Previous studies suggest that nitric oxide (NO) may play a role in sleep regulation, particularly in the homeostatic process. The present studies were undertaken to compare the sleep effects of injecting a NO synthase (NOS) inhibitor when homeostatic sleep pressure is naturally highest (light onset) or when it is at its nadir (dark onset) in rats. Sleep, electroencephalogram delta-wave activity during nonrapid eye movement sleep (NREMS), also known as slow-wave activity (SWA), and brain temperature responses to three doses of the NOS inhibitor Nω-nitro-l-arginine methyl ester (l-NAME; 5, 50, and 100 mg/kg) injected intraperitoneally at light or dark onset were examined in rats (n = 6 to 8). The effects of 5 mg/kg l-NAME were determined in both normal and vagotomized (VX) rats. Light onset administration of 50 mg/kg l-NAME decreased NREMS amounts and suppressed SWA and increased rapid eye movement sleep (REMS) amounts. At dark onset, l-NAME injection also dose dependently suppressed SWA; however, unlike light onset injections, both NREMS and REMS amounts were increased after all three doses. Sleep responses to 5 mg/kg l-NAME were not different in control and VX rats, suggesting that the sleep effects of l-NAME are not mediated through the activation of sensory vagal mechanisms. The present findings suggest that timing of the injection is a major determinant of the sleep responses observed after systemic l-NAME injection in rats.
- Nω-nitro-l-arginine methyl ester
- slow-wave activity
- homeostatic regulation
- suprachiasmatic nucleus
- electroencephalogram power spectrum
sleep regulation relies on the interplay between the homeostatic and circadian processes (5). The homeostatic component manifests itself as an increasing sleep pressure due to prior waking periods, whereas the circadian component is an oscillating process that permits sleep to occur only during low sleep threshold periods. Normally, increased homeostatic pressure coincides with periods of decreased sleep threshold, thus allowing sleep to occur. In nocturnal animals, such as rats, this occurs at the beginning of the light period. A growing number of studies suggest that nitric oxide (NO)-ergic mechanisms play a role in sleep regulation, particularly in the homeostatic component.
The two-process model of sleep regulation postulates that the homeostatic process relies on a neuronal mechanism, the activity of which increases during wakefulness and is dissipated during sleep (5). There is evidence that the activity of NO-generating mechanisms in the brain stem, thalamus, and cortex fits this pattern. For example, NO is released in an activity-dependent manner from neurons in brain stem areas involved in sleep regulation (29). In rats, cortical (6) and thalamic (49) NO levels are the highest during wakefulness, and brain NO synthase (NOS) activity exhibits diurnal variation peaking during the active (dark) period (3). Also, administration of NO donor molecules or the NO precursor l-arginine during the dark phase of the cycle mimics the effects of prolonged wakefulness, i.e., dose dependently increases nonrapid eye movement sleep (NREMS) in rats (26). On the other hand, NOS inhibitors such as Nω-nitro-l-arginine methyl ester (l-NAME) (24, 35, 36, 40) and 7-nitro indazole (7-NI) (7, 17) suppress sleep when administered at light onset in rats.
The most direct evidence, to date, implicating NO-ergic neurotransmission in homeostatic regulation of sleep is that rebound increases in sleep after sleep deprivation (SD) are shorter in duration and lower in amplitude in NOS inhibitor-treated rats than sleep rebounds in control animals (40). Another approach to study the role of NO in the homeostatic components of sleep regulation is to test the effects of a NOS inhibitor when 1) the activity of the homeostatic mechanism is spontaneously at its highest and 2) when the activity of homeostatic mechanisms is the lowest. In rats, homeostatic sleep pressure is the highest at the end of the active period and lowest at the end of the rest period. The aim of the present study was to determine the effects of l-NAME on sleep after light and dark onset administration in rats. The results show that systemic injection of l-NAME suppresses NREMS after light onset injection but promotes NREMS and rapid eye movement sleep (REMS) when injected at dark onset. Our findings are in line with the hypothesis that NO-ergic mechanisms are involved in homeostatic sleep-promoting mechanisms and also suggest that, in addition, they may play a role in arousal mechanisms.
With the use of combined ketamine (87 mg/kg) and xylazine (13 mg/kg) anesthesia, male Sprague-Dawley rats (250–350 g) were implanted with cortical electroencephalographic (EEG) electrodes, nuchal electromyographic (EMG) electrodes, and a calibrated brain thermistor. The EEG electrodes were anchored over the frontal and parietal cortices and the thermistor was placed on the dura over the parietal cortex. After the surgery, the animals were kept in sound-attenuated individual sleep-recording cages for habituation to the experimental conditions. After a 1-wk recovery period, the animals were connected to the recording cables and injected daily for 7 days with isotonic NaCl solution intraperitoneally 5–15 min before light (for experiment I) or dark onset (for experiment II). Animals were kept on a 12:12-h cycle (light onset at 0400) and an ambient temperature of 24 ± 1°C, for at least 2 wk before surgeries, during the recovery and habituation periods and throughout the experimental procedure. Water and food were always available ad libitum. The experiments are consistent with the Guiding Principles for Research Involving Animals and Human Beings issued by the American Physiological Society. The experimental protocol was approved by the Institutional Animal Care and Use Committee of Fordham University.
Experiment I: effect of light onset injection of l-NAME on sleep, slow-wave activity, and brain temperature.
On the control day, animals received 2 ml/kg isotonic NaCl ip and sleep was recorded for 23 h starting at 0400. The next day, the animals received 5 mg/kg (n = 7), 50 mg/kg (n = 6), or 100 mg/kg (n = 6) l-NAME ip (Sigma, St. Louis, MO) dissolved in isotonic NaCl (2 ml/kg). l-NAME is an irreversible inhibitor of all three isoforms of NOS; in rats, single bolus injection of l-NAME suppresses brain NOS activities by 50% for at least 14 h (14). The injections were performed 5–15 min before light onset. Sleep on an additional, recovery, day was also recorded for the 50- and 100-mg/kg doses; the animals received saline intraperitoneally at light onset.
Experiment II: effect of dark onset injection of l-NAME on sleep, slow-wave activity, and brain temperature.
The experimental design was similar to that of experiment I with the exception that all injections were performed 5–15 min before dark onset (1600). Three doses of l-NAME were tested, 5 mg/kg, 50 mg/kg (n = 6), and 100 mg/kg (n = 6). For 50 mg/kg l-NAME, a recovery day was also recorded, when the animals received isotonic NaCl before dark onset. The 5-mg/kg dose was tested on two groups of rats. One group (n = 7) was subjected to bilateral subdiaphragmal vagotomy (VX), and the other (n = 8) consisted of sham-operated rats. Three to 4 wk after the surgeries, the animals were implanted with EEG, EMG, and brain temperature (Tbr) and habituated to the recording conditions. After the experiments, the completeness of the vagotomies was confirmed using the 2-deoxy-d-glucose/neutral red test (9).
EEG, EMG, and Tbr were recorded by computer. EMG activity served the sole purpose of aiding in determining the vigilance states of the animals and was not further quantified. The EEG was filtered below 0.1 Hz and above 40 Hz (Biopac Systems, Branco, CA). The amplified signals were digitized at the frequency of 128 Hz for EEG and EMG and 2 Hz for Tbr. Single Tbr samples were saved on the hard disc in 10-s intervals. Average Tbr was calculated in 3-h time blocks. The vigilance states were determined off-line in 10-s epochs. Time spent in NREMS and REMS was calculated in 3-h time blocks. Also, the number of NREMS and REMS episodes was recorded, and the average episode durations were computed. Online fast Fourier transformation (FFT) analysis of the EEG was also performed in 10-s intervals on 2-s segments of the EEG in 0.5-Hz bands of the 0.5- to 20-Hz frequency range. The EEG power density values were summed in four frequency bands for each 10-s epoch. The spectral data were paired with the vigilance states, and EEG power was computed in each of the four bands separately for each vigilance state. Hourly average δ (0.5–4 Hz)-, θ (4.5–8 Hz)-, α (8.5–12 Hz)-, and β-wave (12.5–20 Hz) activities (μV2) were then calculated for NREMS, REMS, and wakefulness epochs. On the baseline day, average power densities were computed across the entire 23 h for each rat to obtain a reference value for each animal. Power densities in 3-h blocks on the baseline day and the test days were then expressed as a percentage of that reference value.
Statistical comparisons were made between the baseline days and the test days and also between baseline and recovery days. ANOVA was done separately for the dark and light periods. Two-way ANOVA for repeated measures was performed on data averaged in 3-h time blocks for sleep amounts and episode numbers and on 1-h time blocks for Tbr. Samples sizes for sleep episode duration, slow-wave activity (SWA), and θ-wave activity during REMS were not perfectly matched between the baseline and test days. The reason for this is that in some animals there were a few instances when REMS and/or NREMS were absent for an entire 3-h period. For these time blocks, EEG analysis could not be performed, resulting in missing data points that prevented us from performing statistical tests for repeated measures. Therefore, two-way ANOVA for nonrepeated measures was performed on data averaged in 3-h time blocks for sleep episode durations, SWA, and θ-wave activity during REMS. When ANOVA indicated significant treatment effects, Student-Newman-Keuls test was performed a posteriori. ANOVA was also performed on 12-h averages across the light and dark periods between baseline and experimental days for NREMS, REMS, and SWA. Two-way ANOVA was performed to compare the NREMS, REMS, and SWA effects of l-NAME in VX and control rats. We report only the results of the treatment effects from the ANOVAs in Tables 1, 2, and 3 and in the text.
Light Onset Administration of l-NAME
Confirming our previous findings (24, 40), systemic injection of l-NAME at light onset suppressed NREMS in rats (Tables 1 and 2). The lowest dose tested, 5 mg/kg l-NAME, did not elicit any changes in NREMS. REMS amounts, however, were increased in the dark period, 13–23 h after l-NAME injection (Fig. 1); these increases were not accompanied by significant changes in the number or average duration of REMS epochs (Fig. 2). Injection of 50 mg/kg l-NAME decreased NREMS amounts by 52.1 ± 8.6 min in the 12-h period after the injection (Fig. 3). REMS amount was unchanged during the light period, but the average duration of REMS episodes decreased (Figs. 1 and 2). In the subsequent dark period, REMS episode durations significantly increased and there was a strong tendency toward increased number of REMS episodes; these changes together resulted in elevated REMS amounts, this increase, however, did not reach the level of statistical significance (Fig. 1). In the 12-h period after 100-mg/kg l-NAME treatment, NREMS amounts were slightly below baseline levels, although not significantly. REMS amounts did not deviate from baseline during the light period; however, in 12–23 h, REMS was increased by 31.8 ± 9.8 min, ∼122% above baseline (Fig. 3).
Delta-wave activity of the EEG during NREMS, which is also called SWA and is often regarded as a measure of NREMS intensity (5), showed dose-dependent decreases in response to l-NAME. Although there was no change in SWA after the injection of 5 mg/kg l-NAME, both 50 and 100 mg/kg l-NAME elicited decreases in SWA that lasted across the test and recovery days (Fig. 1). Across the 2 days, SWA was suppressed by 16 ± 4.4 and 23.9 ± 3.4% by 50 and 100 mg/kg l-NAME, respectively. EEG θ-wave activity during REMS was suppressed by all three doses of l-NAME; these effects were still evident during the recovery day of the 100-mg/kg dose (Fig. 1). In addition, FFT analysis of all four wave bands across all three vigilance states after 50 mg/kg l-NAME revealed significant increases in α-activity during REMS and wakefulness and increases in the β-activity during wakefulness (Fig. 4). Light onset administration of l-NAME did not have significant effect on Tbr (Fig. 1).
Dark Onset Administration of l-NAME
Dark onset administration of l-NAME induced dose-dependent increases in both NREMS and REMS (Tables 2 and 3 and Fig. 5). In response to 5 mg/kg l-NAME, NREMS and REMS were increased by 22.1 ± 3.6 and 23.7 ± 3.5 min, respectively, during the first 12-h period in sham-operated rats (Fig. 3). The increase in REMS was due to the increased numbers and average durations of REMS episodes (Fig. 6). Vagotomy itself did not have significant effects on baseline sleep amounts and SWA [control vs. VX, 2-way ANOVA, treatment effects; NREMS: F(1,26) = 0.09, not significant (n.s.), REMS: F(1,26) = 0.01, n.s., SWA: F(1,16) = 0.01, n.s.]. Furthermore, NREMS and REMS responses to 5 mg/kg l-NAME in sham-operated rats were not statistically different from those in VX rats (Tukey’s test; REMS: P = 0.642, n.s.; NREMS: P = 0.531, n.s.).
After the injection of 50 mg/kg l-NAME, NREMS amounts were increased by 56.5 ± 9.1 min in 1–12 h, and then returned to baseline for the rest of the recordings (Figs. 3 and 5). Changes in REMS showed a biphasic pattern. In the dark phase after the treatment, REMS increased by ∼250% compared with baseline. In the subsequent light period, REMS returned to baseline levels, but during the next dark period, 25–36 h, a second significant increase in REMS of ∼90% took place. The first phase of REMS increase was due to increased numbers of REMS episodes, and the second phase was due to a combination of increased episode numbers and durations (Fig. 6). The highest dose of l-NAME, 100 mg/kg, also increased NREMS and REMS amounts in the 12-h dark period by 60.9 ± 12.2 and 43.0 ± 16.3 min, respectively (Fig. 3). Both the number and average duration of REMS episodes increased significantly in the dark phase, whereas the number and average durations of NREMS episodes did not change from baseline (Fig. 6).
SWA was dose dependently suppressed after dark onset injection of l-NAME (Figs. 3 and 5). After 5 mg/kg l-NAME, SWA was 10.9 ± 1.9% below baseline levels in 12–23 h (Fig. 3). Fifty milligrams per killigram of l-NAME suppressed SWA by 22.5 ± 1.9% on the test day and by 15.0 ± 5.2% on the recovery day (Fig. 5). After 100 mg/kg l-NAME, SWA was reduced by 26.6 ± 1.8% across the 23-h recording period. Theta-wave activity of the EEG was suppressed during the light phase after each l-NAME treatment (Fig. 5). In addition, FFT analysis of all four wave bands across all three vigilance states after 50 mg/kg l-NAME revealed significant decreases in θ-activity during NREMS and increases in α- and β-activities in REMS and wakefulness (Fig. 4). Dark onset injection of 50 mg/kg l-NAME elicited slight but statistically significant decreases in Tbr in the first 12 h after the treatment (Fig. 5).
Confirming previous findings, both NREMS amounts and SWA were decreased in response to systemic l-NAME treatment during the rest period in rats (24, 36, 40). Consistent also with our prior observations, the SWA responses to l-NAME were long lasting (40); following 50 and 100 mg/kg l-NAME, SWA remained below baseline for at least 36–48 h. Administration of l-NAME at dark onset dose dependently suppressed SWA; however, contrary to the light onset experiments, both NREMS and REMS amounts were increased.
During the light period, the activity of the homeostatic sleep-promoting mechanisms is the highest in rats. Our finding that l-NAME decreased NREMS only when injected at light onset is consistent with the hypothesis that NO-ergic mechanisms play a role in sleep regulation by modulating the homeostatic process. Further support for this hypothesis comes from studies showing that systemic injection of l-NAME suppresses both rebound sleep amounts and intensity seen after SD (40) and cortical (6) and thalamic (49) release of NO are higher during wakefulness than in NREMS. Moreover, NOS activity is highest during behaviorally active periods (3), adding support to the involvement of NO-ergic mechanisms in homeostatic sleep regulation.
Sleep increases in response to dark onset injection of l-NAME cannot be explained by the putative role of NO in homeostatic sleep-promoting mechanisms. NO has also been shown to be a crucial signaling molecule in the suprachiasmatic nucleus (SCN). Glutamate elevates NO levels in the SCN, which, in turn, activate soluble guanylyl cyclase, increasing cyclic guanosine monophosphate levels. Intra-SCN microinjection or in vitro application of glutamate (12, 33) or NO donors (12) to SCN slices elicits phase shifts similar to those produced by in vivo light pulses. In contrast, cerebral injections of chemicals that block this signaling cascade prevent light-induced phase shifts (2, 12, 48). Phase resetting by the light/glutamate/NO pathway is only active during the dark period (19), which is the phase when l-NAME strongly enhances NREMS and REMS. The actions of NO in the SCN suggest that NO-ergic mechanisms may also be part of the circadian process of sleep regulation. We hypothesize that during the dark period, when homeostatic sleep pressure is low, sleep responses to l-NAME are due to its effects on the light/glutamate/NO pathway in the SCN. In fact, preliminary studies have shown that microinjection of NOS inhibitors into the medial preoptic region (41) or anterior diencephalic area (31) increased sleep, particularly REMS, whereas bilateral microinjections of NO donor into the SCN suppressed REMS (42).
The overall action of l-NAME on sleep is likely the net of the effects on the homeostatic and circadian processes. Our results suggest that the activation of those NO-ergic mechanisms involved in the homeostatic and in the circadian component of sleep regulation elicits opposite effects. We hypothesize that during the light period, l-NAME interferes with the sleep-promoting homeostatic effects of NO-ergic mechanisms. In vitro studies indicate that the SCN during this time is not sensitive to NO (19). During the dark period, however, when homeostatic mechanisms are the least active but the SCN is sensitive to l-NAME, the target for l-NAME is likely the SCN, and increased sleep is due to the actions of l-NAME on the circadian process. The site of sleep-suppressing action of l-NAME is not known, although various brain stem structures may possibly play a role. Evidence for the brain stem site of action of NO arose from microinjection experiments where pontine administration of NOS inhibitors decreased NREMS and REMS amounts in rats (21, 25) and in cats (10), and injection of NG-nitro-l-arginine into the medial pontine reticular formation decreased REMS in cats (30). Furthermore, injection of NO donors (10, 21) or the NO precursor l-arginine (21) into the pedunculopontine tegmental nucleus elicited increases in NREMS and REMS. In addition, injection of NOS inhibitors such as l-NAME (35, 36) or 7-NI (7, 36) into the dorsal raphe nucleus resulted in decreased amounts of NREMS or REMS. l-NAME facilitated REMS when it was given before dark or light onset. Dark onset injection of l-NAME enhanced REMS immediately, whereas light onset administration increased REMS after a 12-h delay, i.e., during the next dark phase. Systemic injection of l-NAME likely has an overall REMS-enhancing effect. It is possible, however, that this effect is not manifested during the light periods because REMS is already high.
Irrespective of the timing of the treatment, SWA was always suppressed in response to l-NAME. The changes in EEG power spectrum, however, were not restricted to the δ-wave activity during NREMS; θ-activities during REMS were also suppressed during the light periods after each treatment. Theta-activities during REMS show diurnal variation, and it is likely that θ-suppression was confined to the light phases because during the dark, it was already at its daily nadir. Similar suppressions in θ-activities, restricted mainly to the light phase, were reported in rats in response to systemic injections of 7-NI, an inhibitor of neuronal NOS (15). We hypothesize that this suppression in EEG power may reflect a generalized inhibition of EEG δ- and θ-wave-generating mechanisms, which is, at least in part, independent of sleep. Confirming our previous findings (40), SWA and REMS θ-suppressions lasted for at least 48 h, and some of the sleep effects were also evident as late as 36–40 h after the injection. These long-lasting effects are likely due to the fact that l-NAME is an irreversible inhibitor of NOS (14); for example, systemic bolus injection of l-NAME suppresses NOS activity in the forebrain for at least 24 h (22). We also showed that brain NOS activity is decreased in response to an intraperitoneal bolus injection of 50 mg/kg l-NAME in the hypothalamus, brain stem, and cerebellum by 30–75% in rats (4).
l-NAME elicits vasoconstriction and, as a result, systemic l-NAME treatment results in increased blood pressure. Blood pressure changes may also trigger sleep responses through vagal afferents (38). We studied the effects of 5 mg/kg l-NAME in vagotomized rats to determine whether an intact vagus is required for the sleep effects of l-NAME. The lowest dose was chosen for the VX experiment because, on the one hand, this dose still exhibits significant sleep-promoting actions and, on the other hand, within the dose range we tested, brain NOS activities are the least affected by this dose (22), while it still elicits vasoconstriction and elevated blood pressure (39). Our results that VX did not prevent the REMS-promoting and SWA-suppressing effects of l-NAME suggest that these effects are not likely to be triggered by vagal sensory pathways. l-NAME also elicits vasoconstriction in cerebral blood vessels (43), and both reduced (47) and unchanged (18) cerebral blood flow (CBF) have been reported after systemic injection of NOS inhibitors. It is unlikely, however, that reduced CBF would be responsible for the sleep effects, especially for both wake- and sleep-promoting effects, of l-NAME administration, as reduced CBF per se does not result in changes in sleep or EEG activity. For example, in rats, systemic administration of 10 mg/kg indomethacin reduces CBF by 30–50% (32) but does not affect sleep (37). Similarly, a decrease in CBF induced by indomethacin or hyperventilation does not diminish EEG SWA in humans (28).
NO has been implicated in thermoregulatory mechanisms. Inhibitors of NOS suppress febrile responses to lipopolysaccharide (13, 45) and interleukin 1 (34, 44) (but see also Refs. 1 and 20) and inhibit stress-induced hyperthermia (11) and cold-induced thermogenesis (23). In the present study, a peripheral injection of 50 mg/kg l-NAME at dark onset elicited a 0.2–0.3°C drop in Tbr. This confirms our (24, 40) and others' (11, 34, 46) observation that systemic injection of l-NAME in rats causes minimal decreases in body temperature and is in line with the notion that NO-producing mechanisms may contribute to thermogenesis.
Our present study has its limitations. First, l-NAME is a nonisoform-specific inhibitor of NOS; therefore, the relative contribution of inhibiting neuronal, endothelial, and inducible NOS to the observed effects on sleep and EEG cannot be determined. Both inducible and neuronal NOS have been implicated in sleep regulation and their involvement in REMS-generating mechanisms appears to be different (8). Second, systemic injection of l-NAME suppresses the activity of both peripheral and brain NOS activities. Finally, systemic injection of l-NAME likely inhibits NOS activities in all brain regions that express NOS, therefore the relative importance of specific regions and neuronal circuits in the effects of l-NAME is not clear.
The present studies reveal that the effect of l-NAME on sleep depends on the time of the administration. There has been some controversy in the literature regarding the effects of l-NAME on sleep. Although the majority of the experiments reported thus far have shown that systemic administration of NOS inhibitors such as l-NAME (24, 27, 35, 36, 40), 7-NI (7, 17), or 3 bromo-7-NI (15) suppress sleep, in contrasting reports, increased NREMS and/or REMS amounts were found in response to systemic injection of the NOS inhibitors l-NAME (7) and l-NMMA (16). In the latter studies, the NOS inhibitors were administered at dark onset or during the dark period. These studies further support the present observation that the effects of l-NAME on sleep depend on the timing of the injections, presumably by modulating predominantly the circadian or homeostatic components of sleep regulation.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society