The Convulsive and Electroencephalographic Changes Produced by Nonpeptidic δ-Opioid Agonists in Rats: Comparison with Pentylenetetrazol

Nonpeptidic δ-opioid receptor agonists have been demonstrated to produce convulsions in a number of species, including mice, rats, squirrel, and rhesus monkeys (Comer et al., 1993; Dykstra et al., 1993; Negus et al., 1994; Pakarinen et al., 1995; Hong et al., 1998; Broom et al., 2002b). In rats, δ-opioid agonist-induced convulsions are brief, nonlethal clonic contractions involving mainly the musculature of the head, neck, and forelimbs, lasting 10 to 30 s (Broom et al., 2002b; Jutkiewicz et al., 2004). In mice, the convulsion contains clonic and tonic features that may extend the length of the body (Comer et al., 1993; Hong et al., 1998). These convulsions were prevented with the selective δ-opioid receptor antagonist naltrindole, demonstrating that the convulsive effects were δ-opioid receptor-mediated (Comer et al., 1993; Broom et al., 2002b). In addition, tolerance to the convulsive effects of nonpeptidic δ-opioid agonists was shown to develop after a single administration (Comer et al., 1993; Broom et al., 2002a; Jutkiewicz et al., 2005).

Recently, nonpeptidic δ-opioid agonists demonstrated antidepressant-like effects in the forced swim test in mice and rats (Broom et al., 2002b; Jutkiewicz et al., 2004; Saitoh et al., 2004), and it was demonstrated that the antidepressant activity was also mediated through the δ-opioid receptor (Broom et al., 2002b; Jutkiewicz et al., 2004). It had been proposed that the antidepressant effects of nonpeptidic δ-opioid agonists were produced by the seizurogenic properties of these compounds, similar to nonselective convulsant compounds and treatments (Broom et al., 2002a). Historically, chemical convulsants, such as camphor oil and PTZ, were used for treatment of depression (Fink, 1999), and electroconvulsive therapy is currently used for severe and pharmacotherapy-resistant patient populations with depression (Kraus and Chandarana, 1997; McCall, 2001). In preclinical models, acute PTZ and electroconvulsive shock treatment also had been demonstrated to have antidepressant activity (Porsolt et al., 1977, 1978; Cannizzaro et al., 1993; Suzuki and Masuda, 1999; Zarrindast et al., 2004). This evidence suggests that convulsions alone can have antidepressant-like actions in preclinical models.

However, other data suggest that δ-opioid agonists produce antidepressant-like effects in a manner independent of their convulsant properties. Pretreatment with the anticonvulsant benzodiazepine midazolam and tolerance studies were shown to attenuate the convulsant but not the antidepressant-like effects of the nonpeptidic δ-opioid agonists (+)BW373U86 and SNC80 (Broom et al., 2002a; Jutkiewicz et al., 2005). Although these findings suggested that the δ-opioid agonist-induced antidepressant properties were not dependent on the observed behavioral convulsion, nonconvulsive seizure activity might contribute to the antidepressant effects of δ-opioid agonists. The present study addresses this hypothesis by evaluating potential nonconvulsive seizures with electroencephalographic (EEG) recordings in rats after administration of nonpeptidic δ-opioid agonists at doses that were shown to produce antidepressant-like effects (Jutkiewicz et al., 2004).

Previous studies have evaluated the EEG changes that occurred after administration of peptidic δ-opioid agonists. The selective peptidic δ-opioid agonists [d-Pen², d-Pen⁵]-enkephalin and [d-Pen², l-Pen⁵]-enkephalin administered i.c.v. did not produce EEG seizures, but they did produce complex EEG activity, including initial high-voltage slow wave activity associated with behavioral suppression followed by theta driving associated with behavioral arousal (Tortella et al., 1984). The δ-opioid receptor hexapeptide d-Tyr-Ser-Gly-Phe-Leu-Thr produced myoclonic jerks and epileptic discharges measured in the hippocampus in anesthetized and free-moving rats (Haffmans and Dzoljic, 1983). Likewise, the δ-opioid peptide Tyr-d-Thr-Gly-Phe-Leu-Thr elicited bursts of paroxysmal activity with high voltage and high-frequency activity, but it was associated with immobility and unresponsiveness (Stutzmann et al., 1986). Likewise, intrahippocampal injection of the δ-opioid peptide Ala-deltorphin (d-Ala²-deltorphin) produced naltrindole-reversible, high-frequency spikes or spike waves (epileptic discharges) that lasted 27 to 70 s accompanied by wet dog shakes (De Sarro et al., 1992). Other studies reported that the δ-opioid peptides d-Ala²-d-Leu⁵-enkephalin and Tyr-d-Ser-Gly-Phe-Leu-Thr produced seizures after i.c.v. (Snead, 1986) or medial and lateral thalamic injections (Walker and Yaksh, 1986). Only one study evaluated the EEG changes produced by nonpeptidic δ-opioid agonists and demonstrated that the racemate BW373U86 induced a naltrindole-reversible increase in the slow component of hippocampal theta activity that was associated with locomotor stimulation (Marrosu et al., 1997); however, this study evaluated only low doses of BW373U86 (0.5–2.5 mg/kg) and might not have captured convulsive EEG changes produced by nonpeptidic δ-opioid agonists.

The present study evaluated the EEG responses produced by the acute and repeated administration of the nonpeptidic δ-opioid agonists SNC80 and (+)BW373U86 in Sprague-Dawley rats compared with those produced by the chemical convulsant PTZ. In addition, the effects of antiepileptic drugs, ethosuximide, valproate, gabapentin, phenytoin, and diazepam, on the convulsant properties of the nonpeptidic δ-opioid agonist (+)BW373U86 and PTZ were evaluated.

Materials and Methods

Subjects

Male Sprague-Dawley rats (250–300 g) were obtained from Harlan (Indianapolis, IN) and housed in groups of three rats per cage unless otherwise noted. All animals were fed a standard laboratory diet and maintained on a 12-h light/dark cycle with lights on at 6:30 AM at an average temperature of 21°C. Rats used in the studies were performed in accordance with the Declaration of Helsinki and with the Guide for the Care and Use of Laboratory Animals as adopted and promulgated by the National Institutes of Health. The experimental protocols were approved by the University of Michigan University Committee on the Use and Care of Animals.

Electrode Implantation Surgery

To measure electroencephalograms in rats using a telemetry system, rats were implanted with three-channel radiotransmitters (model F50-EET; Data Sciences International, St. Paul, MN) under ketamine (100 mg/kg i.p.) and xylazine (10 mg/kg i.p.) anesthesia. The transmitter was 4.5 cm in length, 1.5 cm in width, 1 cm in depth, and weighed approximately 13 g. Before implantation, the transmitter was cleaned in ethanol and soaked in sterile saline. An incision in the skin and musculature of the peritoneal cavity was made, and the transmitter was placed inside the peritoneal cavity. The transmitter was attached to the muscle wall of the peritoneum using nonabsorbable nylon suture to prevent the transmitter from shifting after implantation, and the skin over the muscle was closed. The biopotential leads (five) were passed through the peritoneal muscle wall using a 16-gauge needle and threaded subcutaneously, emerging at an incision made in the skin at the base of the head. The head of the rat was placed in a stereotaxic instrument for stability of the head during screw and biopotential lead attachment. After exposing the skull, the bone was cleaned of connective tissue and dried. Five holes were drilled using a micro-drill with 0.7-mm steel burr (Fine Science Tools, Inc., Foster City, CA) for bilateral placement of epidural recording electrodes, which consisted of biopotential leads from the transmitter wrapped around stainless steel slotted, fillister screws (0.8 mm in diameter, 0.12 inches in length; Small Parts, Inc., Miami Lakes, FL). Screws were implanted over the left and right parietal cortex (each side: approximately 1 mm posterior to bregma, 1.5 mm lateral to the midline and 1 mm anterior to lambda, and 1.5 mm lateral to the midline). A reference wire was also placed 1 to 2 mm posterior to lambda. The biopotential leads were prepared by removing approximately 0.5 cm of silicone rubber tubing from the end of each wire, and the wires were stretched to allow easier and more secure wrap around the skull screws. All skull screws and wires were anchored to the skull with dental acrylic cement. Following biopotential lead attachment, the skin incisions on the head were closed with nylon suture. After surgery, rats were singly housed and monitored for at least 7 days for signs of recovery (normal eating, drinking, and defecation) before experimentation.

Data Acquisition and Analysis

Signals detected by the biopotential leads were transmitted to a receiver (RPC-1; Data Sciences International) beneath the rat's homecage. The receiver sent the signal via a cable connector to the Dataquest ART Exchange Matrix (Data Sciences International) that converted the analog signal into digital output that was recorded onto a computer. The signal was filtered for 60-Hz signal, the high bandpass filter was set to 70 Hz, and the low bandpass filter was set to 0.5 Hz. Data analysis was conducted with Somnologica Software and DSI import modules (Medcare Flaga, Reykjavik, Iceland), and waveforms were evaluated at a 30 mm/s recording speed.

Experimental Procedures

Behavioral Observations. For EEG studies, rats were tested on an individual basis in their homecages, and behavioral observations for different rats never overlapped. Individual rats and their digital EEG traces were continuously monitored during the 1- to 2-h test sessions to identify behaviors, movement artifacts, convulsions, nonconvulsive seizures, or other abnormal behaviors that coincided with EEG changes. This telemetry system was previously used as a reliable and reproducible system for evaluating seizures and epileptiform activity (Bastlund et al., 2004). As described previously, seizures were defined as generalized or focal epileptiform activity that occurred continuously for at least 10 s with an amplitude larger than background (Jordan, 1999; Hartings et al., 2003). All EEG experiments were conducted between 8:00 AM and 12:00 PM. During observations, food and water were removed to allow for easier observation. Occasionally, during pretreatment times or during tolerance studies, gentle tapping on the cage or removal of the cage lid was required to maintain the rat in an alert, awake state.

For convulsion observation studies, antiepileptic drugs (AEDs) (ethosuximide, valproate, gabapentin, phenytoin, or diazepam) were administered as 30-min pretreatments to either (+)BW373U86 or the chemical convulsant PTZ. For these studies, only the δ-opioid agonist (+)BW373U86 was used to conserve compound and because less compound was required to produce convulsions in 100% of subjects. After (+)BW373U86 or PTZ administration, rats were observed individually for 30 min each in a clear, plastic observation cage with bedding. Rats were monitored continuously for convulsions, catalepsy, myclonic jerks, lethargy, bouts of stillness (stop-and-stare behaviors), wet dog shakes, and other normal or abnormal behaviors. In addition, a modified Racine (1972) seizure classification scale was used to identify potential seizure-related behaviors and to characterize the intensity of proconvulsant-related behaviors, more specifically, mouth and facial movements (S1); head bobbing or nodding (S2); clonic convulsions of the forelimbs, head, and/or neck with balance maintained (S3); clonic convulsions of the forelimbs, head, and/or neck with rearing while maintaining balance (S4); and clonic convulsions of the forelimbs, head, and/or neck with rearing and falling over (loss of balance) (S5). A score below S3 indicated that no convulsion was observed (indicated by a dotted line at an intensity score of 3 in Fig. 10). After convulsions, catalepsy was measured by a loss of righting reflex and a rod test that required the rat to remove its forepaws from a rod lifted 2 in above the ground within 15 s. All behaviors were averaged across treatment groups.

Drug Treatments. For EEG studies, rats were administered the same dose of s.c. SNC80 (either 3.2, 10, or 32 mg/kg), (+)BW373U86 (10 mg/kg), or PTZ (32 or 56 mg/kg) once per day for 2 to 4 days or until the EEG was indistinguishable from baseline or normal activity. In antagonism experiments, 1.0 mg/kg of the selective δ-opioid receptor antagonist naltrindole was administered s.c. 30 min before the δ-opioid agonist SNC80 (32 mg/kg s.c.). The benzodiazepine midazolam (0.5 mg/kg) was administered s.c. 15 min before 32 mg/kg SNC80 (s.c.). Six rats per drug condition or treatment were used, and each rat was used only once.

For studies with AEDs, either vehicle or one dose of ethosuximide (10, 32, 100, or 320 mg/kg), valproate (32, 100, or 320 mg/kg), gabapentin (32, 100, or 320 mg/kg), phenytoin (10, 32, or 100 mg/kg), or diazepam (0.32, 1, 3.2, or 10 mg/kg) was administered i.p. as a 30-min pretreatment to 32 mg/kg (+)BW373U86 (s.c.) or 56 mg/kg PTZ (s.c.). Six rats per drug condition or treatment were used, and each rat was used only once.

Statistical Analysis

The characteristics of paroxysmal activity (Table 1) were compared with one-way ANOVA and Tukey's post hoc tests for number and duration of paroxysmal bursts and with Student's t tests for frequency and power of bursts with 10 and 32 mg/kg SNC80 (Prism; GraphPad Software Inc., San Diego, CA). The effects of AEDs on (+)BW373U86- and PTZ-induced convulsions (Fig. 10) were compared with two-way ANOVA and Bonferroni's post hoc tests for interaction effects, and the main effects were further analyzed by one-way ANOVA (Prism).

Drugs

SNC80 was dissolved in 8% HCl solution (1 M), and (+)BW373U86 was dissolved in sterile water. Midazolam was diluted in saline, and naltrindole (National Institute on Drug Abuse, Bethesda, MD) was dissolved in sterile water. PTZ and ethosuximide (Sigma-Aldrich, St. Louis, MO) were dissolved in sterile water. Sodium valproate (Sigma-Aldrich) and gabapentin (Tocris Cookson Inc., Ellisville, MO) were dissolved in saline. Phenytoin (Sigma-Aldrich) was dissolved in sterile water with a few drops of 1 M NaOH. Diazepam (Henry-Schein, Melville, NY) was diluted in sterile water. All drugs were administered in volumes of 1 to 2 ml/kg. Ketamine hydrochloride (Vedco Inc., St. Joseph, MO) and xylazine hydrochloride (Fermenta Animal Health Co., Kansas City, MO) were all dissolved in sterile water.

Results

Acute Effects of δ-Opioid Agonists. The present study evaluated the effects of acute and repeated administration of the δ-opioid agonists SNC80 and (+)BW373U86 on EEG measurements in Sprague-Dawley rats. Figure 1a shows 30-min bilateral recordings of a representative SNC80-induced convulsion. The ictal activity generalized bilaterally as measured above parietal cortex regions. After the epileptiform activity and postconvulsive catalepsy, the traces returned to baseline levels as observed before drug administration or immediately after injection. The asterisk on each trace indicates the injection and the correlated EEG artifact that occurred when the rat was removed from the receiver located below the cage. All EEG activity was bilateral; therefore, the later figures show traces from one side of a bilateral recording.

Four of the six rats tested with 32 mg/kg SNC80 demonstrated observable convulsions (Fig. 1, b–e). In rats that convulsed, the convulsion occurred as described previously with repeated clonic contractions of musculature of the head, neck, and forelimb areas, lasting between 10 and 30 s. After the convulsion, there was a period of catalepsy during which the rats failed to demonstrate a righting reflex and did not remove their paws from a rod lifted 1 to 2 inches off of the ground. After catalepsy, rats were very active, exhibiting high levels of walking, rearing, sniffing, exploring, and digging in the cage bedding.

The observed convulsion was correlated with a discharge sequence that increased in amplitude and changed in frequency compared with baseline, consisting of rhythmic spikes and sharp waves initially that progressed to spike-and-wave complexes (Fig. 2). This magnitude and duration of ictal activity was observed only during an overt behavioral convulsion. Spectral analysis determined that these discharge sequences occurred in the delta frequency range (0.5–3.5 Hz).

Before the behavioral convulsion, rhythmic spikes, sharp waves, and occasional spike-and-wave complexes occurred in 1- to 4-s distinct bursts in the 4- to 7-Hz range (Fig. 3a). These bursts of activity were easily identifiable by the abrupt onset and offset of increased amplitude and rhythmic waveforms. In two rats that were tested with the convulsive dose (32 mg/kg SNC80), some bursts of activity occurred before and after the behavioral convulsion and cataleptic period (Fig. 1, d and e), and in the two rats that did not convulse, similar bursts of activity were observed (Fig. 1, f and g). Rats that were tested with 32 mg/kg SNC80 demonstrated a range of 10 to 45 short bursts within 40 min of SNC80 administration (also see Table 1); however, by 1 h postinjection, EEG activity returned to baseline levels. In general, these burst events were not consistently associated with any specific behavioral output, but the rats were generally active in terms of locomotor stimulation.

After the ictal discharge, postictal EEG suppression and slow waves were observed correlating with behavioral catalepsy (Fig. 3, b and c). Catalepsy and the corresponding postictal EEG activity were observed only in rats that experienced a behavioral convulsion. After catalepsy, the trace returned to baseline activity. The selective δ-opioid antagonist naltrindole attenuated the behavioral convulsions produced by 32 mg/kg SNC80 (n = 6) (Fig. 4). A representative trace (Fig. 4b) shows that 1.0 mg/kg naltrindole administered as a 30-min pretreatment to 32 mg/kg SNC80 blocked the ictal activity as well as the bursts of spikes and sharp waves produced by SNC80 demonstrated in Fig. 4a. Naltrindole pretreatment did not alter the baseline EEG trace (data not shown).

The benzodiazepine midazolam (0.5 mg/kg s.c.) also blocked all paroxysmal activity produced by an injection of 32 mg/kg SNC80 (n = 6) (Fig. 5). A representative trace (Fig. 5) demonstrates that midazolam prevented ictal activity as well as epileptiform abnormalities produced by SNC80. This dose of midazolam did not greatly alter EEG activity compared with baseline (data not shown).

Effects of Repeated SNC80 Administration and Comparison with PTZ. The effects of repeated injections of various doses of SNC80 were measured in a group of rats. Figure 6, a and b, shows traces from a rat that received an initial injection of 32 mg/kg SNC80 on day 1 (a) and a second dose 24 h later (b). The observed convulsion and ictal activity from day 1 occurred as described above. However, the second injection of 32 mg/kg SNC80 produced few, if any, changes in EEG waveforms 24 h after the first administration. There was no observed convulsion, and no spikes or sharp waves were identified (Fig. 6b); however, the locomotor activity of the rat remained somewhat elevated. Likewise, there were no EEG changes after administration of the same dose of SNC80 on days 3 and 4 (data not shown).

After an injection of a lower dose 10 mg/kg SNC80, no rats convulsed or demonstrated patterns of ictal activity similar to that observed with 32 mg/kg. Short bursts of epileptiform activity were observed in three of six rats tested on day 1 (representative trace shown in Fig. 6c); however, these events were not correlated with any particular behavior, and all rats were active after the injection. On average, this lower dose of SNC80 produced fewer [F(2,15) = 4.17; p = 0.04] and shorter [F(2,17) = 148.4; p < 0.0001] bursts of epileptiform activity compared with 32 mg/kg SNC80 (Table 1). In addition, these bursts of activity had a higher frequency (p = 0.03) and a trend toward a lower spectral power (p = 0.1) than that observed with 32 mg/kg SNC80. Twenty-four hours later, the second administration of 10 mg/kg SNC80 did not produce epileptiform activity in any rat that was tested (Fig. 6d). Likewise, in rats that did not have spike burst activity on day 1 with 10 mg/kg SNC80 (Fig. 6e), no changes in the EEG waveform were observed on day 2 (Fig. 6f).

With 3.2 mg/kg SNC80, almost no EEG changes were observed after injection on day 1 (Fig. 6g); however, one of six rats had three short bursts of spike activity. Similar to other doses, the second injection of 3.2 mg/kg SNC80 did not produce epileptiform activity on day 2 (Fig. 6h). Most rats were tested with the same dose of SNC80 on days 3 and 4, but no changes in the EEG recordings were observed (data not shown).

Similar to SNC80, the nonpeptidic δ-opioid agonist (+)BW373U86 produced ictal activity and other paroxysmal changes as measured in EEG recordings (Fig. 7). At a dose of 10 mg/kg (+)BW373U86, five of six rats convulsed. A representative trace is shown in Fig. 7a. Before a convulsion or instead of a convulsion, small bursts of rhythmic spikes, sharp waves, and occasional spike-and-wave complexes were recorded. Convulsions were correlated with ictal activity that returned to baseline levels after the convulsion and subsequent cataleptic period. Consistent with the effects of repeated SNC80 injections, a second injection of 10 mg/kg (+)BW373U86 24 h later did not produce epileptiform activity in the EEG recording (Fig. 7b).

In comparison with SNC80-induced epileptiform activity, rats were tested with either 32 or 56 mg/kg PTZ, and EEGs were recorded after the first injection and a second injection 24 h later (Fig. 8). At a dose of 56 mg/kg PTZ, all rats had one behavioral convulsion similar in duration and appearance to that produced by the δ-opioid agonists SNC80 and (+)BW373U86. In addition, these rats demonstrated brief bursts of spike and sharp wave activity before the ictal event; however, unlike the δ-opioid agonists, postconvulsion disturbances in the EEG waveform lasted for 1 to 2 h. Figure 8a shows a representative 60-min trace from a rat that received an injection of 56 mg/kg PTZ. Bursts of epileptiform activity were evident throughout the 60-min recording, although the rats were no longer cataleptic. Unlike the postconvulsion locomotor stimulation observed with SNC80, rats that received 56 mg/kg PTZ were alert, but they demonstrated minimal behavioral stimulation after a convulsion. Twenty-four hours after the first injection (Fig. 8b), the second administration of 56 mg/kg PTZ produced similar observed convulsions and epileptiform activity as observed on day 1. In some rats, repeated administration of 56 mg/kg PTZ produced longer, more severe convulsions compared with the initial observation.

The lower dose of 32 mg/kg PTZ produced behavioral convulsions in three of six rats on day 1. Rats that did not convulse demonstrated myoclonic jerks and bouts of stillness (stop-and-stare behaviors). Figure 8c shows a trace from a rat that received 32 mg/kg PTZ and did not convulse. Bursts of epileptiform activity were evident throughout the 60-min recording, and EEG activity returned to baseline after approximately 1.5 h. A second injection of 32 mg/kg PTZ, 24 h after the initial administration, produced epileptiform activity similar to, or more frequent than, that observed on day 1 (Fig. 8d).

Effects of AEDs on SNC80- and PTZ-Induced Convulsions. In the AED studies, the behavioral manifestations of convulsive activity produced by the δ-opioid agonist (+)BW373U86 and PTZ were compared in terms of number of convulsions produced by a single drug dose, intensity of the proconvulsant and convulsant activity, duration of convulsions, number of myoclonic twitches, bouts of stillness, and wet dog shakes (Fig. 9). PTZ and (+)BW373U86 produced dose-dependent increases in the number of convulsions observed (Fig. 9a), the intensity of seizure-related behaviors (Fig. 9b), duration of convulsions (Fig. 9c), and number of myoclonic twitches (Fig. 9d). PTZ stimulated more bouts of stillness (stop-and-stare behavior) compared with (+)BW373U86 (Fig. 9e), and (+)BW373U86 induced more wet dog shakes than PTZ (Fig. 9f). Doses of 32 mg/kg (+)BW373U86 and 56 mg/kg PTZ produced similar number, intensity, and duration of convulsive activity; therefore, these doses were selected for the studies with the selected anticonvulsant compounds.

The effects of AEDs were evaluated against convulsions produced by PTZ and the nonpeptidic δ-opioid agonist (+)BW373U86 (Fig. 10). Only ethosuximide and diazepam completely attenuated PTZ- and (+)BW373U86-induced convulsions [ethosuximide main effect: F(4,50) = 19.35; p < 0.0001; diazepam main effect: F(4,51) = 16.24; p < 0.0001; Fig. 10, a and m, respectively]. Valproate demonstrated a trend to reduce the average number of convulsions observed per group; however, this effect was not significant [F(4,54) = 2.08; p = 0.096; Fig. 10d]. Gabapentin and phenytoin did not decrease the convulsions produced by either PTZ or (+)BW373U86 but actually increased the number of convulsions observed with PTZ [gabapentin main effect: F(3,39) = 3.16; p = 0.04; phenytoin main effect: F(3,40) = 5.26; p = 0.004; Fig. 10, g and j, respectively].

Ethosuximide, diazepam, and valproate dose-dependently decreased the (+)BW373U86-induced convulsion intensity score to 2 or lower, indicating that an overt convulsion did not occur (Fig. 10, b, e, and n), but only diazepam significantly decreased intensity score produced by (+)BW373U86 [F(4,30) = 6.14; p = 0.001]. All compounds, with the exception of phenytoin, significantly decreased PTZ-induced convulsion intensity scores [ethosuximide: F(4,25) = 12.93; p < 0.0001; valproate: F(4,40) = 7.47; p = 0.0001; gabapentin: F(3,20) = 5.69; p = 0.006; diazepam: F(4,25) = 5.44; p = 0.003; Fig. 10, b, e, and n]. Although gabapentin decreased PTZ-induced convulsion intensity, it did not decrease the frequency of the convulsive event.

Only ethosuximide and diazepam significantly decreased convulsion duration for both (+)BW373U86 and PTZ [ethosuximide: F(4,50) = 7.21; p = 0.0001; diazepam: F(4,51) = 13.11; p < 0.0001; Fig. 10, c and o], but gabapentin decreased the convulsion duration in rats that received (+)BW373U86 only without altering the occurrence of convulsive events [F(3,20) = 4.31; p = 0.02; Fig. 10i].

Discussion

The present study described the ictal activity and other epileptiform abnormalities correlated with SNC80 and (+)BW373U86 administration in Sprague-Dawley rats. Seizure epileptiforms occurring with behavioral convulsions (four of six rats with 32 mg/kg SNC80) included rhythmic spikes and sharp waves that progressed to 3-Hz spike-and-wave discharges persisting for at least 20 s. Short 2- to 3-s bursts of spikes and sharp wave activity occurred in rats that did not convulse; however, these waveforms were not consistently related to an identifiable behavioral pattern or abnormal behaviors. Therefore, nonconvulsive seizures (i.e., prolonged seizure discharges in the absence of overt convulsions) did not occur in rats receiving these δ-opioid agonists.

In terms of dose-dependent EEG changes, the number, duration, and spectral power of the paroxysmal bursts increased, and the frequency decreased with larger SNC80 doses. No convulsions or nonconvulsive seizures were observed with doses lower than 32 mg/kg SNC80. Additionally, the behavioral convulsions and the epileptiform activity produced by SNC80 were attenuated by the selective δ-opioid antagonist naltrindole, demonstrating that these behavioral and seizurogenic effects of SNC80 were mediated by the δ-opioid receptor.

In addition, the short-acting benzodiazepine midazolam prevented the SNC80-induced convulsions and epileptiform abnormalities. Midazolam, similar to other benzodiazepines such as diazepam, can be used to manage or acutely treat status epilepticus, especially in children. These data support and extend previous findings that midazolam attenuated these convulsions through an anticonvulsant action most likely downstream of δ-opioid receptor activation (Comer et al., 1993; Broom et al., 2002a). Midazolam also prevented all EEG changes, specifically paroxysmal bursts, produced by SNC80. In previous studies, midazolam pretreatment blocked δ-opioid agonist-induced convulsions, but it did not prevent the antidepressant-like effects of (+)BW373U86 (Broom et al., 2002a). These data support the hypothesis that the antidepressant-like effects of δ-opioid agonists were not dependent on convulsive episodes or nonconvulsive seizures produced by these compounds.

With repeated administration of δ-opioid agonists, mice and rats became tolerant to the convulsive effects of δ-opioid agonists (Comer et al., 1993; Broom et al., 2002a; Jutkiewicz et al., 2005). Likewise, after a single administration of SNC80 or (+)BW373U86, rats became tolerant to the convulsive effects and epileptiform activity, such that few, if any, EEG changes occurred after a second administration. With the second injection of SNC80 or (+)BW373U86, the rats were active in the cage; however, EEG waveforms resembled baseline activity in most rats tested.

Previous studies demonstrated that midazolam pretreatment and repeated administration altered the convulsive profile but not the antidepressant-like effects of δ-opioid agonists (Broom et al., 2002a; Jutkiewicz et al., 2005). Overall, these data in combination with the present findings suggest that the convulsive and paroxysmal activity produced by δ-opioid agonists in rats are not required for the antidepressant-like effects observed with nonpeptidic δ-opioid agonists. These findings further support the independent nature of the δ-opioid agonist-induced convulsion and antidepressant-like effects.

The seizure episodes produced by δ-opioid agonists were substantially different from those produced by PTZ. For example, EEG waveforms returned to normal baseline activity after δ-opioid agonist-induced convulsions and catalepsy; however, rats receiving PTZ demonstrated epileptiform activity for at least 1 h postinjection. In contrast to δ-opioid agonists, rats did not become tolerant to the PTZ-induced seizures and EEG abnormalities, such that similar or more severe epileptiform activity was observed after a second injection of PTZ. δ-Opioid agonist- and PTZ-induced convulsions have some similar characteristics; however, there are observable differences in postconvulsion recovery, especially in terms of the locomotor stimulation.

PTZ has been used extensively and characterized as a model of absence seizures (for review, see Hosford, 1999), and the similar characteristics between PTZ and δ-opioid agonists might suggest that these compounds have comparable mechanisms of action for producing convulsive behavior. Further investigation with antiepileptic drugs has provided more insight into the potential mechanism of convulsion produced by δ-opioid receptor activation. For example, ethosuximide and valproate, drugs used to clinically treat absence seizures, dose-dependently attenuated the number, intensity, and duration of convulsions produced by the δ-opioid agonist (+)BW373U86 and PTZ. Previous research also demonstrated that ethosuximide and valproate blocked PTZ-induced convulsions at similar doses tested in the present study (Löscher et al., 1991; Mareš et al., 1997; Kumaresan et al., 2000). These data demonstrate that these antiepileptic drugs were able to eliminate important characteristics (i.e., intensity and duration) of the convulsions to produce a complete blockade. Ethosuximide was more effective than valproate, as demonstrated by the complete elimination of convulsions. Additionally, ethosuximide and valproate were more potent at attenuating the occurrence of (+)BW373U86-induced convulsions than PTZ-induced convulsions.

In contrast, gabapentin and phenytoin, antiepileptic drugs used clinically mainly to treat tonic-clonic convulsions, did not decrease the occurrence of (+)BW373U86- or PTZ-induced convulsions. In fact, the highest doses tested increased the number of convulsions that were observed with 56 mg/kg PTZ. Clinically, gabapentin has also been shown to exacerbate absence seizures in some instances (for review, see Manning et al., 2003). Interestingly, gabapentin slightly decreased the duration of the (+)BW373U86 convulsion without altering the severity (intensity), demonstrating that it is possible to minimize some characteristics of the convulsions without eliminating them. Therefore, the most effective treatment is one that eliminates all factors of the convulsion, as observed with ethosuximide and the more general anti-convulsant diazepam.

Absence seizures are thought to be generated by activation of low-threshold T-type Ca²⁺ channels and aberrant corticothalamic rhythms. Ethosuximide is a common and effective treatment for absence seizures that is thought to block these Ca²⁺ channels within the thalamus (for review, see Manning et al., 2003). Previously, studies with intrahippocampal injections of δ-opioid agonist-induced convulsions suggested that the seizure activity involved primarily the limbic system (Haffmans and Dzoljic, 1983; De Sarro et al., 1992); however, the present data might suggest that alternative sites of action and neurocircuits, such as thalamo-cortico circuits, might be involved. Peptidic δ-opioid agonists also were previously suggested to have absence seizure-like activity in rats or a thalamic site of action (Snead and Bearden, 1982; Stutzmann et al., 1986; Walker and Yaksh, 1986). Considering the similarities between PTZ and δ-opioid agonists, these data suggest that (+)BW373U86, like PTZ, might also generate absence-like seizures.

The present research demonstrated that nonpeptidic δ-opioid agonists produced naltrindole-sensitive seizures and epileptiform activity. Doses of SNC80 that did not produce behavioral convulsions failed to produce prolonged (>10-s) changes in EEG recordings, suggesting that this compound did not produce nonconvulsive seizures. In addition, repeated administration of SNC80 and (+)BW373U86 eliminated changes in EEG recordings. These data suggested that δ-opioid agonist-induced convulsions and epileptiform activity might not be required for the antidepressant properties of δ-opioid agonists, further implying that it might be possible to develop a δ-opioid agonist with a large therapeutic index (i.e., wide therapeutic window separating the convulsive and antidepressant-like effects). The present data in conjunction with other findings indicated that the convulsive properties of these compounds were limited, perhaps related to acute receptor activation, and could be eliminated under certain conditions. Further understanding of seizurogenic mechanisms related to δ receptor activation might lead to the design and development of compounds devoid of convulsive activity with therapeutic usefulness in the treatment of a variety of diseases from depression and anxiety to pain management or Parkinson's disease.

In addition, thalamo-cortical circuits may be involved in δ-opioid agonist-induced convulsions because a drug used to treat absence epilepsy, ethosuximide, attenuated convulsions produced by the δ-opioid agonist (+)BW373U86. Further studies should address this hypothesis in detail. Overall, these data suggest that seizures and EEG changes produced by the δ-opioid agonists SNC80 and (+)BW373U86 are different from those produced by PTZ and support previous findings that convulsions and EEG changes are not required for the antidepressant properties of these compounds.