SEARCH

SEARCH BY CITATION

Keywords:

  • adenosine;
  • basal forebrain;
  • nitric oxide;
  • recovery sleep;
  • sleep deprivation

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Sleep homeostasis is the process by which recovery sleep is generated by prolonged wakefulness. The molecular mechanisms underlying this important phenomenon are poorly understood. Here, we assessed the role of the intercellular gaseous signaling agent NO in sleep homeostasis. We measured the concentration of nitrite and nitrate, indicative of NO production, in the basal forebrain (BF) of rats during sleep deprivation (SD), and found the level increased by 100 ± 51%. To test whether an increase in NO production might play a causal role in recovery sleep, we administered compounds into the BF that increase or decrease concentrations of NO. Infusion of either a NO scavenger, 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide, or a NO synthase inhibitor, Nω-nitro-l-arginine methyl ester (L-NAME), completely abolished non-rapid eye movement (NREM) recovery sleep. Infusion of a NO donor, (Z)-1-[N-(2-aminoethyl)-N-(2-ammonioethyl)amino]diazen-1-ium-1,2diolate (DETA/NO), produced an increase in NREM that closely resembled NREM recovery after prolonged wakefulness. The effects of inhibition of NO synthesis and the pharmacological induction of sleep were effective only in the BF area. Indicators of energy metabolism, adenosine, lactate and pyruvate increased during prolonged wakefulness and DETA/NO infusion, whereas L-NAME infusion during SD prevented the increases. We conclude that an increase in NO production in the BF is a causal event in the induction of recovery sleep.

Abbreviations used
ac

anterior commissure

aCSF

artificial cerebrospinal fluid

BF

basal forebrain

cPTIO

2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide

DETA/NO

(Z)-1-[N-(2-aminoethyl)-N-(2-ammonioethyl)amino]diazen-1-ium-1,2-diolate

D-NAME

Nω-nitro-d-arginine methyl ester

EEG

electroencephalogram

EMG

electromyogram

HDB

horizontal limb of the diagonal band of Broca

LDT/PPT

laterodorsal/pedunculopontine tegmental nuclei

L-NAME

Nω-nitro-l-arginine methyl ester

MCPO

magnocellular preoptic area

nBF

outside the BF

NOS

nitric oxide synthase

NOx

NO2 + NO3

NREM

non-rapid eye movement

ox

optic chiasm

REM

rapid eye movement

SD

sleep deprivation

Sleep loss, induced by prolonged wakefulness, produces a decline in cognitive and motor performance (Dinges et al. 1997), mood disturbances, memory deficits (Chee and Choo 2004) and affects immune function (Bryant et al. 2004). These effects are restored by recovery sleep, which is characterized by prolongation and intensification of both the non-rapid eye movement (NREM) and rapid eye movement (REM) components of sleep. Although the two-process model of sleep regulation (Borbely 1982) accurately describes the expected duration of recovery sleep, the molecular mechanisms that underlie this regulation remain less clear. Endogenous sleep factors – substances that accumulate in the brain during prolonged wakefulness – have been suggested to be mediators of homeostatic sleep regulation (Borbely and Tobler 1989). One potential sleep factor is the inhibitory neuromodulator adenosine (Benington and Heller 1995); during prolonged wakefulness extracellular adenosine concentration increases in the basal forebrain (BF) and induces sleep (Porkka-Heiskanen et al. 1997). As adenosine is an indicator of disturbed energy balance (Dunwiddie and Masino 2001), we hypothesized that during sleep deprivation (SD) continuous activity of the waking-promoting cells in the cholinergic region of the BF (Detari et al. 1984; Szymusiak and McGinty 1986; Szymusiak and McGinty 1989) leads to unfavourable changes in energy demand/supply ratio and consequent adenosine release in this area. Supporting our hypothesis, experimentally induced local energy depletion in the BF increased extracellular adenosine concentration and concurrently induced an increase in sleep (Kalinchuk et al. 2003). Recently, it has been shown that NO can inhibit neuronal energy production (Brorson and Zhang 1999; Maletic et al. 2004; Rosenberg et al. 2000) and stimulate adenosine release from forebrain neurones (Rosenberg 2000 et al.), leading us to consider the possible role of NO in the regulation of behavioural state, and specifically in the induction of recovery sleep.

NO is an intercellular signalling molecule that regulates both physiological and pathophysiological processes in the CNS (Garthwaite and Boulton 1995; Gross and Wolin 1995; Keynes and Garthwaite 2004). NO concentrations undergo state-dependent modulation during the sleep–wake cycle both in the thalamus and the cortex (Burlet and Cespuglio 1997; Williams et al. 1997), but there are no measurements of NO concentrations during prolonged wakefulness in any brain area. Several previous studies have shown that intraperitoneal, subcutaneous or intracerebroventricular administration of inhibitors of the NO-synthesizing enzyme NO synthase (NOS) decreases spontaneous sleep (Kapas et al. 1994; Dzoljic et al. 1996; Monti et al. 1999, 2001; Ribeiro et al. 2000; Monti and Jantos 2005; Ribeiro and Kapas 2005; Cavas and Navarro 2006) whereas NO donors increase it (Kapas and Krueger 1996; Monti and Jantos 2004a), suggesting that NO may have a role as a sleep-facilitating agent. Local injections of NOS inhibitors into the pons, including the cholinergic laterodorsal/pedunculopontine tegmental nuclei (LDT/PPT) as well as the dorsal raphe nucleus, have generally also decreased either REM sleep or both NREM and REM sleep (Datta et al. 1997; Leonard and Lydic 1997; Hars 1999; Monti et al. 1999, 2001). Studies employing local manipulations of NO level in the BF have provided controversial results: injections of NOS inhibitors into the BF have been reported to decrease NREM sleep and increase wakefulness (Monti and Jantos 2004b) or have no effect on sleep (Vazquez et al. 2002), and injection of the NO precursor l-arginine or a NO donor have been reported to be ineffective (Monti and Jantos 2004b). Some studies have also suggested that NO has a pro-arousal (Pape and Mager 1992; Marino and Cudeiro 2003) effect. However, the role of NO in the induction of recovery sleep after SD has been addressed in only one study, in which the NOS inhibitor Nω-nitro-l-arginine methyl ester (L-NAME) administered intraperitoneally decreased NREM sleep recovery (Ribeiro et al. 2000).

We hypothesized that release of NO locally in the BF during prolonged wakefulness, either from intrinsic cells or from terminals of projecting neurones such as those from the LDT/PPT, may be critical for the subsequent increase in sleep, and that the NO release is associated with changes in energy metabolism. To test this hypothesis we either decreased the amount of NO produced during prolonged wakefulness or pharmacologically increased it, and measured the effect of these manipulations on metabolites of energy metabolism and sleep. We also measured the concentrations of NO2 and NO3 (collectively termed NOx) in the BF during SD, thereby assessing directly the question of whether NO levels change during prolonged wakefulness.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Animals and surgery

Male Wistar rats (300–400 g) were kept under constant temperature (23.5–24°C) conditions on a 12-h light–dark cycle with lights on at 08.30 hours. Water and food were provided ad libitum. Under general anaesthesia (0.1 mg/kg s.c. medetomidine + 30 mg/kg i.p. pentobarbital), rats were implanted with electrodes for recording electroencephalogram (EEG) and electromyogram (EMG) and with a unilateral guide cannula for microdialysis probes (CMA/11 Guide; CMA/Microdialysis, Stockholm, Sweden) targeting the BF cholinergic area, including the horizontal limb of the diagonal band of Broca (HDB), substantia innominata and magnocellular preoptic area (MCPO) (anterior = − 0.3; lateral = 2.0; vertical = 5; Paxinos and Watson 1998), and as a control, neighbouring areas that do not contain cholinergic cells (Figs 1a and b). After a recovery and adaptation period of 2 weeks, before microdialysis probe insertion, a 30-h baseline EEG recording was obtained for each rat. The experimental protocol was accepted by the Ethical Committee for Animal Experiments at the University of Helsinki and the provincial government of Uusimaa, Finland, and was in accordance with the laws of Finland and the European Union.

image

Figure 1.  Location of microdialysis probe tips. (a) Camera lucida drawing of the microdialysis probe tips in the BF (n = 16, filled circles) or control areas outside the BF (n = 9, open circles). ac, anterior commissure; ox, optic chiasm. (b) Photograph of the track of a representative probe tip located in the BF (HDB area).

Download figure to PowerPoint

Microdialysis experiments

Microdialysis probes (CMA/11, membrane length and diameter 2 and 0.24 mm respectively; CMA/Microdialysis) were inserted into the target areas at least 20 h before the start of experiments (Porkka-Heiskanen et al. 1997) and stayed in place permanently during the 2-week experimental period. If a probe became clogged or dried out, it was replaced by a new one. For experiments, animals were connected to microdialysis leads (EEG/EMG leads combined with microdialysis tubing) for 6 h starting after lights on at 09.00–09.30 hours. Artificial cerebrospinal fluid (aCSF; 147 mm NaCl, 3 mm KCl, 1.2 mm CaCl2, 1.0 mm MgCl2) or solutions of the studied drugs (dissolved in aCSF) were pumped through the microdialysis probe at 1 µL/min. The microdialysis leads were disconnected after 1 h of recovery sleep following SD at 15.30–16.00 hours and replaced by ordinary EEG/EMG recording leads. When metabolite concentrations were measured during 2 h of recovery sleep after SD, microdialysis tubing was disconnected at 16.30–17.00 hours. The 30-min periods during which microdialysis tubing was being replaced were excluded from analysis. Four types of microdialysis experiments were performed.

Experiment type 1

SD was performed between 11.30 and 14.30 hours; changes in metabolite concentrations during SD as well as changes in subsequent sleep after SD (‘recovery sleep’) were measured (Kalinchuk et al. 2003).

Experiment type 2

A drug was infused for 3 h during the spontaneous sleep–wake cycle (11.30–14.30 hours); changes in metabolite concentrations during treatment and in subsequent sleep after treatment were measured.

Experiment type 3

A drug was infused during SD; its effects on changes in metabolite concentrations during SD and on recovery sleep after SD were measured and compared with baseline and the SD effect. Drug infusions to be combined with SD were started 1 h before the SD at 10.30 hours and continued through the deprivation until 14.30 hours.

Experiment type 4

A drug (drug B) was infused simultaneously with another drug, whose effects had already been tested (drug A). Infusion of drug B started 1 h before the infusion of drug A at 10.30 hours and continued for 4 h until 14.30 hours.

Analysis

On each experimental day microdialysis samples for the analysis of metabolites were collected at 30-min intervals from 09.30 hours to 15.30 or 16.30 hours (see above). The first recording was always a baseline aCSF infusion for 6 h. All subsequent experiments were preceded by a daily pretreatment baseline period of aCSF infusion for 2 h, during which samples were collected for metabolite analysis. For each analysed metabolite (NOx, adenosine, lactate and pyruvate), averages of concentrations from two (experiment 3) or three (experiments 1 and 2) samples collected during the pretreatment period and from three samples collected during treatment were compared.

The EEG was recorded for 30 h (continuing for 24 h after the treatment). One animal was used in two to four experiments (including the aCSF baseline run) during a 2-week period with a minimum of 48 h between experiments. The EEG was also recorded during the non-experimental days, and the records were evaluated to ensure that there was no carry-over effect of the drugs on sleep. Baseline was recorded from all animals (n = 25) and SD was performed for 18 animals. For the final analysis, SD data from animals with probes in the BF and outside BF were combined, giving n = 18 for the SD group. In the other groups n = 4–8 (see figure legends).

SD protocol

The animals were trained to human presence for at least 1 week before the experiments. During the daily training sessions the animals were actually handled, by taking them out of the cage and letting them play with the researcher, and then returned to the cage. The daily sessions lasted up to 10 min, and the animals were regarded as trained when they did not show a fear reaction when the researcher entered the room and approached the cage.

Rats were sleep-deprived for 3 h between 11.30 and 14.30 hours using a gentle handling procedure (Franken et al. 1991), which included introduction of new objects into the cages in order to keep the animals occupied and replacing them by new ones when the animals appeared to become sleepy. Any physical contact with animals was excluded. Monitoring of food consumption did not show significant differences between spontaneous wakefulness and SD periods. Continuous monitoring of EEG/EMG during the deprivation period was used to assess the behavioural state of the animals.

EEG recording and analysis

The EEG/EMG signals were amplified and sampled at 104 Hz as described previously (Kalinchuk et al. 2003). EEG recordings were scored using the Spike 2 program (version 5.11; Cambridge Electronic Devices, Cambridge, UK) in 30-s bins semi-automatically for NREM sleep, and manually for REM sleep and wakefulness. Semi-automatic scoring of NREM sleep was performed based on quantification of EEG power in the delta band (0.5–4Hz), sigma band (11–15Hz) and gamma band (30–45Hz) using custom scripts for power spectral analysis as described previously (Stenberg et al. 2003). Scoring of NREM sleep was validated by correlating the results of the semi-automatic scoring with results of manual scoring for 14 records (30 h each); the mean ± SEM correlation was 91.4 ± 0.6%. Manual analysis of wakefulness, NREM sleep (for validation of results of semi-automatic scoring) and REM sleep was performed in accordance with classical criteria: wakefulness was identified by the presence of low-amplitude desynchronized activity in the EEG and high-amplitude activity in the EMG; NREM sleep was identified by the presence of high-amplitude slow waves in the EEG and decreased activity in the EMG compared with wakefulness; REM sleep was distinguished as a state with regular theta activity (5–8Hz) in the EEG and decreased muscle tone compared with wakefulness. The 30-h recordings were divided into 3- and 6-h bins; the amounts of NREM sleep, REM sleep and EEG power in the delta range during NREM episodes (delta power) in each bin during the experimental day were compared with the corresponding time bin on the baseline day and percentage differences were calculated. A period of 18 h after the treatments (14.30–08.30 hours, the shaded area in figures) was used for the final quantitative analysis as this was the period of maximal change from baseline in sleep and delta power.

To compare EEG power density spectra during recovery sleep (n = 13) and NO donor infusion (n = 8), vigilance states were manually scored for 4-s epochs during 6 h after treatment. EEG power spectra (Fast Fourier transform routine, Hanning window) were calculated within the frequency range of 0.4–20 Hz with resolution 0.4 Hz.

Definitions of baselines

Baseline EEG 1 (30 h) was recorded before probe insertion to ensure that the animal had recovered completely from the operation. Baseline EEG 2 (30 h) was recorded before the experiments. The probe was inserted at least 20 h before the start of recording. Microdialysis leads were attached and aCSF was infused for 6 h between 9.30 and 15.30 hours. As there was no difference in sleep between the two baselines, baseline EEG 2 was used when normalizing the sleep data.

Daily pretreatment baseline was determined for metabolite (adenosine, lactate and pyruvate) measurements. The daily baseline values for each animal were monitored throughout the 2-week experimental period to ascertain the stability of the preparation.

HPLC analysis

Adenosine was measured using HPLC coupled to a UV detector or a McPherson fluorimeter [when the infused molecules, i.e. L-NAME and (Z)-1-[N-(2-aminoethyl)-N-(2-ammonioethyl)amino] diazen-1-ium-1,2-diolate (DETA/NO) prevented UV detection]. Details of the adenosine assays have been published previously (Porkka-Heiskanen et al. 1997; Rosenberg et al. 2000). The amounts of lactate and pyruvate were measured as described by Hallström et al. (1989). The detection limits of the assays were 0.8 nm for adenosine (signal to noise ratio 2 : 1), 0.6 µm for pyruvate (3 : 1) and 10 µm for lactate (3 : 1) (Grob 1985). Concentrations of the samples collected during SD or drug infusions were normalized to the mean concentration of samples collected during the baseline pretreatment period when aCSF was infused (= 100%). If a drug interfered with the adenosine assay (2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide, cPTIO), adenosine was not measured.

NOx measurements

As no endogenous source other than NO is known for NO2 and NO3 (collectively NOx) this metabolite has generally been taken as indicative of NO production (Mackenzie et al. 1996). The dietary intake of nitrate can considerably affect the plasma nitrate concentrations, but as these anions do not penetrate the blood–brain barrier, their effect on brain concentrations is not significant (Clark et al. 1996). NOx concentrations were measured using a Nitrate/Nitrite Fluorometric Assay kit (Cayman Chemical Company, Ann Arbor, MI, USA) according to the manufacturer's instructions. The detection limit of the assay was 0.06 µm in the final reaction mixture. The results were calculated as NOx. Concentrations of the samples collected during SD or drug infusions were normalized to the mean concentration of samples collected during the baseline pretreatment period when aCSF was infused (= 100%). Measurement of NOx during DETA/NO infusion was not performed because the values would not reflect correct tissue concentrations.

Materials

To decrease NO concentration in the BF we used either a NO scavenger, cPTIO (potassium salt; Tocris, Ballwin, MO, USA) (Akaike et al. 1993) at concentrations 1, 5 and 10 mm, or a non-selective NOS inhibitor, L-NAME (Sigma, St Louis, MO, USA). After a dose-finding study (data not shown) a concentration of 0.6 mm L-NAME was chosen for the experiments. Nω-nitro-d-arginine methyl ester (D-NAME) (Sigma), an inactive isomer of L-NAME, was infused at 0.6 mm as a control. Infusions were performed both during the normal sleep–wake cycle and during SD.

To increase the local NO concentration in the BF during the spontaneous sleep–wake cycle, we infused the NO donor DETA/NO (Sigma) at 1 mm into the BF (Beltran et al. 2000).

To block effects mediated by adenosine we infused caffeine, a non-specific antagonist of adenosine receptors (Fluka, Basel, Switzerland) at 1 mm, which was estimated to give an effective tissue concentration of 0.1 mm. The effect of adenosine at this dose is predominantly adenosine receptor antagonism, whereas higher doses also act as phosphodiesterase inhibitors (Daly and Fredholm 1998).

According to the instructions of the manufacturer of CMA probes and our previous measurements (Porkka-Heiskanen et al. 1997), the probe recoveries for most substances are 10–15%. Thus, the effective BF concentrations of drugs were estimated to have been about one tenth of those in the infusion solutions. The concentrations indicated above are concentrations in the infusion solution.

Histological verification of the probe locations

After the experiments, the probe tip locations were verified histologically, as described previously (Kalinchuk et al. 2003) (Fig. 1). Figure 1(a) shows the locations of the probe tips in the BF (n = 15) and in the area outside the BF (n = 10). Fig. 1(b) shows a representative probe scar in the brain.

Data presentation

To emphasize the homeostatic component of sleep regulation (Borbely 1982), we normalized the sleep data from experimental days to corresponding time bins from the baseline day – a procedure that eliminated the circadian variation in the data presented. The effects of this procedure on the original data are exemplified in Fig. 4. Because data presented either in 3- or 6-h bins described the results equally, 6-h bins were chosen for presentation. For statistical analysis, three 6-h mean values obtained during the first 18 h after the treatment were averaged and normalized to the respective three 6-h mean values obtained during the baseline day. As the levels of metabolites did not change during the 6-h baseline collection (data not shown), we used a daily baseline protocol to evaluate the changes in metabolite levels during the SD and drug infusions. Concentrations of samples collected during the baseline pretreatment period (average, = 100%) and the experimental period (average) were compared, and the difference was expressed as a percentage increase/decrease compared with baseline. Data are expressed as mean ± SEM.

image

Figure 4.  Effects of SD on NREM sleep, REM sleep and delta power. (a, d, g) Distribution of NREM/REM sleep (minutes) and EEG power density (μV2) in the delta range (0.5–4 Hz) at different time points during a baseline day and during a SD day. (b, e, h) For presentation of the data in 6-h time bins, two 3-h mean values were averaged to obtain 6-h mean values. (c, f, i) The bars show quantitative changes in sleep/delta power during the first 18 h of recovery sleep after the SD period. x-axis open bar, light period; hatched bar, SD during the light period; black bar, dark period. (a–c) SD increased NREM recovery sleep with a maximum at 9–12 h after SD. (d–f) Maximal increase in delta power intensity during NREM sleep was observed at 3–6 h after SD. (g–i) REM sleep was not changed during the first 6 h after SD but was increased at 9–12 h after SD. *p < 0.05, **p < 0.001 versus control (baseline EEG recording). Values are mean ± SEM.

Download figure to PowerPoint

Statistical analysis

Statistical analysis was performed using SigmaStat 3.0 statistical software (SPSS Inc., Chicago, IL, USA). To evaluate the statistical significance of the effects of SD and drug infusions on NREM/REM sleep and delta power for comparisons of two groups we used the Mann–Whitney rank sum test and for comparisons between more than two groups we used one-way anova followed by Student–Newman–Keuls post hoc test (for normally distributed values) or Kruskal–Wallis one-way anova on ranks, followed by Dunn's post hoc test (for non-normally distributed values). Comparison of EEG power spectra was performed using anova on ranks for repeated measures followed by Holm–Sidak post hoc test.

To evaluate the effects of different treatments on concentrations of adenosine, lactate, pyruvate and NOx before and after the treatments we used a paired t-test or Wilcoxon signed rank test (for non-normally distributed data).

Results

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Effects of SD

NOx concentration in the BF

We hypothesized that production of NO in the BF during SD is causally related to the generation of recovery sleep. As an initial test of this hypothesis, we measured the concentration of NO2 + NO3 (NOx), oxidative degradation products of NO, in microdialysis fluid taken from a probe implanted into the BF. NOx concentrations in samples collected before the SD (during the daily baseline collection, average of three samples) were compared with those measured during the 3 h of SD (average of three samples). The basal concentration of NOx measured in samples collected before SD was 0.6 ± 0.1 µm. During the 3 h of SD, NOx concentrations increased by 100 ± 51% compared with baseline predeprivation values (n = 7, paired t-test, t = 2.589, p < 0.05) (Fig. 2a; BF). Separate analysis of samples collected at 1, 2 and 3 h of SD revealed increasing concentrations of NOx during the course of SD, with the highest value reached by the end of SD (Fig. 3a). Concentrations gradually returned to baseline during 2 h of recovery sleep.

image

Figure 2.  Changes in NOx, adenosine, lactate and pyruvate concentrations in the BF and outside the BF during SD. In the BF area (BF; n = 7) concentrations of NOx (a), adenosine (b), lactate (c) and pyruvate (d) during SD were significantly higher than before SD; outside the BF (nBF; n = 7) they were not changed. Values are mean ± SEM. *p < 0.05 versus control (predeprivation baseline level).

Download figure to PowerPoint

image

Figure 3.  Changes in NOx and adenosine concentrations in the BF during SD and recovery sleep. Values are mean ± SEM of seven rats. (a) The concentrations of both adenosine (a) and NOx (b) increased gradually in the course of SD and then gradually decreased during the recovery sleep.

Download figure to PowerPoint

Adenosine, lactate and pyruvate concentrations in the BF

The basal predeprivation concentration of adenosine in microdialysis samples was 3.1 ± 0.5 nmol/L. During the 3 h of SD the adenosine concentration increased by 254 ± 70% compared with baseline predeprivation values (n = 7, paired t-test, t = 5.179, p < 0.01) (Fig. 2b; BF). In agreement with previous data (Porkka-Heiskanen et al. 1997), the adenosine level gradually increased during the course of SD reaching its maximum by the third hour and returned to the baseline level during 2 h of recovery sleep (Fig. 3b), closely resembling the pattern of NOx in the course of SD and recovery (Fig. 3a).

In agreement with our previous data (Kalinchuk et al. 2003), the concentration of lactate and pyruvate in samples collected during SD were increased by 45 ± 8% (paired t-test, t = 5.020, p < 0.01) and 35 ± 5% (paired t-test, t = 5.371, p < 0.01) respectively, compared with the predeprivation level (Figs 2c and d; BF).

Adenosine, lactate, pyruvate and NOx concentrations outside the BF

The basal level of NOx detected outside BF did not differ significantly from the level in the BF (0.5 ± 0.1 µm; t-test, t = 0.611, p > 0.5) but, in contrast to the effect observed in the BF, outside the BF SD did not induce any changes in NOx (n = 7, non-significant decrease by 2 ± 16% compared with predeprivation level) (Fig. 2a; nBF).

In agreement with our previously published data (Kalinchuk et al. 2003), adenosine, lactate and pyruvate concentrations were not changed during SD outside the BF (Figs 2b–d; nBF).

Sleep and delta power (recovery sleep)

In order to characterize the effect of SD on the generation of NREM and REM sleep we combined data collected from animals with microdialysis probes located in the BF and outside the BF (n = 18). The effect of the SD on sleep was evaluated by comparing the EEG recording obtained on the SD day to that obtained on the baseline day (30-h recording). SD significantly increased subsequent NREM sleep, which was increased by 32 ± 3% compared with baseline (Mann–Whitney rank sum test, T = 551.000, p < 0.01) (Figs 4a–c) and led to a 44 ± 8% increase in delta power (0.4–4Hz) intensity during NREM sleep (T = 352.000, p < 0.001) (Figs 4d–f). SD also increased REM sleep by 41 ± 7%, compared with baseline (T = 494.000, p < 0.001) (Figs 4g–i).

In summary, SD increased both NREM and REM sleep as well as delta power (recovery sleep) after the cessation of the 3-h SD period. During the deprivation, adenosine, lactate, pyruvate and NOx concentrations were raised in microdialysates collected from the BF area, but not in those collected outside the BF area.

Effect of increase in NO concentration in the BF

If production of NO in the BF during SD plays a causal role in the generation of recovery sleep, then we would expect that instillation of NO donor into the BF should mimic the effect of SD. For this purpose we infused the NO donor, DETA/NO, through the microdialysis probe (experiment type 2). The effect of drug on metabolite concentrations was evaluated by comparison of average concentrations from pretreatment (= 100%) and treatment periods. The effect of the drug on sleep was evaluated by comparing the EEG record obtained on the drug infusion day to that obtained on the baseline day (30-h recording) and SD day. The distributions of spectral power during spontaneous sleep, recovery sleep and NO-induced sleep were also compared.

Adenosine, lactate and pyruvate

During DETA/NO infusion into the BF, the adenosine concentration increased by 312 ± 89% compared with the preinfusion baseline level (n = 7, paired t-test, t = 2.807, p < 0.05) (Fig. 5a; BF), to a level similar to that encountered during SD. The levels of lactate and pyruvate were significantly increased by 54 ± 12% (paired t-test, t = 2.795, p < 0.05) and 31 ± 4% (paired t-test, t = 4.723, p < 0.05), respectively (Fig. 5b; BF).

image

Figure 5.  Changes in metabolite concentrations, sleep and delta power after DETA/NO infusion into the BF and outside the BF, and after pretreatment with caffeine. (a–c) Changes in metabolite concentrations during the DETA/NO infusion. Concentrations of adenosine (a), lactate (b) and pyruvate (c) during the 3 h of DETA/NO infusion into the BF (n = 7) and outside the BF (n = 7) were significantly increased compared with pretreatment baseline values. *p < 0.05 versus baseline. (d–f) Changes in sleep/delta power and REM sleep after the DETA/NO infusion and DETA/NO + caffeine infusion. Probes were placed in the BF or outside the BF (nBF). Infusion of DETA/NO into the BF (n = 8) induced significant increases in NREM sleep (p < 0.001) (d) and delta power (p < 0.05) (e), which were similar to increases observed in recovery sleep after SD. Pretreatment with caffeine (DETA + caffeine BF, n = 4) blocked the DETA/NO effect. REM sleep after DETA/NO infusion into the BF was significantly decreased (p < 0.05) (f). Infusion of DETA/NO outside the BF (n = 7) did not change time spent in NREM sleep (d), delta power during NREM sleep (e) or time spent in REM sleep (f). Values are mean ± SEM.

Download figure to PowerPoint

Sleep and delta power

The treatment increased NREM sleep by 35 ± 4% compared with the baseline (n = 8, one-way anova, F = 24.575, p < 0.001; Student–Newman–Keuls post hoc test, q = 8.531, p < 0.001) (Fig. 5d). This effect was not quantitatively different from that of SD (Student–Newman–Keuls post hoc test, q = 0.748, p > 0.05).

Infusion of DETA/NO into the BF increased delta power during NREM sleep by 46 ± 21% (Kruskal–Wallis one-way anova on ranks, H2 = 19.966, p < 0.001; Dunn's post hoc test, Q = 2.577, p < 0.05) (Fig. 5e). Comparison with the effect induced by SD showed no difference between treatments (Dunn's post hoc test, Q = 0.267, p > 0.05).

Analysis of EEG power spectra during the first 6 h after SD or DETA/NO infusion into the BF revealed that both treatments induced increases in slow wave activity in the range of 0.4–1.6Hz, showing that SD and DETA/NO induced similar changes in the power spectrum, increasing the low range of delta power, which is a typical change in sleep after SD (Cirelli et al. 2005) (Fig. 6).

image

Figure 6.  Effect of SD and DETA/NO infusion into the BF on EEG power spectra in NREM sleep. Values were obtained by normalizing the average power density during NREM sleep in the first 6 h after SD (n = 13) or DETA/NO infusion (n = 8) to the average power in the same frequency bin during the baseline EEG recording (= 100%). Both SD and DETA/NO infusion induced an increase in the low-frequency range of delta activity (0.4–1.6Hz) compared with baseline; there were no differences between the two experimental groups. Values are mean ± SEM. *p < 0.05, **p < 0.001 versus baseline.

Download figure to PowerPoint

In contrast to the effect on NREM sleep, infusion of DETA/NO into the BF led to significant decrease in REM sleep by 30 ± 9% (one-way anova, F = 19.869, p < 0.001; Student–Newman–Keuls post hoc test, q = 3.157, p < 0.05) (Fig. 5f).

Effect of increase in NO concentration outside the BF

Adenosine, lactate and pyruvate

Infusion of DETA/NO into the non-BF areas induced similar effects as in the BF area: adenosine concentration was increased by 265 ± 133% compared with baseline (n = 7, Wilcoxon signed rank test, T = 27.000, p < 0.05), and concentrations of lactate and pyruvate were increased by 34 ± 6% (paired t-test, t = 4.894, p < 0.05) and 30 ± 5% (paired t-test, t = 2.915, p < 0.05) respectively (Figs 5a–c; nBF).

Sleep and delta power

Infusion of DETA/NO into the non-BF areas did not induce any significant changes in NREM or REM sleep (Figs 5d–f).

Subsequent NREM sleep differed from the response evoked by DETA/NO infusion into the BF (n = 7, one-way anova, F = 62.728, p < 0.001; Student–Newman–Keuls post hoc test, q = 14.310, p < 0.001) but did not differ from control (Student–Newman–Keuls post hoc test, q = 1.574, p > 0.05) (Fig. 5d).

Delta power during NREM sleep after DETA/NO infusion outside the BF was not changed compared with the control (decrease by 10 ± 18%; Kruskal–Wallis one-way anova on ranks, H2 = 17.437, p = 0.004; Dunn's post hoc test, Q = 1.171, p > 0.05) but was significantly different from the effect observed after DETA/NO infusion on to the BF (Dunn's post hoc test, Q = 3.549, p < 0.05) (Fig. 5e).

The amount of REM sleep did not differ from that of the control (one-way anova, F = 8.175, p < 0.05; Student–Newman–Keuls post hoc test, q = 0.539, p > 0.05) and significantly differed from the effect of DETA/NO infusion into the BF (Student–Newman–Keuls post hoc test, q = 4.469, p < 0.05) (Fig. 5f).

Effects of adenosine receptor antagonist infusion into the BF on DETA/NO-induced sleep

In order to verify that the effect of DETA/NO on NREM sleep can be mediated by adenosine we continuously infused the non-specific antagonist of adenosine receptors, caffeine (1 mm), into the BF for 4 h, starting 1 h before DETA/NO infusion (experiment type 4). Pretreatment with caffeine was able to block the increase in NREM sleep, which was significantly different from effects induced by DETA/NO infusion (n = 4, one-way anova, F = 10.961, p < 0.001; Student–Newman–Keuls post hoc test, q = 6.078, p < 0.001) and SD (Student–Newman–Keuls post hoc test, q = 6.271, p < 0.001) (Fig. 5d).

Pretreatment with caffeine also diminished the increase in delta power compared with the effect of DETA/NO infusion (one-way anova, F = 7.596, p < 0.05; Student–Newman–Keuls post hoc test, q = 4.664, p < 0.05) (Fig. 5e), and the effect of SD (Student–Newman–Keuls post hoc test, q = 5.357, p < 0.001).

In summary, infusion of the NO donor DETA/NO into the BF and areas outside the BF during the spontaneous sleep–wake cycle, a treatment that locally increases NO concentration, induced increases in adenosine, lactate and pyruvate concentrations. However, subsequent increases in NREM sleep and delta power were observed only when DETA/NO was infused into the BF area. In contrast to the effect of SD, REM sleep was decreased after DETA/NO infusion into the BF, whereas infusion outside the BF area had no effect on subsequent REM sleep. The effects of DETA/NO on NREM sleep and increase in delta power were blocked by caffeine, a non-specific adenosine antagonist.

Effect of decreasing NO in the BF during SD

We wanted to test whether NO production in the BF during SD is necessary for the generation of recovery sleep. We approached this in two ways, one testing the effect of pharmacological inhibition of NOS and second testing the effect of a NO scavenger instilled during SD.

Effect of a NOS inhibitor L-NAME

The effects of L-NAME infusion into the BF were compared with those of D-NAME (the non-active isomer of L-NAME) infusion into the BF and infusion of L-NAME outside the BF area. Drug infusions started 1 h before SD and continued for 4 h throughout SD (experiment type 3).

Infusion of L-NAME at 0.6 mm into the BF during SD inhibited the increase in NOx concentration characteristic of SD: the level was actually decreased by 14 ± 22% compared with the pretreatment baseline level (n = 6, paired t-test, t = 0.850, p > 0.05) (Fig. 7a). Infusion of D- NAME was ineffective and did not prevent the increase in NOx level in the course of SD (117 ± 58% above baseline; n = 6, Wilcoxon signed rank test, T = 0.036, p < 0.05).

image

Figure 7.  Effect of L-NAME infusion into the BF and outside the BF during SD on metabolite concentrations, NREM/REM recovery sleep and delta power. (a–d) Changes in metabolite concentrations during SD accompanied by L-NAME or D-NAME infusions. L-NAME infusion into the BF (n = 6) prevented increases in NOx (a), adenosine (b), lactate (c) and pyruvate (d) levels during SD, whereas D-NAME infusion had no effect. L-NAME infusion outside the BF (n = 6) did not induce significant changes in adenosine concentration. The level of pyruvate was significantly decreased during L-NAME infusion outside the BF compared with pretreatment baseline values. *p < 0.05 versus baseline. (e–g) Changes in sleep/delta power and REM sleep after SD accompanied by L-NAME or D-NAME infusion. Infusion of L-NAME into the BF during SD (n = 7) significantly reduced NREM recovery sleep (e) and delta power (f) compared with SD (both p < 0.05) whereas D-NAME did not block the SD effect (p > 0.05). (g) REM sleep recovery was significantly decreased after L-NAME infusion into the BF during SD (p < 0.05). Infusion of L-NAME outside the BF (n = 7) during SD did not effect NREM recovery sleep, delta power or REM recovery sleep compared with SD (all p > 0.05). Values are mean ± SEM.

Download figure to PowerPoint

L-NAME infusion into the BF prevented the adenosine increase (no difference from predeprivation value; Wilcoxon signed rank test, T = 0.361, p > 0.5) whereas during D-NAME infusion the adenosine level was significantly increased by 288 ± 65% (paired t-test, t = 3.424, p < 0.05) (Fig. 7b).

Infusion of L-NAME into the BF prevented accumulation of lactate and pyruvate during SD whereas D-NAME did not change the increase in metabolite concentrations; pyruvate reached 29 ± 10% and was significantly different from the pretreatment level (paired t-test, t = 4.240, p < 0.05); lactate could not be measured because the drug interfered with the assay (Figs 7c and d).

Infusion of L-NAME into the BF during SD significantly inhibited NREM sleep recovery compared with SD (n = 7, Kruskal–Wallis one-way anova on ranks, H3 = 27.070, p < 0.001; Dunn's post hoc test, Q = 3.464, p < 0.05) whereas D-NAME did not influence recovery sleep: NREM was significantly increased by 24 ± 4% compared with baseline (Dunn's post hoc test, Q = 3.068, p < 0.05) and did not differ from the effect of SD on NREM sleep (Dunn's post hoc test, Q = 0.810, p > 0.05) (Fig. 7e).

Infusion of L-NAME into the BF during SD significantly inhibited the increase in delta power compared with SD between 0 and 18 h after treatment (n = 7, one-way anova, F = 5.645, p < 0.01; Student–Newman–Keuls post hoc test, q = 4.215, p < 0.05) (Fig. 7f). After infusion of D-NAME, delta power was still increased by 39 ± 14% compared with baseline. The response was different from that after L-NAME infusion (Student–Newman–Keuls post hoc test, q = 3.093, p < 0.05) and was not different from that of SD alone (Student–Newman–Keuls post hoc test, q = 0.472, p > 0.05).

Infusion of L-NAME also inhibited REM sleep recovery as compared with SD (Kruskal–Wallis one-way anova on ranks, H2 = 12.480, p < 0.05; Dunn's post hoc test, Q = 2.700, p < 0.05) decreasing to baseline level (Dunn's post hoc test, Q = 0.089, p > 0.05) (Fig. 7g).

Effect of the NO scavenger cPTIO

Similar to the L-NAME protocol, cPTIO infusion started 1 h before SD and continued for 4 h up to the end of SD (experiment type 3). The effects on sleep of three different doses were tested.

Infusion of cPTIO at 10 mm during SD into the BF totally prevented the increase in NOx concentrations: in fact, the NOx level was significantly decreased by 73 ± 22% compared with pretreatment baseline values (p < 0.01) (data not shown).

Infusion of cPTIO decreased NREM sleep recovery in a concentration-dependent manner (Fig. 8a). After infusion of cPTIO at the lowest dose (1 mm), sleep was still significantly increased by 22 ± 2% compared with baseline (n = 7, Kruskal–Wallis one-way anova on ranks, H4 = 33.481, p < 0.001; Dunn's post hoc test, Q = 3.102, p < 0.05) and did not differ from the effect of SD (Dunn's post hoc test, Q = 0.824, p > 0.05). Infusion of cPTIO at 5 mm substantially blocked recovery sleep (increase by 12 ± 2%), which was not different from baseline (n = 7, Dunn's post hoc test, Q = 1.733, p > 0.05). Infusion of cPTIO at 10 mm totally inhibited recovery sleep compared with SD (n = 7, Dunn's post hoc test, Q = 1.173, p < 0.05). Similarly, infusion of cPTIO dose-dependently decreased the increase in delta power after SD (Fig. 8b). cPTIO blocked the increase in delta power compared with SD at 5 mm (one-way anova, Student–Newman–Keuls post hoc test, q = 4.059, p < 0.05) and 10 mm (Student–Newman–Keuls post hoc test, q = 5.219, p < 0.05). In both cases delta power was no different from that of the control (Student–Newman–Keuls post hoc test, all q < 0.754, p > 0.05).

image

Figure 8.  Effect of cPTIO infusion into the BF during SD on NREM/REM recovery sleep and delta power. Infusion of cPTIO at 1, 5 and 10 mm (n = 7) induced dose-dependent suppression of NREM recovery sleep (a) and delta power (b). Infusion of cPTIO at 1 mm induced a significant decrease in REM sleep recovery (c); infusions of cPTIO at 5 and 10 mm decreased REM but the changes were not significant. Values are mean ± SEM.

Download figure to PowerPoint

cPTIO also inhibited REM recovery sleep. The effect was significant at 1 mm (Kruskal–Wallis one-way anova on ranks, H4 = 16.356, p < 0.05; Dunn's post hoc test, Q = 3.174, p < 0.05); at higher concentrations, similar reductions were observed, but they were not significantly different from SD (Dunn's post hoc test, all Q < 1.597, p > 0.05) (Fig. 8c). They were also not significantly different from the control.

Effect of NO decrease in the non-BF area during SD

Effect of the NOS inhibitor L-NAME

In contrast to the BF, the NOx level outside the BF, which was not increased during SD alone (Fig. 2a), was significantly decreased after L-NAME infusion by 31 ± 7% (n =6, paired t-test, t = 4.454, p < 0.05) (Fig. 7a).

In contrast to the BF, outside the BF there was no increase in adenosine with SD. Infusion of L-NAME into the non-BF areas during SD induced a small and insignificant decrease in the level of adenosine (by 23 ± 19% compared with predeprivation value; n = 6, Wilcoxon signed rank test, T = 0.418, p > 0.05) and lactate (by 27 ± 20%, paired t-test, t = 1.867, p > 0.05). The concentration of pyruvate was significantly decreased by 38 ± 10% (paired t-test, t = 4.943, p < 0.05) (Figs 7b–d).

The increase in NREM sleep after L-NAME infusion outside the BF during SD was significant compared with the control (by 33 ± 2%; n = 7, Kruskal–Wallis one-way anova on ranks, H3 = 27.082, p < 0.001; Dunn's post hoc test, Q = 3.338, p < 0.05) and did not differ from recovery sleep after normal SD (Dunn's post hoc test, Q = 0.415, p > 0.05) (Fig. 7e).

L-NAME infusion outside the BF was not effective in decreasing delta power (increase by 43 ± 20% compared with baseline; one-way anova, F = 3.523, p < 0.05; Student–Newman–Keuls post hoc test, q = 3.642, p < 0.05); the effect was not different from that of SD (Student–Newman–Keuls post hoc test, q = 9.104, p > 0.05) (Fig. 7f).

Infusion outside the BF did not block REM recovery sleep, that increased by 44 ± 19%(compared with SD; Kruskal–Wallis one-way anova on ranks, H2 = 10.401, p < 0.05; Dunn's post hoc test, Q = 0.525, p > 0.05) (Fig. 7g).

In summary, treatment with L-NAME in the BF during SD blocked the increase in NOx concentrations, prevented adenosine, lactate and pyruvate accumulation in the BF, and abolished the development of NREM recovery sleep, increase in delta power and REM sleep recovery. Infusion of L-NAME outside the BF was not effective in blocking recovery sleep. Similarly, scavenging NO in the BF during SD by cPTIO blocked the development of NREM sleep recovery, decreased delta power and diminished REM sleep recovery in a dose-dependent manner. These studies indicate that NO production in the BF is necessary for the generation of recovery sleep.

Effects of decreased NO in the BF on spontaneous sleep

Finally, we tested the effects of the NOS inhibitor L-NAME on spontaneous sleep, and measured the respective changes in metabolite concentrations in the BF and outside. The results were confirmed by infusion of the NO scavenger cPTIO into the BF. Drugs were infused for 3 h between 11.30 and 14.30 hours (experiment type 2).

Effect of the NOS inhibitor L-NAME

Infusion of L-NAME at 0.6 mm into the BF during the spontaneous sleep–wake cycle induced a significant decrease in adenosine level by 38 ± 7% compared with pretreatment baseline level (n = 6, paired t-test, t = 3.312, p < 0.05) (Fig. 9a; BF)

image

Figure 9.  Effect of L-NAME infusion into the BF and outside the BF on spontaneous sleep. (a–c) Changes in metabolite concentrations during L-NAME infusion. Infusions of L-NAME into the BF (n = 6) and outside the BF (nBF; n = 6) decreased adenosine in the BF (a), and lactate and pyruvate in both BF and outside BF. *p < 0.05 versus pretreatment baseline. (d–f) Changes in NREM sleep/delta power and REM sleep after the L-NAME infusion. NREM sleep (d) and delta power (e) were significantly decreased after infusions of L-NAME into the BF (n = 5, both p < 0.01) but not outside (n = 5, p > 0.05). (f) REM sleep was not changed after L-NAME infusions into either area (all p > 0.01). Values are mean ± SEM.

Download figure to PowerPoint

After infusion of L-NAME into the BF the lactate level was significantly decreased by 38 ± 6% (Wilcoxon signed rank test, T = 21.000, p < 0.05) and the pyruvate level was significantly decreased by 37 ± 10% (Wilcoxon signed rank test, T = 21.000, p < 0.05) (Figs 9b and c; BF).

L-NAME significantly decreased NREM sleep by 17 ± 3% (n = 5, Mann–Whitney test, T = 15.00; p < 0.01) (Fig. 9d) and decreased delta power by 58 ± 4% (Fig. 9e) compared with baseline (Mann–Whitney test, T = 15.00; p < 0.01), whereas the decrease of 25 ± 12% in REM sleep was not significant (Mann–Whitney test, T = 21.00; p > 0.05) (Fig. 9f).

Effect of the NO scavenger cPTIO

Infusion of 10 mm cPTIO into the BF decreased NREM sleep by 19 ± 7% compared with baseline (n = 4, Mann–Whitney test, T = 26.000, p < 0.05) (Fig. 10) but the decrease in delta power of 22 ± 19% was not significant (Mann–Whitney test, T = 22.00; p > 0.05). cPTIO at 1 mm slightly decreased NREM sleep by 8.2 ± 3.2%; the difference was not statistically significant (data not shown).

image

Figure 10.  Effect of cPTIO infusion into the BF on spontaneous sleep. Data were obtained by normalizing cPTIO infusion day values for NREM sleep and REM sleep to the baseline day values (= 100%) in 6-h bins for each rat. Three 6-h mean values from the period 18 h after the 3-h cPTIO infusion (shown by the grey shaded area) were averaged and taken for final quantitative analysis. cPTIO infusion at 10 mm (n = 4) induced a significant decrease in NREM sleep (p < 0.05) and a non-significant decrease in REM sleep (p > 0.05). Values are mean ± SEM.

Download figure to PowerPoint

Infusion of cPTIO at 10 mm decreased REM sleep by 27 ± 12%. The difference was not statistically significant (Mann–Whitney test, T = 22.00; p > 0.05) (Fig. 10).

Effects of decreased NO outside the BF on spontaneous sleep

Effect of the NOS inhibitor L-NAME

Infusion of L-NAME into the non-BF area decreased the adenosine level by 38 ± 7% but the change was not statistically significant (n = 6, paired t-test, t = 2.086, p = 0.105). Infusion of L-NAME outside the BF induced a significant decrease in lactate level by 36 ± 1% (Wilcoxon signed rank test, T = 21.000, p < 0.05) and the pyruvate level was significantly decreased by 23 ± 7% (paired t-test, t = 3.961, p < 0.05) (Fig. 9a–c; nBF).

L-NAME infusion outside the BF did not induce significant changes in NREM sleep (decreased by 1 ± 4% compared with baseline; n = 5, Mann–Whitney test, T = 27.00; p > 0.05) and in delta power (increased by 11 ± 5%; Mann–Whitney test, T = 45.00; p > 0.05) (Figs 9d and e). The change in REM sleep was also non-significant (decrease by 11 ± 18%; Mann–Whitney test, T = 33.00; p > 0.05) (Fig. 9f).

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The main finding of the present study is that the production of NO in the basal forebrain during SD is necessary and sufficient to explain the production of NREM recovery sleep. NOx concentration in the BF rose during SD and fell during recovery sleep; infusion of a NO donor into the BF produced an increase in NREM sleep comparable to that produced by SD, and infusion of a NOS inhibitor or a NO scavenger into the BF prevented the generation of NREM recovery sleep. Remarkably, these effects were restricted to the BF; outside this structure the treatments failed to have effects on sleep. Moreover, increases in adenosine, lactate and pyruvate concentrations in the BF preceded the increase in sleep, whereas inhibition of recovery sleep was not associated with an increase in metabolite concentrations, suggesting a connection between energy metabolism and production of recovery sleep in the basal forebrain area.

We have previously established that SD is accompanied by an increase in extracellular adenosine concentration in the BF (Porkka-Heiskanen et al. 1997; Kalinchuk et al. 2003). In the present study, an NO-associated increase in NREM sleep was accompanied by increased concentrations of adenosine, whereas inhibition of NO synthesis or use of an NO-scavenging compound was associated with unchanged levels of adenosine and no NREM sleep recovery, suggesting that the effect of NO on NREM sleep recovery is mediated through an increase in extracellular adenosine in the BF. This view was further confirmed by the observation that blocking of adenosine receptors with caffeine prevented the DETA/NO-induced increase in sleep. An increase in extracellular adenosine level is a signal of discrepancy between energy demand and availability (Dunwiddie and Masino 2001). We have previously shown that experimentally induced energy depletion in the BF increases local adenosine, lactate and pyruvate levels and induces NREM sleep (Kalinchuk et al. 2003). Inhibition of energy production as a consequence of local NO production may be a specific pathway for generating adenosine release in the vicinity of structures capable of producing sleep. In vitro, NO donors stimulate glycolysis, increasing adenosine, lactate and pyruvate levels, and inhibit oxidative phosphorylation, resulting in depletion of total energy production and a decrease in the ATP/ADP ratio (Rosenberg et al. 2000; Maletic et al. 2004). Neurones and astrocytes respond differently to the increase in NO: astrocytes are able to maintain their energy production by increasing glycolysis by activating 6-phosphofructo-1-kinase, whereas neurones appear to be unable to do this (Almeida et al. 2004). Active inhibition by NO of energy production in a specific brain region – the BF – may be considered as a specific mechanism for induction of sleep in response to prolonged wakefulness. We propose that the concentration of NO increases during prolonged wakefulness, resulting in inhibition of local energy metabolism and production of adenosine, which is the signal for NREM sleep induction.

NO production undergoes state-dependent modulations during the sleep–wake cycle both in the thalamus and the cortex, probably due to state-dependent changes in the activity of neuronal NOS-containing projection neurones (Burlet and Cespuglio 1997; Williams et al. 1997); the NO values in both structures were lower during sleep than during waking, suggesting an activity-related NO increase in the brain. NO concentrations during SD have not been assessed previously. Our results, showing an increased level of NO in the BF during SD, support the hypothesis that NO production is critical for homeostatic sleep regulation. One previous study has shown a decrease in recovery sleep after L-NAME administration during SD (Ribeiro et al. 2000). In that study the recovery sleep was attenuated but not completely abolished as in the present study. Systemic administration of the drug in the previous study, as opposed to local administration in the present study, may explain the partially different outcomes of the two studies.

Our results are in agreement with those of many previous studies, in which either intracerebroventricular or systemic routes of administration of NO donors or NOS inhibitors were used, most often during the spontaneous sleep–wake cycle (reviewed in Gautier-Sauvigne et al. 2005). In the present study, administration of both the NOS inhibitor and the NO scavenger during spontaneous sleep decreased NREM sleep, indicating that a certain NO level is required for the appearance of a normal amount of NREM sleep, and further that the effects of the drugs during SD may have two components: effects on the mechanisms of recovery sleep production (sleep homeostasis) and effects on spontaneous sleep. Remarkably, L-NAME decreased spontaneous sleep only when administered into the BF area, indicating that the effect of NO on spontaneous sleep is also at least partially localized. Two previous studies, employing local microinjections of NOS inhibitors into the BF during spontaneous sleep, reported conflicting data (Vazquez et al. 2002; Monti and Jantos 2004b); one reported changes in sleep and the other did not. The discrepancies can most probably be explained by the sharp localization of the effect, as shown in the present study. Previous studies have also shown that local administration of L-NAME into the pons modulates sleep (Datta et al. 1997; Leonard and Lydic 1997; Hars 1999). NOS is co-localized with acetylcholine in most of the BF nuclei, as well as in the LDT/PPT nuclei that project to the BF (Vincent and Kimura 1992). Both the BF and the LDT/PPT area of the pons contain cholinergic neurones, which regulate the vigilance state by release of acetylcholine (Jones 1991; Leonard and Lydic 1997; Hars 1999). As it has been shown that NO modulates acetylcholine release (Vazquez et al. 2002; Leonard and Lydic 2005), it is possible that the effects of NO on sleep are mediated through cholinergic neurones (Leonard and Lydic 2005).

Adenosine, through A1 receptors, is an inhibitory neuromodulator (Fredholm 1995). Clinically, adenosine is used for neuroprotection in connection with seizures (Boison 2005) and as anti-arrhythmic treatment of supraventricular tachycardia (Hutchinson and Scammells 2004). Adenosine is also a powerful vasodilator (Biaggioni 2004). It stimulates neuronal activity through A2A receptors (Fredholm 1995). Adenosine appears to have a specific function in the regulation of recovery sleep (Porkka-Heiskanen et al. 1997). The site-specificity of the effects in the present study is in agreement with earlier studies, in which the increase in adenosine levels during SD was found to be restricted to the BF (Porkka-Heiskanen et al. 2000; Kalinchuk et al. 2003). The sites at which adenosine, lactate, pyruvate and NO increases were found during SD closely correspond to the area where the cholinergic cells of the BF are situated – the HDB, substantia innominata and MCPO. The present results suggest that recovery sleep may be regulated through waking-active cholinergic cells, whereby adenosine acting on A1 receptors would decrease their firing rate (Rainnie et al. 1994). However, sleep homeostasis was not affected in mice with A1 receptor knockout (Stenberg et al. 2003), and a specific lesion of cholinergic cells did not affect sleep (Kapas et al. 1996; Gerashcenkko et al. 2001). Another possibility is that adenosine could disinhibit the inhibitory GABAergic neurones of the nearby ventrolateral preoptic nucleus (Sherin et al. 1996; Chamberlin et al. 2003) thus promoting sleep. A2A receptor agonist injection into the subarachnoid space promoted sleep (Scammell et al. 2001), suggesting that adenosine could also modulate sleep through activation of A2A receptors. In addition, the effects of caffeine were recently shown to be mediated through A2A receptors (Huang et al. 2005). It is probable that during SD these mechanisms work in synchrony: the increased adenosine level inhibits the waking-active neurones and disinhibits sleep-active neurones through A1 receptors, while activating sleep-active neurones through A2A receptors. The sleep response to increased adenosine levels continues for several hours (Porkka-Heiskanen et al. 1997). We have previously shown that increased adenosine levels in the BF increase the expression of A1 receptors, providing one possible mechanism for the prolonged sleep-promoting effect (Basheer et al. 2001).

An alternative explanation that we considered is that the drugs used in this study might have non-specific effects on sleep that subtract from recovery sleep without actually affecting the fundamental process. For example, infusion of glutamate might block recovery sleep but this does not necessarily mean that glutamate plays a specific role in homeostatic sleep regulation. We do not think that the effects of the drugs used in this study are non-specific because: (i) drugs of varied chemical structure were found to act in a predictable way based on their effect on NO levels: a NOS inhibitor and a NO scavenger inhibited recovery sleep, whereas a NO donor mimicked recovery sleep; (ii) NO2/NO3 levels increased with SD, indicative of an association between NO production and recovery sleep; and (iii) the effects of drugs blocking recovery sleep were long-lasting. The effects were typically at a maximum not during or immediately after administration, which was during the light/rest period, when a stimulant would be expected to have maximal effect, but during the subsequent dark/active period (Figs 7 and 8). Interestingly, reduction of NO levels in the BF also had long-lasting effects on spontaneous sleep (Figs 9 and 10), suggesting that NO production is a consequence of waking per se and is central in the generation of sleep drive or Process S.

The present results, in addition to confirming NO as a powerful sleep-facilitating agent, provide strong evidence that NO is a critical part of the homeostatic sleep control mechanism regulating effects of prolonged wakefulness, and is, in fact, necessary and sufficient for the generation of recovery sleep. The results further support the hypothesis that local energy depletion in the BF, as reflected in increases in adenosine, lactate and pyruvate levels specifically in this area during SD, may be the initiator of the chain of events that culminate in the induction of recovery sleep. Adenosine, one of the metabolites indicative of energy depletion, and which during SD accumulates in the extracellular space where it can activate adenosine receptors, appears to be the key molecule in the final induction of recovery sleep.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

We thank Mr Ernst Mecke, Mrs Pirjo Saarelainen and Mrs Sari Levo-Siitari for excellent technical assistance. The work was funded by NIH grants P50 HL60292 and P01 HD18655, the Academy of Finland, European Union grant MCRTN-CT-2004–512362, Finska Läkaresällskapet, the Sigrid Juselius Foundation and the European Sleep Research Society Sanofi-Synthelabo Research Award to AK.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  • Akaike T., Miyazaki K., Ueda S. and Maeda A. (1993) Antagonistic action of imidazolineoxyl N-oxides against endothelium derived relaxing facto/NO through a radical reaction. Biochemistry 32, 827832.
  • Almeida A., Moncada S. and Bolanos J. P. (2004) Nitric oxide switches on glycolysis through the AMP protein kinase and 6-phosphofructo-2-kinase pathway. Nat. Cell Biol. 6, 4551.
  • Basheer R., Halldner L., Alanko L., McCarley R. W., Fredholm B. B. and Porkka-Heiskanen T. (2001) Opposite changes in adenosine A1 and A2a receptor mRNA following sleep deprivation. Neuroreport 12, 15771580.
  • Beltran B. B., Mathur A., Duchen M. R., Erusalimsky J. D. and Moncada S. (2000) The effect of nitric oxide on cell respiration: a key to understanding its role in cell survival or death. Proc. Natl Acad. Sci. USA 19, 1460214607.
  • Benington J. H. and Heller H. C. (1995) Restoration of brain energy metabolism as the function of sleep. Prog. Neurobiol. 45, 347360.
  • Biaggioni I. (2004) Clinical and molecular pharmacologic characteristics of adenosine-induced vasodilation. Clin. Pharmacol. Ther. 75, 137139.
  • Boison D. (2005) Adenosine and epilepsy: from therapeutic rationale to new therapeutic strategies. Neuroscientist 11, 2536.
  • Borbely A. A. (1982) A two process model of sleep regulation. Human Neurobiol. 1, 195204.
  • Borbely A. A. and Tobler I. (1989) Endogenous sleep-promoting substances and sleep regulation. Physiol. Rev. 69, 605670.
  • Brorson J. R. and Zhang H. (1999) Nitric oxide acutely inhibits neuronal energy production. J. Neurosci. 19, 147158.
  • Bryant P. A., Trinder J. and Curtis N. (2004) Sick and tired: does sleep have a vital role in the immune system? Nat. Rev. Immunol. 4, 457467.
  • Burlet S. and Cespuglio R. (1997) Voltammetric detection of nitric oxide (NO) in the rat brain: its variations throughout the sleep–wake cycle. Neurosci. Lett. 226, 131135.
  • Cavas M. and Navarro J. F. (2006) Effects of selective neuronal nitric oxide synthase inhibition on sleep and wakefulness in the rat. Prog. Neuropsychopharmacol. Biol. Psychiatry 30, 5667.
  • Chamberlin N. L., Arrigoni E., Chou T. C., Scammell T. E., Greene R. W. and Saper C. B. (2003) Effects of adenosine on gabaergic synaptic inputs to identified ventrolateral preoptic neurons. Neuroscience 119, 913918.
  • Chee M. W. L. and Choo W. C. (2004) Functional imaging of working memory after 24 h of total sleep deprivation. J. Neurosci. 24, 45604567.
  • Cirelli C., Huber R., Gopalakrishnan A., Southard T. L. and Tononi G. (2005) Locus coeruleus control of slow-wave-homeostasis. J. Neurosci. 25, 45034511.
  • Clark R. S., Kockhanek P. M., Obrist W. D., Wong H. R., Billiar T. R., Wisniewski S. R., Stephen R. and Marion D. W. (1996) Cerebrospinal fluid and plasma nitrite and nitrate concnetrations after head injury in humans. Crit. Care Med. 24, 12431251.
  • Daly J. W. and Fredholm B. B. (1998) Caffeine – an atypical drug of dependence. Drug Alcohol Depend. 51, 199206.
  • Datta S., Patterson E. H. and Siwek D. F. (1997) Endogenous and exogenous nitric oxide in the pedunculopontine tegmentum induces sleep. Synapse 27, 6978.
  • Detari L., Juhasz G. and Kukorelli T. (1984) Firing properties of cat basal forebrain neurones during sleep–wakefulness cycle. Electroencephal. Clin. Neurophysiol. 58, 362368.
  • Dinges D. F., Pack F., Williams K., Gillen K. A., Powell J. W., Ott G. E., Aptowicz C. and Pack A. I. (1997) Cumulative sleepiness, mood disturbances and psychomotor vigilance performance decrements during a week of sleep restricted to 4–5 hours per night. Sleep 20, 267.
  • Dunwiddie T. V. and Masino S. A. (2001) The role of regulation of adenosine in the central nervous system. Annu. Rev. Neurosci. 24, 3155.
  • Dzoljic M. R., De Vries R. and Van Leeuwen R. (1996) Sleep and nitric oxide: effects of 7-nitro indazole, inhibitor of brain nitric oxide synthase. Brain Res. 718, 145150.
  • Franken P., Dijk D. J., Tobler I. and Borbely A. A. (1991) Sleep deprivation in rats – effects on EEG power spectra, vigilance states, and cortical temperature. Am. J. Physiol. 261, R198R208.
  • Fredholm B. B. (1995) Adenosine receptors in the central nervous system. News Physiol. Sci. 10, 122128.
  • Garthwaite J. and Boulton C. L. (1995) Nitric oxide signaling in the central nervous system. Annu. Rev. Physiol. 57, 683706.
  • Gautier-Sauvigne S., Colas D., Parmantier P., Clement P., Gharib A., Sarda N. and Cespuglio R. (2005) Nitric oxide and sleep. Sleep Med. Rev. 9, 101113.
  • Gerashcenkko D., Salin-Pascual R. J. and Shiromani P. J. (2001) Effects of hypocretin–saporin injections into the medial septum on sleep and hippocampal theta. Brain Res. 913, 106115.
  • Grob R. L. (1985) Modern Practice of Gas Chromatography. John Wiley & Sons, New York.
  • Gross S. S. and Wolin M. S. (1995) Nitric oxide: pathophysiological mechanisms. Annu. Rev. Physiol. 57, 737769.
  • Hallström Å., Carlsson A., Hillered L. and Ungerstedt U. (1989) Simultaneous determination of lactate, pyruvate nad ascorbate in microdialysis samples from rat brain, blood, fat and muscle using high-performance liquid chromatography. J. Pharmacol. Methods 21, 113124.
  • Hars B. (1999) Endogenous nitric oxide in the rat pons promotes sleep. Brain Res. 816, 209219.
  • Huang Z. L., Qu W. M., Eguchi N., Chen J. F., Schwarzschild M. A., Fredholm B. B., Urade Y. and Hayaishi O. (2005) Adenosine A2A, but not A1, receptors mediate the arousal effect of caffeine. Nat. Neurosci. 8, 858859.
  • Hutchinson S. A. and Scammells P. J. (2004) A(1) adenosine receptor agonists: medicinal chemistry and therapeutic potential. Current Pharmacol. Design 10, 20212039.
  • Jones B. E. (1991) Paradoxical sleep and its chemical/structural substrates in the brain. Neuroscience 40, 637656.
  • Kalinchuk A., Urrila A.-S., Alanko L., Heiskanen S., Wigren H.-K., Suomela M., Stenberg D. and Porkka-Heiskanen T. (2003) Local energy depletion in the basal forebrain increases sleep. Eur. J. Neurosci. 17, 863869.
  • Kapas L. and Krueger J. M. (1996) Nitric oxide donors SIN-1 and SNAP promote nonrapid-eye-movement sleep in rats. Brain Res. Bull. 41, 293298.
  • Kapas L., Fang J. and Krueger J. M. (1994) Inhibition of nitric oxide synthesis inhibits rat sleep. Brain Res. 664, 189196.
  • Kapas L., Obal F. Jr, Book A. A., Schweitzer J. B., Wiley R. G. and Krueger J. M. (1996) The effects of immunolesions of nerve growth factor-receptive neurons by 192 IgG-saporin on sleep. Brain Res. 712, 5359.
  • Keynes R. G. and Garthwaite J. (2004) Nitric oxide and its role in ischaemic brain injury. Curr. Mol. Med. 4, 179191.
  • Leonard T. O. and Lydic R. (1997) Pontine nitric oxide modulates acetylcholine release, rapid eye movement sleep generation, and respiratory rate. J. Neurosci. 17, 774785.
  • Leonard T. O. and Lydic R. (2005) Nitric oxide synthase inhibition decreases pontine acetylcholine release. Neuroreport 6, 15251529.
  • Mackenzie I. M. J., Ekangaki A., Young J. D. and Garrard C. S. (1996) Effect of renal function on serum nitrogen oxide concentartions. Clin. Chem. 42, 440444.
  • Maletic S. D., Dragicevic-Djokovic L. M., Ognjanovic B. L., Zikic R. V., Stajn A. S. and Spasic M. B. (2004) Effects of exogenous donor of nitric oxide – sodium nitroprusside on energy production of rat reticulocytes. Physiol. Res. 53, 439447.
  • Marino J. and Cudeiro J. (2003) Nitric oxide-mediated cortical activation: a diffuse wake-up system. J. Neurosci. 23, 42994307.
  • Monti J. M. and Jantos H. (2004a) Effects of l-arginine and SIN-1 on sleep and waking in the rat during both phases of the light–dark cycle. Life Sci. 75, 20272034.
  • Monti J. M. and Jantos H. (2004b) Microinjection of the nitric oxide synthase inhibitor L-NAME into the lateral basal forebrain alters the sleep/wake cycle of the rat. Prog. Neuropsychopharmacol. Biol. Psychiatry 28, 239247.
  • Monti J. M. and Jantos H. (2005) A study on the brain structures involved in the qcute effects of fluoxetine on REM sleep in the rat. Int. J. Neuropsychopharmacol. 8, 7586.
  • Monti J. M., Hantos H., Ponzoni A., Monti D. and Banchero P. (1999) Role of nitric oxide in sleep regulation. effects of L-NAME, an inhibitor of nitric oxide synthase, on sleep in rats. Behav. Brain Res. 100, 197205.
  • Monti J., Jantos H. and Monti D. (2001) Increase of waking and reduction of NREM and REM sleep after nitric oxide synthase inhibition: prevention with GABA-A or adenosine A1 receptor agonists. Behav. Brain Res. 123, 2335.
  • Pape H. C. and Mager R. (1992) Nitric oxide controls oscillatory activity in thalamocortical neruons. Neuron 9, 441448.
  • Paxinos G. and Watson C. (1998) The Rat Brain in Stereotaxic Coordinates. Academic Press, New York.
  • Porkka-Heiskanen T., Strecker R. E., Bjorkum A. A., Thakkar M., Greene R. W. and McCarley R. W. (1997) Adenosine: a mediator of the sleep-inducing effects of prolonged wakefulness. Science 276, 12651268.
  • Porkka-Heiskanen T., Strecker R. E. and McCarley R. W. (2000) Brain site specificity of extracellular concentration changes during sleep deprivation and spontaneous sleep: an in vivo microdialysis study. Neuroscience 99, 507517.
  • Rainnie D. G., Grunze H. C., McCarley R. W. and Greene R. W. (1994) Adenosine inhibition of mesopontine cholinergic neurons: implications for EEG arousal. Science 263, 689692.
  • Ribeiro A. C. and Kapas L. (2005) Day- and nighttime injection of a nitric oxide synthase inhibitor elicits opposite sleep responses in rats. Am. J. Physiol. 289, R521R531.
  • Ribeiro A., Gilligan J. G. and Kapas L. (2000) Systemic injection of a nitric oxide synthase inhibitor suppresses sleep responses to sleep deprivation in rats. Am. J. Physiol. 278, R1048R1056.
  • Rosenberg P. A., Li Y., Le M. and Zhang Y. (2000) Nitric oxide-stimulated increase in extracellular adenosine accumulation in rat forebrain neurons in culture is associated with ATP hydrolysis and inhibition of adenosine kinase activity. J. Neurosci. 20, 62946301.
  • Scammell T. E., Gerashcenko D., Mochizuki T., McCarthy M. T., Estabrooke I. V., Sears C. A. and Saper C. B. (2001) An adenosine A2a agonist increases sleep and induces Fos in ventrolateral peroptic neurons. Neuroscience 107, 653663.
  • Sherin J. E., Shiromani P. J., McCarley R. W. and Saper C. B. (1996) Activation of ventrolateral preoptic neurons during sleep. Science 271, 216219.
  • Stenberg D., Litonius E., Halldner L., Johansson B., Fredholm B. B. and Porkka-Heiskanen T. (2003) Sleep and its homeostatic regulation in mice lacking adenosine A1 receptor. J. Sleep Res. 12, 283290.
  • Szymusiak R. S. and McGinty D. J. (1986) Sleep-related neuronal discharge in the basal forebrain of cats. Brain Res. 370, 8292.
  • Szymusiak R. S. and McGinty D. J. (1989) Sleep–waking discharge of basal forebrain projection neurons in cats. Brain Res. Bull. 22, 423430.
  • Vazquez J., Lydic R. and Baghdoyan H. A. (2002) The nitric oxide synthase inhibitor NG-nitro-l-arginine increases basal forebrain acetylcholine release during sleep and wakefulness. J. Neurosci. 22, 55975605.
  • Vincent S. R. and Kimura H. (1992) Histochemical mapping of nitric oxide synthase in the rat brain. Neuroscience 46, 755784.
  • Williams J. A., Vincent S. R. and Reiner P. B. (1997) Nitric oxide production in rat thalamus changes with behavioral state, local depolarization, and brainstem stimulation. J. Neurosci. 17, 420427.