.01). WAG/Rij rats as well as ACI rats that were EC housed had more Type 2 SWD than those WAG/Rij or ACI rats that were IC housed. The anova for the ‘mean duration of Type 2 SWD’ revealed a main effect for strain (F1,43 = 11.86, P < 0.001). ACI rats had longer Type 2 discharges than WAG/Rij rats, and this was independent of housing.

Figure 4.

Mean (± SE) for (a) number of Type 2 SWD per strain (WAG/Rij vs. ACI) and housing condition (EC vs. IC); and (b) mean duration of Type 2 SWD in seconds per strain (WAG/Rij vs. ACI) and housing condition (EC vs. IC). EC housed rats had more Type 2 SWD than IC housed rats. ACI rats had longer Type 2 SWD than WAG/Rij rats. *P < 0.01; **P < 0.001 indicates level of significance for differences between housing condition or strain, respectively.

Discussion

About 85% of the WAG/Rij rats had Type 1 SWD, but none of the ACI rats had it. We have not seen any environmental effect on the number of animals showing Type 1 SWD. The absence of SWD in some WAG/Rij rats at three months of age is in line with earlier data reported by our lab (Coenen & Van Luijtelaar 1987) and points to an incomplete penetrance at this age. At six months of age, all WAG/Rij rats show SWD (Coenen & Van Luijtelaar 1987; Van Luijtelaar & Coenen 1986).

Not only the development of Type 1 SWD but also the number of Type 1 SWD was strain dependent, because WAG/Rij rats had, independent of housing, a significant higher number of Type 1 SWD compared to ACI rats. Nevertheless, regarding the mean duration of Type 1 SWD, EC housed WAG/Rij rats had longer discharges than IC housed WAG/Rij rats. Hence, the expression of the phenotype is at least partly dependent on housing, with EC housing leading to an increase in the length of the discharge in those rats showing Type 1 SWD. Our observation that number and duration of Type 1 SWD are differently influenced by strain or housing strengthens the assumption that number and duration are two independent factors affecting SWD.

The present findings about Type 1 SWD are in line with earlier research, both genetic and pharmacological. By cross-breeding WAG/Rij rats with ACI rats, Peeters et al. (1990, 1992) have found evidence that different genes are responsible for the development as well as for the control of the number and duration of the Type 1 discharges. Also, by applying remacemide or its active metabolite FPL 12495, it was found that the number and duration of Type 1 SWD were influenced in different ways (Van Luijtelaar & Coenen 1995).

Neither strain nor housing had an influence on the number of rats that show Type 2 SWD at three months of age. This is in contrast to the strain difference found for the number of rats having Type 1 SWD. The absence of a strain effect on the number of rats showing Type 2 SWD might indicate that the presence of Type 2 SWD in the EEG is a common phenomenon in rats. However, possible changes in the number of affected rats by age should also be considered. In addition, Type 2 SWD have been described in these two strains only (Midzianovskaia et al. 2001), hence further research including more strains is needed before one can draw a firm conclusion.

Strain and housing had, however, an influence on the number and duration of Type 2 SWD. Rats that were EC housed had more Type 2 SWD than those that were IC housed. In addition, for those rats that show Type 2 SWD, ACI rats had longer discharges than WAG/Rij rats. These findings imply that, and this is comparable to Type 1 SWD, the number and duration of Type 2 SWD are independent variables that are controlled by different processes. Furthermore, our data indicate that the expression of Type 1 and Type 2 SWD are oppositely influenced by strain and housing. While the number of Type 1 SWD depends on strain, the number of Type 2 SWD depends on housing. In contrast, housing influences mean duration of Type 1 SWD, while strain influences mean duration of Type 2 SWD.

Hence, the current results further underline that Type 1 and Type 2 SWD are different phenomena. Next to differences in their morphology (Van Luijtelaar & Coenen 1986) Type 1 and Type 2 SWD also differ in their spatial distribution and pharmacological properties. Whereas the Type 1 SWD show a more fronto-parietal distribution in the EEG of WAG/Rij rats, generalized across both hemispheres, the Type 2 SWD are more localised with a parieto-occipital distribution. Furthermore, by applying the dopamine agonist apomorphine, Type 2 SWD increased, whereas Type 1 SWD were completely suppressed. Haloperidol, the prototype dopamine antagonist, promoted Type 1 SWD (Midzianovskaia et al. 2001). It is interesting to note, however, that both types of paroxysms react the same to the administration of GABAergic agents (Coenen et al. 1995; Bouwman et al. 2003).

It was found that EC housing caused a prolongation of Type 1 SWD in WAG/Rij rats and an increase of Type 2 SWD in both strains. This implies that EC housing can lead to changes in the brain that influence the expression of the two phenotypes. However, EC housing leads to numerous and complex changes in the brain (Rampon et al. 2000; Renner & Rosenzweig 1987; Van Praag et al. 2000) and the current data cannot answer which changes in particular could have caused the observed differences in the phenotypes. Therefore, it would be interesting to study how far EC housing of the two strains will lead to changes in such factors as the GABAergic or dopaminergic system, known to be involved in the expression of Type 1 and Type 2 SWD (Danober et al. 1998).

EC housing combines many different factors such as a stimulus complex and always changing environment, social interactions and physical activity. Hence, it cannot be answered unambiguously which factors of enrichment have influenced the expression of Type 1 and Type 2 SWD. It has been shown (Ferchmin & Bennett 1975; Rosenzweig et al. 1978), however, that it seems to be rather the combination of the different factors leading to the enrichment effects, than a given factor alone (Van Praag et al. 2000).

Independent of which housing factors have caused the observed changes in mean duration of Type 1 SWD and number of Type 2 SWD, it is interesting that the results are contrary to what was found in convulsive epilepsy models. Housing rats in an EC condition delayed (Auvergne et al. 2002), or even inhibited (Young et al. 1999) the occurrence of convulsive seizures. In WAG/Rij rats, however, EC housing led to a prolongation of Type 1 SWD and hence to a ‘worsening’ of absence seizures in those rats showing the trait. Given the fact that many different rat strains develop absence seizures (Willoughby & Mackenzie 1992), it would be interesting to study how far the worsening of absence seizures by EC housing can be generalized to other strains with a large number of SWD. Nevertheless, the opposite effects of EC housing on convulsive and absence seizures fit nicely into the common observation that anti-epileptic drugs such as tiagabine or vigabatrin, but also carbamazepine and phenytoin, have opposite effects on convulsive and absence seizures (Danober et al. 1998; Peeters et al. 1988; Vergnes et al. 1984).

In summary, the results of the present study show that the development and the expression (number and mean duration) of Type 1 and Type 2 SWD at three months of age are differently influenced by strain and housing. This supports earlier findings, showing that Type 1 and Type 2 SWD are independent phenomena and that number and mean duration of these paroxysms are controlled by different mechanisms. Finally, different from what was found for convulsive seizures, enriched housing led to worsening of absence seizures in WAG/Rij rats, which underlines the opposite characteristics of convulsive and non-convulsive types of epilepsies.

Acknowledgments

This research was supported by the Dutch Organization for Scientific Research (NWO), Grant 425-20-401. The valuable help of Willeke Arentz, Jean-Paul Dibbets, Joeri van der Kloet, Hans Krijnen, Gerard van Oijen and the late Willie van Schaijk is kindly acknowledged.

Ancillary