Urotensin II Modulates Rapid Eye Movement Sleep through Activation of Brainstem Cholinergic Neurons

Introduction

Urotensin II (UII) is a neuropeptide that was initially isolated from urophysis extracts of several species of fish (Pearson et al., 1980), with a core structure that is conserved through mollusks to mammals (Conlon et al., 1996, 1997; Chartrel et al., 1998; Coulouarn et al., 1998, 1999). UII has potent vasoconstrictor activity in the periphery (Russell et al., 2001; Bohm and Pernow, 2002). However, intracerebroventricular injection or intra-arterial injection of UII induces hypotensive and bradycardiac effects in rats (Gibson et al., 1986), suggesting that central UII plays a role in cardiovascular homeostasis through its specific receptor in blood vessels and in the CNS (Ames et al., 1999; Liu et al., 1999; Mori et al., 1999; Nothacker et al., 1999). Recent studies have shown that UII receptor mRNA colocalizes with choline acetyltransferase in the mesopontine tegmental area, including the pedunculopontine tegmental (PPT) and the lateral dorsal tegmental (LDT) nuclei (Clark et al., 2001).

The distribution of UII receptor mRNA in the cholinergic PPT and LDT neurons suggests that, in addition to its vascular actions, the UII system may be involved in the regulation of the sleep-wake cycle. Previous studies in cats and rats have demonstrated the role of PPT and LDT nuclei in sleep regulation. The injection of cholinergic agonists into the medial pontine reticular formation, a target region of the PPT and LDT, induces a rapid eye movement (REM)-like state (George et al., 1964; Baghdoyan et al., 1984; Quattrochi et al., 1989; Bourgin et al., 1995) including cortical desynchronization, hippocampal theta rhythm, muscle atonia, reduction of REM sleep onset, and increase in total time of REM sleep (Vanni-Mercier et al., 1989; Yamamoto et al., 1990; Vertes et al., 1993). These and other studies (for review, see Steriade and McCarley, 1990) clearly indicate that cholinergic neurons in the PPT are important regulators of REM sleep.

Considering these studies and the observation that cholinergic PPT and LDT neurons express UII receptor mRNA, we tested the hypothesis that UII can significantly influence cholinergic PPT and LDT neuron activity and alter REM sleep patterns in rats. In addition, because UII induces changes in blood pressure (Gibson et al., 1986), and because it has been proposed that the hypnogenic properties of some putative sleep factors could be related to changes in blood pressure (Feinberg and Campbell, 1999), we also measured cortical blood flow (CBF) in animals treated with UII.

Our results show that both local injection of UII (0.6 pmol) into the PPT nucleus and intracerebroventricular administration of UII (0.6 nmol) dramatically increase the amount of REM sleep in rats. Relative cerebral blood flow was unaffected when UII was administered into the PPT nucleus. Furthermore, using whole-cell recordings, we demonstrate that UII selectively excites mesopontine cholinergic PPT neurons by activating a slow inward current. Collectively, our data strongly suggest that UII can function to modulate REM sleep by enhancing the excitability of mesopontine cholinergic neurons, independent of its effects on the cerebral vasculature.

Materials and Methods

Animals and surgery. Thirty-two adult male Sprague Dawley rats (250-300 g) were implanted under halothane anesthesia (1-2%), with a standard set of stainless-steel screw electrodes for chronic sleep recordings. The electroencephalogram (EEG) was recorded from electrodes placed on the frontal (two electrodes) and parietal (two electrodes) bone over the hippocampus [anteroposterior (AP), -2.46; mediolateral (ML), ±2.0]. A fifth EEG electrode was placed over the cerebellum to ground the animal and to reduce signal artifacts. Two wire electrodes were inserted in the neck musculature to record postural tone through electromyographic (EMG) activity. Insulated leads from the EEG and EMG electrodes were then soldered to a miniconnector that was cemented to the skull with dental acrylic. In addition to the EEG and EMG electrodes, a chronic guide stainless-steel cannula was implanted aseptically into the right lateral ventricle of 10 rats. The guide cannula was placed according to the stereotaxic coordinates of the atlas of Paxinos and Watson (1986) [AP, -0.34; ML, +1.4; dorsoventral (DV), +2.8]. A second group of rats (n = 10) was implanted bilaterally in the pars dissipatus region of the PPT nucleus (AP, -8.0, ML, ±2.0; DV, -5.7) (Paxinos and Watson, 1986). A third cohort of rats (n = 8) was implanted bilaterally with guide cannulas into the PPT pars compacta (PPT-pc) region (AP, -8.3; ML, ±2.0; DV, -7.0), which has a dense population of cholinergic neurons and is the region in which the electrophysiological recordings were conducted. To evaluate the specificity of effects of UII on PPT cholinergic neurons, four rats were implanted with a chronic guide stainless-steel cannula into the right locus ceruleus (LC) (AP, -12.35; ML, +1.1; DV, +7.45, using a posterior angle of 20° from vertical).

After surgical implantation and appropriate wound closure, rats were housed in individual Plexiglas recording cages placed in environmentally controlled chambers (Tech/Serv model EPC-010; BRS/LVE, Laurel, MD). The animals were allowed sufficient time to recover before the study. During anesthetic recovery, animals were observed in a clean single cage, and their normal righting capability and locomotion were monitored. The ambient temperature was maintained at 25°C ± 1, and a 12 h light/dark cycle was maintained throughout the 10 d recovery period and the subsequent experiment period. Food and water were available ad libitum.

Sleep recordings. Rats were housed in individual recording cages that were maintained at the same temperature and light/dark cycle described above. To record sleep-wake states, rats were connected to commutators with flexible cables, allowing their unrestricted movement within the cage, and they were habituated to the recording cages for 96 h. The EEG and EMG signals were amplified in a Grass Instruments (Quincy, MA) model 7D polygraph, filtered in a frequency range of 0.30-75 Hz and sampled at 256 Hz. The EEG and electromyogram were displayed on a computer monitor and stored with a resolution of 128 Hz for off-line scoring of sleep-wake states and spectral analysis, using software supplied by Kissei Comtec (Irvine, CA). The EEG and electromyogram were recorded over 6 h. The recording chambers contained a mini video camera for continuously observing animal behavior during the recording sessions.

Experimental protocol. Six days after surgery, the patency and free drainage of the guide cannula were assessed in animals intracerebroventricularly implanted by microinjection of 5.0 μl of pyrogen-free isotonic saline (PFS; Abbott Laboratories, North Chicago, IL) over a 5 min period. The patency for cannulas implanted into the PPT nucleus was assessed by microinjection of 0.5 μl of PFS over a 10 min period. Procedures for minimizing potential contamination by extraneous pyrogens were used for all drug solutions and vehicles.

The sleep recordings on rats implanted with an intracerebroventricular cannula began 5 h after the onset of the light period [circadian time (CT) 5; CT0, lights on] observing the following schedule: on day 1, rats were injected intracerebroventricularly with 5.0 μl of PFS vehicle at CT5, and recordings were continued for 6 h; on day 3, the same rats were challenged at CT5 with 0.6 nmol of UII in 5.0 μl of vehicle. Injections were made over a 5 min period. UII was obtained from Bachem (Torrance, CA).

The cohort of rats implanted with cannulas in the PPT nucleus pars dissipatus and four rats implanted with cannulas into the PPT-pc were recorded according to the following schedule: on the first day, rats received a bilateral injection of 0.5 μl of PFS vehicle into the PPT nucleus at CT5, and recordings were continued for 6 h; on the third and fifth days, rats were injected bilaterally at CT5 into the PPT nucleus with 0.6 and 6.0 pmol, respectively, of UII in 0.5 μl of vehicle.

The other four rats implanted with cannulas into the PPT-pc received, on day 1, a bilateral injection of 0.3 μl of PFS. On day 3, 3 ng of Cpa-D-Cys-Pal-D-Trp-Lys-Val-Cys-Cpa-NH₂ (SB-710411), a UII receptor antagonist (Behm et al., 2002, 2004), was injected and followed 30 min later by a 0.6 pmol injection of UII. On day 5, animals were treated with 3 ng of SB-710411.

For both intracerebroventricular and microinjection experiments, UII treatments were done at CT5, and 6 h recordings were obtained. We skipped 1 d between treatments to ensure that the effects of the previous treatment had disappeared. All microinjections were made over a 10 min period.

To determine the spread of intracellular injections, four animals were injected with 0.2 μl of biotinylated UII (Phoenix Pharmaceuticals, Belmont, CA) into the PPT. The rats were killed 2 h after injection under deep anesthesia with 5% halothane by intracardial perfusion with 4% paraformaldehyde in phosphate buffer. The brain was removed, sliced, and stained for NADPH diaphorase to visualize the PPT nucleus. The injected peptide was visualized with ABC peroxidase and DAB/H₂O₂. In all instances, the biotinylated peptide could be detected in the PPT area (see Fig. 1).

Rats implanted for LC injections received, on day 1, a microinjection of PFS, whereas on day 3 they were challenged with 6.0 pmol of UII. Injection time and recording conditions were the same as those for other groups.

Infusion sites were verified histologically in 25 μm frontal-brain sections stained with cresyl violet and methylene blue. A set of sections was stained for NADPH diaphorase, which stains cholinergic neurons in the LDT/PPT (Vincent and Kimura, 1992), to verify that the injection site was in close proximity to the PPT.

Sleep and EEG data analysis. The polygraphic data were classified in 15 s epochs as the following states of vigilance: wakefulness (W), slow-wave sleep 1 (SWS1), SWS2, and REM sleep, as described previously (Lerner et al., 1994; Cravatt et al., 1995). Percentage of total time spent in W, SWS1, SWS2, and REM sleep per hour was calculated. The latency to the onset of the first SWS2 and REM sleep episodes and the frequency and duration of the individual SWSs and REM sleep episodes were also calculated. In addition to standard sleep analysis, EEG spectral analyses in the different sleep-wake stages were performed by Fourier fast transformation analysis using 4 s epochs, giving 1.0 Hz bins from 0 to 30 Hz. Each bin was named after its lower limit. The spectral power was divided into wave bands corresponding to δ (0.5-4.0 Hz), θ (4.5-12.0 Hz), β (15.0-20.0 Hz), and γ (30.0-60.0 Hz), averaged over the epoch and divided by bandwidth to give power density for each band. Epochs with artifacts in the polygraph records were discarded. In W, SWS1, SWS2, and REM sleep, only scoring epochs that were both preceded and followed by the same stage were included in the analysis. Because the absolute power is quite variable from rat to rat, we normalized the data according to Mendelson and Bergmann (1999) and then compared the mean power in δ, θ, β, and γ bands between groups through a repeated-measures ANOVA, with Scheffé's F test used for post hoc comparisons. In addition, unpaired, bidirectional t tests were applied to assess the significance of mean EEG power differences in the intracerebroventricular study.

Results

Discussion

We hypothesized that the UII system could be involved in the regulation of REM sleep based on the fact that UII receptors colocalize with ChAT in the mesopontine tegmental area, including PPT and LDT nuclei (Clark et al., 2001). Our results show that both local PPT and intracerebroventricular administration of low doses of UII dramatically increased the total time of REM sleep in rats. This effect could be accounted for by a significant increase in the number of REM episodes. Likewise, our results showed that UII selectively excites PPT cholinergic neurons at their resting membrane potential by activating a slow inward current (Fig. 7), supporting the hypothesis that increments in REM sleep caused by UII could be a result of neuroexcitation of mesopontine cholinergic neurons. In addition, the absence of a UII-related increase in REM sleep in both animals pretreated with SB-710411 and rats treated with UII in the LC strongly suggests that UII regulates REM sleep through activation of UII receptors located in PPT neurons and that its effects are not mediated through metabolic products.

Both the LDT and PPT send cholinergic projections to the REM sleep, generating neurons located in the pontine reticular formation (Mitani et al., 1988). These inputs, when stimulated, produce EPSPs in the pontine reticular formation neurons (Imon et al., 1996; Homma et al., 2002). Because of the stimulatory effect of UII on PPT neurons, it is likely that UII induces acetylcholine release in the pontine reticular formation and induces REM sleep. The magnitude of the increase in REM recorded in this study is at least as large as that recorded when carbachol is injected into the oral pontine reticular nucleus of the rat, which induces an increase in REM sleep episodes with no change in episode length or latency to REM onset (Bourgin et al., 1995). Microinjection of carbachol into the medial gigantocellular tegmental field produces an increase in ACh release for 7 h that has been related to PPT/LDT activation (Lydic et al., 1991; Lydic and Baghdoyan, 1993), whereas glutamatergic stimulation of PPT increases REM sleep up to 6 h after injection (Datta et al., 2001). These results mirror the parameters of the UII-induced increase in REM state. The slow onset of the UII-induced increase in REM sleep in our studies could be caused by the slow diffusion rate of UII in the brain parenchyma, as described previously for hypocretin (Hcrt) (Bourgin et al., 2000).

There are significant differences between the UII-related REM sleep effects and the activation of PPT neurons by microinjection of excitatory amino acids such as l-glutamate. Glutamatergic stimulation of the PPT in rats reduces REM sleep latency and increases REM sleep because of an increase in both duration and number of episodes of REM sleep (Datta et al., 2001). We did not observe significant changes in the latency or duration of REM sleep episodes in the animals treated with UII, suggesting that UII produces a stimulation that is different from the glutamatergic stimulation of the mesopontine cholinergic neurons. It is unlikely that the observed excitation of the PPT nucleus is mediated by glutamate, because the effect of UII was independent of ionotropic blocking of ionotropic glutamate receptors (Fig. 8). Furthermore, breakdown of UII in the doses used here would give rise to a maximal concentration of glutamate of 4 μm, a concentration that would be ineffective at activating glutamate receptors.

Our results also show that UII affects REM sleep at lower doses and is not effective at higher doses. A possible explanation for this result could be the desensitization of UII responses, as suggested by our whole-cell recordings. This is supported by studies showing that UII receptors located in the aorta of rats desensitize after a double treatment at intervals of 5 h, with different doses of human UII (Camarda et al., 2002). Moreover, UII receptors heterologously expressed in cell lines desensitize after repeated UII treatment (H.-P. Nothacker, personal communication). Additional studies are necessary to establish the dose dependency and the molecular base for the observed phenomenon.

Spectral analysis of cortical EEGs showed a differential EEG activation effect between local PPT and intracerebroventricular treatment of UII in freely moving animals. Thus, injection of UII into the PPT nucleus significantly increases γ power in W and hippocampal theta rhythm during REM sleep, suggesting an activation of the PPT nucleus (Vertes et al., 1993; Kinney et al., 1998; Nowacka et al., 2002). In contrast, intracerebroventricular injection of UII increased high-frequency EEG bands during W and REM sleep, suggesting UII-induced cortical activation through cholinergic neurons. This is consistent with the fact that activating cholinergic thalamic afferents produces cortical activation by suppressing slow cortical waves between 0.3 and 4.0 Hz and spindle wave oscillations (11.0-14.0 Hz) (Hu et al., 1989; Curro Dossi et al., 1991; Steriade et al., 1993).

Another key finding of this study was that UII excited bNOS-immunopositive PPT neurons by activating a slow inward current without affecting recorded bNOS-immunonegative neurons, at least at the membrane potentials studied (near rest). This suggests that local application of UII specifically targeted the cholinergic subpopulation of the PPT, as expected from the distribution of UII receptor mRNA (Clark et al., 2001). In unclamped conditions, UII depolarized cholinergic neurons and promoted firing. Blocking action potential discharge and antagonizing fast glutamatergic, GABAergic, and glycinergic neurotransmission did not attenuate the response to UII. This is consistent with a direct postsynaptic action of UII and, considering our immunochemical findings, implies that the functional UII receptors reside on cholinergic PPT neurons.

We also found that input resistance increased during UII-evoked excitation in both voltage-clamp and bridge-mode recordings. This suggests that UII acts by blocking a hyperpolarizing conductance, which is activated around the resting potential, e.g., a potassium current. Additional work will be necessary to identify the underlying current(s). Nevertheless, this elevation of input resistance should also enhance the effectiveness of synaptic inputs, which may further increase PPT neuron excitation.

The UII-evoked increase in PPT neuron discharge rate will also promote the release of ACh at PPT target structures, such as the thalamus and reticular formation. Indeed, UII may have a general role in promoting ACh release, because, in the periphery, UII stimulates contractions of the ileum by activating myenteric neurons to release ACh (Horie et al., 2003) and acts on frog motoneuron terminals to increase spontaneous ACh release (Brailoiu et al., 2003). Although it is not known whether UII acts at central cholinergic terminals, such an action could also facilitate ACh release and deserves additional investigation.

The slow excitatory action of UII was reminiscent of the excitatory action of Hcrt on mesopontine cholinergic neurons in the LDT (Burlet et al., 2002), although Hcrt activates a broader range of neurons. In the LDT, Hcrt directly activates both cholinergic and noncholinergic neurons, as well as glutamatergic afferents, which synapse on both types of neurons (Burlet et al., 2002). Thus, although Hcrt stimulates the entire reticular activating system (for review, see Siegel, 2004), UII actions appear restricted to the cholinergic component. This difference in cellular sensitivity to the two peptides may have considerable functional importance, because local application of Hcrt enhances waking and suppresses REM sleep (Xi et al., 2001), whereas UII enhances REM sleep.

Previous studies have shown that UII possesses vasoactive properties in fish and mammals (Pearson et al., 1980; Gibson et al., 1986; Ames et al., 1999). We characterized the consequences of UII on regional cerebral blood flow by LD flowmetry (Skarphedinsson et al., 1988; Haberl et al., 1989a,b; Lindsberg et al., 1989; Muir et al., 1992). Although LD flowmetry is not quantitative, it monitors relative changes in regional cerebral blood flow quite accurately (Iadecola, 1997). Our results showed that local PPT administration of UII did not alter the CBF in halothane-anesthetized animals, suggesting that the effect of UII on REM sleep is not dependent on the CBF effects. In contrast, intracerebroventricular administration of UII increased CBF. The cortical hemodynamic effect could be associated with the UII activation of different brain areas involved in cardiovascular regulation (Martin and Haywood, 1992; Kubo et al., 1997; Comer et al., 1998; Haywood et al., 2001). In fact, local injection of UII into the paraventricular nucleus, arcuate nucleus, and noradrenergic cells in the lower part of the medulla induces cardiovascular effects (Lu et al., 2002). Thus, UII may have a neurogenic role in the regulation of cardiovascular physiology, and this may be the cause of the intracerebroventricular application-induced blood flow effects shown here. Despite these findings, an increase in cortical blood flow is not seen with local injections into the PPT. It is, therefore, likely that the blood flow increases are not a result of mesopontine cholinergic activation and would be a result of UII acting on noncholinergic neuronal sites (Lu et al., 2002). However, additional studies are needed to confirm that UII does not induce changes in local blood flow.

Together, the results are consistent with a role for UII in the regulation of REM sleep by activating mesopontine cholinergic neurons directly and independently of the cardiovascular effects of UII.