Info

Copyright © 2006, Elsevier Inc. All rights of reproduction in any form reserved.

perform and marked brain swelling could introduce injury to the cortex. Anticonvulsant drugs should be avoided. In general within 2 to 4 weeks sharp waves can be detected over the injection site. Within 1 to 2 months, epileptiform activity and clinical seizures will begin to be observed. These are initially focal and frequently progress to Jacksonian-type marches. Secondary generalized seizures are not uncommon.

Characteristics and Defining Features Behavioral and Clinical Features

Monkeys that received alumina gel applications to the left cerebral hemispheres to produce seizure foci were analyzed for behavioral and EEG changes. Injections into the precentral and postcentral gyri did not cause any abnormalities in the EEG tracings initially. Within 4 to 8 weeks, however, abnormalities appeared on the EEG tracings and focal seizures began in the contralateral upper extremity corresponding to the alumina gel-injected area of the sensori-motor cortex that is the homologous region representing that extremity. Seizure activity was observed in most monkeys after 4 weeks, and the seizure frequency was two or three per day. The duration of each seizure ranged from a few seconds on electrocorticography (ECoG), up to 1 to 2 minutes behaviorally when the monkey's focal seizure generalized. The epileptic animals demonstrated an ECoG consistent with spike-and-wave discharges in the area immediately adjacent to the alumina gel injection. As indicated by previous studies, the seizures begin with a tonic-clonic movement of the hand contralateral to the cortical injection site, followed by tonic-clonic movements of the ipsilateral upper extremity. These tonic-clonic events often generalize to the lower extremity and the face, leading to the monkey's falling in the cage and to unconsciousness. Although seizure activity spreads to both upper extremities, there is no evidence of a mirror focus in the contralateral homotopic cortex in monkeys; Harris and Lockard (1981) did not show this region as having independent seizure activity.

Electroencephalographic Features

The ECoG showed spike-and-wave activity around the electrodes overlying the injected sensorimotor cortex. This spike-and-wave activity corresponded to the behavioral manifestations described in the preceding section.

Neuropathology

Cell Loss

Blocks of tissue from epileptic cortex and a homologous area in the contralateral nonepileptic cortex were analyzed in light and electron microscopic preparations. As a result of the injections into the sensorimotor cortex, the area bordered by the injection site often shows a disruption in the laminar pattern in Nissl-stained sections. It is remarkable that adjacent to the injection site, normal-looking pyramidal neurons are seen throughout the cortical layers, suggesting that the amount of neuronal loss is minimal outside the injection site. In monkeys with chronic seizures a significant loss of neurons was observed that ranged from 11 to 61% at the focus compared with the normal side (Ribak et al., 1989). Details about the loss of g-aminobutyric acid (GABA)ergic neurons are provided later in a separate section of this chapter.

Reactive Gliosis

The glial changes consist of thick pial-glial membranes (Harris, 1980), hypertrophied neuroglial cell processes containing increased numbers of filaments, and increased numbers of intracellular glial connections of both the gap junction and desmosome type. In layers IV and V of cortex, hypertrophied astrocytes are seen in organized bands where high-potassium ion fluxes have been demonstrated to be associated with seizures (Harris, 1980). Furthermore, analysis of the surface of pyramidal cell bodies in layer V showed a statistically significant increase in apposition by astrocytic processes (Ribak et al., 1982) (Figure 1). For example, at the focus, 50% of the soma of pyramidal cells was apposed by glia, whereas 22% in the contralateral normal cortex were apposed by glia (see Figure 1). In addition, the area in the neuropil occupied by glial profiles showed a 50% increase at the focus relative to the normal cortex (Ribak et al., 1982). Therefore reactive gliosis is a well-documented neuro-pathological finding in this model of focal epilepsy.

Loss of GABAergic Neurons

To test the hypothesis that a decrease in the number of inhibitory GABAergic neurons could lead to seizure activity, several studies were performed in this model of epilepsy, beginning in the late 1970s (Ribak et al. 1979). Initially light microscopic analysis of glutamate decarboxylase (GAD) immunocytochemistry indicated a significant decrease in the number of GAD-positive terminals at the epileptic focus compared with the contralateral normal cortex. Because many of these terminals are apposed to the surfaces of layer V pyramidal cells, it was suggested that these pyramidal neurons were inhibited less and therefore were more likely to be engaged in excitatory circuitry and bursting (Ribak et al., 1979). Subsequently, Houser et al. (1986) analyzed GABAergic axon terminals in monkeys before seizure onset and reported a decrease in GABAergic terminals ranging from 14 to 22%, whereas the initial studies of chronic monkeys reported a loss of 58 to 62% (Ribak et al., 1979).

The GABAergic somata were analyzed (Ribak et al. 1986) using an antibody that recognizes GAD within somata (Oertel et al., 1981). The results indicated a 35 to 52% significant loss of small GABAergic somata at the foci of

FIGURE 1 Bar graphs showing the number of synapses per 10 ||m-length of somal surface of 90 layer V pyramidal somata (left) and the amount of glial apposition to these layer V pyramidal somata (right). The number of axosomatic synapses was greatest in the nonepileptic cortex (N) and least at the focus (F). The parafocus (P) displayed an intermediate number. The opposite result was found for glial apposition. The error bars represent standard error of the mean (SEM). (Reprinted with permission from Ribak et al., 1982.)

FIGURE 1 Bar graphs showing the number of synapses per 10 ||m-length of somal surface of 90 layer V pyramidal somata (left) and the amount of glial apposition to these layer V pyramidal somata (right). The number of axosomatic synapses was greatest in the nonepileptic cortex (N) and least at the focus (F). The parafocus (P) displayed an intermediate number. The opposite result was found for glial apposition. The error bars represent standard error of the mean (SEM). (Reprinted with permission from Ribak et al., 1982.)

chronic epileptic monkeys. These data suggested that the loss of GABAergic terminals at epileptic foci is due to the loss of neuronal somata that give rise to these axons. Similar to the study by Houser et al. (1986), Ribak et al. (1989) analyzed the cortex next to the alumina gel application for a loss of GABAergic somata before onset of epileptic activity in monkeys. They observed a 23 to 44% reduction in the number of GABAergic neurons at the site of the developing focus. Nissl preparations were also used to determine the selectivity of this loss, and statistical tests showed no differences in the number of total neurons between this and the other site. In contrast, chronically epileptic monkeys had a significant loss of neurons that ranged from 11 to 61% at the focus, compared with the normal side.

To determine whether the loss of GABAergic terminals was due to an actual degeneration of these terminals or simply to a loss of GAD immunostaining within the terminals, electron microscopy was utilized. GABAergic axon terminals were identified in electron microscopic preparations by the fact they form symmetric synapses. In two studies (Ribak, 1985; Ribak et al., 1982), pyramidal neurons were analyzed for a loss of GABAergic synapses on both their cell bodies and axon initial segments (Figure 2). The results showed an 80% loss of such axon terminals on the cell body (see Figure 1) and a more severe loss of axon terminals forming symmetric synapses on the axon initial segment. These findings indicated that the previous reported loss of GAD-positive axon terminals was due to their degeneration and that two GABAergic cell types are particularly affected at the epileptic focus: the basket and chandelier cells.

Biochemical Findings

The biochemical data for alumina gel-treated epileptic monkeys are consistent with the immunocytochemical results stated previously herein. Bakay and Harris (1981) reported decreased GABA receptor binding, GABA concentration, and GAD activity at the epileptic focus. They suggested that these alterations in presynaptic GABA indices were probably due to the degeneration of GABAergic axon terminals at the epileptic focus. Furthermore, they reported that seizure frequency was correlated with the loss of GAD activity. These results support the GABA hypothesis for focal epilepsy.

Response to Antiepileptic Drugs

Anticonvulsant drugs have been screened using the alumina gel-injected monkeys. Two of these drugs, pheny-toin and phenobarbital, have been shown to protect alumina gel-injected monkeys from developing secondary generalized tonic clonic seizures. In addition both drugs decrease seizure frequency and severity relative to placebo-treated animals (Lockard et al., 1976a, b). However, if cessation of drug treatment occurs in the monkeys injected with alumina gel, seizure frequency and severity are increased relative to placebo-treated monkeys (Lockard et al., 1976b).

FIGURE 2 Electron micrographs of axon initial segments from layer III pyramidal cells from normal cortex (a) and the epileptic focus (b). The normal axon initial segment in (a) displays two symmetrical synapses (arrows) formed with axon terminals. Two characteristic features of axon initial segments are shown: a fasciculation of microtubules (M) and a dense undercoating that is continuous with the coating over a pinocytotic vesicle (V). In the focus, reactive astrocytes (A) were seen where axon terminals were found in the normal cortex. The initial segment showed normal features including microtubules (M) and a dense undercoating (arrow). Bar = 1 mm. (Reprinted with permission from Ribak, 1985.)

FIGURE 2 Electron micrographs of axon initial segments from layer III pyramidal cells from normal cortex (a) and the epileptic focus (b). The normal axon initial segment in (a) displays two symmetrical synapses (arrows) formed with axon terminals. Two characteristic features of axon initial segments are shown: a fasciculation of microtubules (M) and a dense undercoating that is continuous with the coating over a pinocytotic vesicle (V). In the focus, reactive astrocytes (A) were seen where axon terminals were found in the normal cortex. The initial segment showed normal features including microtubules (M) and a dense undercoating (arrow). Bar = 1 mm. (Reprinted with permission from Ribak, 1985.)

Other drugs that have been screened in the alumina gel injected monkeys include valproic acid. Results from these studies indicate that the efficacy of valproic acid in decreasing seizure frequency and duration correlates with plasma levels and that plasma levels show drastic fluctuations. When plasma levels are high, relatively few EEG interictal spikes are seen, and when they are low, relatively high numbers of interictal spikes are observed (Lockard et al., 1977). Thus valproic acid plasma levels inversely correlate with EEG interictal spike activity. In these monkeys, there is a high correlation between the number of interictal spikes and occurence seizures (Lockard, 1980).

Benzodiazepines and their derivatives are common antiepileptic treatments in humans. One of these, clon-azepam, has been evaluated in the alumina gel-injected monkey. When administered to achieve plasma levels above

29.9ng/ml and 59.9ng/ml, clonazapam caused seizures to decrease in severity and frequency or disappear, respectively (Lockard et al., 1979). It is pertinent that while being treated with clonazapam, the monkeys exhibited improved performance on a lever-pressing behavioral task. However, if drug treatment was stopped, seizure activity increased and a deficit in lever pressing behavior was seen (Lockard et al., 1979).

Limitations

In general the alumina gel model can be introduced into any nonhuman primate species, sex, or age. Seizures almost always develop when the appropriate injection technique is used, and in the rare incidences when they do not develop, repeat injections are usually successful. A properly prepared animal will produce stable spontaneous seizures for many years. However, in an occasional subject, status epilepticus will develop and anticonvulsants are needed to maintain viability. This is more likely to occur with large injections and subpial spread.

There is a relatively high variability as to the seizure rates, with rates as low as one per month or as high as 10 to 20 per day, even though the same injection technique, species, age, and sex of monkey are used. To get an accurate seizure frequency, 24-hour video monitoring is required. A movement-detection device or ECoG monitoring technique allows for the videotape review to be streamlined (Lockard and Ward, 1980). Like the human condition, seizures in monkeys may cluster in relation to estrus and may increase with behavioral or physical stress. These features are particularly useful in a model of focal neocortical epilepsy. Although seizures can be generated in neocortical areas other than sensorimotor cortex, they are more difficult to document and far more variable in their development and maintenance. Such seizures may model a variety of epileptic activities, which might result from dysgenesis, vascular abnormality, or a lesion or tremor. The disadvantages of this method are that it is very expensive and the number of subjects that can be involved in any one project is relatively small; the monitoring techniques are complex and the studies require long-term commitment of resources. The advantages are that the model involves a nonhuman primate model with anatomic, biological, pharmacologic, and elec-trophysiologic properties much closer to the human than those found in rodent models.

Insights into Human Disorders

The neuropathological and biochemical results from the analysis of the sensorimotor cortex of monkeys injected with alumina gel indicate a causal role for GABAergic neuronal loss. This model best replicates most of the features of human focal epilepsy. This model is also a chronic one and follows a time course in its development similar to post-traumatic epilepsy in humans (Wyler and Ward, 1984). That is, the EEG abnormalities and clinical seizures develop at a similar time in this model and in humans with posttraumatic epilepsy. Based on the data from this experimental model, we suggest that people with this type of epilepsy probably display a reduction in the number of GABAergic neurons and synapses at the epileptic focus. This hypothesis is supported by the pharmacologic data showing that GABAergic agonists are effective to varying degrees at limiting the frequency, duration, and severity of seizures in monkeys injected with alumina gel.

Was this article helpful?

0 0

Post a comment