The Effect of Electrical Stimulation and Lesioning of the Anterior Thalamic Nucleus on Kainic Acid–Induced Focal Cortical Seizure Status in Rats

Authors


Address correspondence and reprint requests to Dr. S. Takebayashi at Department of Neurosurgery, Asahikawa Medical College, 2-1, Midorigaoka-Higashi, Asahikawa 078-8510, Japan. E-mail: tikurin@asahikawa-med.ac.jp

Abstract

Summary: Purpose: The present study aimed to clarify the effect of electrical stimulation and lesioning of the anterior nucleus of the thalamus (ANT) on kainic acid (KA)–induced focal cortical seizures in a rat model. To address the mechanism underlying these anticonvulsant actions, cerebral glucose metabolism after ANT electrical stimulation and lesioning was also examined.

Methods: Wistar rats were divided into five major groups: control (n = 9), unilateral (n = 9), and bilateral (n = 9) ANT electrical stimulation, and unilateral (n = 9) and bilateral (n = 9) ANT lesioning. After KA injection, average clinical-seizure frequencies in each group were measured. Electrical stimulation of ANT was introduced after induction of seizure status epilepticus. Stimulation was on for 30 min and off for 30 min per 60-min cycle. Local cerebral glucose utilization (LCGU) was also measured by using [14C] 2-deoxyglucose autoradiography in three groups of rats: control (n = 7), bilateral ANT stimulation (n = 7), and bilateral ANT lesioning (n = 7).

Results: Unilateral ANT electrical stimulation and lesioning significantly reduced clinical seizure frequency, compared with control animals. Strikingly, no animals treated with bilateral ANT procedures demonstrated any clinical seizure. LCGU was markedly increased in the sensorimotor cortex, striatum, thalamus, mammillary body, and midbrain tegmentum of control group rats after KA injection, but no increase in LCGU was noted in rats treated with bilateral ANT lesioning or stimulation.

Conclusions: The electrical stimulation and lesioning of ANT suppressed focal cortical clinical seizures induced by KA injection. Additionally, an analysis of cerebral metabolic changes indicated that these procedures might suppress the function as amplifier and synchronizer of seizure activity.

Electrical stimulation of deep brain structures for the treatment of medically intractable epilepsy has been a subject of recent focus (Fisher et al., 1992; Velasco et al., 1993, 2001; Mirski et al., 1997; Benabid et al., 2000; Loddenkemper et al., 2001; Blumenfeld, 2002; Hodaie et al., 2002; Hamani et al., 2004; Kerrigan et al., 2004). Deep brain stimulation (DBS) promises a potentially reversible and minimally invasive therapeutic tool for treating epilepsy, compared with classic resective surgery. Consequently DBS may provide a new strategy in patients who are not eligible for resective surgery. In particular, electrical stimulation of the anterior nucleus of the thalamus (ANT) is receiving increased attention as a novel surgical therapy for epilepsy (Mirski et al., 1997; Hodaie et al., 2002; Hamani et al., 2004; Kerrigan et al., 2004).

The ANT projects largely to the cingulate gyrus, whence it further projects to limbic structures and wide regions of neocortex (Nolte, 1999), so the stimulation of this relatively small anatomic region can influence physiologic activity throughout areas of cortex.

The rationale underlying electrical stimulation of the ANT as an effective treatment of epileptic seizures is based on both animal data and limited preliminary human studies (Mirski et al., 1997; Hodaie et al., 2002; Hamani et al., 2004; Kerrigan et al., 2004). The exact mechanism of the anticonvulsant effect of ANT electrical stimulation is not fully understood, and studies monitoring the anticonvulsant benefits of this procedure in focal seizure models are scarce, compared with systemic convulsant models such as pentylenetetrazol (PTZ) or pilocarpine administration (Mirski et al., 1984, 1986a, 1986b, 1986c, 1987, 1994, 1997; Hamani et al., 2004).

Therefore we used a rat model of focal cortical seizures induced by intracortical kainic acid (KA) injection to determine the anticonvulsant effects of electrically stimulating the ANT. We tested the effects of both unilateral and bilateral electrical stimulation and introduced lesioning of the ANT as a comparison procedure. In addition, we monitored changes in cerebral glucose metabolism after lesioning or during electrical stimulation of the ANT, to elucidate a mechanism for the anticonvulsant action.

METHODS

Adult male Wistar rats weighing 280–320 g were used in these experiments; they were housed in a temperature-controlled room (20 ± 3°C) with a 12-h light/12-h dark cycle. All animals were provided with water and food ad libitum. The experimental protocol was approved by the Animal Care and Management Committee of Asahikawa Medical College.

Experiment 1: The effect of electrical stimulation and lesioning of the anterior nucleus of the thalamus

Animal groups

Animals were divided into five major groups of nine rats each, as follows.

  • 1Control group (n = 9); rats received KA injections without implantation of electrodes into the ANT.
  • 2Unilateral electrical stimulation group (n = 9); rats received KA injections after undergoing unilateral implantation of electrodes into the ANT. Unilateral implantation was always introduced on the left side.
  • 3Bilateral electrical stimulation group (n = 9); rats received KA injections after undergoing bilateral implantation of electrodes into the ANT.
  • 4Unilateral lesioning group (n = 9); rats received KA injections after undergoing unilateral lesioning of the ANT. Unilateral lesionings were always introduced on the left side.
  • 5Bilateral lesioning group (n = 9); rats received KA injections after undergoing bilateral lesioning of the ANT.

Stereotactic procedures

Rats were fixed in a stereotactic apparatus under intraperitoneal pentobarbital anesthesia (40 mg/kg). Rectal temperature was monitored and maintained by using a warm bulb lamp throughout the experiment. A stainless-steel chemitrode (0.6 mm in outer diameter) with an inner stainless-steel needle (0.3 mm in diameter) was inserted stereotactically into the left sensorimotor cortex (SMC) through a burr hole for KA injection and electroencephalogram (EEG) recording, according to the atlas of Paxinos and Watson (1986). A stainless-steel screw was also placed on the dura over the right SMC for recording the EEG. Stereotactic coordinates for the SMC were 1.0 mm anterior to the bregma, ±2.5 mm lateral to the sagittal suture, and 1.0 mm deep to the dura mater. An additional screw was placed in the frontal sinus for the nonspecific electrode. A ground electrode was placed in the midline, posterior to the occipital suture.

For electrical stimulation, insulated bipolar twisted-wire electrodes (0.2 mm in diameter with the cut surface exposed) were stereotactically inserted into the left or bilateral ANT. Stereotactic coordinates for the ANT were 2.1 mm posterior to the bregma, ±1.6 mm lateral to the sagittal suture, and 5.7 mm deep to the dura mater.

Thalamotomy (microlesioning of the ANT) was performed by delivering a 3-mA current for 10 s, through a radiofrequency generator (Ugo Basile lesion-making device, model 3500; Ugo Basile, Comerio, Italy). All of the electrodes were connected to a socket with six channels, which was fixed to the skull with the chemitrode by using resin dental cement. After surgery, rats were placed under a warming lamp until they were able to move around the cage, and then were transferred back to the animal-housing facility to recover for ≥5 days.

KA-induced focal cortical seizure

At 5–7 days after surgery, the inner needle was removed, and a stainless-steel injection cannula (0.3 mm in outer diameter) was inserted into the chemitrode in the left SMC. KA (Sigma, St Louis, Mo, U.S.A.) was dissolved in phosphate buffer solution (0.2 M, pH 7.4) at a concentration of 2 μg/μl. One μl of KA solution was then administered into the left SMC over a 3-min period. These procedures prevented leakage of KA solution into the surrounding structures. The effectiveness of this method was confirmed initially by using KA solution containing 0.2% methylene blue. It was also confirmed histologically that diffusion of the dye with KA solution was 1.0 mm in diameter and limited within the left SMC. Immediate changes after KA injection into a unilateral SMC have been described elsewhere (Yamamoto et al., 1995; Hashizume and Tanaka, 1998). In brief, focal cortical status epilepticus with occasional secondarily generalized seizures (SGSs) occurred 60–90 min after KA injection. Facial or forelimb clonus, rearing, and falling were observed and lasted for 4–8 h. These clinical and EEG seizures gradually subsided within 24 h. Rat behavior and EEGs were continuously monitored and recorded by a video-EEG system for ∼24 h (Video-EEG, Nihon Koden Neurofax Tokyo, Japan). The EEG filter setting was as follows: The high-cut and low-cut filter were 50 Hz and 1.5 Hz, respectively. The time-constant was 0.1. Alternating current (a.c.) filters (50 Hz) were on. Video recordings were made by using a color video camera, and synchronization of video and EEG was assured by time-stamping the video image. Rats were remonitored intermittently over a period of ∼1 month to detect spontaneous seizures.

Electrical stimulation

Electrical stimulation was performed with an electrical stimulator (Nihon Koden SEN-7203, Tokyo, Japan) at a frequency of 130 Hz and a squared biphasic pulse of 100 μsec. The intensity was gradually increased until a marked increase in motor and vigilant behavior in each of the tested animals was observed; after this point, it was decreased to 70% of the behavioral threshold, within the range of 140–500 μA.

Stimulation parameters, including stimulation frequency, intensity, and duration, were similar to values used in prior rat and human studies of ANT stimulation (Mirski et al., 1997; Hodaie et al., 2002; Hamani et al., 2004; Kerrigan et al., 2004). Because of a high incidence of stimulation-related technical difficulties, most rats underwent only a single stimulation trial. Electrical stimulation of the ANT was begun at ∼60–90 min after KA injection, when clinical seizure frequency became more than one per 10 min. The pulse was on for 30 min and off for 30 min in every 60-min cycle over a total of 4–8 h, and the mean seizure frequencies in these two modes were compared.

Ictal assessment

Motion or stimulation artifacts were frequently observed on EEG recordings, making the frequency and duration of EEG seizures difficult to calculate precisely by using a spike detector. Furthermore, assessment of motor seizure frequency was only approximate because of the inability to characterize precisely all subtle and confusing behavior consisting of mouth or eye clonus, head bobbing, and sniffing. Consequently, we defined a clinical motor seizure by the simultaneous presence of both clear motor manifestation (facial or limb clonus, rearing and falling) and multiple spikes on EEG. We judged the event as an electrical seizure when only multiple spikes on EEG were seen without motor manifestation. In this study, we used the change in clinical seizure frequency to assess the antiepileptic effect of ANT stimulation or lesioning. SGS frequency was also counted to estimate seizure severity instead of using the Racine scale (Racine, 1972), which originally scored limbic seizure. Mean seizure frequency was determined by dividing the total number of clinical motor seizures observed by the duration of monitoring, and groups were compared. EEGs were reviewed without blinding the EEG reviewer.

Histology

At 1 month after KA injection, the rats were deeply anesthetized with an overdose of pentobarbital (100 mg/kg), and the brain was removed and fixed with 10% formalin. Consecutive coronal sections were cut and stained with hematoxylin and eosin. The locations of the lesioning and stimulating electrode tip were verified to be all centered in the target portion of the ANT. Animals with misplaced lesions or intracerebral hemorrhage or infection were excluded from the analysis.

Statistical analysis

The Student's or Welch's t-test and paired t-test were used. Statistical significance was set at p < 0.05. Values in the test are displayed as means ± standard deviation.

Experiment 2: Cerebral glucose metabolism in electrical stimulation and lesioning of the ant

In total, 21 rats were used to investigate the metabolic changes induced by ANT lesioning or electrical stimulation by using autoradiographic techniques. Animals were divided into three groups of seven:

  • 1Control group (n = 7); rats received KA injections without implantation of electrodes of the ANT.
  • 2Bilateral electrical stimulation group (n = 7); rats received KA injections after undergoing bilateral implantation of electrodes into the ANT.
  • 3Bilateral lesioning group (n = 7); rats received KA injections after undergoing bilateral lesioning of the ANT.

All groups were treated as described in the experiment 1. At 5–7 days after stereotactic surgery, a polyethylene catheter was inserted into the unilateral femoral artery and vein under light halothane anesthesia. The lower body and limbs were restrained by a plaster cast to prevent catheter removal. Four hours after KA injection, when clinical focal cortical seizure status epilepticus was induced, 25 μCi of [14C]2-deoxy-d-glucose ([14C]2-DG, Amersham, Piscataway, NJ, U.S.A.) was intravenously injected. Arterial blood samples were collected at 2, 3, 5, 10, 15, 20, 25, 30, 35, and 45 min after the radioactive tracer injection. The rats were decapitated as soon as possible after the last blood collection, and the brain was removed quickly. In the bilateral electrical stimulation group, stimulation was continuously performed from the KA injection to decapitation. Brain specimens were immediately frozen and sectioned into 20-μm thick serial coronal sections by using a cryostat. The sections were dried at 60°C. Dried sections and [14C]methyl metacrylate standards (Amersham) were consecutively placed on x-ray film (Kodak, SB-5) in a cassette and exposed for 7 days. Blood samples were centrifuged, and plasma [14C] radioactivity and glucose concentration were measured (LS6500; Beckman Instruments, Fullerton, CA, U.S.A., and Seralizer, Miles, Elkhart, IN, U.S.A.). The optical density of the autoradiogram was measured, and the tissue [14C] concentration was calculated on an image analyzer (the Inquiry Autoradiography System, Loats Associates Westminster, MD, U.S.A.).

Local cerebral glucose utilization (LCGU) was determined in various brain structures in each group; it was calculated by using the equation of Sokoloff et al. (1977). Dried coronal sections at the level of the ANT were also stained by hematoxylin and eosin to verify the location and size of electrode tip and lesioning. The data were averaged in two consecutive slices to decrease the error due to slice thickness, and comparatively analyzed with Welch's or Student's t-test. Statistical significance was set at p < 0.05. Values in the test are displayed as means ± standard deviation.

RESULTS

Experiment 1: Effect of electrical stimulation and lesioning of the ANT

Electrophysiological effects of ANT electrical stimulation and lesioning

  • 1Control groupMultiple spikes were recorded from the electrode in the left SMC at 30–60 min after KA injection, with subsequent multiple spikes appearing in the right SMC recording (Fig. 1A). All control animals treated with KA alone (n = 9) developed focal cortical seizure status epilepticus with occasional SGSs. Mean seizure frequency was 27.2 ± 9.4 times/h, and mean SGS frequency was 9.4 ± 10.3 times/h (Fig. 2A). However, spontaneous seizures were not detected for the next 30 days by video-EEG monitoring.
  • 2Unilateral electrical stimulation groupAnimals subjected to unilateral electrical stimulation of the ANT (n = 9) exhibited a significantly decreased mean seizure frequency (11.7 ± 3.3 times/h), compared with control animals, and had a mean SGS frequency of 3.0 ± 3.5 times/h (Fig. 2A). In comparison between on and off stimulation, mean seizure frequency was significantly lower with stimulation on (8.8 ± 3.2) than with stimulation off (14.6 ± 4.0), and mean SGS frequencies were 3.0 ± 3.5 versus 1.2 ± 1.6 (off vs. on) (Fig. 2B).
  • 3Bilateral electrical stimulation groupNo animals with bilateral ANT electrodes, even in the unstimulated mode, developed clinical seizures after cortical injection of KA (Fig. 2A). Thus no DBS results were obtained for this group. However, EEG seizure activities were observed in those animals without any obvious clinical motor seizures (Fig. 1B).
  • 4Unilateral lesioning groupAnimals subjected to unilateral lesioning (n = 9) also exhibited significantly decreased mean seizure frequency (4.6 ± 4.4) and SGS frequency (1.2 ± 2.0) (Fig. 2A).
  • 5Bilateral lesioning groupAnimals subjected to bilateral lesioning demonstrated intermittent EEG epileptiform activity, but without clinical motor seizures (Fig. 2A).
Figure 1.

Ictal electroencephalograms (EEGs) in KA-treated control animals. A: Multiple spikes were recorded from the electrode in the left SMC at 30–60 min after KA injection, with subsequent multiple spikes appearing in the right SMC recording. B: Electroencephalograms (EEGs) in the bilateral DBS group. Electroencephalographic seizure activities were observed in the bilateral SMC without any obvious clinical motor seizures. The EEG was recorded during electrical stimulation-off period.

Figure 2.

Figure 2.

A: Bar graphs describing the mean clinical motor seizure frequency and secondarily generalized seizure (SGS) frequency in each of the studied groups. Bars, standards deviations; *statistically significant results. B: Bar graphs describing the mean clinical seizure frequency and SGS frequency between stimulation-off and -on periods in the unilateral DBS group. Bars, standards deviations, *statistically significant results. C: Bar graphs describing the mean clinical seizure frequency and SGS frequency between electrical stimulation-off and -on periods in the Non-ANT targeted group. Bars, standards deviations.

Figure 2.

Figure 2.

A: Bar graphs describing the mean clinical motor seizure frequency and secondarily generalized seizure (SGS) frequency in each of the studied groups. Bars, standards deviations; *statistically significant results. B: Bar graphs describing the mean clinical seizure frequency and SGS frequency between stimulation-off and -on periods in the unilateral DBS group. Bars, standards deviations, *statistically significant results. C: Bar graphs describing the mean clinical seizure frequency and SGS frequency between electrical stimulation-off and -on periods in the Non-ANT targeted group. Bars, standards deviations.

Figure 2.

Figure 2.

A: Bar graphs describing the mean clinical motor seizure frequency and secondarily generalized seizure (SGS) frequency in each of the studied groups. Bars, standards deviations; *statistically significant results. B: Bar graphs describing the mean clinical seizure frequency and SGS frequency between stimulation-off and -on periods in the unilateral DBS group. Bars, standards deviations, *statistically significant results. C: Bar graphs describing the mean clinical seizure frequency and SGS frequency between electrical stimulation-off and -on periods in the Non-ANT targeted group. Bars, standards deviations.

Missed targets of ANT

The percentage of ANT “hits” was 45.0% (36/80) for rats in which the electrode or lesion location could be verified histologically. In particular, 14 rats with unilateral stimulation electrodes positioned outside the target portion of the ANT (i.e., ANT misses) were used to determine whether the effects of stimulation were specific to the ANT. These animals showed no significant decrease in mean seizure frequency (25.8 ± 14.1) or mean SGS frequency (5.9 ± 4.4), compared with control animals (Fig. 2A). Furthermore, onset of electrical stimulation did not produce a significant difference in mean seizure frequency between stimulation on (24.0 ± 13.7) and off (27.7 ± 15.4), or in mean SGS frequency, which was 5.0 ± 3.8 with stimulation on and 7.2 ± 5.4 with stimulation off (Fig. 2C). For the most part, ANT “misses” were mainly located anterior or lateral to the target portion of the ANT (Fig. 3), although three rats with electrodes positioned posterior or medial to the target portion (such as at the mediodorsal nucleus) exhibited a decreased mean seizure frequency (12.1 ± 3.2), compared with control animals.

Figure 3.

Location of electrode tips in Non-ANT targeted group shown schematically [modified from the Atlas of Paxinos and Watson (1986)]. For the most part, electrode "misses" were located anterior or lateral to the target portion of ANT (black dot), although three rats had electrodes positioned posterior or medial to the target portion (black square).

Histopathologic findings

The locations and sizes of the ANT-stimulating electrode tips were as indicated in Fig. 4A. The diameter and volume of lesions were approximately 0.4–0.5 mm and 0.1–0.2 mm3, respectively. Although small gliotic lesions were observed around the electrode tips, neither degeneration nor neuronal cell loss due to the electrical stimulation was observed.

Figure 4.

Photomicrographs of coronal sections in rat brains. A: The locations and approximate sizes of the stimulating electrode tip were as indicated (arrow). B: The locations and approximate sizes of the ANT microlesion were as indicated (arrowhead). C: The location of the target portion of ANT is illustrated. H-E staining, original magnification is x5 in A and B.

The locations and sizes of the ANT microlesions were as indicated in Fig. 4B. The diameter and volume of lesions were ∼0.6–0.8 mm and 0.3–0.5 mm3, respectively. A small gliotic lesion with neuronal cell loss was found in the area where lesioning in the ANT occurred.

Experiment 2: Regional Cerebral Blood Glucose Utilization

The autoradiograms using [14C]2-DG are shown in Figs. 5–9, and the LCGU measurements for each group are summarized in Table 1.

Figure 5.

[14C]2-deoxy-d-glucose autoradiograms of coronal sections in rat brains from (A) control; (B) bilateral DBS group; and (C) bilateral lesioning group. Comparison of coronal sections is at the level of the sensorimotor cortex (SMC). Arrowhead, SMC; arrow, caudate-putamen (CP); Cg, cingulate cortex.

Figure 6.

[14C]2-Deoxy-d-glucose autoradiograms of coronal sections in rat brains from (A) control; (B) bilateral DBS group; and (C) bilateral lesioning group. Comparison of coronal sections is at the level of the anterior nucleus of the thalamus (ANT). CP, caudate-putamen; GP, globus pallidus; central th., central thalamus; HP, hippocampus.

Figure 7.

[14C]2-Deoxy-d-glucose autoradiograms of coronal sections in rat brains from (A) control; (B) bilateral DBS group; and (C) bilateral lesioning group. Comparison of coronal sections is at the level of the posterior thalamus. Arrow, ventral posterolateral (VPL) and ventral posteromedial (VPM) thalamic nucleus; PF, parafascicular nucleus.

Figure 8.

[14C]2-Deoxy-D-glucose autoradiograms of coronal sections in rat brains from (A) control; (B) bilateral DBS group; and (C) bilateral lesioning group. Comparison of coronal sections is at the level of the mammillary body. Arrow, mammillothalamic tract (MTT); arrowhead, mammillary body (MB).

Figure 9.

[14C]2-Deoxy-D-glucose autoradiograms of coronal sections in rat brains from (A) control; (B) bilateral DBS group; and (C) bilateral lesioning group. Comparison of coronal sections is at the level of the midbrain tegmentum.

Table 1. Local cerebral glucose utilization in each group
StructuresControl group (n = 7)DBS group (n = 7)Lesioning group (n = 7)
RightLeftRightLeftRightLeft
  1. Values expressed as mean ± SEM (μmol/100 g/min).

  2. ap < 0.05.

  3. bp < 0.01 (Welch's or Student's t-test, Comparison between KA control group, DBS group, and lesioning group).

  4. ANT, anterior nucleus of the thalamus; VP, ventroposterior nucleus; PF, parafascicular nucleus;

  5. VTN, ventral tegmental nucleus; DTN, dorsal tegmental nucleus.

Cerebral cortex
 Sensorimotor164.1 ± 25.9296.0 ± 18.9 95.3 ± 13.8b138.0 ± 17.3b 94.1 ± 23.1b126.0 ± 7.9b
 Cingulate216.1 ± 25.2333.1 ± 29.8107.6 ± 19.5b128.3 ± 22.9b100.7 ± 24.9b121.9 ± 26.0b
Extrapyramidal
 Caudate-putamen  123 ± 7.85  341 ± 12.4 95.1 ± .2a  164 ± 91.3b 89.0 ± 37.5b165.0 ± 38.2b
 Globus pallidus 95 ± 4.1  155 ± 9.2 90.4 ± 6.7108.0 ± 11.6b 94.9 ± 17.3107.0 ± 18.4b
 Substantia nigra 72.2 ± 12.3 75.5 ± 9.6 68.8 ± 9.8 70.5 ± 15.6 65.3 ± 0.2 68.6 ± 7.6
Limbic system
 Dorsal hippocampus129.1 ± 15.1138.6 ± 23.1115.1 ± 20.5116.4 ± 22.3113.7 ± 11.0117.3 ± 9.3
 Mammillary body160.7 ± 18.195.3 ± 5.2b88.6 ± 18.3b
 Medial septal n. 59.8 ± 9.9 58.3 ± 9.2 55.9 ± 10.2 58.6 ± 5.6 58.5 ± 3.8 59.8 ± 3.2
 Lateral septal n. 55.8 ± 8.6 56.9 ± 5.4 54.0 ± 5.5 55.5 ± 6.3 53.5 ± 5.6 54.5 ± 5.8
 Accumbens n. 56.8 ± 3.5 59.9 ± 3.3 55.9 ± 6.3 60.6 ± 6.5 52.2 ± 4.5 53.3 ± 4.2
 Entorhinal 70.5 ± 3.3 73.2 ± 3.8 68.9 ± 8.8 70.3 ± 7.7 67.2 ± 2.9 68.4 ± 3.8
 Piriform 79.2 ± 10.2 81.0 ± 8.8 75.6 ± 5.5 78.5 ± 5.5 74.0 ± 5.0 75.0 ± 5.2
 Thalamus 
 ANT150.6 ± 19.3154.4 ± 8.6109.4 ± 12.1b110.4 ± 15.0b 89.6 ± 20.2b 90.3 ± 19.0b
 Central thalamus168.0 ± 20.1113.3 ± 14.0b93.0 ± 19.4b
 VP122.7 ± 4.6205.7 ± 38.0 97.4 ± 5.7b 99.9 ± 10.8b 94.6 ± 26.6b 96.9 ± 15.9b
 PF165.0 ± 25.0123.9 ± 15.0103.1 ± 6.7b102.0 ± 8.4 97.2 ± 26.6b 91.3 ± 20.6b
Brainstem
 VTN129.9 ± 15.4 94.9 ± 12.6b87.7 ± 22.4b
 DTN104.3 ± 18.988.7 ± 11.781.9 ± 24.5
 Superior colliculus 92.8 ± 10.289.6.3 ± 10.6 93.5 ± 10.6 89.5 ± 6.4 88.4 ± 5.4 84.4 ± 5.5
 Inferior colliculus 98.9 ± 12.2 91.3 ± 8.8 95.3 ± 9.6 92.3 ± 8.6 88.4 ± 6.8 88.9 ± 5.6
 Cerebellum 45.6 ± 6.8 48.9 ± 5.6 49.2 ± 4.4 42 ± 3.3 46.1 ± 5.6 47.2 ± 4.5

In the control group, marked increases in LCGU were measured in the cerebral cortex, extrapyramidal system, thalamus, hypothalamus, and brainstem. In the cerebral cortex, the increased LCGU was particularly obvious in the left SMC, cingulate, and contralateral cortices (Fig. 5A). In the extrapyramidal system, LCGU was increased in the lateral part of the caudate-putamen (CP) and globus pallidus (GP) (Figs. 5A and 6A). In the thalamus, the LCGU was increased not only in the bilateral medial and caudal part of the ANT but also in the central thalamic regions such as rhomboid and reuniens nucleus (Fig. 6A). An increased LCGU was also found in the entire posterior thalamus, especially the ventral posterolateral nucleus (VPL), ventral posteromedial nucleus (VPM), and parafascicular nucleus (PF) (Fig. 7A). Outside of these regions, increases in LCGU were also found in the mammillothalamic tract (MTT), mammillary body (MB), midbrain tegmentum (Figs. 8A, 9A). A moderate increase in LCGU was seen in the dorsal hippocampus (Fig. 6A), but no increase in the other limbic structures, such as amygdala, entorhinal cortex, pyriform cortex, nucleus accumbens, and septal nucleus.

In contrast, the animals subjected to bilateral ANT electrical stimulation showed slightly increased LCGU in the ANT and central thalamus, but overall, this metabolism was significantly diminished in comparison with that of the control group (Fig. 6B). LCGU was not increased at all in the posterior thalamus, as it was in the control group (Fig. 7B). Similarly, no increase in LCGU was found in the MTT, MB, or midbrain tegmentum (Figs. 8B and 9B). Moderate increases in LCGU were demonstrated in the SMC, CP, GP, and dorsal hippocampus, although it was still significantly reduced compared with the control group except for dorsal hippocampus (Figs. 5B and 6B).

In the bilateral ANT lesions group, cerebral metabolic changes in each anatomic region observed were similar to those in the bilateral ANT electrical stimulation group (Figs. 5C–9C).

DISCUSSION

This study demonstrated several key findings regarding electrical stimulation and lesioning of the ANT for the control of epileptic seizures. First, both electrical stimulation and lesioning of the ANT suppressed clinical focal cortical seizures induced by an intracortical KA injection. In particular, bilateral ANT procedures had a powerful anticonvulsant effect, because KA injection did not induce motor seizures in these animals. Second, the antiepileptic effects of stimulation were site specific for the ANT, strengthening the importance of site-selective ANT rather than a general “subcortical DBS” stimulus. No sham stimulation was performed in the present study, because individual differences in seizure frequency were marked. Instead, animals were used as their own controls by comparing time on and off stimulation. Although it was unclear whether the electrical-stimulation effect persisted during the stimulation-off period, onset of electrical stimulation produced a definite and significant difference in mean seizure frequency between stimulation-on and -off situations.

Our results are consistent with previous studies on lesioning of the ANT and its afferents (Mirski and Ferrendelli, 1984, 1986a, 1987; Hamani et al., 2004). In both animal investigation and limited human trials, anticonvulsant benefits have been demonstrated either through application of γ-aminobutyric acid (GABA) agonists to inhibit neuronal firing or of high-frequency electrical stimulation to the ANT (Mirski et al., 1986b, 1997; Hodaie et al., 2002; Hamani et al.,2004; Kerrigan et al., 2004). However, few studies have involved focal seizure models, compared with systemic administration of the convulsants. Hamani et al. (2004) showed that bilateral ANT stimulation, and particularly ANT lesioning, had a protective role against seizures induced by an intraperitoneal pilocarpine injection, although unilateral ANT stimulation and lesioning were not effective. They also noted that because various pathways involved in generalization of seizures and part of the ANT efferents have bilateral projections, it is not surprising that only bilateral ANT procedures were effective. However, the data presented here suggest that ipsilateral ANT stimulation and lesioning are also effective in the focal KA cortical seizure models.

In our bilateral ANT electrode-insertion group, the KA injections did not induce motor seizures, even in the absence of stimulation. These beneficial effects are thought to be due to insertion of the DBS electrodes. In a study of prolonged ANT-DBS in patients with medically intractable epilepsy, the authors (Hodaie et al., 2002) suggested that an initial lesioning or so-called “microthalamotomy” effect may be present, defined as a reduction in or abolition of symptoms with insertion of the electrode. Another possible explanation for this finding may be related to the volume of tissue involved with the insertion of the DBS electrodes in our study, which had a relatively large diameter in rat brain compared with those used in clinical practice. It seems very possible that the lesion created by the DBS electrode tract contributes significantly to the anticonvulsant effect.

The finding that EEG foci remained active without clinically relevant motor manifestations suggests that electrical stimulation and lesioning in the ANT may impede the propagation of afterdischarges from the focus to other regions of the brain. Kusske et al. (1972) showed that appropriate lesioning of the ANT with alumina foci in monkey motor cortex disassociated the cortical ictal EEG phenomenon from the motor manifestations. Hashizume et al. (1998) examined the effect of multiple subpial transection (MST) on EEG and cerebral glucose metabolism by using KA-induced focal cortical seizures. Although MST suppressed horizontal propagation to the neighboring cortical area, clinical seizures were not completely inhibited because vertical interactions between the focus and the subcortical area were preserved. Although ANT electrical stimulation and lesioning might block the vertical interaction between the cortical focus and subcortical structures, the precise anticonvulsant mechanisms remain to be defined fully.

In experiment 2, we assessed cerebral glucose metabolism after electrical stimulation and lesioning of the ANT by [14C]2-DG autoradiography, to investigate the mechanism underlying these anticonvulsant actions. To our knowledge, no previous study has examined changes in cerebral glucose metabolism during electrical stimulation of the ANT.

In our KA control group, which showed focal cortical seizure status epilepticus and occasionally SGSs, selective enhancements in LCGU were observed in the sensorimotor cortex (SMC), caudate-putamen (CP), globus pallidus (GP), thalamus (ANT, central thalamus, and posterior thalamus), mammillothalamic tract (MTT), mammillary body (MB), midbrain tegmentum, and dorsal hippocampus. These results suggest that seizure activities propagated to those structures.

In this focal cortical seizure model, no increase in LCGU was seen in limbic structures other than the dorsal hippocampus, such as amygdala, entorhinal cortex, pyriform cortex, nucleus accumbens, and septal nucleus; thus the propagation mode of seizure activity was thought to be different from the limbic seizure models. The hypermetabolic area observed here was comparable with that reported by Yamamoto et al. (1995), who studied the metabolic anatomy of KA-induced focal cortical seizures by using [14C]2-DG autoradiography. In another study of the propagation of penicillin-induced focal cortical seizures, increased LCGU was observed in the ipsilateral motor cortex, basal ganglia, thalamus, substantia nigra, and contralateral cerebellar cortex (Collins et al., 1976). A significant increase of LCGU in substantia nigra and contralateral cerebellar cortex was not always observed in our KA control group, although this difference from the previous study might be attributable to differences in the timing of the LCGU measurement. When cerebral glucose metabolism was examined in a state of minimal seizure activity induced by the systemic combined administration of the convulsant PTZ and the anticonvulsant ethosuximide (ESM) by using [14C]2-DG autoradiography (Mirski and Ferrendelli, 1986a), the selective metabolic activated areas were interestingly consistent with those in our focal cortical seizure model.

Several articles demonstrate that the synchronization of seizure activity within a brain region can be driven by local circuitry once the seizure has started, indicating that “pacemakers” can occur locally (Wong and Prince, 1990; White and Price, 1993). The observed enhancement of metabolic activity supports the hypothesis that neuroanatomic circuits described later may be involved in the expression of KA-induced focal cortical seizures. That is, when the motor cortex was initially activated by the KA intracortical injection followed by seizure-activity propagation from the motor cortex to the CP, GP, and thalamus (Nauta and Mehler, 1966; Kemp and Powell, 1970; Gloor et al., 1977; Royce, 1978, 1983), the corticothalamic circuit was driven via the thalamocortical pathway (Scheibel and Scheibel, 1966; Skinner and Lindsley, 1967; Jones and Leavitt, 1974; Robertson and Kaitz, 1981; Sakai and Tanaka, 1984) (Fig. 10). If this corticothalamic circuit is unilaterally and repeatedly driven, clinically epilepsia partialis continua may occur. Finally, when the MB and midbrain tegmentum were activated via the MTT and the mammillary circuit (Nauta, 1958; Cowan et al., 1964; Cruce, 1977; Mirski and Ferrendelli, 1984, 1986a), seizure activity from the brainstem was also initiated to become an occasional SGS. It was also possible to involve the Papez circuits (Papez, 1937): hippocampal formation (HPF) to MB to ANT to cingulate cortex to HPF, because increased LCGU was observed in all structures of these circuits.

Figure 10.

Proposal for the functional anatomy showed that the neuroanatomic circuits are involved in the expression of KA-induced focal cortical seizures. *1, thalamocortical pathway; *2, mammillothalamic tract. Solid arrow, corticothalamic circuit; double-line arrow, mammillary circuit; broken-line arrow, Papez circuit; ANT, anterior nucleus of the thalamus; CP, caudate-putamen; GP, globus pallidus; HP, hippocampus.

The ANT is unique in its neuroanatomic associations with other subcortical regions and the cortex. The major afferents to the ANT are from cingulate cortex (Domesick, 1969, 1972), the HP via the fornix (Guillery, 1956; Nauta, 1956), and from the MB via the MTT (Fry and Cowan, 1972; Watanabe and Kawana, 1980). Principal efferent projections of the ANT are to the presubiculum of the HPF (Domesick, 1970) and to the cingulate gyrus (Domesick, 1969, 1972). Through its anatomic connections, this diencephalic nucleus is closely associated with the HP, MB, brainstem, and cingulate cortex. The ANT is thought to function as a relay structure to amplify and synchronize seizure activities in these circuits (Fig. 10). The increased LCGU in the central thalamus and the entire posterior thalamus suggests also that the ANT cooperates with the central thalamus and posterior thalamus to amplify and synchronize seizure afterdischarges. The ANT may therefore be a key structure not only in these circuits, but also in the intrathalamic pathways (Jones and Leavitt, 1974).

In our bilateral ANT electrical stimulation and lesioning group, although LCGU was slightly increased in the ANT and central thalamus, and moderately increased in the SMC, CP, GP, and HP, no increases of LCGU was found in posterior thalamus, MTT, MB, and midbrain tegmentum. From the results presented here, we propose that the electrical stimulation and lesioning of the ANT may therefore act to inhibit the described functions of this anatomic structure.

The therapeutic effects of ANT-DBS were similar to those produced by making lesions in our experiments 1 and 2; hence it was postulated that the predominant effect of stimulation was inhibition of neural activity. Low-frequency stimulation of ANT reduced the threshold for early EEG bursts, leading to generation of recruiting rhythms, synchronized EEG activity, and was proconvulsant (Steriade, 1997), whereas the site-specific high-frequency (100 Hz) ANT electrical stimulation in the rat was effective in desynchronized cortical EEG and increasing the clonic seizure threshold (Mirski et al., 1994, 1997).

Two general hypotheses are used to explain the mechanisms of DBS-induced inhibition: (a) stimulation-induced alterations in the activation of voltage-gated currents of neurons near the stimulating electrode (depolarization blockade) (Beurrier et al., 2001; Bikson et al., 2001); or (b) indirect regulation of neuronal output via the activation of axon terminals that make synaptic connections with neurons near the stimulating electrode (synaptic inhibition) or both (Dostrovsky et al., 2000; Wu et al., 2001). Ziai et al. (2005) demonstrated that ANT stimulation may act to alter neurotransmitter release, particularly in the serotonergic system, and subsequent local neuronal inhibition and excitation.

In this study, we selected the medial and caudal pole of the ANT, adjacent to the mediodorsal nucleus, as the stereotactic target portion for electrical stimulation and lesioning. This portion of the ANT was shown to be involved in the critical transition from interictal to ictal states on the preliminary glucose-metabolism study in the focal cortical seizure models. Multiple forms of epilepsy have different initiation sites, and therefore understanding the functional anatomy of seizure initiation and synchronization is important, as is choosing the proper target for each type of epilepsy to be treated.

In summary, this study demonstrated that both electrical stimulation and lesioning of the ANT had anticonvulsant effects in clinical focal cortical motor seizures induced by KA injection. A novel and important aspect of this work was the anticonvulsant effect observed in focal cortical seizure models, which are considered good models for studying human focal cortical epilepsy, implying that a well-defined focus of onset exists.

Because the ANT is an important site for memory and mental processing, the possibility of reversible and minimally invasive therapy with DBS is more attractive for application in humans. Although we used an open-loop stimulation approach, closed-loop stimulations triggered by automated seizure detection for seizure blockage may be more beneficial and informative in future studies (Osorio et al., 2002, 2005).

In addition, glucose-metabolism studies suggested that the ANT functions as a relay structure to amplify and synchronize seizure activities in focal cortical seizures and that the ANT electrical stimulation and lesioning may therefore act to inhibit the described functions of this anatomic structure. Recently, Lado (2006) showed that ANT stimulation exacerbated seizure frequency in rats with chronic epilepsy after status epilepticus induced by systemic KA administration. Our acute focal epilepsy model of status epilepticus was used also in rats, although this model is limited with regard to translation and may not be as applicable to humans, in whom seizures are frequently are chronic and spontaneous. The role of ANT electrical stimulation and lesioning in the treatment of epilepsy must be further evaluated experimentally, particularly with respect to other models of focal epileptic seizure, anticonvulsant effects at chronic stages, behavioral effects, and the most suitable target portion of the ANT for clinical treatment.

Ancillary