Kindling Stimuli Delivered at Different Times in the Sleep-Wake Cycle

Transcription

1 BASIC RESEARCH Kindling Stimuli Delivered at Different Times in the Sleep-Wake Cycle Pei-Lu Yi, MSc 1 ; Chon-Haw Tsai, MD, PhD 2 ; Jaung-Geng Lin, MD, PhD 3 ; Cheng-Chun Lee, MD, PhD 2 ; Fang-Chia Chang, PhD 2 1Department of Nursing, Jen-Teh Junior College of Medicine, Nursing and Management, Miaoli, Taiwan; 2 Neuroscience Laboratory, Department of Neurology, China Medical University Hospital, Taichung, Taiwan; 3 Acupuncture Research Center, China Medical University, Taichung, Taiwan Study Objectives: Clinical and experimental observations argue that sleep disturbances are common and coexist in patients with epilepsy. Our previous observations have suggested that neural-immune interactions between corticotropin-releasing hormone (CRH) and interleukin-1 (IL-1) are involved in the regulation of physiological sleep-wake behavior. In the present study, we determined the involvement of CRH and IL-1 in the alteration of sleep-wake activities in rats when amygdala kindling was given either at the beginning of the light period (light-onset kindling) or at the beginning of the dark period (dark-onset kindling). Design: The analysis of sleep-wake activities was performed before and after full-blown kindling, and the actions of CRH and IL-1 were tested by pharmacological blockade. Setting: Neuroscience Laboratory at China Medical University Hospital. Participant and Interventions: Male Sprague-Dawley rats were implanted with electroencephalogram (EEG) electrodes and an intracerebroventricular (ICV) guide cannula. Kindling stimuli delivered via a bipolar electrode placing in the right central nucleus of the amygdala. Measurement and Results: Amygdala kindling induced diverse effects on sleep. Slow-wave sleep (SWS) and rapid eye movement (REM) sleep decreased during the first 12-hour light period when rats were kindled at light onset. When dark-onset kindling was given, SWS increased but REM sleep was not altered during the first 12-hour dark period. After light-onset kindling, the circulating corticosterone concentrations increased and were blocked by ICV administration of the CRH receptor antagonist, astressin or α-hcrh. The ICV administration of the CRH antagonist blocked the light-onset kindling-induced decrease of SWS in a dose-dependent manner. After dark-onset kindling, IL-1 mrna expression in the hippocampus and cortex increased. The dark-onset kindling-induced SWS was blocked in a dose-dependent manner by ICV administration of IL-1 receptor antagonist (IL-1 ra). The slow-wave activity during SWS was enhanced regardless of when kindling occurred, but both CRH antagonists and IL-1 ra had little effect on the alteration of slow-wave activity. Conclusion: These observations argue that amygdala-kindling-induced sleep-wake alterations are modulated by central increases in CRH or IL-1. Key Words: Sleep, amygdala, kindling, epilepsy, CRH, and IL-1 Citation: Yi PL; Tsai CH; Lin JG et al. Kindling stimuli delivered at different times in the sleep-wake cycle. SLEEP 2004;27(2): INTRODUCTION CLINICAL AND EXPERIMENTAL OBSERVATIONS SUGGEST THAT THE RELATIONSHIP BETWEEN SLEEP AND EPILEPSY IS COMPLICATED AND RECIPROCAL. Much attention has been focused on the influence of sleep, including slow-wave sleep (SWS) and rapid-eye-movement (REM) sleep, on interictal epileptiform discharges and seizure occurrence. While REM sleep suppresses seizure activity, non-rem (NREM) sleep facilitates it. 1-5 For example, an increase in REM sleep by microinjection of a cholinergic agonist, carbachol, into the pontine reticular formation decreases seizure activity. 4 In addition, NREM sleep provokes seizure-discharge propagation via neural generators of synchronous electroencephalogram (EEG) oscillations, and neural generators of asynchronous neuronal discharge patterns of waking and REM sleep can reduce EEG seizure. 5 Nonetheless, disturbances in the sleep-wake continuum in patients and animals with epilepsy, while common, are often overlooked. Patients with epilepsy experience more daytime sleepiness compared with control patients, 6 and children with epilepsy experience poor quality of sleep, anxiety about sleep, and sleepdisordered breathing. 7 While experimental amygdala kindling can promote wakefulness and light sleep (see review 8 ), one study observed the appearance of increases in deep SWS and decreases in light SWS and wakefulness after the first full-blown seizure. 9 These conflicting results Disclosure Statement This work was supported by National Science Council grant NSC B and China Medical University Hospital grant DMR Submitted for publication March 2003 Accepted for publication October 2003 Address correspondence to: Fang-Chia Chang, PhD, Neuroscience Laboratory, Department of Neurology, China Medical University Hospital, No. 2, Yu-Der Road, Taichung 404, Taiwan; Tel: , ext. 7063; Fax: ; fchang@ SLEEP, Vol. 27, No. 2, may be due to the manipulation of kindling at different circadian time points. The present study therefore was systematically designed to explore the nature of kindling-induced alterations in sleep at 2 circadian time points: one kindled time point was 5 minutes prior to the light period (light-onset kindling), and the other was 5 minutes before the dark period (dark-onset kindling). A second goal of the present study was to determine whether corticotropin-releasing hormone (CRH) and interleukin-1 (IL-1) are involved in the kindling-induced sleep regulations. This hypothesis derives from the following observations. Certain neuropeptides and cytokines (ie, CRH, neuropeptide Y, IL-1, and tumor necrosis factor [TNF]) may be involved in seizure-associated processes and secondary epileptogenesis CRH promotes age-dependent limbic seizures, 15 augments bicuculline-induced epileptiform activity, 10 and exacerbates cocaine-kindled seizures. 11 Cytokines may also be involved in adaptive mechanisms associated with epileptic activity and in neuroprotection. 12,14 Transforming growth factor-β (TGF-β), for example, which is implicated in neuroprotection via the metabotropic glutamate receptor, is upregulated during the first 3 weeks after status epilepticus throughout the hippocampus. 16 On the other hand, exogenous application of IL-1β prolongs kainate-induced hippocampal EEG seizures. 13 Epileptogenic cortical rhythms also increase CRH and IL-1 levels in several brain areas. Kindling increases CRH mrna expression in the paraventricular nucleus of the hypothalamus, 17 hippocampus, 18 and cortexes 19 and also might up regulate IL-1β, TNF-α and TGF-β1 mrna in the cortex, amygdala, and hippocampus. 12,14 We have previously hypothesized that CRH and IL-1 are both involved in the regulation of physiological sleep-wake activity, based in part on observations that the central or systemic administration of CRH receptor antagonists reduces spontaneous waking and that IL-1 mediates an increase in SWS, which is induced by the CRH receptor blockade. 22 The present study further elucidates the contribution of CRH and IL-1 to sleep alterations induced after amygdala kindling, by the administration of CRH receptor antagonists, astressin, or α-helical CRH 9-41 (α-hcrh), or the IL-1 receptor antagonist (IL-1ra).

2 Here we report that CRH mediates the reduction in SWS and the increase in wakefulness induced by light-onset kindling, whereas darkonset kindling increases SWS via enhancement of IL-1. MATERIALS AND METHODS Substances Stock solutions of astressin (Bachem, Torrance, Calif), α-hcrh (Tocris, Bristol, UK) and human recombinant IL-1ra (Bachem, Torrance, Calif) were dissolved in pyrogen-free saline (PFS). These stock solutions were stored at -30 C until use. The doses of the substances used in these experiments were as follows: for α-hcrh, 0.26, 1.3, and 6.5 nmol (1, 5, and 25 µg, respectively); for astressin, 0.14, 0.7, and 3.5 nmol (0.5, 2.5, and 12.5 µg, respectively); for IL-1ra, 0.01, 0.1, and 0.25 µg. The doses of α-hcrh and astressin are the same as those used in previous studies of the effects of CRH on spontaneous sleep. 20 Animals Male Sprague-Dawley rats ( g; National Laboratory Animal Breeding and Research Center, Taiwan) were used in present study. Rats were anesthetized by intraperitoneal injection of ketamine and xylazine (87 mg/kg and 13 mg/kg, respectively), and intraperitoneally injected with butorphanol tartrate and penicillin G benzathine to reduce pain and avoid infection. The EEG screw electrodes and a guide cannula directed into the lateral ventricle were surgically implanted as previously described. 23 The insulated leads from the EEG electrodes were routed to a Teflon pedestal (Plastics One, Roanoke, Va). Additional implantation of a bipolar electrode (model #MS303/9, Plastics One, Roanoke, Va) was placed in the right central nucleus of the amygdala; the coordinates, adopted from the Paxinos and Watson rat atlas, were as follows: AP, -2.0 mm from bregma; ML, 4.0 mm; DV, 8.0 mm. The Teflon pedestal and bipolar electrode were then cemented to the skull with dental acrylic (Cranioplastic cement and Cyanoacrylate gel, Plastics One, Roanoke, Va). The incision was treated topically with polysporin (polymixin B sulfate bacitracin zinc), and the rats were allowed to recover for at least 1 week before the initiation of experiments. The rats were then housed in individual recording cages in an environmentally controlled chamber (COCONO model #LE-539; Ron-Fong Technology Corporation, HsinChu, Taiwan). Each chamber possessed 2 cages. The chambers were maintained at 23 C ± 1 C with a 12:12-hour light:dark cycle (20 watt x 6 tubes illumination). Food and water were available ad libitum. Two days after EEG electrode implantation, the pedestal was connected to the amplifier system via a connector cable ( cable, Plastics One, Roanoke, Va). On the third postsurgical day, we administered 400 ng ICV angiotensin II (Tocris), which causes a drinking response mediated by structures in the preoptic area, 24 to assess the free drainage of the ICV cannula. All rats were again injected with angiotensin at the end of each experimental protocol; only data from those rats that elicited a positive drinking response were included in the subsequent analyses. The animals were habituated by daily handling and ICV injections of PFS with daily kindling manipulation (see next section) timed to coincide with scheduled experimental administrations until kindling seizure was apparent. Kindling Manipulation Seven days after the operation, the rats were subjected to a kindling procedure. A stimulator-isolator unit (A360 Stimulus Isolator, World Precision Instruments, Sarasota, Fla) triggered by a main stimulator (Accupulser A310, World Precision Instruments, Sarasota, Fla) was used to deliver kindling stimuli. Each kindling stimulus was simultaneously applied once a day to the right central nucleus of the amygdala when the animals were allowed to move freely in their cages. The stimulus was a train of biphasic pulses (1-millisecond duration each) of 50 Hz for 1 second. The intensity for each stimulus, which ranged from 50 to 200 µa, was determined as a threshold for the first appearance of the concurrent SLEEP, Vol. 27, No. 2, afterdischarge. The EEG recordings were observed 20 minutes before, during, and 20 minutes after the kindling stimulation. The afterdischarges are defined as spikes of more than 2 Hz and a voltage at least 3 times higher than the prestimulation baseline EEG. 25 The kindling stimulation was performed once a day, when animals were awake, 5 minutes prior to either the light onset (light-onset kindling) or the dark onset (dark-onset kindling). Criteria for establishing the presence of a kindled seizure required behavioral seizures that attained stage 5 by Racine s criteria. 26 Racine-assessed kindling seizures were classified into 5 stages as follows: stage 1, mouth and facial twitches; stage 2, clonic head movement; stage 3, unilateral forelimb clonus followed by contralateral clonus; stage 4, clonic rearing; stage 5, loss of postural control. 26 Three to 4 weeks was required until Racine s stage 5 seizures developed. The secondary epileptiform EEG was defined as the EEG with high-amplitude sharp waves of increased frequency and polyspikes. The criteria of secondary epileptiform EEG are similar to those of afterdischarge, with spikes of more than 2 Hz and amplitude at least 3 times higher than baseline EEG. Apparatus and Recording Signals from the EEG electrodes were connected to amplifiers (Colbourn Instruments, Lehigh Valley, Penn; model V75-01). The EEG signals were amplified at a gain of 5,000, and the analog bandpass was filtered between 0.1 and 40 Hz (frequency response: ± 3 db; filter frequency roll off: 12 db/octave). These conditioned signals (EEG) were subjected to analog-to-digital conversion with 16-bit precision at a sampling rate of 100 ks per second (NI PCI-6033E; National Instruments, Austin, Tex). The digitized EEG waveform was stored as binary computer files until subsequent analysis. Determination of vigilance state was analyzed by visual scoring of 12- second epochs using custom software (ICELUS, Dr. Opp) written in LabView for Windows (National Instruments). 23 The animal s behavior was categorized into SWS, REM sleep, or waking based upon previously defined criteria. 23 Briefly, SWS is characterized by large-amplitude EEG slow waves and high power density values in the delta frequency band ( Hz). During REM sleep, the amplitude of the EEG is reduced, and the predominant EEG power density occurs within the theta frequency ( Hz). During waking, the rats are generally active, and the amplitude of the EEG is similar to that observed during REM sleep, but power density values in the delta frequency band are generally greater than those in the theta frequency band. Experimental Protocols Five distinct groups of rats were used. Rats in Groups 1 through 4 (n = 8 for each group) were used to determine the behavioral alterations induced by amygdala kindling occurring either at the beginning of the dark period or at the beginning of the light period and to determine the effects of CRH receptor antagonists (astressin and α-hcrh) and the IL- 1ra on kindling-induced alterations in sleep. After the rats had adapted to the 12:12-hour light:dark cycle, a 24-hour undisturbed baseline EEG was recorded. Rats in Group 1 then received ICV administration of PFS 20 minutes prior to the beginning of the light period on the next day, and a 23-hour continuous EEG was recorded as a vehicle control. Then, kindling manipulation was performed. Until consistent Racine s stage 5 seizures appeared, rats were given PFS injections combined with kindling stimuli prior to the light onset on the first recording day. The ICV administration of 3 doses of astressin, 0.14, 3.5, and 0.7 nmol, prior to the kindling stimulation was initiated 20 minutes before the beginning of the light period of the 12:12-hour light:dark cycle on the subsequent second, third, and fourth recording days, respectively. All recordings began at the light onset and continued for 23 hours. Rats in Group 2 underwent a protocol similar to that used on the rats in Group 1, except that they were given α-hcrh (0.26, 6.5, and 1.3 nmol, respectively) instead of astressin. Rats in Group 3 received all injections and kindling manipulation prior to the beginning of the dark period of the light:dark cycle, in

3 which rats were given ICV administered IL-1ra in doses of 0.01, 0.25, and 0.1 µg, and the EEG recording began from the onset of the dark period. Rats in Group 4 were used to determine the effects of 12.5 µg astressin/25 µg α-hcrh (n = 4) on dark-onset kindling-induced sleep alterations and the effects of 0.25 µg IL-1ra (n = 4) on light-onset kindling-induced sleep disturbances. A diagram of the experimental protocol is delineated in Figure 1. The volume for all injections was 3 µl, and each injection was given over approximately a 2-minute period. The doses of the CRH antagonists and IL-1ra used in this study are known to have effects on physiological sleep, but they last only for a few hours after ICV administration. 20,27 The short-term effect of these substances suggests that the 1-day washout period used in the current protocol is suitable for the purpose of multiple-doses administration. Rats in Group 5 (n = 48) were used to determine the effects of kindling on circulating corticosterone concentrations and cytokine mrna expression in the brain in both light and dark periods. An ICV guide cannula and a stimulation electrode were surgically implanted in these animals. They were maintained under the same conditions in the same environmental chambers as those used in the behavioral studies (see above). These rats were habituated by daily handling and ICV injection of vehicle (PFS). Kindling protocol was the same as for those in the behavioral studies. On the experimental day, these rats were kindled and injected with PFS, 3.5 nmol of astressin, or 6.5 nmol of α-hcrh at either the dark onset or the light onset. Those that received PFS were decapitated at 1 or 2 hours (n = 6 for each group) after injection, and those that received astressin or α-hcrh were decapitated 1 hour after injection (n = 6 for each group). Trunk blood was collected into Vacutainer tubes containing EDTA and centrifuged for 15 minutes at 1500 rpm (approximately 400 x g) at 4 C. Plasma was then aliquoted and stored at -80 C until assay. The brains were immediately removed from the skull after decapitation and placed on an ice-cold surface. The meninges were then removed, and the hippocampus and a section of the parietal cortex were dissected out. All tissues were immediately placed in liquid nitrogen and stored at -80 C until mrna extraction and assay. Corticosterone Radioimmunoassay Total plasma corticosterone was measured by radioimmunoassay using rabbit antiserum raised against corticosterone-3-carboxymethyloxime BSA and [ 125 I]corticosterone (ICN Biomedicals, Calif). The antiserum has very low cross-reactivity with other glucocorticoids and their products (less than 0.4%). Intraassay coefficients of variation are between 4.4% and 10.3% and intraassay coefficient of variations are between 6.5% and 7.2% (manufacturer s specifications). Ribonuclease Protection Assay for Cytokine mrna Expression We used a ribonuclease protection assay as previously described 22 to detect and quantify specific mrna species. Briefly, total RNA from each tissue sample was extracted using TRI Reagent (Sigma, St. Louis, MO, USA). A rat-specific cytokine multiprobe template kit (rck-1, BD PharMingen, San Diego, Calif, USA) for the T7 polymerase-directed synthesis of high-specific-activity, [ 35 P]-labeled antisense RNA probes, was used to detect mrna species, including IL-1α, IL-1β, tumor necrosis factor (TNF)-β, IL-3, IL-4, IL-5, IL-6, IL-10, TNF-α, IL-2, and gamma interferon (IFNγ). A probe specific for the ribosomal protein L32 was added to the template to serve as an internal standard. Yeast trna as a negative control and a rat mrna from stimulated splenocytes as a positve control were also added to each assay. The probe set was then hybridized in excess to target RNA, after which the free probe and other single-stranded RNA were digested with RNases. The remaining RNase-protected probes were purified and resolved on 5% denatured polyacrylamide gels, and the images were developed by Phosphor Imager SI (Molecular Dynamics, Buckinghamshire, UK). The images were then processed using ImageQuant (Molecular Dynamics). The intensity of a band in the image was directly proportional to the amount of radioactivity within the band. The optic density values obtained from each band were normalized against the optic density obtained from the L32 band in the same sample by the following expression: [(optic density of the sample band/optic density of the L32 band ) x 100]. Statistical Analyses Behavioral Experiments All values are presented as mean ± SEM. One-way analyses of variance for the duration of each vigilance state (SWS, REM sleep, waking) were performed across the two 12-hour time blocks comprising the 23- hour recording period. The main effect consisted of manipulation (PFS + kindling, astressin + kindling, α-hcrh + kindling, and IL-1ra + kindling). If statistically significant differences were detected, posthoc multiple comparisons were made to determine which condition or conditions contributed to the effect. An α level of P.05 was taken as indicating a statistically significant difference between 2 different manipulations. Circulating Corticosterone Concentrations and Cytokine mrna Expression Figure 1 The diagram of experimental protocol. The dark and open portions of bars represent the dark and light periods of the 12:12-hour light:dark cycle. Arrow indicates the manipulation, // represents the kindling procedure for developing stage 5 seizure, and RD stands for recording day. PFS refers to pyrogen-free saline; K, kindling stimulation; A1, 0.5 µg astressin; A2, 12.5 µg astressin; A3, 2.5 µg astressin; α1, 1 µg α-hcrh; α2, 25 µg α- hcrh; α3, 5 µg α-hcrh; IL (interleukin)-1ra 1, 0.01 µg IL-1ra; IL-1ra 2, 0.25 µg IL-1ra; IL-1ra 3, 0.1 µg IL-1ra. SLEEP, Vol. 27, No. 2, Circulating corticosterone concentrations are presented as mean ± SEM in unit of ng/ml. Cytokine mrna expression is depicted as optic density values. Independent t tests were performed on each sample to determine if values obtained after administration of vehicle (PFS) differed from those obtained after kindling. An α level of P.05 was taken as indicating a statistically significant difference between the manipulations.

4 RESULTS Effects of Amygdala Kindling on the Alterations of Sleep-Wake Activity Application of kindling at light onset reduced the total time spent in SWS during the subsequent 12-hour light period from 56.4% ± 2.1% to 41.3% ± 1.7% (n = 16 [Group 1 and 2 combined]; P <.05). The amount of REM sleep was also suppressed from 9.2% ± 0.6% to 3.9% ± 0.5% after light-onset kindling (n = 16 [Group 1 and 2 combined]; P <.05). The light-onset kindling increased the amount of wakefulness (Figure 2). Sleep-wake architecture was unchanged in the subsequent dark period. In contrast, application of kindling at dark onset increased the amount of SWS from 23.8% ± 2.2% to 32.1% ± 1.6% during the subsequent 12- hour dark period (n = 8; P <.05). The amount of REM sleep was not altered during the dark period, and there was also a mirrored reduction in the amount of wakefulness (Figure 3). However, the time spent in REM sleep in the following light period was suppressed from 5.1% ± 0.7% to 2.5% ± 0.4% after dark-onset kindling (P < 0.05). Figure 2 Effects of astressin (left panel) and α-helical corticotropin-releasing hormone (α-hcrh) (right panel) on amygdala kindling-induced sleep-wake alterations. Left panel: Individual data points depict the mean ± SEM values obtained from each group. The shaded areas depict the mean ± SEM range of values obtained from animals injected with pyrogen-free saline (PFS) (n = 16, Group 1 and 2 combined). These values were compared with those obtained from animals after intracerebroventricular administration of PFS + light-onset kindling ( ; n = 16 [Group 1 & 2 combined]) or 12.5 µg astressin + light-onset kindling ( ; n = 8 [Group 1]). The injections were given 20 minutes before the light onset, and kindling stimuli were performed 5 minutes prior to the light period. The dark portion of the bar on the x-axis represents the dark period of the 12:12-hour light:dark cycle. The inset depicts the mean ± SEM amount of time spent in slow-wave sleep (SWS) during the first 12-hour light period. From left to right: gray bar, PFS injection only; filled bar, PFS + kindling; hatched bars, 0.5 µg astressin + kindling and 2.5 µg astressin + kindling; open bar, 12.5 µg astressin + kindling. Right panel: The shaded areas depict the mean ± SEM range of values obtained from animals injected with PFS (n = 16, Group 1 & 2 combined). These values were compared with those obtained from animals after intracerebroventricular administration of PFS + light-onset kindling ( ; n = 16 [Group 1 & 2 combined]) or 25 µg α-hcrh + light-onset kindling ( ; n = 8 [Group 2]). The injections were given 20 minutes before the light onset and kindling stimuli were performed 5 minutes prior to the light period. The inset depicts the mean ± SEM amount of time spent in SWS during the first 12-hour light period. From left to right: gray bar, PFS injection only; filled bar, PFS + kindling; hatched bars, 1 µg α-hcrh + kindling and 5 µg α-hcrh + kindling; open bar, 25 µg α-hcrh + kindling. #depicts statistically significant difference from control (PFS only) values and * represents statistically significant difference from PFS + kindling. WAKE refers to wakefulness; REMS, rapid-eye-movement sleep. SLEEP, Vol. 27, No. 2,

5 Analyses of sleep-architecture parameters across hours 1 to 12 revealed that the reduction in SWS during the light period induced by the light-onset kindling stimulus was primarily due to a decrease in SWS bout duration, although the number of SWS bouts increased. The suppression of REM sleep induced by light-onset kindling was primarily due to a decrease in the number, rather than the length, of REM sleep bouts. The number of transitions from 1 state of vigilance to another during the 12-hour light period increased significantly after kindling stimulation, indicative of sleep fragmentation (Table 1). In addition to alterations in sleep quantity and architecture, EEG slow wave activity (SWA) during SWS was enhanced by light-onset amygdala kindling (Figure 4, left panel). Increase in SWS across the 12-hour dark period induced by dark-onset kindling stimulation was primarily due to the increase in the bout duration of SWS. The number of transitions also increased significantly, indicating that sleep was fragmented despite the enhancement of SWS (Table 1). The EEG SWA during SWS also increased after darkonset kindling stimulation (Figure 4, right panel). Effects of CRH-Receptor Antagonists on Kindling-induced Sleep-Wake Alterations ICV administration of astressin 20 minutes prior to the beginning of the light period dose-dependently blocked both the decrease in SWS and the increase in wakefulness induced by light-onset kindling (Figure 2, left panel and inserted bar graph). The time spent in SWS increased from 41.3% ± 1.7% (n = 16) after kindling stimulation to 52.6% ± 2.2% (n = 8; P < 0.05) after the administration of 3.5 nmol of astressin with kindling stimulation. REM sleep was not altered by astressin (Figure 2, left panel). Similar results were observed when the rats received 6.5 nmol α- hcrh ICV (Figure 2, right panel). We observed that high doses of ICVadministered astressin (12.5 µg) or α-hcrh (25 µg) did not significantly change the dark-onset kindling-induced sleep alteration. The amounts of SWS obtained after dark-onset kindling, astressin + dark-onset kindling, and α-hcrh + dark-onset kindling were 33.2% ± 2.1%, 33.4% ± 1.8%, and 32.8% ± 2.5% (n = 4), respectively. The enhancement of SWS during the light period after administration of 3.5 nmol of astressin was primarily due to the increase in bout duration of SWS; SWS bout duration increased from 2.3 ± 0.3 minutes after kindling stimulation to 3.0 ± 0.4 minutes after astressin with kindling stimulation (Table 1; P <.05). There was no significant change in any sleep-architecture parameter after administration of α-hcrh. The enhancement of SWA during SWS induced after light-onset kindling stimulation was not altered by either astressin or α-hcrh (Figure 4). Changes in Circulating Corticosterone Concentrations After Kindling Stimulation Circulating corticosterone concentrations in plasma 1 hour after lightonset kindling stimulation increased from 199 ± 32 ng/ml to 563 ± 43 ng/ml (n = 6; P <.05). The ICV administration of 3.5 nmol of astressin and 6.5 nmol of α-hcrh significantly reduced corticosterone concentrations to 220 ± 63 ng/ml and 312 ± 71 ng/ml, respectively (n = 6 for each group; P <.05). Concentrations of circulating corticosterone during the dark period did not differ between the control and the dark-onset kindled animals; the normal circadian concentration of corticosterone at 1 hour after dark onset was 703 ± 99 ng/ml, and the concentration obtained after 1 hour dark-onset kindling was 675 ± 110 ng/ml (n = 6). Effects of CRH Receptor Antagonists on Secondary Epileptiform EEG Figure 3 Effects of interleukin-1 receptor antagonist (IL-1ra) on dark-onset amygdala kindling-induced sleep-wake alterations. Individual data points depict the mean ± SEM values obtained from 8 rats. The shaded areas depict the mean ± SEM range of values obtained from animals injected with pyrogen-free saline (PFS) (n = 8). These values were compared with those obtained from animals after intracerebroventricular administration of PFS + dark-onset kindling ( ) or 0.25 µg IL-1ra + dark-onset kindling ( ). The injections were given 20 minutes before the dark onset and kindling stimuli were performed at 5 minutes prior to the dark period. The dark portion of the bar on the x-axis represents the dark period of the 12:12-hour light:dark cycle. The inset depicts the mean ± SEM time spent in slowwave sleep (SWS) during the first 12-hour dark period. From left to right: gray bar, PFS injection only; filled bar, PFS + kindling; hatched bars, 0.01 µg IL-1ra + kindling and 0.1 µg IL-1ra + kindling; open bar, 0.25 µg IL-1ra + kindling. # depicts statistically significant difference from control (PFS only) values and * represents statistically significant difference from PFS + kindling. WAKE refers to wakefulness; REMS, rapid-eye-movement sleep. SLEEP, Vol. 27, No. 2, Our results indicate that light-onset kindling stimulation induced secondary epileptiform EEG, which occupied 2.6% ± 0.6% of the total 12- hour light period. The α-hcrh (6.5 nmol) suppressed secondary epileptogenesis; the time spent in epileptiform EEG during the light period was significantly reduced to 1.1% ± 0.3% after ICV administration of α- hcrh (n = 8; P < 0.05). Although astressin (3.5 nmol) had no significant effect on the secondary epileptiform EEG during the light period, ICV administration of astressin suppressed epileptogenesis during the subsequent dark period (Figure 5, upper panel). Effects of IL-1ra on Kindling-induced Sleep-Wake Alterations The ICV administration of IL-1ra 20 minutes prior to the beginning of the dark period dose-dependently blocked both the increase in SWS and

6 the decrease in waking induced by dark-onset amygdala kindling (Figure 3 and inserted bar graph). The total time spent in SWS was reduced from 32.1% ± 1.6% after kindling stimulation to 21.6% ± 1.5% after administration of IL-1ra with kindling stimulation (Figure 3, P <.05). During the dark period, REM sleep was suppressed after injecting IL-1ra prior to kindling, although REM sleep was not altered after dark-onset kindling. Furthermore, we observed that ICV administration of 0.25 µg IL- 1ra prior to the light period did not significantly alter light-onset-induced sleep alteration; the amounts of SWS were 42.3% ± 1.5% and 41.5% ± 2.3% (n = 4) after light-onset kindling and IL-1ra + light-onset kindling, respectively. Analyses of sleep-architecture parameters across the 12-hour dark period revealed that the blockade of dark-onset kindling-induced SWS enhancement by IL-1ra was primarily due to reduction in SWS bout duration (Table 1). The number of transitions from 1 state of vigilance to another during the 12-hour dark period increased even more after application of IL-1ra, indicating that sleep was more fragmented (Table 1). The IL-1ra did not alter the enhanced EEG SWA during SWS induced by dark-onset kindling; however, the IL-1ra partially decreased SWA during the subsequent light period (Figure 4). Kindling-induced Alterations of Cytokine mrna Expression values represent basal levels of cytokine mrna expression at hour 1 and hour 2 of the dark period of the 12:12-hour light:dark cycle. The IL-1α mrna expression in the hippocampus and the cortex increased 2 to 3 fold after 1 and 2 hours post-dark-onset kindling during the dark period (Table 2, Figure 6). The IL-1β mrna expression increased 3 to 4 fold in the cortex after 1 and 2 hours postkindling stimulation (Table 2, Figure 6). The IL-6 and TNF-α mrna expression was not consistently altered by kindling stimuli. On the other hand, levels of cytokine mrna expression in the hippocampus and cortex during the light period were not altered by light-onset kindling (data not shown). Effects of IL-1ra on Secondary Epileptiform EEG Our results indicate that dark-onset kindling stimulation also induced secondary epileptiform EEG activity, which occupied 1.5% ± 0.4% of the dark period of the 12:12-hour light:dark cycle. A high dose of IL-1ra (0.25 µg) enhanced kindling-induced secondary epileptogenesis during the first 12-hour dark period; the time spent in epileptiform EEG was significantly enhanced to 4.3% ± 1.1% after ICV administration of 0.25 µg IL-1ra (Figure 5, lower panel; P <.05). However, there was no change in the amount of epileptiform EEG activity during the subsequent light period. Low levels of IL-1α, IL-1β, IL-6, and TNF-α mrna expressions were detected in the hippocampus and the cortex taken from rats injected with PFS at 1 and 2 hours after dark onset. (Table 2, Figure 6). Those DISCUSSION Table 1 Effects of light-onset and dark-onset amygdala kindling on sleep-wake architecture of rats Bouts, no./h Bout duration, min Transitions, no./h Manipulation Hour L:D cycle Wake SWS REM sleep Wake SWS REM sleep PFS 1-12 L 4.6 ± ± ± ± ± ± ± 3.2 PFS + light-onset kindling 1-12 L 7.8 ± 0.3* 10.7 ± 0.4* 1.25 ± 0.4* 5.1 ± ± 0.3* 1.2 ± ± 5.0* 12.5 µg astressin + light-onset kindling 1-12 L 6.5 ± ± ± ± ± ± ± µg α hcrh + light-onset kindling 1-12 L 7.5 ± ± ± ± ± ± ± 8.2 PFS 1-12 D 4.6 ± ± ± ± ± ± ± 3.0 PFS + dark-onset kindling 1-12 D 5.3 ± ± ± ± 1.6* 2.7 ± 0.2* 0.9 ± ± 2.9* 0.25 µg IL-1ra + dark-onset kindling 1-12 D 6.3 ± ± ± ± ± ± ± 10.2 Values are provided as mean ± SEM. Differences were detected by 1-way analyses of variance within the indicated time blocks. *denotes a statistically significant difference between values obtained after intracerebroventricular administration of pyrogen-free saline (PFS) and those obtained after amygdala kindling with pretreated vehicle. denotes a statistically significant difference between values obtained after amygdala kindling with pretreated vehicle (PFS) and those obtained after amygdala kindling with pretreated substances (either 12.5 µg astressin, 25 µg α-helical corticotropin-releasing hormone (hcrh), or 0.25 µg interleukin [IL]-1ra) Period of the light:dark cycle immediately prior to which injections were given: D = dark period, L = light period. WAKE refers to wakefulness; SWS, slow-wave sleep; REM sleep, rapid eye movements sleep Table 2 Expressions of cytokine mrna in hippocampus and cortex 1 hour and 2 hours after dark-onset amygdala kindling during dark period Hippocampus 1H Hippocampus 2H Cortex 1H Cortex 2H Control Kindling Control Kindling Control Kindling Control Kindling IL-1α 0.78 ± ± 0.40* 1.08 ± ± 0.23* 1.02 ± ±.32* 1.05 ± ±.43* (6/6) (6/6) (6/6) (6/6) (6/6) (6/6) (6/6) (6/6) IL-1β 2.50 ± ± ± ± ± ±.52* 2.01 ± ±.24* (6/6) (6/6) (6/6) (6/6) (6/6) (6/6) (6/6) (6/6) IL ± ± ± 1.05 Not Not Not 1.09 (3/6) (3/6) (1/6) (3/6) detected detected detected (1/6) TNF-α 1.60 ± ± ± ± ± ± ± 0.51 (4/6) (4/6) (3/6) (3/6) (4/6) (3/6) (1/6) (2/6) Values are mean ± SEM optical density ratios obtained from brain tissue of rats sacrificed either 1 hour (1H) or 2 hours (2H) after intracerebroventricular administration of pyrogen-free saline (PFS) with (kindling) or without (control) dark-onset amygdala kindling. *denotes statistically significant differences of values obtained from rats given kindling prior to dark onset as compared to values obtained from rats injected with PFS only (P.05). Numbers in parentheses indicate the number of samples in which optical density values for specific mrnas exceeded threshold criteria out of the total number of samples obtained. IL refers to interleukin; TNF, tumor necrosis factor The results from this study revealed an intimate interplay between epilepsy and sleep-wake behavior. The effects of seizure on sleep patterns have been controversial. Van Sweden has suggested that experimental amygdala kindling induces disturbed sleep by favoring wakefulness and light sleep. 8 Raol and Meti, however, have reported that there is an increase in deep SWS, a decrease in light SWS, and a decrease in the quiet wakefulness after a full-blown seizure. 9 Furthermore, REM sleep has been demonstrated to be unaltered after full-blown kindling stimulation, 28,29 whereas other evidence has shown that kindling seizures suppress REM sleep. 30,31 Results from Gigli and Gotman 32 have delineated that full-blown kindling reduces the percentage of REM sleep and the number of entries into REM sleep and increases the percentage of wakefulness. Reasons for these confusing results might be due to the different kindled brain structures, ie, hippocampus or amygdala, or due to the different stimulation time points during the 24-hour circadian rhythm. Therefore, we propose that kindling stimuli occurring at different time points in the sleep-wake cycle differ in the ways in which they alter sleep patterns. We kindled the central nucleus of the amygdala at the beginning of either the light period or the dark period when rats were awake. We SLEEP, Vol. 27, No. 2,

7 continued the kindling procedure until a consistent Racine s stage 5 seizure occurred. Our results indicate that amygdala kindling at different circadian time points produces diverse alterations in sleep-wake activity; light-onset kindling decreases both SWS and REM sleep during the major sleep period, whereas dark-onset kindling enhances SWS but does not alter REM sleep during the major wake period. These results suggest that the diversity of sleep changes is dependent upon the timing of seizure occurrence. Evidence that daytime seizures significantly decrease REM sleep the following night and that the decrease in REM sleep is more pronounced when seizures occur at night has been demonstrated in humans by Bazil et al. 33 In our study, light-onset kindling decreased REM sleep during the light (rest) period and dark-onset kindling suppressed REM sleep in the following light (rest) period, results consistent with Bazil s observations. The EEG seizures may increase neuropeptides and cytokines in several brain regions. The CRH mrna in the paraventricular nucleus of the hypothalamus increases after a kindling seizure, 17 and amygdala-kindled seizures elevate CRH and CRH binding protein (CRH-BP) in the dentate gyrus of the hippocampus in rats. 18 Furthermore, mrna expressions of CRH, CRH-BP, and CRH receptor type-1 (CRH-R1) are significantly enhanced in cortical tissue obtained from postmortem brains of children with generalized epilepsy. 19 In addition to the alteration of neuropeptides, kindling may also modulate cytokines in the central nervous system. Amygdala kindling up regulates IL-1β, IL-1 receptor type 1 (IL- 1R1), TNF-α, and TGF-β mrna expressions in the parietal, prefrontal, and piriform cortexes and in the amygdala and hippocampus. 14 Limbic seizures induced either by electrical stimulation of the hippocampus or intrahippocampal injection of kainic acid or bicuculline methiodide enhance IL-1β, IL-6, and TNF-α mrna in the hippocampus. 12 Cytokine concentrations in the central nervous system have also been altered by other seizure models in addition to the kindling. 13,34,35 Under normal circumstances, hypothalamic-pituitary-adrenal axis activity tracks the release of CRH from the paraventricular nucleus of the hypothalamus. Our results show that circulating corticosterone concentrations, which are lowest in the rest (light) period, increased during the rest period when light-onset kindling was applied. Corticosterone concentrations were not altered by dark-onset kindling during the active (dark) period because corticosterone concentration is normally at its highest during the Figure 4 Effects of intracerebroventricular (ICV) administration of astressin α-helical corticotropin-releasing hormone (α-hcrh) (left panel) and interleukin (IL)-1ra (right panel) on kindling-induced increases in slow-wave activity (SWA) during slow-wave sleep (SWS). Left panel: Kindling stimulation increased SWA during SWS, but astressin and α-hcrh had little effect on the increase in SWA during SWS induced by light-onset kindling. Values are mean ± SEM obtained from animals after ICV administration of pyrogen-free saline (PFS) only ( ), PFS + light-onset kindling ( ) or 12.5 mg astressin (upper panel) / 25 µg α-hcrh (lower panel) + light-onset kindling ( ). The inset depicts the mean ± SEM values of SWA during SWS in the first 12-hour light period. Gray bars, animals ICV injected with PFS; filled bars, same animals injected ICV with PFS + light-onset kindling; open bars, same animals injected ICV with either astressin (upper inset) or α-hcrh (lower inset) + light-onset kindling. * P <.05 versus PFS injection only. Right panel: Dark-onset kindling manipulation increased SWA during SWS. The ICV administration of IL-1ra did not alter the increase in SWA during SWS induced by dark-onset kindling during the dark period but partially suppressed SWA in the subsequent light period, although there was no statistical difference. Values are mean ± SEM obtained from animals after ICV administration of PFS only ( ), PFS + dark-onset kindling ( ), or 0.25 µg IL-1ra + dark-onset kindling ( ). SLEEP, Vol. 27, No. 2,

8 dark period. On the other hand, we also found that IL-1α and IL-1β mrna expressions in the hippocampus and the parietal cortex were enhanced at 1 and 2 hours after dark-onset kindling stimuli during the dark period, but light-onset kindling did not consistently alter IL-1 mrna expression. Therefore, our results suggest that amygdala kindling simultaneously stimulates the CRH and IL-1 systems in the brain. Of course, there are also data suggesting alterations in some classic neurotransmitters after EEG seizures. For example, amygdala-kindled seizures decrease GABA sensitivity in the dorsal raphe, which might reflect long-term neuronal changes associated with seizures. 36 GABA has also been implicated in sleep generation (for a review, see Mendelson 37 ); therefore, it would be intriguing to determine the involvement of other neurotransmitters in the nature of kindling seizures in future experiments. Both CRH and cytokines (ie, IL-1) in the central nervous system play important roles in the regulation of sleep-wake behavior. We have previously hypothesized that CRH is involved in the regulation of physiological sleep-wake activity, based in part on observations that central or systemic administration of CRH receptor antagonists reduces spontaneous waking. 20,21,38 Other evidence supports the hypothesis that CRH contributes to spontaneous wakefulness. For example, CRH is found in cerebrospinal fluid where, in humans and other primates, it exhibits circadian rhythms. 39,43-45 In addition, circadian fluctuation of plasma adrenocorticotropic hormone and cortisol or corticosterone that are lowest prior to the major sleep time and peak around the beginning of the active period in humans, 46,47 rhesus monkeys, 48 and rats 49,50 is temporally associated with waking. On the other hand, the somnogenic properties of IL-1 are well documented (reviewed by Krueger et al 51 and Opp et al 52 ). IL-1 administered centrally into rats during their active (dark) period is particularly effective in increasing SWS and reducing wakefulness. IL-1β mrna expression in rat brain is highest during the light period of the light:dark cycle, the period when rats sleep the most, and lowest during the dark period, when rats are most active. Based upon the aforementioned evidence, we hypothesized that amygdala kindling simultaneously stimulates the synthesis and release of both CRH and IL-1 in the central nervous system, which in turn modulates epileptogenesis and Figure 5 Effects of corticotropin-releasing hormone (CRH) receptor antagonists and interleukin (IL)-1 receptor antagonist (IL-1ra) on kindling-induced secondary epileptogenesis. Upper panel: CRH receptor antagonist, 12.5 µg astressin (open bars), or 25 µg α-helical CRH (α-hcrh)(hatched bars) statistically suppressed light-onset kindling-induced epileptiform electroencephalogram (EEG). Light-onset amygdala kindling was given 5 minutes before the light onset. Filled bar denotes the values obtained after PFS + light-onset kindling. *P <.05 versus pyrogen-free saline (PFS) + light-onset kindling. Lower panel: IL- 1ra enhanced dark-onset kindling-induced epileptiform EEG. Dark-onset amygdala kindling was given 5 minutes prior to the dark onset. Filled bar: PFS + dark-onset kindling; hatched bar: 0.1 µg IL-1ra + dark-onset kindling; open bar: 0.25 µg IL-1ra + dark-onset kindling. *P <.05 vs PFS + dark-onset kindling. SLEEP, Vol. 27, No. 2, Figure 6 Alterations in cytokine mrna expression 2 hours after dark-onset amygdala kindling. This is a representative image of a 5% denatured polyacrylamide gel obtained by a phosphorimager on which double-stranded ribonuclease-protected fragments were resolved. Total mrna was extracted from brain tissue of rats that had been injected intracerebroventricularly with pyrogen-free saline (PFS) and kindling (K) or without kindling (C) prior to dark onset. These animals were sacrificed 2 hours after the dark onset. The ribonuclease protection assay template consisted of radiolabeled antisense RNA probes for interleukin (IL)-1α (protected probe size, 403 nucleotides), IL-1β (361 nucleotides), tumor necrosis factor (TNF)β (322 nucleotides), IL-3 (286 nucleotides), IL-4 (256 nucleotides), IL-5 (226 nucleotides), IL-6 (202 nucleotides), IL-10 (181 nucleotides), TNFα (160 nucleotides), IL-2 (142 nucleotides), and gamma interferon (IFNγ, 129 nucleotides). Cont. RNA refers to control RNA obtained from stimulated rat splenocytes (positive control); Temp. RNA, template; Hippo, hippocampus; Cort, cortex.

9 SLEEP, Vol. 27, No. 2, sleep-wake activities. Blockade of light-onset kindling-induced decreases in SWS and increases in wakefulness by CRH receptor antagonists supports this hypothesis. Further support comes from the fact that CRH blockers attenuate the enhancement of circulating corticosterone concentrations induced by light-onset kindling. These results argue that increased CRH after light-onset kindling contributes to sleep-wake alteration in rats. The fact that suppression of REM sleep that was induced by kindling was not altered after the administration of astressin or α- hcrh suggests that mechanisms other than CRH may be involved in REM-sleep alteration. In addition to sleep regulation, CRH antagonists also reduced secondary epileptic activity after kindling in our study. Although our previous results suggested that CRH antagonist has an effect on physiological sleep-wake regulation, 20 there is the possibility of a secondary effect on sleep due to an influence on seizure susceptibility after administration of CRH antagonist. Our previous results have shown that the effective doses of astressin and α-hcrh given at dark onset reduced waking with different time courses; astressin reduced waking with a 6- to 7-hour delay, but α-hcrh reduced waking limited to the first 2 hours after it was administered. 20 However, both antagonists given at light onset did not alter sleep. 20 Our present results show that both antagonists block kindling-induced changes in sleep during the light period and that the effects are immediate during hour 1 and thereafter. This observation suggests that blocking CRH receptors alters kindling-induced effects on sleep by mechanisms at least somewhat independent of sleep-regulation mechanisms per se. Nevertheless, the fact that CRH receptor antagonists normalize sleep of kindled rats is clear. Our data also indicate that IL-1ra, which has little effect on sleep during the dark period, dose-dependently blocked the dark-onset kindlinginduced increase in SWS and decrease in wakefulness. The administration of IL-1ra enhanced secondary epileptic activity after kindling. This evidence may also suggest that the impact of IL-1ra on sleep disruption reflects a mechanism independent of sleep regulation (ie, seizure susceptibility). Our results indicate that dark-onset kindling increases IL-1 concentrations in the brain during the dark period when IL-1 expression is at its lowest, which may contribute to the increase in SWS and the decrease in waking. Daytime kindling only increases limited IL-1, while IL-1 concentration is high during the rest period, and this minimal increase in IL-1 may not produce additional somnogenic action. REM sleep was not altered by dark-onset kindling during the first 12-hour dark period, but it was suppressed in the subsequent light period. Application of IL-1ra did not significantly block REM-sleep suppression in the following light period, indicating that IL-1 may not be involved in this mechanism. These results are consistent with previous observations that administration of IL-1 enhances SWS but does not alter REM sleep in rats. 27,53 Surprisingly, IL-1ra reduced REM sleep during the dark period after dark-onset kindling, although REM sleep was not altered by kindling itself. This observation suggests that other mechanisms regulated by IL-1 in the central nervous system may become predominant in REM-sleep generation after dark-onset kindling stimulation. For example, IL-1 stimulates growth hormone secretion through a hypothalamic effector, growth hormone-releasing hormone. Both growth hormone and growth hormone-releasing hormone exhibit REM-sleep promotion (see Krueger et al 54 ). If the growth hormone / growth hormone-releasing hormone system induced by IL-1 were to become predominant in REMsleep generation after dark-onset kindling, administration of IL-1ra might be able to reduce the IL-1 effect and subsequently reduce REM sleep mediated by the growth hormone /growth hormone-releasing hormone system. However, this hypothesis needs to be tested in future studies. Analyses of sleep architecture parameters across the 12-hour period after kindling stimulation indicates that amygdala kindling, regardless of when the stimuli were given, increased sleep fragmentation. This result is in agreement with previous observations by Shouse and colleagues 55 that amygdala kindling in cats enhances sleep fragmentation. One clinical observation also showed that temporal lobe or frontal lobe seizures induce dark-onset sleep fragmentation and daytime sleepiness. 56 In addition to alterations in sleep quantity and architecture, our results also demonstrated that EEG SWA during SWS significantly increased following dark-onset and light-onset kindling. The EEG SWA during SWS is thought to reflect both sleep debt and intensity and is a characteristic homeostatic response to sleep deprivation and sleep fragmentation. 57 As such, the increase in SWA during SWS that we observed could conceivably be a response to sleep fragmentation and may not be relevant to either the CRH or IL-1 systems. It is well documented that CRH exhibits a proconvulsant effect, as well as limbic seizures that resemble the seizures induced by the rigid glutamate analogue, kainic acid, and by rapid amygdala kindling. 15 CRH also augments bicuculline-induced interictal-like bursts of population spikes to a greater extent in slices in rats 10 and exacerbates cocaine-kindled seizure development. 11 Our results demonstrate that the time spent in epileptiform EEG induced by light-onset kindling increased and that the CRH receptor antagonists, astressin and α-hcrh, reduced secondary epileptogenesis. The results we observed further confirm that increased CRH concentrations in the brain induced after amygdala kindling may exacerbate secondary epileptogenesis. Nonetheless, it remains controversial whether the enhancement of central cytokines by EEG seizure is neuroprotective or epileptogenic. There is growing evidence supporting the neuroprotective actions of cytokines. For example, TGF-β which has been implicated in metabotropic glutamate receptor 3-mediated neuroprotection, is up regulated during the first 3 weeks after status epilepticus throughout the hippocampus. 16 On the other hand, the fact that intrahippocampal injection of human recombinant IL-1β enhances seizure activity induced by kainate 13 indicates that cytokines may contribute to epileptogenesis. IL-1β has been reported to inhibit postsynaptic NMDA receptors, 58 but another study demonstrated the neurotoxicity of TNF-α or IL-1β via NMDA receptor-mediated nitric oxide production. 59 The results of the current study show that the epileptiform EEG induced by dark-onset kindling is less than that induced by lightonset kindling and that the blockade of IL-1 actions by central administration of IL-1ra dose dependently enhances dark-onset kindlinginduced epileptiform EEG, suggesting that increased IL-1 in the central nervous system after dark-onset kindling may possess a protective role to suppress secondary epileptogenesis. Collectively, our results provide the contribution of CRH and IL-1 to the seizure-induced sleep-wake alterations, which may reflect patients with seizures who feel drowsy and have increased sleepiness during day and disrupted sleeps during the night. In addition to the sleep regulations, we demonstrated the epileptogenic effect of CRH and the epileptosuppressive effect for IL-1. Therefore, the use of CRH receptor antagonists and IL-1 agents would be beneficial to improve seizure control and seizure-associated sleep disruptions. However, none of these peptide analogs crosses the blood-brain barrier when they are administered systemically. Developing a nonpeptide compound that reach central nervous system receptors is required for achieving this treatment purpose. ACKNOWLEDGEMENTS The technical assistance of Ms. Meng-Ping Hsiao and Mr. Hsu-Jan Liu is gratefully acknowledged. We thank Dr. Mark R. Opp and Mr. Jeffrey Conrad for reading this manuscript. This work was supported by National Science Council grant NSC B and China Medical University Hospital grant DMR REFERENCES 1. Herman ST, Walczak TS, Bazil CW. Distribution of partial seizures during the sleepwake cycle: differences by seizure onset site. Neurology 2001;56: Ferrillo F, Beelke M, Nobili L. Sleep EEG synchronization mechanisms and activation of interictal epileptic spikes. Clin Neurophysiol 2000;111:S Ferrillo F, Beelke M, De Carli F, et al. Sleep-EEG modulation of interictal epileptiform discharges in adult partial epilepsy: a spectral analysis study. Clin Neurophysiol 2000;111: Kumar P, Raju TR. Seizure susceptibility decreases with enhancement of rapid eye movement sleep. Brain Res 2001;922: Shouse MN, Farber PR, Staba RJ. Physiological basis: how NREM slepp components