Abstract
Pregabalin, an analog of γ-aminobutyric acid (GABA) that does not interact with GABA receptors, is in development as an analgesic, an anticonvulsant, and an anxiolytic. We evaluated the potential somnogenic actions of pregabalin in rats and compared it to those of triazolam, a widely used hypnotic. Pregabalin increased the duration of nonrapid eye movement sleep (NREMS) and decreased rapid eye movement sleep (REMS) after either dark onset or light onset administration. Triazolam increased duration of NREMS and had no effect on duration of REMS. Pregabalin markedly increased the duration of NREMS episodes and decreased the number of NREMS episodes. Power spectrum analysis revealed pregabalin-induced dose-dependent increases in relative delta power after administration. In contrast, triazolam decreased electroencephalographic power density in low frequency bands. Results suggest that pregabalin is a potential sleep modulating agent.
Several molecules in the benzodiazepine class are used as somnogenic agents. These substances, e.g., triazolam, increase duration of nonrapid eye movement sleep (NREMS), primary stage II in humans (Pakes et al., 1981;Tan et al., 1998). The enhanced duration of sleep episodes (sleep consolidation) induced by benzodiazepines is thought to be responsible for the feelings of restoration following short-term use of these benzodiazepine-based somnogenic agents. However, an enigma associated with their use has been the observation that benzodiazepine-based agents inhibit electroencephalographic (EEG) power in the low frequency range (0.5–10 Hz) (Tan et al., 1998). In experimental sleep-deprivation studies, recovery sleep is characterized by enhanced EEG delta wave (0.5–4 Hz) activity and reduced REMS (Pappenheimer et al., 1975). This enhanced EEG delta wave activity is thought to reflect sleep intensity and indeed EEG delta wave activity is the key variable used to model process S in the widely accepted two-process model of sleep regulation (Borbély and Achermann, 1999). Such observations suggest that benzodiazepine-based somnogenic agents do not induce physiological sleep. Furthermore, their long-term use is associated with confounds, such as sleep disruption, memory loss, addiction, and multiple drug interaction, that limit their usefulness (Kirkwood, 1999).
Pregabalin (3-isobutyl GABA), the pharmacologically activeS-enantiomer of 3-aminomethyl-5-methyl-hexanoic acid, has anticonvulsant, analgesic, and anxiolytic activity in many animal models (Bialer et al., 1999; Bryans and Wustrow, 1999; Kinsora et al., 1999; Field et al., 2001). It was synthesized as a lipophilic analog of GABA capable of penetrating the blood-brain barrier (Bryans and Wustrow, 1999). Pregabalin has a novel, although incompletely understood, mechanism of action. Pregabalin significantly improved sleep in several recently completed clinical studies of neuropathic pain (U. Sharma, D. Iacobellis, C. Glessner, M. Hes, L. LaMoreaux, R. Allen, and R. Pool, unpublished observations). Therefore, we thought it useful to investigate whether pregabalin had direct sleep-modulating effects.
Triazolam is a short half-life 1,4-benzodiazepine analog that is a potent sleep inducer in various mammalian species and a widely used hypnotic in humans (Pakes et al., 1981). The purpose of this study was to evaluate the effects of pregabalin on spontaneous sleep in rats and compare it to the somnogenic actions of triazolam. We report that pregabalin is a novel sleep modulator in that it enhances NREMS and EEG slow wave (0.5–4.0 Hz) activity and inhibits REMS.
Materials and Methods
Agents.
Pregabalin was a gift from Parke-Davis Warner-Lambert (Ann Arbor, MI). Triazolam was purchased from Sigma (St. Louis, MO). Pregabalin was dissolved in water just before its administration. Because triazolam is relatively insoluble in water, the mixture of triazolam and water was sonicated and then administered as a uniform suspension.
Animals and Surgery.
Male Sprague-Dawley rats weighing 250 to 360 g were used. The rats were kept on a 12:12 h light/dark cycle (lights on at 9:00 AM) at 23 ± 2°C ambient temperature. They had free access to water and food during the experiment. Surgeries were performed under ketamine-xylazine (87 and 13 mg/kg, respectively) anesthesia. Stainless steel jewelry screws for EEG recording were placed over the frontal and parietal cortices. An electromyogram (EMG) electrode was implanted in the dorsal neck muscles. A calibrated 30 kΩ thermistor (model 44008; Omega Engineering, Stanford, CT) was placed on the dura mater over the parietal cortex to measure brain temperature (Tbr). The leads from the EEG and EMG electrodes and the thermistor were routed to a Teflon pedestal. The pedestal and leads were immobilized and attached to the skull with dental acrylic (Duz-All; Coralite Dental Products, Skokie, IL).
Experimental Protocols.
A total of 98 rats was used. In all experiments, rats were acclimatized to a gavage tube (Poppe & Sons, Inc., New Hyde Park, NY) by inserting it once per day for 4 days prior to the control day and injecting 1.5 ml of sterile water intragastrically. On the control day, all rats received 1.5 ml of sterile water intragastrically. All agents were administered in 1.5 ml of water on the day after the control day; each rat received only one dose of drug. In experiment I, each rat received one dose of pregabalin just before dark onset (between 8:30 and 9:00 PM): 3 mg/kg (n = 10), 10 mg/kg (n = 10), 30 mg/kg (n = 10), and 100 mg/kg (n = 10). In experiment II, pregabalin was given just before light onset (between 8:30 and 9:00 PM) at the doses of 3 mg/kg (n = 8), 10 mg/kg (n = 8), 30 mg/kg (n = 8), and 100 mg/kg (n = 8). In experiment III, triazolam was administered just before light onset at doses of 0.5 mg/kg (n = 8), 1.5 mg/kg (n = 8), and 4.5 mg/kg (n = 10). After the administration of the drugs, EEG, EMG, and Tbr were recorded for the next 23 h. Latency to sleep was defined as the time to the first episode of NREMS after the recordings were begun.
Recording and Analyses.
After a recovery period of at least 1 week, rats were moved to sleep recording chambers (Hot Pack, Philadelphia, PA). The rats were allowed relatively unrestricted movement inside the recording cages. A flexible tether connected the electrodes and thermistor leads to an electronic swivel (SL6C; Plastics One, Roanoke, VA). The leads from the swivel were routed to Grass 7D polygraphs in an adjacent room. The EEG was filtered below 0.1 Hz and above 35 Hz. The amplified signals were digitized at a frequency of 128 Hz for the EEG and EMG. Tbr data were saved on a computer in 10-s intervals. Some of the Tbr values were lost because of technical problems, and therefore the sample sizes for Tbr are less than those for sleep data. For scoring purposes the records were displayed on a computer screen; the screen contained the EEG in 10-s and 20-min epochs, and the fast Fourier transformation analysis, brain temperature and EMG in 10-s epochs. The individual scoring the records was unaware of the experimental treatment. The data collection and analysis programs were developed by Dr. J. Fang.
The vigilance states of wakefulness, NREMS and REMS were determined off-line in 10-s epochs using criteria previously reported (Krueger et al., 1993). Briefly, wakefulness is characterized by fast low-amplitude EEG waves, gradually increasing Tbr, and a high incidence of gross body movements. NREMS is associated with slow high-amplitude EEG waves, slowly decreasing Tbr, and lack of body movements. In contrast, REMS is characterized by fast low-amplitude EEG waves, appearance of rhythmic theta EEG, rapidly increasing Tbr at REMS onset, and lack of body movements. The amount of time spent in each vigilance state was calculated every 3 h.
On-line Fourier analysis of the EEG was performed. Three-hour averages of the EEG delta wave activity during NREMS, also called EEG slow wave activity (SWA), were determined as previously described (Kushikata et al., 1999). Moreover, power spectrum analysis during NREMS was performed for the 0.5 to 25 Hz frequency range for the initial 12 h in experiments II and III; 3-h time blocks were used for these analyses. Because of the EEG signal quality, data from two rats in experiment II were excluded from this power spectrum and EEG SWA analyses. This, however, did not affect the ability to score sleep stages. In addition, the number of NREMS and REMS episodes, the mean episode length, and length of sleep cycles (REMS-REMS interval; defined as the time between the onset of a REMS episode lasting 30 s to the next) were determined using a computer program. Twelve and 11-h time blocks were used for these statistical analyses.
Statistical Analysis.
Two-way ANOVA for repeated measures followed by the Student-Newman-Keuls test was used for the analyses of time-spent and episode in each vigilance state, EEG SWA, and Tbr. Paired t tests were used for the comparison of the latency data. For power spectrum analysis data, the actual EEG power density values in 3-h time blocks were summed in four frequency bands: delta (0.5–4.0 Hz), theta (4.5–8.0 Hz), alpha (8.5–12.0 Hz), and beta (12.0–25.0 Hz) wave activities. For the statistical analysis, the average power of each frequency band throughout the initial 12-h control recording period in each rat was normalized to 100%. Then all EEG power data were converted to percent values. Two-way ANOVA for repeated measures was performed for these normalized frequency data. A significance level of P < 0.05 was accepted.
Results
Experiment I: Effect of Pregabalin on Spontaneous Sleep after Dark Onset Administration.
A 3 mg/kg dose of pregabalin given at dark onset did not affect duration of NREMS or REMS. The 30 and 100 mg/kg doses of pregabalin given at dark onset induced increases in NREMS [ANOVA values for the 23-h postinjection period; 30 mg/kg: treatment effect F(1,9) = 9.99, P = 0.0115 with time-treatment interaction F(7,63) = 4.31,P = 0.0006; 100 mg/kg: treatment effectF(1,9) = 74.19, P < 0.0001]. These increases were evident from 4 h to about 12 h after the injection (Fig. 1). If the analyses are confined to the 12-h dark period immediately following the injection, 10 mg/kg, 30 mg/kg, and 100 mg/kg doses of pregabalin increased NREMS significantly [ANOVA, treatment effect; 10 mg/kg:F(1,9) = 5.65, P = 0.0414; 30 mg/kg:F(1,9) = 14.01, P = 0.0046; 100 mg/kg:F(1,9) = 31.38, P = 0.0003] (Table1). The NREMS-promoting effects were mainly due to increases in the duration of NREMS episodes [ANOVA, treatment effect for the 23-h postinjection period; 10 mg/kg:F(1,9) = 10.21, P = 0.0109; 30 mg/kg:F(1,9) = 9.71, P = 0.0176; 100 mg/kg:F(1,9) = 29.99, P = 0.0004]; the number of NREMS episodes decreased significantly after the dose of 100 mg/kg [ANOVA, treatment effect for the 23-h period;F(1,9) = 8.39 P = 0.0176] (Table2). However, pregabalin did not affect the latency of NREMS after any dose (Table 2).
In contrast to its effects on NREMS, pregabalin inhibited REMS in doses of 30 mg/kg and 100 mg/kg [ANOVA, treatment effect for the 23-h postinjection period; 30 mg/kg: F(1,9) = 19.37,P = 0.0017; 100 mg/kg: F(1,9) = 69.26,P < 0.0001]. The effects on REMS occurred during the first 12 h after injection (Table 1). REMS loss resulted from a decreased duration and number of REMS episodes [ANOVA, treatment effect for duration of REMS episodes during the 23-h postinjection period; 30 mg/kg: F(1,9) = 5.97, P = 0.0371; 100 mg/kg: F(1,9) = 13.44, P = 0.0052; ANOVA, treatment effect for the number of REMS episodes during 23 h; 100 mg/kg: F(1,9) = 18.79, P = 0.0019] (Table 3). However, there were no significant differences in REMS-REMS intervals. Pregabalin induced increases in EEG SWA. Significant effects were observed after all doses [ANOVA; 3 mg/kg: treatment effect F(1,9) = 9.98,P = 0.0115; 10 mg/kg: time-treatment interactionF(7,63) = 3.03, P = 0.0082; 30 mg/kg: treatment effect F(1,9) = 24.30, P = 0.0008 with time-treatment interaction F(7,63) = 15.20,P < 0.0001; 100 mg/kg: treatment effectF(1,9) = 69.95, P < 0.0001 with time-treatment interaction F (7,63) = 6.50,P < 0.0001] (Fig. 1).
Experiment II: Effect of Pregabalin on Spontaneous Sleep after Light Onset Administration.
The effects of light onset injections of pregabalin were similar to those observed after dark onset injections. The higher two doses of pregabalin increased NREMS time [ANOVA for the 23-h postinjection period; 30 mg/kg: treatment effectF(1,7) = 14.19, P = 0.0070; 100 mg/kg: treatment effect F(1,7) = 19.88, P = 0.0029 with time-treatment interaction effects F(7,49) = 8.97,P < 0.0001]. These effects were prominent during the initial 12-h light period (Table 1). The NREMS-promoting effects were due to marked increases in the duration of NREMS episodes [ANOVA, treatment effect for the 23-h postinjection period; 10 mg/kg:F(1,7) = 12.50, P = 0.0096, 30 mg/kg:F(1,7) = 67.60, P < 0.0001, 100 mg/kg:F(1,7) = 149.0, P < 0.0001]. This effect was observed after all doses of pregabalin during the initial 12-h light period (Table 2). The number of the NREMS episodes were significantly decreased and this effect was dose-dependent [ANOVA, treatment effect for the 23-h postinjection period; 3 mg/kg:F(1,7) = 17.17, P = 0.0043; 10 mg/kg:F(1,7) = 20.53, P = 0.0027; 30 mg/kg:F(1,7) = 44.36, P = 0.0003; 100 mg/kg:F(1,7) = 70.70, P < 0.0001] (Table 2). The effects on NREMS episodes were stronger than those in experiment I because significant effects were observed even after the lower doses. The latency to NREMS was not affected by pregabalin (Table 2). Pregabalin inhibited REMS duration [ANOVA for 23 h; 30 mg/kg: treatment effect F(1,7) = 14.24, P = 0.0070 with time-treatment interaction F(7,49) = 5.33,P = 0.0001; 100 mg/kg: treatment effectF(1,7) = 27.93, P = 0.0011 with time-treatment interaction F(7,49) = 8.57, P< 0.0001] (Table 1).
The inhibitory effect on REMS was due to a reduction in the number and duration of episodes [ANOVA, treatment effect for duration of REMS episodes during the 23-h postinjection period; 3 mg/kg:F(1,7) = 7.12, P = 0.0320; 30 mg/kg:F(1,7) = 6.75, P = 0.0355; 100 mg/kg:F(1,7) = 28.78, P = 0.0010. ANOVA, treatment effect for number of REMS episodes during the 23-h postinjection period, 100 mg/kg: F(1,7) = 6.09, P = 0.0431] (Table 3). Pregabalin increased EEG SWA [ANOVA for the 23-h postinjection period; 3 mg/kg: treatment effect F(1,7) = 27.96, P = 0.0011; 10 mg/kg: treatment effectF(1,6) = 7.74, P = 0.0319 with time-treatment interaction F(7,42) = 4.80, P= 0.0005; 30 mg/kg: treatment effect F(1,7) = 52.33 with time-treatment interaction F(7,49) = 9.41, P< 0.0001; 100 mg/kg: treatment effect F(1,6) = 27.10,P = 0.0020 with time-treatment interactionF(7,42) = 15.50, P < 0.0001] (Fig.2). Power spectrum analysis revealed significant pregabalin-induced increases in relative EEG delta power and decreases in relative beta power. The maximum effect on EEG delta power was observed 4 to 6 h after the administration. The decreasing effect on EEG beta power persisted throughout the initial 12 h. (Fig. 3, Table4).
Pregabalin did not affect Tbr after any of the doses (data not shown), and the normal changes in Tbr associated with state changes persisted in pregabalin-treated rats. Pregabalin also did not induce gross abnormal behavior in the sense that animals appeared to behave normally when handled at the end of the experiments.
Experiment III: Effects of Triazolam on Spontaneous Sleep after Light Onset Administration.
Triazolam significantly increased NREMS after the doses of 1.5 and 4.5 mg/kg [ANOVA, treatment effect for the 23-h postinjection period; 1.5 mg/kg: F(1,7) = 8.85,P = 0.0207; 4.5 mg/kg: F(1,9) = 12.28,P = 0.0067]. However, triazolam had no effect on REMS or the number or duration of REMS episodes (Tables5 and 6). The increase in NREMS time was due to significant increases in the duration of NREMS episodes [ANOVA, treatment effect for the 23-h postinjection period; 1.5 mg/kg: F(1,7) = 6.58, P = 0.0373; 4.5 mg/kg: F(1,9) = 53.7, P < 0.0001] (Table7). Although we did not observe clear dose-response relationships, 1.5 mg of triazolam significantly reduced latency to NREMS (paired t test: P = 0.0233) (Table 7). EEG SWA decreased after the 1.5 and 4.5 mg/kg triazolam doses [ANOVA for the 23-h postinjection period; 1.5 mg/kg: time-treatment interaction, F(7,49) = 2.55,P = 0.0253; 4.5 mg/kg: treatment effect,F(1,9) = 17.55, P = 0.0023] (Fig.4). Triazolam inhibited EEG power densities in all bandwidths and especially in the theta and alpha power frequencies (Fig. 5, Table 4). These effects were most prominent during the first 3 h after triazolam treatment (Fig. 5). Triazolam did not affect Tbr after any of the doses.
Discussion
The purposes of the present study were to characterize the actions of pregabalin on sleep and to compare the effects of pregabalin and triazolam on the sleep-awake cycle and EEG power spectrum of rats. These two drugs had some similarities in their somnogenic actions but distinct differences were also apparent. The major similarity is that both pregabalin and triazolam increased NREMS. Both substances did this by increasing the duration of NREMS episodes, suggesting effects on sleep consolidation. Pregabalin had a relatively greater effect than triazolam on the time spent in NREMS and duration of NREMS episodes. Some of these differences may be due to pharmacokinetic differences between the two compounds. Neither pregabalin nor triazolam consistently altered the latency to sleep onset. This may in part be due to the relatively short latency to sleep onset of 20 to 30 min observed in rats under normal conditions.
The major differences between pregabalin and triazolam were their effects on EEG SWA and EEG power spectrum. Pregabalin enhanced, while triazolam inhibited, EEG low frequency (0.5–4 Hz) power. These actions of pregabalin resemble the effects that sleep deprivation has on subsequent sleep (Pappenheimer et al., 1975; Borbély et al., 1984). The actions of pregabalin on NREMS episode length, sleep cycle length, and inhibition of REMS are consistent with this notion. However, although EEG delta-wave amplitudes are thought to reflect the intensity of NREMS in physiological sleep, some studies have indicated that the mechanism responsible for NREMS and EEG SWA are different. For example, removal of basal forebrain cholinergic neurons has little effect on duration of NREMS but decreases EEG power (Kapás et al., 1996). Similarly, electrolytic lesions of the preoptic area reduce NREMS and EEG SWA; the reduction in NREMS is transitory while the reduction in SWA persists (Shoham et al., 1989). Regardless, current results clearly indicate that the effects of pregabalin on the EEG more closely resemble those occurring after sleep deprivation than do the effects of triazolam on the EEG.
In addition to the effects on the sleep-awake cycle and EEG reported in this paper, pregabalin produces anticonvulsant, anxiolytic, and antihyperalgesic effects in animal models (Bialer et al., 1999; Bryans and Wustrow, 1999; Kinsora et al., 1999; Field et al., 2001). The mechanism(s) by which pregabalin produces these different pharmacological effects is not known with certainty but is the topic of extensive research. Although pregabalin is a GABA analog, it does not interact with GABAA and GABAB receptors directly (Suman-Chauhan et al., 1993). Pregabalin activates the GABA synthetic enzyme, glutamic acid decarboxylase in vitro, but the concentrations needed for this effect are too high to be therapeutically relevant (Taylor et al., 1992). Pregabalin binds with high affinity to the alpha-2-delta subunit of voltage-gated calcium channels (Gee et al., 1996). This action may be related to reported effects on decreased neurotransmitter release (Schlicker et al., 1985; Dooley et al., 2000) but these effects usually require sustained depolarization for inhibition to be observed. These alpha-2-delta binding sites are thought to be responsible for the observed distribution in the central nervous system of [3H]gabapentin (Taylor et al., 1993), which are widely distributed in the brain with high levels in the cerebral cortex and hippocampus (Hill et al., 1993) where they can potentially influence EEG activity. Gabapentin elevates, through an unknown mechanism, concentrations of GABA in the occipital lobe of epilepsy patients (Petroff et al., 1996). It is unknown, however, if pregabalin induces a similar effect.
Even if direct or indirect modulation of the GABAergic system mediates some of the effects of pregabalin, there is not a consistent pattern of effects of GABAergic agents on the EEG. Both GABAA and GABAB receptors have been reported to mediate inhibitory mechanisms involved in the generation of EEG synchronization (Juhasz et al., 1994; Lancel et al., 1996). However, there are differences in the EEG modulatory effects of direct acting GABAA agonists, e.g., muscimol and 4,5,6,7-tetrahydroisoxazolo(5,4-c)pyridin-3-ol (THIP), and the effects of the benzodiazepines, which act as indirect modulators of GABAA activity. For example, although both direct acting GABAA agonists and benzodiazepines, increase NREMS, the former increase EEG slow wave activity, whereas the latter decrease EEG slow wave activity (Lancel et al., 1996; Faulhaber et al., 1997; Lancel and Faulhaber, 1996; Lancel, 1997; ).
The reasons why benzodiazepines inhibit EEG SWA remain unclear. Interestingly, the reduction of EEG SWA by benzodiazepines may not be mediated by the GABAA-benzodiazepine receptor complex (Borbély et al., 1991); e.g., flumazenil, a benzodiazepine receptor antagonist, does not inhibit benzodiazepine-induced suppression of EEG SWA, while other sleep parameters are antagonized (Gaillard and Blois, 1989). Another characteristic of benzodiazepines on EEG spectra is an increase in sigma activity (11–16Hz) (Lancel, 1999). Triazolam also markedly increases sigma activity in humans (Johnson et al., 1983; Aeschbach et al., 1994; Tan et al., 1998). We observed a nonsignificant increase in EEG fast wave (>20 Hz) during the initial 3 h after the high dose of triazolam. This was not unexpected since it was previously reported that intraperitoneal injection of triazolam in rats enhanced EEG power in the higher frequencies only during the initial 1 or 2 h after treatment (Edgar et al., 1991). It is also possible that oral administration of triazolam is less effective than intravenous or intraperitoneal administration because of its absorption or extensive liver first-pass metabolism; the sleep effects of triazolam after oral administration are different from those observed after intravenous injection in rabbits (Scherschlicht and Marias, 1983).
Another finding in our investigation is that high doses of pregabalin inhibit REMS as a result of a decrease in the number and duration of REMS episodes. One of the mechanisms of REMS inhibition is thought to be due, in part, to a drug-induced inhibition of EEG desynchronization rather than to the disruption of the REMS generating process (Borbély et al., 1991). Since pregabalin enhances synchronization of EEG slow waves, inhibition of EEG desynchronization may also be a possible mechanism of pregabalin REMS inhibition. Benzodiazepines also inhibit REMS and the effects of triazolam on REMS depend on the experimental conditions. For instance, in humans triazolam inhibits REMS (Pakes et al., 1981; Aeschbach et al., 1994; Tan et al., 1998). In rabbits, the effect depends on the route of administration.Scherschlicht and Marias (1983) reported that REMS increased significantly after oral administration of triazolam, while there was no significant effect after intravenous administration during a 6-h recording period. In rats, triazolam inhibits REMS (Edgar et al., 1991;Gandolfo et al., 1994) or has no effect on REMS (Mendelson and Monti, 1993; Mendelson, 1998).
Some of the differences between the effects of direct acting GABAA agonists and benzodiazepines, may be due to the different degree and location of receptor activation. Directly acting GABAA agonists may stimulate all GABAA receptors, whereas the activation induced by benzodiazepines will depend upon receptor subtype distribution and level of GABA release. In conclusion, we have demonstrated a novel sleep modulating effect of pregabalin. The effects to consolidate sleep and increase EEG slow wave activity suggest the possibility that pregabalin may act to induce more restorative sleep.
Acknowledgments
We thank Richard A. Brown for excellent technical assistance and Dr. J. L. Werth for comments on the manuscript.
Footnotes
-
This study was supported by Warner-Lambert Co., Ann Arbor, MI.
- Abbreviations:
- GABA
- γ-aminobutyric acid
- REMS
- rapid eye movement sleep
- NREMS
- non-REMS
- EEG
- electroencephalographic
- EMG
- electromyogram
- Tbr
- brain temperature
- SWA
- slow wave activity
- ANOVA
- analysis of variance
- Received March 20, 2001.
- Accepted August 21, 2001.
- The American Society for Pharmacology and Experimental Therapeutics