Veratridine-Induced Wet Dog Shake Behaviour and Apoptosis in Rat Hippocampus


Author for correspondence: Sameer Otoom, Royal College of Surgeons in Ireland, Medical University of Bahrain, P.O. Box 15503, Adliya, Kingdom of Bahrain (fax +973 17583600, e-mail,


Abstract: We have previously evaluated veratridine as an in vitro model of seizure using conventional electrophysiological recordings in rat hippocampal CA1 pyramidal neurones. The aim of this investigation is to further characterize this convulsant as an in vivo model of seizure. Veratridine was administered intraperitoneally to male Fisher rats in a dose range of 100–400 μg/kg. Within 5 min. after the injections, the animals entered a quiescent period which was followed 10–15 min. later by facial automatism (washing), grooming, masticatory jaw movement and profuse salivation. This phenomenon was followed by the development of wet dog shake and forelimb clonus. The time (mean±S.E.M.) for the onset of induction of these shakes for all tested doses was 31.65±2.85 min. and the number of shakes (mean±S.E.M.) 30 min. after the onset was 17.2±2.85. The onset and number of wet dog shakes induced by veratridine was dose-dependent. No rat death was recorded until 2 weeks after the experiments. Histopathological studies of animals 2 weeks after veratridine administration showed evidence of apoptosis in the hippocampus. Our results indicate that veratridine produced a behavioural pattern of a limbic seizure which mimics temporal lobe epilepsy in man. Based on our previous findings in vitro and of this investigation in vivo, veratridine can be used as an experimental tool to evaluate potential antiepileptic drugs effective against this type of limbic behaviour.

We have characterized the proconvulsant activity of veratridine in vitro in rat brain slices and conventional electrophysiological intracellular techniques in hippocampal CA1 pyramidal neurones (Tian et al. 1995). Veratridine (0.03–0.2 μM) caused no changes in the passive membrane parameters including the resting membrane potential and input resistance. These concentrations also induced a relatively slow, large, synaptic-independent potential called the slow depolarizing after-potential. Higher doses of veratridine (over 0.1 μM) induced bursting, or seizure-like activity after applying a brief super threshold intracellular stimulation (Tian et al. 1995). The duration of seizure-like activity bursting which can be up to 10 sec. was dependent on the amplitude of slow depolarizing after-potential, and not on the stimulus strength or duration. At 0.3 μM or higher concentration, veratridine induced spontaneous rhythmic bursting. Both veratridine-induced evoked or spontaneous rhythmic bursting were sensitive to membrane potential changes and independent of synaptic transmission (Otoom et al. 1998). This was evident when seizure activity persisted after complete blockade of evoked synaptic potential with kynurenic acid.

The mechanism of veratridine-induced bursting activity in hippocampal CA1 pyramidal neurones revealed that veratridine accentuated the depolarizing rectification so that a zero or negative slope appears in the current-voltage curve of untreated neurones. Both the veratridine-induced bursting activity and the negative slope resistance were blocked by low concentrations of tetrodotoxin or by raising the calcium concentration of the superfusion medium (Alkadhi & Tian 1996; Otoom et al.1998). Using single-electrode voltage clamp technique, we showed that veratridine enhanced a subthreshold persistent (slowly inactivating) sodium current. This current is thought to be responsible for the slow depolarizing phase of bursting activity, the development of negative slope resistance and the induction of bursting in rat hippocampal CA1 pyramidal neurones (Alkadhi & Tian 1996; Otoom et al. 1998).

We have also tested different drugs on this model. Therapeutic concentrations of lamotrigine and carbamazepine inhibited both evoked and spontaneous bursting induced by veratridine. This inhibition was accompanied by an increase in the firing threshold of the bursting but no change in the resting membrane potential and membrane input resistance (Otoom & Alkadhi 2000a; Otoom & Nusier 2001). Similarly, therapeutic concentrations of phenytoin inhibited both evoked and spontaneous bursting in a voltage-dependent fashion but produced an increase in the membrane input resistance (Otoom & Alkadhi 2000b). Interestingly, valproic acid produced a biphasic effect on this model. Therapeutic concentrations of valproic acid inhibited both evoked and spontaneous bursting in a voltage-dependent manner without affecting membrane resting potential or input resistance (Otoom & Alkadhi 2000b). However, large concentrations of valproic acid (5 mM or more) enhanced rather than inhibited the epileptiform activity induced by veratridine (Otoom & Alkadhi 1999). During the proepileptic phase of valproid acid, a membrane depolarization accompanied by a decrease in membrane input resistance were evident. The voltage-dependent proepileptic effect of valproic acid was blocked by tetrodotoxin but not by the calcium channel blockers, diltiazem or omega-conotoxin GVIA (Otoom & Alkadhi 1999). Seizure-like activity of veratridine was also sensitive to the action of the anaesthetic agent propofol (Otoom & Hasan 2004) but not to the antiepileptic drug ethosuximide even at high concentrations (Otoom & Alkadhi 2000b).

Systemic administration of kainic acid, 4-aminopyridine, picrotoxin and bicuculline in the rat results in the development of a limbic seizure characterized by excitotoxic syndrome, consisting of automatisms known as wet dog shake. This behavioural effect was concomitantly associated with seizure activity changes in the hippocampus recorded in the electroencephalogram, which rapidly propagated to the other structures (Turski et al.1985; Kubova et al. 1990). Wet dog shakes are associated with status epilepticus in the kindled animals in the amygdala fast kindling model (Rodrigues et al. 2005). It is documented that while the presence of WDS is an index of partial limbic seizures, its absence indicates a fully generalized state (Rondouin et al. 1987). Wet dog shake was also observed after morphine withdrawal which was also associated with suppression of rapid eye movement sleep and a decline in electroencephalographic spectral power during slow-wave sleep episodes (Young & Khazan 1985).

Based on our previous findings which show veratridine as an in vitro model of epilepsy, this investigation has been carried out to characterize the behavioural effects of veratridine in vivo.

Materials and Methods

Animals. Male Fisher rats weighing 190–220 g were housed under standard laboratory conditions with a natural light-dark cycle (lights on 7 a.m.–7 p.m.) for at least one week prior to the experiments. The animals were maintained on rat chow-pellets and tap water ad libitum in a climate-controlled environment (23±1 °). Each rat was placed in a plexiglass cage (25×35×20 cm). The observations were carried out in a well-lit quiet room. Wet dog shakes were counted after intraperitoneal injection of veratridine over a period of 30 min. after their onset. Some experiments were videotaped.

Drugs. Veratridine was obtained from Sigma Chemicals-USA and dissolved in ethanol 50 mg/ml of solution.

Histopathology studies. The rats were decapitated after 2 weeks of veratridine. The brains were taken out and fixed in 10% neutral formalin solution; the paraffin-embedded tissue blocks were processed by conventional method. The blocks were cut in to 2.5–4 μm thick sections and were stained by routine haematoxylin and eosin, and examined using light microscope (Olympus BX 41) under 100 and 400×magnification. The sections were examined for any morphological evidence like apoptosis which is characterized by shrunken cytoplasm, disintegrated nucleus and pyknosis.

Statistical analysis. Data are expressed as mean±S.E.M. and analyzed using one-way analysis of variance (ANOVA). P values of less than 0.05 were considered statistically significant.


Veratridine in doses of 100–400 μg/kg was administered intraperitoneally to the rats. Four to five min. after the injections, the animals entered a quiescent period. Ten–15 min. after this period, the rats began to have excessive grooming, face washing, masticatory jaw movement, and profuse salivation followed by the development of wet dog shakes involving the head and trunk. The latency (mean±S.E.M.) of these shakes for all tested doses was 31.65±2.85 min. and the number (mean±S.E.M.) of shakes 30 min. after the onset was 17.2±2.85. Although increasing doses of veratradine resulted in a decrease in latency for the onset of wet dog shakes (table 1), the difference was not statistically significant between the four tested doses. However, the number of shakes induced by veratridine 30 min. after the onset was dose-dependent and a statistically significant difference between the doses was elicited (fig. 1). After the appearance of wet dog shakes, the rats also developed forelimb clonus and status epilepticus. The number of shakes mostly disappeared after 3–4 hr of the onset of their development. The rats were observed for 2 weeks after completing the experiments for any behavioural abnormalities. None of the rats treated with veratridine died during this period of observation. Histopathological sections from the CA1/CA2 pyramidal cells of the hippocampus of the rats 2 weeks after veratridine exposure showed the presence of apoptotic cells (fig. 2).

Table 1.  Effect of different doses of veratridine on the onset of wet dog shakes.
Dose (μg/kg)
(N=5 for each dose)
Onset of wet dog shakes (min.)
Figure 1.

Dose-dependency of veratridine-induced wet dog shakes (WDS). Each bar represents the mean of 5 observations in rats. Vertical lines are S.E.M. There is a statistical significant difference in the number of WDS with increasing veratridine dose (one-way ANOVA, P-value<0.05).

Figure 2.

Photomicrograph of rat brain showing apoptotic cells in the CA1/CA2 pyramidal cells of the hippocampus (haematoxylin & eosin staining, magnification ×100).


Animal models of seizures not only helped in understanding the physiological and pathological changes in human epilepsy but they also had an important role in the discovery of antiepileptic drugs (Engel 1992). As many of the physiological changes of clinical epilepsy are still not fully understood, the future need for developing new animal models continues to be important and necessary. Our experiments characterized the behavioural effects of veratridine in vivo and showed that it produced a dose-dependent increase in the number of wet dog shakes. This behavioural effect was accompanied by morphological changes in the hippocampus presented as neuronal apoptosis.

The wet dog shakes were previously observed during the generation of limbic seizure and described as ictal and postictal event (Kleinrok & Turski 1980). This behaviour which is characterized by shakes of head and the trunk was observed in different models of partial epilepsy such as limbic kindling stimulation or the intralimbic injection of kainic acid or quisqualic acid (Kleinrok & Turski 1980). Models of generalized epilepsy such as convulsions induced by electroshock, or by the injection of chemicals such as pentylenetetrazol failed to generate wet dog shakes in rats. The conclusion derived from these studies that wet dog shakes correlates with the partial limbic seizures and its number can be used as an index of the involvement of limbic structure, in particular, the hippocampus in evoked seizure. Its absence indicated a fully generalized state (Rondouin et al. 1987).

Many studies suggest that wet dog shake in rats is mediated by 5-HT2 receptors (Lucki et al. 1984; Fone et al. 1991). Intrathecal injection of 5-HT agonists such as 2,5-dimethoxy-α,4-dimethylbenzene ethamine hydrochloride produced a dose-dependent back muscle contraction and wet dog shake in Wistar rats. This behaviour was attenuated by pretreating the rats with 5-HT antagonists such as ritanserin (Fone et al. 1991). The 5-HT2A receptor agonist 1-(2,5-dimethoxy-4-iodophenyl)-2-aminopropane induced wet dog shake in a dose-dependent manner that was blocked by the 5-HT2A selective antagonist ketanserin (Kitamura et al. 2002). When these receptors were up-regulated by chronic administration of ACTH (Kuroda et al. 1992), wet dog shake induced by 5-HT2A receptor agonist was enhanced. On the other hand, the reduction of the density of these receptors by chronic administration of imipramine (Peroutka & Snyder 1980), attenuated 5-HT2A receptor agonist-induced wet dog shake (Kitamura et al. 2002).

Moreover, wet dog shake behaviour was also induced by intracerebroventricular injection of the cholinomimetic drug carbachol in rats which was not related to increase in the activity of central serotonergic mechanisms (Turski et al. 1981a). However, this response was abolished by the administration of atropine suggesting the involvement of muscarinic receptors in this behaviour (Turski et al. 1981b). Similar results were obtained by intrahippocampal pilocarpine given via microdialysis to freely moving rats. Pilocarpine-induced wet dog shake was accompanied by an increase in the level of glutamate and dopamine in the hippocampus and was also blocked by atropine (Smolders et al. 1997).

We found that veratridine produced behavioural effects similar to those induced by kainic acid which is considered a model of complex partial seizure in human temporal lobe epilepsy (Engel 1992). Acute injection of kainic acid in rodents causes features of limbic seizure characterized by facial automatism such as eye blinking and chewing, head bobbing, loss of posture, scratching and wet dog shake that is followed by forelimb clonus leading to status epilepticus (Ben-Ari 1985). The induction of this seizure was accompanied by an increase in the level of mRNA encoding sodium channel subunits in the hippocampus but not in the neocortex of rat (Bartolomei et al. 1997). Kainic acid produced lesions in the brain that are characterized by loss of γ-aminobutyric acid (GABA)-ergic interneurones in the dentate hilus and mossy fibers sprouting in the dentate gyrus. Similar to our finding using veratridine, kainic acid was also associated with cell death in the area CA1 and CA3 of the hippocampus (Tauck & Nadler 1985).

Our results indicate that veratridine produced evidence of apoptosis in the hippocampus which is a known structure that show epilepsy-related damage. This structure contains muscarinic and N-methyl-D-aspartate (NMDA) receptors. It is also rich in glutamatergic and GABAergic input (Cotman & Monaghan 1986). Veratridine was shown to produce calcium-dependent death in rat hippocampal neurones and cerebellar granular cells (Pauwels et al. 1989). This is explained by veratridine depolarization of excitable cells and preventing inactivation of sodium channels that leads to an increase in the influx of sodium and calcium, thus cell death (Jordan et al. 2000). Additionally, veratridine apoptosis may be related to NMDA receptor-mediated excitotoxicity through an increase in the release of glutamate (Smolders et al. 1997). It has been shown that NMDA receptor antagonists such as MK-801 at nanomolar concentrations protect against glutamate-induced neuronal cell death (Pauwels et al. 1989). Moreover, veratridine-induced apoptotic death was also demonstrated in bovine chromaffin cells in vitro which share common features with neurones such as the embryological origin and the process of catecholamine release (Le Douarin 1986). The mechanism of this death was related to an increase in calcium, mitochondrial dysfunction and the production of superoxide (Maroto et al. 1994).

In conclusion, results from this in vivo study that characterized the behavioural pattern induced by veratridine support our previous results performed in vitro on this convulsant. Thus, we suggest that the wet dog shakes induced by veratridine together with the known mechanism of its action as a convulsant agent make it a potential experimental tool to evaluate drugs against this type of limbic behaviour.


This study was partially supported by a grant from the Arabian Gulf University-Kingdom of Bahrain, Number 17/2004. The authors like to thank Professor Ali Satir and Dr. Urmil Brahmi, Department of Pathology, Arabian Gulf University for their assistance.